After the first part of the review with an unboxing and a teardown of the MINIX NEO Z100-0dB mini PC, we tested the fanless Intel N100 mini PC with Windows 11 Pro in the second part, and we are now ready to report our experience with Linux, and more specifically Ubuntu 22.04, on the MINIX Z100-0dB mini PC in the final and third part of the review.

We will go through features testing, run some benchmarks to evaluate the performance in Linux, perform storage and network performance tests, check the cooling ability of the mini with a stress test, and also check its power consumption under different scenarios. We will also compare the MINIX Z100-0dB to the actively cooled GEEKOM Mini Air12 mini PC that uses the same Intel Processor N100 CPU.

Ubuntu 22.04 installation

We will install Ubuntu 22.04 alongside Windows 11 Pro in dual boot configuration, so we first shrank the Windows partition to give space for Ubuntu, before installing the Linux distribution from a USB flash drive with a Ubuntu 22.04.4 ISO.

The installation went smoothly with WiFi properly detected. We had to press the Esc key to enter the BIOS and select the USB flash drive for the installation.

Ubuntu 22.04 system information

Going to the About section in the Settings shows we have a “MINIX Technology Limited MINIX NEO Z100-0dB” mini PC with a quad-core Intel N100 processor, 16GB of RAM, and a 512.1 GB M.2 running the latest Ubuntu 22.04.4 LTS.

We can get a few more details in a terminal window…

aey@Z100-0dB-CNX:~$ cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=22.04
DISTRIB_CODENAME=jammy
DISTRIB_DESCRIPTION="Ubuntu 22.04.4 LTS"
aey@Z100-0dB-CNX:~$ uname -a
Linux Z100-0dB-CNX 6.5.0-21-generic #21~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Feb  9 13:32:52 UTC 2 x86_64 x86_64 x86_64 GNU/Linux
aey@Z100-0dB-CNX:~$ free -mh
               total        used        free      shared  buff/cache   available
Mem:            15Gi       1.9Gi        10Gi       480Mi       3.4Gi        12Gi
Swap:          2.0Gi          0B       2.0Gi
aey@Z100-0dB-CNX:~$ df -mh
Filesystem      Size  Used Avail Use% Mounted on
tmpfs           1.6G  2.4M  1.6G   1% /run
/dev/nvme0n1p5  271G   14G  244G   6% /
tmpfs           7.7G   77M  7.7G   1% /dev/shm
tmpfs           5.0M  4.0K  5.0M   1% /run/lock
efivarfs        192K  119K   69K  64% /sys/firmware/efi/efivars
/dev/nvme0n1p1  1.5G   60M  1.5G   4% /boot/efi
tmpfs           1.6G  136K  1.6G   1% /run/user/1000

… and the inxi utility:

aey@Z100-0dB-CNX:~$ inxi -Fc0
System:
  Host: Z100-0dB-CNX Kernel: 6.5.0-21-generic x86_64 bits: 64
    Desktop: GNOME 42.9 Distro: Ubuntu 22.04.4 LTS (Jammy Jellyfish)
Machine:
  Type: Desktop System: MINIX product: MINIX NEO Z100-0dB v: N/A
    serial: <superuser required>
  Mobo: N/A model: N/A serial: <superuser required>
    UEFI: American Megatrends LLC. v: 5.27 date: 11/23/2023
CPU:
  Info: quad core model: Intel N100 bits: 64 type: MCP cache: L2: 2 MiB
  Speed (MHz): avg: 700 min/max: 700/3400 cores: 1: 700 2: 700 3: 700
    4: 700
Graphics:
  Device-1: Intel driver: i915 v: kernel
  Display: wayland server: X.Org v: 1.22.1.1 with: Xwayland v: 22.1.1
    compositor: gnome-shell driver: gpu: i915 resolution: 1600x900~60Hz
  OpenGL: renderer: Mesa Intel Graphics (ADL-N)
    v: 4.6 Mesa 23.2.1-1ubuntu3.1~22.04.2
Audio:
  Device-1: Intel driver: snd_hda_intel
  Sound Server-1: ALSA v: k6.5.0-21-generic running: yes
  Sound Server-2: PulseAudio v: 15.99.1 running: yes
  Sound Server-3: PipeWire v: 0.3.48 running: yes
Network:
  Device-1: Intel driver: iwlwifi
  IF: wlo1 state: up mac: e0:c2:64:ee:50:e3
  Device-2: Realtek RTL8125 2.5GbE driver: r8169
  IF: enp1s0 state: down mac: 8c:32:23:06:4b:fc
Bluetooth:
  Device-1: Intel AX201 Bluetooth type: USB driver: btusb
  Report: hciconfig ID: hci0 state: up address: E0:C2:64:EE:50:E7 bt-v: 3.0
Drives:
  Local Storage: total: 476.94 GiB used: 13.08 GiB (2.7%)
  ID-1: /dev/nvme0n1 model: NEO-J51-SSD-512GB size: 476.94 GiB
Partition:
  ID-1: / size: 270.46 GiB used: 13.02 GiB (4.8%) fs: ext4
    dev: /dev/nvme0n1p5
  ID-2: /boot/efi size: 1.5 GiB used: 59.5 MiB (3.9%) fs: vfat
    dev: /dev/nvme0n1p1
Swap:
  ID-1: swap-1 type: file size: 2 GiB used: 0 KiB (0.0%) file: /swapfile
Sensors:
  System Temperatures: cpu: 53.0 C mobo: N/A
  Fan Speeds (RPM): N/A
Info:
  Processes: 243 Uptime: 56m Memory: 15.39 GiB used: 2.66 GiB (17.3%)
  Shell: Bash inxi: 3.3.13

Inxi reports the system features a quad-core Intel Processor N100 CPU clocked up to 3,400 MHz with NEO-J51-SSD-512GB NVMe SSD, a Realtek RTL8125 2.5GbE controller, and an Intel AX201 wireless module for WiFi and Bluetooth. The CPU temperature at idle is shown to be 53°C.

Ubuntu 22.04 Benchmarks on the MINIX Z100-0dB mini PC

We’ll start Linux benchmarks with the sbc-bench.sh script:

aey@Z100-0dB-CNX:~/Downloads/sbc-bench-master$ sudo ./sbc-bench.sh -r
Starting to examine hardware/software for review purposes...

sbc-bench v0.9.64

Installing needed tools: apt-get -f -qq -y install gcc make build-essential powercap-utils curl git sysstat links mmc-utils smartmontools stress-ng p7zip, tinymembench, ramlat, mhz, cpufetch, cpuminer. Done.
Checking cpufreq OPP. Done.
Executing tinymembench. Done.
Executing RAM latency tester. Done.
Executing OpenSSL benchmark. Done.
Executing 7-zip benchmark. Done.
Throttling test: heating up the device, 5 more minutes to wait. Done.
Checking cpufreq OPP again. Done (11 minutes elapsed).

Results validation:

  * Measured clockspeed not lower than advertised max CPU clockspeed
  * No swapping
  * Background activity (%system) OK
  * Too much other background activity: 0% avg, 6% max -> https://tinyurl.com/mr2wy5uv
  * Powercap detected. Details: "sudo powercap-info -p intel-rapl" -> https://tinyurl.com/4jh9nevj

Full results uploaded to http://sprunge.us/dAZ2v0

# MINIX Technology Limited MINIX NEO Z100-0dB  / N100

Tested with sbc-bench v0.9.64 on Wed, 06 Mar 2024 14:43:35 +0700. Full info: [http://sprunge.us/dAZ2v0](http://sprunge.us/dAZ2v0)

### General information:

    Information courtesy of cpufetch:
    
    Name:                Intel(R) N100
    Microarchitecture:   Alder Lake
    Technology:          10nm
    Max Frequency:       3.400 GHz
    Cores:               4 cores
    AVX:                 AVX,AVX2
    FMA:                 FMA3
    L1i Size:            64KB (256KB Total)
    L1d Size:            32KB (128KB Total)
    L2 Size:             2MB
    L3 Size:             6MB
    Peak Performance:    435.20 GFLOP/s
    
    N100, Kernel: x86_64, Userland: amd64
    
    CPU sysfs topology (clusters, cpufreq members, clockspeeds)
                     cpufreq   min    max
     CPU    cluster  policy   speed  speed   core type
      0        0        0      700    3400       -
      1        0        1      700    3400       -
      2        0        2      700    3400       -
      3        0        3      700    3400       -

15759 KB available RAM

### Clockspeeds (idle vs. heated up):

Before at 63.0°C:

    cpu0: OPP: 3400, Measured: 3390 

After at 71.0°C:

    cpu0: OPP: 3400, Measured: 3391 

### Performance baseline

  * memcpy: 9572.6 MB/s, memchr: 15573.3 MB/s, memset: 8552.2 MB/s
  * 16M latency: 115.2 103.1 113.5 113.7 115.9 109.3 102.3 106.0 
  * 128M latency: 128.4 119.8 128.5 120.8 128.1 125.2 118.2 121.0 
  * 7-zip MIPS (3 consecutive runs): 12324, 9829, 9903 (10680 avg), single-threaded: 3938
  * `aes-256-cbc     940228.66k  1179035.97k  1218735.96k  1229265.58k  1231844.69k  1232142.34k`
  * `aes-256-cbc     919321.40k  1179155.71k  1218922.75k  1229503.49k  1232041.30k  1232743.08k`

### PCIe and storage devices:

  * Realtek RTL8125 2.5GbE: Speed 5GT/s (ok), Width x1 (ok), driver in use: r8169
  * 476.9GB "NEO-J51-SSD-512GB" SSD as /dev/nvme0: Speed 8GT/s (ok), Width x4 (ok), 0% worn out, drive temp: 54°C
  * 16MB SPI NOR flash, drivers in use: spi-nor/intel-spi

### Challenging filesystems:

The following partitions are NTFS: nvme0n1p3,nvme0n1p4 -> https://tinyurl.com/mv7wvzct

### Swap configuration:

  * /swapfile on /dev/nvme0n1p5: 2.0G (0K used)

### Software versions:

  * Ubuntu 22.04.4 LTS (jammy)
  * Compiler: /usr/bin/gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 / x86_64-linux-gnu
  * OpenSSL 3.0.2, built on 15 Mar 2022 (Library: OpenSSL 3.0.2 15 Mar 2022)    

### Kernel info:

  * `/proc/cmdline: BOOT_IMAGE=/boot/vmlinuz-6.5.0-21-generic root=UUID=29792f2d-4917-48a2-b3ac-d2f9d3c90049 ro quiet splash vt.handoff=7`
  * Vulnerability Spec store bypass:    Mitigation; Speculative Store Bypass disabled via prctl
  * Vulnerability Spectre v1:           Mitigation; usercopy/swapgs barriers and __user pointer sanitization
  * Kernel 6.5.0-21-generic / CONFIG_HZ=250

Waiting for the device to cool down...................................... 50.0°C

While the CPU temperature never exceeds 80°C, the performance is impacted after a while with the 7-zip benchmark starting at 12,324 MIPS, but eventually dropping to 9,829 and 9,903 MIPS for the second and third runs with an average of 10,680 MIPS.

Let’s check the power limits as recommended in the script:

aey@Z100-0dB-CNX:~$ sudo powercap-info -p intel-rapl
enabled: 1
Zone 0
  name: package-0
  enabled: 1
  max_energy_range_uj: 262143328850
  energy_uj: 7170855116
  Constraint 0
    name: long_term
    power_limit_uw: 6000000
    time_window_us: 27983872
    max_power_uw: 6000000
  Constraint 1
    name: short_term
    power_limit_uw: 12000000
    time_window_us: 2440
    max_power_uw: 0
  Constraint 2
    name: peak_power
    power_limit_uw: 78000000
    max_power_uw: 0
  Zone 0:0
    name: core
    enabled: 0
    max_energy_range_uj: 262143328850
    energy_uj: 5362964664
    Constraint 0
      name: long_term
      power_limit_uw: 0
      time_window_us: 976
  Zone 0:1
    name: uncore
    enabled: 0
    max_energy_range_uj: 262143328850
    energy_uj: 39667134
    Constraint 0
      name: long_term
      power_limit_uw: 0
      time_window_us: 976

PL1 is set to 6W and PL2 to 12W just as in Windows 11 Pro, and much lower than the power limits set in the Mini Air12 mini PC (15W/25W) which may impact the CPU performance somewhat.

Let’s carry on with Geekbench 6.2.2 single-core and multi-core tests

The MINIX Z100-0dB mini achieved 1,243 points in the single-core test and 3,189 points in the multi-core test.

The Intel iGPU was tested with Unigine Heaven Benchmark 4.0 at 1920×1080 resolution. The system rendered the scene at 11.7 FPS on average with a 294-point score.

Next up is YouTube 4K and 8K video playback in Firefox.

A 4K 30 FPS VP9 video played perfectly smoothly with no frame dropped after watching the video for 5 minutes.

8K 30 FPS was all good as well with just 9 frames dropped out of 11,352 during a 6-minute test.

The MINIX Z100-0dB mini PC has no problem playing a 4K 60 FPS YouTube video in Ubuntu 22.04, although the number of dropped frames went up to 203 out of 24,981.

Going up to 8K 60 FPS was another story with the system struggling to play the video that ended up being unwatchable with 8,289 frames dropped out of 18,288, or a 45% drop rate.

The YouTube streaming experience was similar in Windows and Linux for 2160p30, 2160p60, and 4320p30, but the 8K60 (4320p60) video was more or less watchable in Windows 11 with 6.7% dropped frames, while it was not in Ubuntu 22.04 with 45% of frames dropped.

We ran Speedometer 2.0 to evaluate the web browsing performance in the latest version of Firefox in Ubuntu 22.04.

The MINIX fanless mini PC rendered the test at 146 runs per minute with the 10 iterations of the test ranging between 136.5 and 152.8 runs/s.

Comparison of MINIX Z100-dB0 Ubuntu 22.04 benchmarks against other mini PCs

Let’s now compare the Ubuntu 22.04 benchmark results for the MINIX Z100-0dB fanless mini PC against other Alder Lake-N hardware platforms, namely the GEEKOM Mini Air12 (Intel Processor N100), Blackview MP80 (Processor N97) mini PC, Weibu N10 Core i3-N305 mini PC, and UP 7000 fanless single board computer (Intel Processor N100).

Here are the basic specifications of the five systems under test.

	MINIX Z100-0dB	GEEKOM Mini Air12	Blackview MP80	Weibu N10	UP 7000 SBC
SoC	Intel Alder Lake-N Processor N100	Intel Alder Lake-N Processor N100	Intel Alder Lake-N Processor N97	Intel Core i3-N305 Alder Lake-N	Intel Processor N100 Alder Lake-N
CPU	4 cores up to 3.4 GHz	4 cores up to 3.4 GHz	4 cores up to 3.6 GHz	8 cores up to 3.80 GHz	4 cores up to 3.4 GHz
GPU	24EU Intel HD Graphics	24EU Intel HD Graphics	24 EU Intel UHD Graphics	32EU Intel HD Graphics	24EU Intel HD Graphics
Memory	16 GB DDDR4-3200	16GB DDR5-4800	16GB LPDDR5	8GB DDR4-3200	8GB LPDDR5
Storage	512GB NVMe SSD	512GB NVMe SSD	512GB NVMe SSD	512GB NVMe SSD	64 GB eMMC flash
Default OS	Windows 11 Pro	Windows 11 Pro	Windows 11 Pro	Windows 11 Pro	None

All benchmark results are in Ubuntu 22.04, except for the Blackview MP80 (N97) mini PC that was running Fedora 39.

	MINIX Z100-0dB	GEEKOM Mini Air12	Blackview MP80 (N97)	Weibu N10 Core i3-N305	UP 7000 SBC
OS	Ubuntu 22.04	Ubuntu 22.04	Fedora 39	Ubuntu 22.04	Ubuntu 22.04
sbc-bench.sh
- memcpy	9,572.6 MB/s,	10,459.3 MB/s	8989.0 MB/s	9,949.4 MB/s	8,277.6 MB/s
- memset	8,552.2 MB/s	10,665.4 MB/s	12881.2 MB/s	8,991.6 MB/s	13,678.5 MB/s
- 7-zip (average)	10,680	13,940	13,230	17,615	12,890
- 7-zip (top result)	12,324	13,976	13,270	20,002	13,178
- OpenSSL AES-256 16K	1,232,743.08k	1,233,283.75k	1302822.91k	1,377,211.73k	1,231,454.21k
Geekbench 6 Single	1,243	1,213	1,251	1,177 (Geekbench 5)	-
Geekbench 6 Multi	3,189	3,272	3,141	4,856 (Geekbench 5)	-
Unigine Heaven score	294	303	404	451	258
Speedometer (Firefox)	146	149	152	-	-

The Geekbench 5 results for the Weinu N10 mini PC are just there for reference since Geekbench 5 and Geekbench 6 results can’t be compared directly. The MINIX NEO Z100-0dB has the lowest score in most of the tests, so there’s a cost to going fanless, although most users will probably not notice. All benchmarks were performed in a room with a 28C to 30C ambient temperature, so it might be a tougher test than for most users. Having said that the UP 7000 fanless SBC with the same Intel N100 SoC did perform somewhat better, except in the Unigine Heaven 4.0 benchmark (lower) and single-core benchmarks (about the same).

NVMe SSD storage and USB performance

Storage performance may be important to overall system performance, especially during boot and launching programs, so we also tested the 512GB M.2 NVMe SSD that ships with the mini PC using iozone3:

aey@Z100-0dB-CNX:~$ sudo iozone -e -I -a -s 1000M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2 
	
                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
         1024000       4   147453   251479   348949   349401    51702   249240                                                                
         1024000      16   472785   696410   883857   885519   126532   656920                                                                
         1024000     512   657154   972524  1528673  1523950   939217  1354287                                                                
         1024000    1024  1459188  1510039  1529597  1532754  1207208  1350257                                                                
         1024000   16384  1545730  1541024  1593991  1597533  1618665  1516253                                                                

iozone test complete.

The sequential read speed was 1,593 MB/s, the sequential write speed was 1,545 MB/s, and the SSD also delivered decent random I/O performance, so it’s acceptable, but not quite to the level of higher-end SSDs found in more expensive mini PCs or even the SSD in the GEEKOM Mini Air12.  For reference, the read speed was 2,066 MB/s and the write speed was 1,585 MB/s in Windows 11 using CrystalDiskMark.

To double-check the actual speed of the USB ports, we connected the ORICO M234C3-U4 USB 3.0 NVMe SSD enclosure to each of the USB 3.0 ports and a Seagate USB Expansion drive to the USB 2.0 ports while relying on lsusb and iozone3 command line utilities to confirm the speed.

Here’s the output for the USB-C port on the front panel:

aey@Z100-0dB-CNX:/media/aey/EXT4-REVIEW$ lsusb -t | grep uas
    |__ Port 1: Dev 2, If 0, Class=Mass Storage, Driver=uas, 10000M
aey@Z100-0dB-CNX:/media/aey/EXT4-REVIEW$ sudo iozone -e -I -a -s 1000M -r 16384k -i 0 -i 1
	
                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
         1024000   16384   965648   976617   853478   856033                                                                                  

iozone test complete.

… and the top USB 2.0 port on the rear panel:

aey@Z100-0dB-CNX:/media/aey/USB3_EXT4$ lsusb -t | grep uas
    |__ Port 2: Dev 6, If 0, Class=Mass Storage, Driver=uas, 480M
aey@Z100-0dB-CNX:/media/aey/USB3_EXT4$ sudo iozone -e -I -a -s 1000M -r 16384k -i 0 -i 1

                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
         1024000   16384    41796    42573    43116    43235                                                                                  

iozone test complete.

The results of all 5 ports can be summarized as follows:

• “Front panel” with power button
  • USB-C – 10 Gbps – 853 MB/s read speed, 965 MB/s write speed
  • USB-A (Top) – 10 Gbps – 867 MB/s read speed, 994 MB/s write speed
  • USB-A (Bottom) – 10 Gbps – 851 MB/s read speed, 982 MB/s write speed
• “Rear panel” with DC jack
  • USB-A  (top) – 480 Mbps – 43 MB/s read speed, 41 MB/s write speed
  • USB-A (bottom) – 480 Mbps – 42 MB/s read speed, 43 MB/s write speed

All ports work as advertised. I wrote “front panel” and “rear panel” in quotes because it’s difficult to assign a rear and front panel the way the mini PC was designed.

Networking (2.5GbE and WiFi 6)

Network throughput was measured with iperf3, and we started testing the 2.5GbE Ethernet port using UP Xtreme i11 Edge mini PC at the other end.

• Download

aey@Z100-0dB-CNX:~$ iperf3 -t 60 -c 192.168.31.12 -i 10 -R
Connecting to host 192.168.31.12, port 5201
Reverse mode, remote host 192.168.31.12 is sending
[  5] local 192.168.31.41 port 53734 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate
[  5]   0.00-10.00  sec  2.74 GBytes  2.35 Gbits/sec                  
[  5]  10.00-20.00  sec  2.74 GBytes  2.35 Gbits/sec                  
[  5]  20.00-30.00  sec  2.74 GBytes  2.35 Gbits/sec                  
[  5]  30.00-40.00  sec  2.74 GBytes  2.35 Gbits/sec                  
[  5]  40.00-50.00  sec  2.74 GBytes  2.35 Gbits/sec                  
[  5]  50.00-60.00  sec  2.74 GBytes  2.35 Gbits/sec                  
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.03  sec  16.4 GBytes  2.35 Gbits/sec    0             sender
[  5]   0.00-60.00  sec  16.4 GBytes  2.35 Gbits/sec                  receiver

iperf Done.

• Upload

aey@Z100-0dB-CNX:~$ iperf3 -t 60 -c 192.168.31.12 -i 10 
Connecting to host 192.168.31.12, port 5201
[  5] local 192.168.31.41 port 49706 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-10.00  sec  2.74 GBytes  2.35 Gbits/sec    0    663 KBytes       
[  5]  10.00-20.00  sec  2.74 GBytes  2.35 Gbits/sec    0    663 KBytes       
[  5]  20.00-30.00  sec  2.74 GBytes  2.35 Gbits/sec    0    988 KBytes       
[  5]  30.00-40.00  sec  2.74 GBytes  2.35 Gbits/sec    0    988 KBytes       
[  5]  40.00-50.00  sec  2.74 GBytes  2.35 Gbits/sec    0   1.45 MBytes       
[  5]  50.00-60.00  sec  2.74 GBytes  2.35 Gbits/sec    0   1.45 MBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.00  sec  16.4 GBytes  2.35 Gbits/sec    0             sender
[  5]   0.00-60.04  sec  16.4 GBytes  2.35 Gbits/sec                  receiver

iperf Done.

• Full-duplex

aey@Z100-0dB-CNX:~$ iperf3 -t 60 -c 192.168.31.12 -i 10 –bidir
Connecting to host 192.168.31.12, port 5201
[  5] local 192.168.31.41 port 39506 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-10.00  sec  2.74 GBytes  2.35 Gbits/sec  183    641 KBytes       
[  5]  10.00-20.00  sec  2.74 GBytes  2.35 Gbits/sec    0    641 KBytes       
[  5]  20.00-30.00  sec  2.74 GBytes  2.35 Gbits/sec    0   2.12 MBytes       
[  5]  30.00-40.00  sec  2.74 GBytes  2.35 Gbits/sec    0   2.12 MBytes       
[  5]  40.00-50.00  sec  2.74 GBytes  2.35 Gbits/sec    0   2.12 MBytes       
[  5]  50.00-60.00  sec  2.74 GBytes  2.35 Gbits/sec    0   2.12 MBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.00  sec  16.4 GBytes  2.35 Gbits/sec  183             sender
[  5]   0.00-60.04  sec  16.4 GBytes  2.35 Gbits/sec                  receiver

iperf Done.

Excellent performance, and here the MINIX Z100-0dB is quite better than the GEEKOM Mini Air12 which is limited by its gigabit Ethernet interface.

We then switched to WiFi 6 using Xiaomi Mi AX6000 router.

• Download

aey@Z100-0dB-CNX:~$ iperf3 -t 60 -c 192.168.31.12 -i 10 -R
Connecting to host 192.168.31.12, port 5201
Reverse mode, remote host 192.168.31.12 is sending
[  5] local 192.168.31.208 port 52208 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate
[  5]   0.00-10.00  sec  1.09 GBytes   937 Mbits/sec                  
[  5]  10.00-20.00  sec  1.10 GBytes   941 Mbits/sec                  
[  5]  20.00-30.00  sec  1.09 GBytes   940 Mbits/sec                  
[  5]  30.00-40.00  sec  1.10 GBytes   942 Mbits/sec                  
[  5]  40.00-50.00  sec  1.10 GBytes   941 Mbits/sec                  
[  5]  50.00-60.00  sec  1.10 GBytes   942 Mbits/sec                  
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.04  sec  6.57 GBytes   940 Mbits/sec    0             sender
[  5]   0.00-60.00  sec  6.57 GBytes   940 Mbits/sec                  receiver

iperf Done.

• Upload

aey@Z100-0dB-CNX:~$ iperf3 -t 60 -c 192.168.31.12 -i 10
Connecting to host 192.168.31.12, port 5201
[  5] local 192.168.31.208 port 50828 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-10.00  sec  1.08 GBytes   931 Mbits/sec   83   1.59 MBytes
[  5]  10.00-20.00  sec  1.09 GBytes   940 Mbits/sec    3   1.75 MBytes
[  5]  20.00-30.00  sec  1.10 GBytes   942 Mbits/sec    3   1.74 MBytes
[  5]  30.00-40.00  sec  1.09 GBytes   941 Mbits/sec    5   1.66 MBytes
[  5]  40.00-50.00  sec  1.10 GBytes   942 Mbits/sec    1   1.77 MBytes
[  5]  50.00-60.00  sec  1.09 GBytes   941 Mbits/sec    6   1.77 MBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.00  sec  6.56 GBytes   939 Mbits/sec  101             sender
[  5]   0.00-60.05  sec  6.56 GBytes   938 Mbits/sec                  receiver

iperf Done.

Again, excellent performance with 940 Mbps and 938 Mbps download/upload speed, and faster than the already-impressive 800 Mbps we got with the GEEKOM Mini Air12. As usual, WiFi performance is quite better in Linux, as the Z100-0dB (only) delivered a download speed of 686 Mbps and a 581 Mbps upload speed in Windows 11 Pro.

We also successfully tested Bluetooth with a file transfer to an Android smartphone.

Thermal performance testing with a stress test

We ran a stress test while checking the CPU temperature to see how well the MINIX Z100-0dB fanless mini PC manages to cool the Intel Processor N100 quad-core CPU.

While testing in a room with a 28°C ambient temperature, the CPU temperature never exceeded 66°C while the CPU frequencies hovered around 2,600 to 2,665 MHz. That’s cooler than the actively-cooled GEEKOM Mini Air12 whose N100 CPU operated at around 77°C for the same test, but also slower since the CPU frequency was 2,900 MHz. This can be explained by the lower thermal limit set in the MINIX Z100-0dB mini PC and also shows up in slightly lower benchmark scores.

MINIX Z100-0dB power consumption in Ubuntu 22.04

The power consumption was measured with a wall power meter:

• Power off
  • 0.9 – 1.0 Watt (shutdown command)
  • 1.4 to 1.5 Watts (after removing and inserting the power plug)
• Idle –  7.5 to 7.6 Watts
• Web browsing – 8.2 to 13.9 Watts
• Video playback – 14.7 to 17.9 Watts (Youtube 8Kp60 in Firefox)
• Stress test on all four cores – 17.2 to 19.8 Watts

Note: The mini PC was connected to WiFi 6, two USB RF dongles for a wireless keyboard and mouse combo, and a VGA monitor (through an HDMI to VGA adapter) during measurements.

Conclusion

MINIX Z100-0dB fanless mini PC works great with Ubuntu 22.04 and we had no issues with any of the features. As we’ve seen in Windows, benchmark results are also a bit lower in Linux than the actively-cooled GEEKOM Mini Air12 mini PC with the same Intel Processor N100 CPU due to lower power limits, and the SSD is not quite as fast, but the MINIX Z100-0dB does have its advantages too (besides the fanless design) with better networking performance thanks to a 2.5GbE port and higher WiFi 6 throughput close to 1 Gbps thanks to the Intel AX201 wireless module.

YouTube video playback worked fine up to 4Kp60 and 8Kp30 in Firefox, but the video was unwatchable at 8Kp60 with close to half of the frames dropped. Cooling is adequate too with no CPU thermal throttling detected under a stress test with the maximum recorded temperature being 66°C in a room at 28°C. That’s because power limits come into play before the CPU thermal limits, and the CPU runs at 2.6 GHz under heavy loads.

We’d like to thank MINIX for sending us a sample of the NEO Z100-0dB mini PC for review. The model showcased here with 16GB RAM and 512GB SSD is sold for $269.69 on Amazon and you’ll also find it on the MINIX online store along with the cheaper 8GB/256GB model going for $239.

CNXSoft: This article is a translation – with some additional insights – of the original review on CNX Software Thailand by Suthinee Kerdkaew.

The post MINIX Z100-0dB review – Part 3: A fanless Intel N100 mini PC tested with Ubuntu 22.04 appeared first on CNX Software - Embedded Systems News.

The Spartan UltraScale+ FPGA family is the latest inclusion to AMD’s Cost-Optimized portfolio, a series of FPGAs designed to balance cost, power, and form factor with affordability. The UltraScale+ FPGA family is designed for cost-sensitive, low-power applications requiring high I/O count and substantial security.

Devices in the Spartan UltraScale+ family offer a high I/O to logic cell ratio for FPGAs built in 28nm and lower process technology (the highest in the industry, according to AMD), consume up to 30% less power than compared to the previous generation, and feature robust security features that outclass the rest of the Cost-Optimized portfolio.

This FPGA family is built on the same UltraScale+ architecture as previous Artix and Zynq products. They are the first AMD UltraScale+ FPGAs to feature a hardened DDR memory controller and PCIe Gen4 x8 support, “providing both power efficiency and future-ready capabilities for customers.”

AMD Spartan UltraScale+ specifications:

• System Logic Cells – Up to 218K
• Memory
  • On-Chip Memory: Block RAM for low latency, high throughput, and UltraRAM
  • External Memory: LPDDR4x and LPDDR5 up to 4266 Mb/s (hard MC), and DDR4 (soft MC) IP up to 2400 Mb/s
• I/O
  • I/O Count: 572
  • Types:
    • High Density I/O (HDIO) up to 3.3V,
    • High-Performance I/O (HPIO) up to 1.8V,
    • XP5IO up to 1.5V, supporting 3200 Mb/s MIPI and 1800 Mb/s LVDS
  • PCIe: Gen4 x8
  • Transceivers – Up to 8 GTH transceivers supporting up to 16.3 Gb/s
• Security – NIST-approved post-quantum cryptography, unique device identification, permanent tamper penalty for device protection, and side-attack protection via DPA countermeasures

The Spartan UltraScale+ FPGA series is primarily targeted for embedded vision, healthcare, industrial networking, robotics, and video applications according to the press release. The high I/O count will enable the FPGAs to interface with a wide range of sensors and coupled with the programmable logic make it possible to control the sensors in real time and with low latency.

On the software end, the Spartan UltraScale+ FPGA family (along with the rest of AMD’s FPGA and adaptive SoCs lineup) is supported by AMD’s Vivado Design Suite and Vitis Unified Software Platform. This allows both hardware and software designers to leverage the “productivity benefits of these tools via a single designer cockpit from design to verification.”

AMD Spartan UltraScale+ samples and evaluation kits are expected to be available for sampling and evaluation in the first half of 2025. Documentation is currently available and tools support starting with the AMD Vivado Suite in Q4 2024. To learn more about the Spartan UltraScale+ FPGA family and what it offers, be sure to visit the product page.

The post AMD announces the Spartan UltraScale+ FPGA family for cost-sensitive and IO-intensive applications appeared first on CNX Software - Embedded Systems News.

PineBerry Pi was the first company to launch a Raspberry Pi HAT+ board making use of the Raspberry Pi 5’s PCIe FPC connector with the Hatdrive! M.2 HAT for the Raspberry Pi 5 along with the HatDrive! Bottom expansion board that goes underneath the Raspberry Pi 5 SBC.

The company has now launched five additional HAT+ boards for the Raspberry Pi 5 with various PCIe (NVMe/AI modules) and networking (2.5GbE, GbE) options.

Here’s the list with basic specifications:

• HatNET! 2.5G –  2.5GbE for the Raspberry Pi 5 using a Realtek RTL8125BG controller, dedicated MAC address range (i.e. not randomly assigned), and an RJ45 jack with Activity and Link LEDs
• HatDrive! NET 1G – Gigabit Ethernet (Realtek RTL8111H) and M.2 M-Key socket for NVMe 2230,2242 SSD. Implemented through an ASMedia PCIe Gen 2 switch and booting with the SSD is not supported at this time. You may want to follow the relevant GitHub issue if that’s important to you.
• HatDrive! AI – Two M.2 PCIe sockets: one M-Key socket for a 2230/2242 NVMe SSD and one E-Key socket for a Google Coral Edge TPU module or other compatible M.2 module.  Implemented through an ASMedia PCIe Switch with the same limitation as above.
• HatDrive! Dual –  Same as above but with two M.2 M-Key PCIe sockets for 2230/2242 NVMe SSDs. RAID-1 is supported for data redundancy.
• HatBRICK! Commander – Adds two PCIe FPC connectors to the Raspberry Pi 5 through an ASMedia PCIe switch to daisy chain as many PCIe HAT+ boards as needed. A 5V DC input connector is also included if additional power is required

Besides the five new boards, PineBerry Pi also changed the design of the original HatDrive! TM1 HAT+ with the new HatDrive! TM1S HAT+ adding one opening to provide easier access to the MIPI CSI/DSI connectors and a custom CNC-milled metal bracket for easier SSD installation.

All the PineBerry Pi HAT+ boards ship with an FPC PCIe ribbon cable and relevant accessories. You’ll find them on the company’s online store for €16,99 to €45,99 depending on the selected model. Some of the new boards are for pre-order so they’ll only start shipping around March 20. There might be a benefit in terms of compactness when using HAT+ boards, but there’s also a price premium, most notably for the HatNET! 2.5G (45.99 Euros) since you can find RTL8156BG USB 3.0 dongles with the same performance going for $20 and up on Amazon.

Via Twitter/X

The post PineBerry Pi launches five additional HAT+ boards for the Raspberry Pi 5 with 2.5GbE, GbE, M.2 NVMe, and more appeared first on CNX Software - Embedded Systems News.

Shortly after I wrote about the Mustool MT13S 2-in-1 thermal imager and multimeter, Xinfrared asked me if I wanted to review the Xtherm II TS2+ thermal imager for smartphones. They offer versions that work for Android or iOS smartphones, so the company sent me the Android version of the Xtherm II TS2+ for review. After listing the key features and specifications, I’ll go through an unboxing, and report my experience using the thermal imager with the OPPO A98 5G smartphone running Android 14.

Xtherm II TS2+ specifications

• Minimum focus – 8mm
• Resolution – 256×192
• Pixel Pitch – 12μm
• FOV – 44.9° x 33.4°
• Image Frame Rate – 25Hz
• NETD (Noise Equivalent Temperature Difference) – ≤40mK@25°C, F#1.0
• MRTD (Minimum Resolvable Temperature Difference) – ≤500mK@25°C,F#1.0
• Temperature Range
  • Measurements- -20°C ~ +450°C with ±2°C or ±2% reading accuracy
  • Operating – -20°C ~ +50°C
• Temperature Correction – Manual/automatic
• Power Consumption – <350mW
• Dimensions – 26×26×24.2mm
• Weight – Less than 18grans

Supported OS including Android, Harmony OS, and iOS, but not the hardware will change too since some devices use a USB-C port, while Apple devices rely on a Lightning connector. The Xtherm application can capture photos or videos, and the company also offers an SDK for secondary development.

Unboxing

I was expecting to receive a small USB-C device, so I was confused when I received the relatively large package.

There’s a bag inside…

… with a bunch of accessories including another small bag with the thermal imager, a smartphone and thermal imager holder, a female USB-C to male USB-C cable, a female USB-C to male micro USB cable, a lanyard, a user guide, and two Xinfrared stickers.

The thermal imager itself is tiny with the lens protected with a plastic cover.

The outer ring can be rotated to manually adjust the focus point. More on that later.

Thermal imager and smartphone holder

While it’s possible to connect the Xinfrared Xtherm II TS2+ directly to your smartphone’s USB Type-C port, the provided smartphone holder can be convenient for some use cases, notably to mount the kit to a tripod.

Xtherm II TS2+ review with the Xtherm app on Android

We’ll need to install the Xtherm app for Android and insert the thermal imager into the USB-C port of the smartphone to get started. But the first time, I only had a black screen with controls, and it looked like the thermal imager was not detected.

That’s because the OTG port is disabled by default in Android for security reasons. So I went to the settings, searched for “USB” or “OTG”, and enabled the OTG connection. Note that if the USB-C OTG connection is not used for over 10 minutes, it will be automatically disabled. In practice, that means that each time I need to use the app, the OTG connection must be manually enabled before inserting the thermal camera.

 

Here’s what the interface looks like:

It’s basically just like a camera app. My first “photo” was of my review table. It shows the temperature at the center (31.9°C) and the highest temperature (42.6°C) which happens to be the power supply fpr the TP-Link switch in the center-right of the image. We can also see a Xiaomi Mi AX6000 router that’s fairly warm on the left and another power supply on the right.

I also turned the camera around to take some selfies. The left one is with the aircon with the air coming out at about 8°C where I’ll yellow because I’m the hottest part of the capture, the center one is with the review table, and in the last one, I mainly point at the ceiling that’s fairly warn due to a hot afternoon in Thailand.

I also went through some practical checking out ice cubes in the freeze (as well as -20.7°C) and boiled some water (100.8°C). As you can see from the left screenshot it’s also possible to enable the camera from the smartphone as well. But note that this feature is useless for closer shots since the smartphone camera points to another area.

I also tried the video capture function while starting to boil the water. There aren’t any flames in the shot, and the yellow parts just represent hot air.

The main reason I was interested in a thermal imager was for single board computer and mini PC reviews. So the next step was to connect a Raspberry Pi 5 SBC, remove the active cooler, and check the temperature at idle.

The top of the CPU was pretty cool at just 29.5°C (we’ll explain why below), and the hottest part was the DA9091 PMIC (45.4°C). Since the Xtherm II TS2+ thermal imager can focus as close as 8mm, I moved closer to the PMIC and circuit around after adjusting the focus ring until it would not turn anymore.

The hottest part was not the PMIC itself, but one of the components around it.

I then started a stress test on all four cores…

The metal cover on the Broadcom BCM2712 processor never really got hot, but we can see heat flowing from the sides. Again the DA9091 PMIC was the hottest shown as reaching up to 80.7°C above. [Update: That’s because the top of the processor is a bright reflective material with low emissivity, so the temperature measurement is not valid. So I stuck some black tape on top of the processor, ran the stress test for around 15 minutes, and took another measure, and this time around the CPU temperature was indeed shown to be 84.1°C.

]

I then fitted the active cooler back on top of the Raspberry Pi 5 while the stress test was running and waited a little while before checking again.

Although we can see still some heat coming out of the power circuitry, the hottest parts of the board are now covered up, and the hottest part is not in the middle of the cooling fan.

Next up I switched to the Radxa Rock 5B SBC with the fan disconnected. At idle, the hottest part is the Realtek RTL8125B 2.5GbE controller at 44.5°C.

If I start a stress test on the eight cores of the Rockchip RK3588 processor, the area under the fan gets hotter.

If we connect the fan, the temperature around it drops, and the Realtek RTL8125B controller becomes the hottest part on the board again.

I changed the focus point again to have a closer look at the networking part (Ethernet and WiFi) of the Rock 5B board.

I also did a quick test on the MINIX NEO Z100-0dB fanless mini PC while it was idle (left)…

… and under higher loads (right).

The Xtherm Android app is pretty easy to use and also comes with various settings for the thermal imager notably the temperature switch to select between -20°C to +120°C and +120°C to +450°C measurement ranges. It’s also possible to mirror, flip, and rotate the image depending on the orientation of the thermal imager.

Conclusion

Xinfrared Xtherm II TS2+ thermal imager is very easy to use with the Android app and seems fairly accurate based on the ice and boiling water tests I did. It can also be useful to test PCB heat dissipation with a minimal focus length of just 8mm allowing the user to take close-up thermal images of the boards.

I’d like to thank Xinfrared for sending the Xtherm II TS2+ for review. You’ll find the kit reviewed there on Aliexpress for $323.53 including shipping, but the SHOUHOU301 coupon code will bring that to under $300. It’s also available on Amazon ($339) and the official store ($339). Update: The coupon code CNXSOFT5 will lower the price by 5% on the Amazon and Xinfrared stores.

The post Xtherm II TS2+ review – A 256×192 thermal imager tested with an Android smartphone appeared first on CNX Software - Embedded Systems News.

The Waveshare RP2040-LCD-0.99-B rounded display is the latest display module by Waveshare. This board is built around the Raspberry Pi RP2040 MCU and hosts a 128×115 pixels 65K color IPS LCD, along with a QMI8658C IMU, all inside a CNC metal case with an acrylic dull-polish bottom plate.

The board also hosts a USB Type-C connector, an LDO, a 2MB NOR-Flash, a SH1.0 6PIN connector (adapting 4x GPIO pins), and a BOOT button.

Waveshare RP2040-LCD-0.99-B rounded display module specification:

• Microcontroller – Raspberry Pi RP2040 dual-core Arm Cortex M0+ processor up to 133 MHz with 264KB SRAM
• Storage – 2MB on-board flash memory
• Display – 0.99-inch 128×115 pixel 65K color IPS LCD
• USB – 1x USB 1.1 Type-C port (host/device) for power and programming using drag-and-drop via USB mass storage
• Sensor – QMI8658C IMU
• Expansion – 4x GPIO via SH1.0 6PIN connector
• Power Management – Low-power sleep and dormant modes
• Dimension – 33 x 9.8 mm
• Construction:
  • CNC metal case
  • Acrylic dull-polish bottom plate

As the display is built around the Raspberry Pi RP2040 MCU it supports the official Pico C/C++ and MicroPython SDKs and is also compatible with the Arduino IDE.

This is not the first rounded display module we have covered, previously we have written about displays like the MaTouch ESP32-S3, T-RGB ESP32-S3, and ESP32-S3 Round SPI TFT. Feel free to check these out if you are looking for a specific rounded display product.

The company also provides a detailed interface diagram of the hardware circuit board, showing essential components such as the QMI8658C IMU, W25Q16JVUXIQ 2MB NOR Flash, RT9013-33GB LDO, SH1.0 6-PIN GPIO connector, and a BOOT button.

Waveshare provides detailed outline dimensions of the boards. Additionally, they offer an extensive Wiki page for further information about the product.

The Waveshare RP2040-LCD-0.99-B rounded display can be purchased on Aliexpress for $11.90 plus shipping, Amazon US ($23.29), or the company’s online store for $16.99 excluding shipping.

The post Raspberry Pi RP2040-powered 0.99″ rounded display is housed in CNC metal case appeared first on CNX Software - Embedded Systems News.

When we first wrote about the Golioth IoT development platform with ESP32 and nRF9160 devices support in 2022, we noted they offered a free Dev Tier account for up to 50 devices, 10 MB of LightDB data with a 7-day retention policy, and other limitations.

The company has now decided to remove many of the limitations from the free developer tier without any limit to the number of IoT devices and also added other benefits:

• Unlimited Device Connections: Empowering developers to scale projects without constraints.
• Over-the-air (OTA) Device Firmware Updates (DFU) with 1GB monthly bandwidth
• 1,000,000 Monthly Log Messages up to 200MB
• Free data retention
  • Time series – 30 days
  • Logs: 14 days

The main limitations compared to paid plans are that only one project can be created and a single seat (single loading) is available, so it’s not possible to have a team of users with different permissions like in the Team or Enterprise paid plans, which makes sense for a developer account. Support is different as well since only community support is provided for the free developer tier, and you would not get a dedicated engineer from Golioth working on your deployment.

A free account may still be charged if the OTA update bandwidth and logging storage limits are exceeded with additional usage charged at $0.35/MB for OTA downloads and $0.20/MB for logging. The company also charges for data ingestion and routing with data streamed through Golioth at $1/GB to LightDB Stream and $0.40/MB for data streamed out to third-party services.

The company explains they dropped their 50 IoT devices limit due to “sleepy devices” such as soil sensors that often operate with minimal data transmission, and to grow the IoT developer community making use of the Golioth IoT platform. If you’d like to give it a try, you can do so on the Golioth website after checking out whether your IoT devices are supported.

The post Golioth expands its free tier for developers with unlimited IoT devices, OTA updates, 1GB bandwidth appeared first on CNX Software - Embedded Systems News.

Formuler Z mini is a 4K Android 12 TV dongle powered by the new RealTek RTD1325 SoC with a Mali-G57 MC1, 4K AV1 video processing unit, and support for features such as Quick Media Switching (QMS) from the HDMI 2.1 specifications.

The Android TV dongle also comes with 2GB LPDDR4 memory, 8GB eMMC flash, an HDMI 2.1 port up to 4K resolution, WiFi 5 and Bluetooth 5.0 connectivity, Widevine L1 DRM, and a Bluetooth voice remote control.

Formuler Z mini specifications:

• SoC – RealTek RTD1325
  • CPU – TBD. Likely quad-core Cortex-A55 or Cortex-A53
  • GPU – Arm Mali-G57 MC1
  • VPU – 4Kp60 10-bit AV1 video decoder
• System Memory – 2GB LPDDR4 @ 2400MHz
• Storage – 8GB eMMC flash
• Video Output – HDMI 2.1 up to 4K resolution (3840 x 2160) with 4K HDR (HDR10+, HLG), QMS support
• Networking – WiFi 5 2×2 MIMO and Bluetooth 5.0
• DRM – Widevine L1, L3
• Power Supply – 5V/1A via USB port
• Power Consumption – Standby: 0.55W; max: 5W
• Dimensions – 83 x 60 x 12.5 mm

The Formuler runs Android 12 with the company’s MYTVOnline3 media-viewing client and supports Widevine Level 1 DRM to watch videos from apps such as Netflix. While video output is limited to 4K resolution it still features an HDMI 2.1 port implementing QMS so that users can switch between different frame rates without a black screen transition.

We found very little public information about the RealTek RTD1325 processor, but Formuler highlights that its Mali-G57 GPU is about twice as powerful as competing solutions with SoCs equipped with the Mali-G31 GPU (such as Amlogic S905X4), so the user interface could be smoother or with extra animations. There are also some patchsets in the Linux kernel mailing list related to supporting peripherals for the RTD1325, so we should learn more soon enough.

The product page for the Formuler Z mini has more information about the dongle, but no price information. However, AndroidTVBox.eu says the recommended retail price will be 124 Euros.

The post RealTek RTD1325 SoC shows up in Formuler Z mini 4K Android 12 dongle appeared first on CNX Software - Embedded Systems News.

MYiR MYC-LT527 is a compact System-on-Module (SoM) based on Allwinner T527 octa-core Arm Cortex-A55 SoC with a 2 TOPS AI accelerator, up to 4GB RAM, 32GB flash, and a land grid array (LGA) comprised of 381 pads with a range of interfaces for displays and cameras, networking, USB, and PCIe, and more.

The company also introduced the MYD-LT527 development board to showcase the capabilities of the Allwinner T527 CPU module suitable for a range of applications such as industrial robots, energy and power management, medical equipment, display and controller machines, edge AI boxes and boards, automotive dashboards, and embedded devices that require media and AI functionalities.

MYC-LT527 Allwinner T527 System-on-Module

MYC-LT527 specifications:

• SoC – Allwinner T527
  • CPU
    • Octa-core Arm Cortex-A55 processor with four cores @ 1.80 GHz and four cores @ 1.42GHz
    • E906 RISC-V core up to 200 MHz
  • DSP – 600MHz HIFI4 Audio DSP
  • GPU – Arm Mali-G57 MC1 GPU
  • AI accelerator –  Optional, up to 2 TOPS NPU
• System Memory – 2GB or 4GB LPDDR4
• Storage – 16GB or 32GB eMMC flash, 32Kbit EEPROM
• 381-pad LGA package with
  • Display interfaces
    • 1x HDMI 2.0
    • 1x eDP
    • 2x LVDS with dual link or 2x RGB
    • 2x MIPI-DSI
  • Camera interfaces
    • 16-bit Parallel CSI
    • 4+4-lane, 4+2+2-lane, or 2+2+2+2-lane MIPI-CSI
  • Audio
    • 3x audio outputs
    • 3x audio inputs
    • 4x I2S/PCM
    • 1x SPDIF
  • Networking – 2x RGMII/RMII
  • USB
    • 1x USB 2.0 DRD
    • 1x USB 2.0 Host
    • 1x USB 3.1 DRD (multiplexed with PCIe 2.1)
  • PCIe – 1 x PCIe 2.1, RC mode (multiplexed with USB3.1)
  • Analog
    • 24x 12-bit GPADC
    • 2x 6-bit LRADC
    • 2x DACs and 3x ADCs
  • Other I/Os
    • 2x SDIO 3.0
    • 10x UART, 2x CAN Bus
    • 9x I2C, 30x PWM, 4x SPI
    • 2x CIR RX and 1x CIR TX
    • Up to 138x GPIO
• Power supply – 5V/3A
• Dimensions – 45 x 43mm (12-layer PCB design)
• Temperature Range – -40 to +85°C (industrial grade) or -20 to 70°C (extended temperature)

MYiR says Android 13 is ready now, a Yocto-based Linux 5.15 image with Qt will become available next month (April 2024), and Ubuntu/Debian images should be ready by May 2024. Sadly, the software is already outdated before launch with Uboot 2018.07, Linux 5.15.104 (March 2022), and the source code for the drivers is also provided to customers.

It’s not the first time we have written about the Allwinner T527 as we quickly mentioned it in our article about the Allwinner A523 tablet processor since at the time we noted it was the same design as the T527 (automotive market) and R828/MR828 (Smart Home market) processors. This also explains why Android 13 is ready before Linux since the tablet market requires Android support.

MYD-LT527 development board

MYD-LT527 specifications:

• SoM – MYC-LT527 system-on-module as described above
• Storage – MicroSD card slot
• Video Output
  • HDMI port
  • Single-channel LVDS compatible with MYIR’s MY-LVDS070C LCD module with capacitive touchscreen
  • Mini DP interface
  • MIPI DSI interface
• Video Input – 2x 24-pin MIPI CSI connectors compatible with MYIR’s MY-CAM003M camera module
• Audio – 3.5mm Headphone/Mic jack
• Networking
  • 2x Gigabit Ethernet RJ45 jacks
  • WiFi 5 and Bluetooth 5.2 module
• USB – 1x USB 3.0 host port, 2x USB 2.0 OTG ports
• Serial ports
  • 2x TTL serial ports (UART3/7 brought out from RPi expansion interface)
  • 2x TTL serial ports (UART2/4/5/6 brought out from Mi Fans Pi expansion interface)
  • 2x CAN Bus interfaces
• Expansion
  • 40-pin “RPI interface” header with GPIO/I2C/UART/SPI/CAN compatible with Raspberry Pi standard 40-pin GPIO header and  MYIR’s MY-WIREDCOM RPI Module with RS485, RS232, and CAN Bus
  • 40-pin “Mi Fans Pi interface” header with  GPIO/I2C/UART/SPI/USB/PWM
• Misc
  • 4x buttons (Reset, Wake up, FLASH, USER)
  • RTC Battery Interface
• Power Supply – 12V/3A via DC power barrel jack
• Dimensions – 120 x 90mm (6-layer design)
• Temperature Range
  • Industrial grade – -40 to +85°C or -30 to +85°C with the WiFi/BT module
  • Extended – -20 to +70°C

MYiR Tech is selling the MYC-LT527 module for $40 to $89 depending on the selected configuration and the development kit for $79 to $119. More details and purchased links can be found on the product page.

The post Allwinner T527 System-on-Module features octa-core Cortex-A55 CPU, 2 TOPS AI accelerator appeared first on CNX Software - Embedded Systems News.

ThingPulse ePulse Feather C6 is a new ESP32-C6 development board with WiFi 6, BLE5, Zigbee, Thread, and Matter connectivity that follows the Adafruit Feather form factor and supports LiPo battery charging through a charger IC and a fuel gauge.

Just like its predecessor called ePulse Feather, the ePulse Feather C6 is optimized for low power consumption but replaces an ESP32-WROVER-E-N8R8 module but an ESP32-C6-MINI-1 with a 802.15.4 radio for Zigbee, Thread, and Matter, as well as 2.4 GHz WiFi 6 and Bluetooth 5.3 LE connectivity.

ePulse Feather C6 specifications:

• Wireless module – ESP32-C6-MINI-1
  • SoC – ESP32-C6H4 32-bit RISC-V microprocessor up to 160 MHz with 320KB ROM, 512KB HP SRAM, 16KB LP SRAM, 4MB flash
  • Wireless – 2.4 GHz WiFi 6 with Target Wake Time (TWT) support, Bluetooth 5.3 LE and Mesh, 802.15.4 radio with Zigbee, Thread, and Matter
  • PCB Antenna
• USB – 1x USB Type-C port for power/charging and programming
• Expansion – 12-pin + 16-pin headers with up to 17x GPIOs, 2x USB pins, 1x fuel guage pin
• Power management
  • Supply Voltage
    • 5V via USB-C port
    • 3.3 to 6V via Vin pin
  • XC6220B33 LDO
  • LiPo battery support via 2-pin JST connector
  • TP4065 charging IC and MAX17048 fuel gauge
• Power Consumption – Deep sleep: around 18uA above 3.3V or about 33uA below 3.3V
• Dimensions – PCB: 50.8 x 24.4mm; with ESP32-C6-MINI-1 and USB-C port: 57.4 x 24.4mm (Adafruit Feather and breadboard compatible)
  
• Weight – About 12 grams
• Certifications – FCC and CE

Another change compared to the previous ePulse Feather board is that it does not include an external UART chip for programming. as ESP32-C6 handles that directly through the USB-C data lines connected to pins 12 and 13. ThingPulse does not provide any software as the board can be programmed as a standard ESP32 C6 Dev Module with the ESP-IDF or Arduino IDE and most of its differentiating features are hardware-related. You’ll find the PDF schematics, STEP file, and pinout diagram on GitHub.

ThingPulse is selling the ePulse Feather C6 board for $14.95 on its online store, but they also told us CNX Software readers can get a 15% discount with the coupon code cnx-feather-c6 bring the price down to $12.70 before shipping.

The post ePulse Feather C6 – An ESP32-C6 development board with Adafruit Feather form factor, LiPo battery support appeared first on CNX Software - Embedded Systems News.

STMicro’s new STM32WBA series, starting with the STM32WBA52, STM32WBA54, and STM32WBA55 devices, is a family of Arm Cortex-M33 wireless microcontrollers with Bluetooth LE 5.4, Zigbee, Thread, and Matter connectivity that achieved the SESIP (Security Evaluation Standard for IoT Platforms) Level 3 security certification and should make them compliant with US Cyber Trust Mark and EU Radio Equipment Directive (RED) regulations due to become mandatory in 2025.

The 100MHz STM32WBA54 and STM32WBA55 microcontrollers come with up to 1MB of flash memory, support Arm TrustZone architecture isolating secure processes and storage,  and incorporate background autonomous mode, flexible power-saving states, and analog and digital peripherals found in STMicro STM32U5 ultra-low-power MCUs.

STM32WBA54 and STM32WBA55 specifications:

• MCU core – Arm Cortex-M33 at 100MHz with FPU and DSP
• Memory – Up to 128KB SRAM
• Storage – Up to 1 MB flash
• Wireless
  • Bluetooth Low Energy 5.4
    • Long Range (LR) capable
    • Up to 2 Mbps
    • Bluetooth LE audio support
    • -96 dBm sensitivity
    • Up to 20 links for multipoint communication
    • Embedded balun + matching to lower design cost
  • 802.15.4 radio
    • Zigbee, OpenThread, Matter
    • 106dBm RF link
  • +10 dBm output power with low power consumption
  • External PA support to get ultra-wide communication distance
  • Concurrent mode
• Peripherals
  • 2x SPI, 2x I2C, up to 35x GPIOs
  • 2x USART, LPUART
  • Audio – 1-ch SAI
  • Analog – 12-bit ADC up to 2 Msps, 2x ULP comparator
  • Capacitive Touch
  • 32-bit timer, 4x 16-bit timers (1x for motor control), 2x 16-bit ULP timers
•  Security
  • Arm TrustZone
  • AES, PKA side attack resistant
  • RTC active tamper enabled
  • SESIP Level 3 target certification: compliance with the US Cyber Trust Mark and the EU Radio Equipment Directive (RED)
  • Secure radio update and stack firmware with SBSFU/SFI
  • Authenticated firmware upgrade system
• Power management
  • Supply Voltage – 1.71 V to 3.6V
  • LDO/DCDC support
  • Switched-mode power supplies (SMPS) for low-power applications (STM32WB55 only)
• Power Consumption
  • < 140 nA standby mode
  • 0.9 μA in ULP-mode with full RAM
  • Down to 1.35μA with RTC and 64KB of RAM.
  • 30 µA/MHz when running at 100MHz
  • 5.2 mA BLE Radio Tx current
• Package – 32 to 59 pins QFN/BGA packages
• Temperature Range – Up to 105°C
• Process – Built using 40nm process technology

The STM32WBA52 microcontroller is a cost-down version with a Bluetooth LE 5.4 radio only, fewer peripherals, and different power consumption as shown in the table below. Note: the STM32WBA52 was tested with a 1.8V LDO and the STM32WBA55 with a 3V SMPS which explains the much different numbers in Amperes.

STMicro STM32WBA55 is a suitable candidate for Matter border routers, while Bluetooth LE Audio and Bluetooth Auracast support means STM32WBA32 will be found in portable audio devices such as earbuds.

The STM32WBA series adds to the existing STMicro STM32 wireless microcontrollers such as the STM32WB Cortex-M4/M0+ MCUs and STM32WB09 Cortex-M0+ microcontrollers building upon the BlueNRG series as shown in the illustration below.

STMicro offers both software development tools such STM32CubeWBA with a connectivity stack, examples, and peripheral drivers,  STM32CubeMX for code generation and power calculation, STM32CubeMonRF for performance monitoring and ratio testing, and STM32CubeProg to program the application to the target, as well as STM32WBA52 and STM32WBA55 development boards with Arduino headers, such as the STM32WBA55G-DK1 Discovery kit and NUCLEO-WBA55CG Nucleo-64 board.

STMicro says samples are available now, but pricing requests must be made through the company’s sales office. The NUCLEO-WBA52CG development board is available now for $61.25 on STMicro eStore, An STM32WBA module with necessary external components including power supply and antenna-balancing circuitry should also become available in June 2024. More details may be found on the product page. Note: right now only the STM32WBA52 is shown, but details about the STM32WBA55 should be added by March 12 according to the press release.

Thanks to TLS for the tip.

The post STM32WBA microcontrollers with Bluetooth LE 5.4, Zigbee, Thread, and Matter to comply with US and EU Cybersecurity regulations appeared first on CNX Software - Embedded Systems News.

Samsung has just started sampling its 256 GB SD Express microSD card (aka microSD Express card) with a sequential read speed of up to 800 MB/s through its PCIe/NVMe interface.

The SD Express standard was first introduced in the SD specification ver 7.0 in 2018 adding pads for PCIe/NVMe interfaces in full-sized SD cards before the microSD Express standard was introduced in the SD 7.1 specification the following year. Since then, the SD association further improved the standard promising ever faster SD Express speeds up to 4GB/s and microSD Express speeds up to 2GB/s.

Saying the adoption of the new SD Express and microSD Express standards has been slow would be an understatement, and the Samsung 256GB microSD Express card looks to be the world’s first such storage device. Looking at full-size SD Express cards, I could only find one on Amazon for $59.99 with 256GB capacity, 820 MB/s read speed, and 500 MB/s write speed. It also ships with an SD Express card reader simply because I’m not aware of any devices supporting the new standard… At least those cards are backward compatible so they’ll work on any SD card/microSD card reader but at a slower speed, for example up to 104MB/s with a UHS-I card reader. ADATA also offers two models with 256GB or 512GB capacities at higher prices.

Besides the chicken and egg issue of launching a new storage standard, those microSD Express can get very hot, and Samsung claims its Dynamic Thermal Guard (DTG) technology helps maintain the optimum temperature for the SD Express microSD card even when they are used for a long period of time. Performance-wise, 800MB/s is four times faster than the best UHS-1 memory cards and 1.4 times faster than recent SATA SSDs capable of up to 560 MB/s transfer rates.

Samsung says the 256GB SD Express microSD card will be available for purchase later this year, but did not provide pricing information at this time. The company did say it was developed in collaboration with a customer so we can expect a consumer device such as a smartphone or laptop with an integrated microSD Express reader around the same time.

Via Liliputing

The post Samsung unveils 256GB microSD Express card with up to 800MB/s read speeds appeared first on CNX Software - Embedded Systems News.

Collabora has been working on the Panthor open-source GPU kernel driver for the third-generation Arm Valhall GPU (Arm Mali-G310, Mali-G510, Mali-G610, and Mali-G710) for around two years, and the code has just been merged in drm-misc meaning it should be part of the upcoming Linux 6.10 release sometime in July 2024.

Many regular readers must already be familiar with the Panfrost open-source driver for Arm Mali GPUs as we’ve covered its development progress over the years. Panthor is a new kernel driver specific to the 3rd gen Valhall GPUs that still relies on the Panfrost driver residing in userspace, as explained by Boris Brezillon from Collabora. Furthermore, the existing Gallium “Panfrost” driver in Mesa has also received a merge request adding support for those GPUs (10th gen Arm Mali = 3rd gen Arm Mali Valhall) meaning popular targets such as the Rockchip RK3588 SoC with an Arm Mali-G610 MP4 GPU will soon have 100% open-source graphics support in Linux.

The announcement on the Collobora website provides a few more details about the involvement of various persons and organizations in the project, including Arm which recently claimed: “Panfrost is now the GPU driver for the Linux community“.  For the Panthor project specifically, Arm not only provided documentation about the GPU but also technical support with two Arm engineers assigned as co-maintainers of the kernel driver.

The current Panthor driver supports OpenGL ES, but work in the Vulkan API has also started and we may get a release by the end of the year. Besides the driver itself, Collabora will also work on tools for the 3rd gen Valhall GPUs such as performance counters, support for devcoredump, and command stream tracing. Deeper technical details about the implementation of the Panthor driver can be found in an earlier post on Collabora with information about the new Command Stream Frontend and the new uAPI.

The post Panthor open-source driver for Arm Mali-G310, Mali-G510, Mali-G610, and Mali-G710 GPUs to be part of Linux 6.10 appeared first on CNX Software - Embedded Systems News.

Atreyo AG-702 is an industrial OpenWrt gateway based on the MT7628 processor with dual Ethernet and WiFi connectivity as well as a built-in LTE and GNSS modem supporting two SIM cards.

The gateway is equipped with isolated RS485 and RS232 interfaces, two digital isolated inputs, one relay output, and a USB host port that can be used to connect a flash drive or USB converters to various other interfaces. It is housed in an anodized aluminum enclosure and supports a wide 14-60V DC input range suitable for industrial settings.

Atreyo AG-702 specifications:

• SoC – MediaTek MT7628 MIPS processor at 580MHz
• System Memory – 256MB
• Storage – 32MB eMMC flash + 512MB NAND flash with ExtRoot support (overlay)
• Networking
  • 100Mbps Ethernet WAN port
  • 100Mbps Ethernet LAN port
  • LTE/GPRS with dual SIM, 1x external antenna
  • 2.4 GHz WiFi with 1x external antenna
  • GNSS with active antenna support
• USB – 1x USB 2.0 Type-A port
• Serial – Isolated RS485 and RS232
• Expansion – 2x isolated digital input, 1x relay output
• Misc – LED indicators for system, I/Os, WiFi, SIM card, LTE, signal strength
• Power supply – 14 to 60V DC via terminal block, Passive PoE support
• Dimensions – 88 x 87 x 35mm (Aluminium housing with DIN rail mounting option)
• Weight – 240 grams

The AG-702 is said to run OpenWrt 23.05 with the Atreyo Environment V1.01b and Linux kernel 5.15.71. The default firmware supports various features such as VPN (OpenVPN, WireGuard, etc..), ModBus (TCP slave, TCP master, RTU master, RTU gateway),  and remote management through a Web UI, SSH, SNMP, MQTT(s), and so on. Atreyo also provides an SDK allowing users to develop their own application(s) for the gateway. Additional information about the hardware and software can be found on the documentation website.

Atreyo told CNX Software the AG-702 industrial OpenWrt gateway is available now with pricing starting at $125 without cellular and GNSS connectivity. I can also see they have offices in India and Poland, so European companies should also be able to easily source the gateway. More details may be found on the product page.

The post Industrial OpenWrt gateway features MediaTek MT7628 SoC, WiFi, LTE, and GNSS connectivity appeared first on CNX Software - Embedded Systems News.

Japanese company Yuridenki-Shokai Co. Ltd will soon launch the Kaki Pi single board computer based on the just-announced Renesas RZ/V2H Arm microprocessor with a powerful 80 TOPS AI accelerator, with Raspberry Pi-inspired form factor and features such as the 40-pin GPIO header, the same PCIe 3.0 connector as found in the Raspberry Pi 5, and four 22-pin MIPI CSI connectors that look to be compatible with the Raspberry Pi cameras.

The board also comes with up to 8GB LPDDR4,  a microSD card for the OS, a 22-pin MIPI DSI connector for a display, a gigabit Ethernet port, two USB 3.0 ports, two CAN Bus connectors, and other interfaces that make it suitable for robotics applications such as Autonomous Mobile Robots (AMR) and HSR (Human Support Robots) as well as IoT projects.

Kaki Pi specifications:

• SoC – Renesas RZ/V2H
  • CPU/MCU cores
    • Quad-core Arm Cortex-A55 processor up to 1.8 GHz
    • Dual-core Arm Cortex-R8 real-time microcontroller up to 800 MHz
    • Arm Cortex-M33 microcontroller core up to 200 MHz for system management
  • GPU- Arm Mali-G31 GPU
  • ISP – Mali C55
  • NPU
    • DRP-AI3 up to 8 TOPS (INT8) or 80 TOPS (Sparse)
    • DRP dynamically reconfigurable processor (STP4)
• System Memory – 2GB, 4GB, or 8GB LPDDR4 memory up to 1,333MHz
• Storage – microSD card slot
• Video Output – 22-pin 4-lane MIPI DSI connector
• Video Inputs – 4x 22-pin 4-lane MIPI CSI connectors
• Networking – Gigabit Ethernet RJ45 ports
• USB – 2x USB 3.0 ports
• Expansion
  • PCIe 3.0 x1 FFC connector compatible with the one in Raspberry Pi 5
  • 40-pin GPIO header (mostly) compatible with the Raspberry Pi header with 26x GPIO, UART/ I2C, SPI, I2S, PWM, 5V, 3.3V, and GND.  Note: GPIOs work at 3.3V, output up to 16mA
  • 2x CAN bus connector
• Misc – RTC, fan header, Reset button, DIP switch for configuration
• Power Supply
  • 7.5V to 24V DC input via 5.5/2.1mm DC jack; 12V/2.4A power supply provided
  • Via USB-C port (no info about voltages)
  • Optional PoE with addon connected to 4-pin header
• Dimensions – 85 x 56 x 19 mm (similar to Raspberry Pi Model B)

Yuridenki will provide a Yocto Linux SDK for the Kaki Pi SBC, the company also mentions support for the ROS2 robotic operating system and an unnamed RTOS, and the two AI accelerators in the RZ/V2H microprocessor can help offload OpenCV computer vision workloads.

According to an article on PR Times (Japanese), the Kaki Pi will be launched at the end of April, but I could not find pricing information and whether it will be available outside of Japan at this time. Having said that, I’m confident it will be much cheaper than the Renesas RZ/V2H-EVK evaluation kit that goes for close to $1,200 US. Further information may be found on the official website.

Thanks to Will Whang for the tip.

The post Kaki Pi is a Raspberry Pi-inspired Renesas RZ/V2H AI SBC with four camera connectors, a PCIe 3.0 interface appeared first on CNX Software - Embedded Systems News.

In today’s hyper-connected world, reliable and high-speed wireless connectivity is no longer a luxury, but a necessity. WiFi standards have evolved rapidly. WiFi 6, which has been around for only a few years, is now being replaced by WiFi 7 or IEEE 802.11be. Equipment manufacturers are now rushing to phase in WiFi 7 products.

Compex has released its latest WiFi 7 module WLE7002E25 to help shorten time to market. The Compex WLE7002E25 is a dual-band concurrent 2×2 2.4GHz and 2×2 5GHz radio based on Qualcomm’s Waikiki high-performance WiFi 7 chipset which is capable of over 10Gbps raw data rate. It is in a standard mini PCIe form factor.

WiFi 7 in a Standard Mini PCIe Form Factor

The standard mini PCIe footprint means that WLE7002E25 is readily compatible with existing systems and designs, particularly customers whose devices have been using WiFi 4 or WiFi 5 standard mini PCIe modules. This simplifies integration since device manufacturers do not need to make modifications to their hardware to support a new generation WiFi module, thus reducing development costs and accelerating time-to-market.

The diagram below shows the dimensions differences between Qualcomm’s high-performance Waikiki M.2 E-key reference design and Compex’s standard-size mini PCIe module using the same Waikiki chipset.

Lower Power Consumption, Impressive Transmit Power

Compex’s RF engineers set themselves the twin targets of condensing the Qualcomm reference design by about 50% to an industrial standard form factor, as well as reducing its power consumption significantly. Compared with the high-performance Qualcomm reference design, Compex managed to achieve the size and consumption targets with a little loss in performance. The following is a comparison of the RF power outputs and consumption between the Compex WLE7002E25 mini PCIe module and the Qualcomm WK03.2 reference board.

	2.4GHz Max. Transmit Power (dBm)	5GHz Max. Transmit Power (dBm)	Max. Power Consumption (W)
Compex WLE7002E25	22	21	8.0
Qualcomm WK03.2	23	21	13.3

Multi-Link Operation (MLO)

An important differentiator between WiFi 7 and WiFi 6 is MLO (Multi Link Operation). This allows a client to connect to the AP through multiple bands simultaneously. This has a significant advantage in reducing latency. Latency is of critical importance to interactive applications such as games, robotics, real-time translation, teleconferencing, etc…

The Compex WLE7002E25 supports MLO between its 2.4GHz and 5GHz bands. It also implements MLO signals between adapters to support MLO between the WLE7002E25 and other WiFi 7 modules such as the Compex WLE7000E6 4×4 6GHz module.

Open-source Ath12k driver support

Compex Systems recognize the importance of supporting customers’ product integration with the open-source ath12k driver. Ath12k integration support for non-Qualcomm platforms such as x86 embedded boards is also available.

Competitive Pricing

The Compex WLE7002E25 is only marginally more expensive than WiFi 6 products of similar configurations. This is to encourage direct migration from WiFi 4 and WiFi 5 to WiFi 7, skipping the WiFi 6 development time and costs.

Investing in the Future

The Compex WLE7002E25 represents more than just a WiFi 7 mini PCIe module; it’s an investment in the future. With its cutting-edge technology, comprehensive features, and diverse application potential, the Compex WLE7002E25 stands poised to rewrite the rules of wireless connectivity. Contact Compex here.

The post Compex WLE7002E25 is a dual-band dual-concurrent WiFi 7 module in standard mini PCIe size (Sponsored) appeared first on CNX Software - Embedded Systems News.

The RZ/V2H is the latest addition to the Reneasas family of 32-bit and 64-bit microprocessors. Renesas RZ/V2H microprocessor features four Arm Cortex-A55 application processing cores, dual Arm Cortex-R8 real-time processing cores, and one Cortex-M33 core for system management, as well as an AI accelerator with up to 80 TOPS (sparse) of performance.

The RZ/V2H MPU includes the third generation of Renesas’ dynamically reconfigurable processor for AI (DRP-AI3) accelerator, a staple of the RZ/V series. The DRP-AI accelerator can run complex image AI models at a power efficiency of 10 TOPS per watt (TOPS/W), as much as 10 times higher than conventional microprocessors. Additionally, the RZ/V2H has another DRP that can be used in image processing and robotics applications.

The board also comes with PCIe Gen3, USB 3.2, and Gigabit Ethernet for high-speed applications such as autonomous robotics and factory automation machine vision. It offers a performance boost over previous products such as the RZ/G2L MPU, which allows for processing more data in real-time and reduces the need for cloud computing platforms.

Renesas RZ/V2H specifications:

• CPU
  • 4x 64-bit Arm Cortex-A55 cores for application processing (up to 1.8 GHz)
  • 2x 32-bit Arm Cortex-R8 cores for real-time processing (up to 800 MHz)
  • 1x 32-bit Arm Cortex-M33 core for system management (up to 200 MHz)
• Memory
  • 6MB on-chip SRAM with ECC
  • External DDR memory interface
  • 2-channel memory controller for LPDDR4-3200 or LPDDR4X-3200 with a 32-bit bus width
  • xSPI interface
  • SDHI (eMMC/SD)
• AI accelerator engines
  • AI accelerator (dynamically reconfigurable processor for AI (DRP-AI3) – 8 TOPS/dense, 80 TOPS/sparse
  • Dynamically reconfigurable processor (DRP)
• Video & Graphics
  • GPU – Arm Mali-G31 (optional)
  • Mali C-55 image signal processor (ISP) (optional)
  • Image scaling unit (ISU)
  • Video codec unit (VCD)
• Display – MIPI DSI
• Camera – MIPI CSI-2
• Network – 2x Gigabit Ethernet
• USB – 1x USB 2.0 (Host, OTG), 1x USB 3.2 Gen2 (2ch: Host-only)
• PCIe – PCIe Gen3
• Other Peripherals
  • 6x CAN/CAN FD (compliant with ISO11898-1)
  • 3x SPI, 9x I2C, 1x I3C
  • 8x 12-bit ADC
• Supply Voltage –  3.135V – 3.465V
• Package – BGA, 19mm x 19mm, 1368-pin, 0.5mm pitch
• Temperature Range – -40 to 125°C

Renesas has partnered with SEGGER to provide support for the RZ/V2H via the J-link debug probe. The company has also released a library of pre-trained models and an AI Linux software development kit (SDK) to aid rapid development. There is an evaluation board, RZ/V2H-EVK, that can be used to test AI applications early in the design phase.

The V2H microprocessor and its evaluation kit are available now, with the latter priced at $1,176. Pricing for the RZ/V2H can be found on the product page. You can request free samples on the website and learn more about the microprocessor and its development tools by checking out the product page and press release.

Thanks to TLS for the tip.

The post Renesas RZ/V2H Cortex-A55/R8/M33 MPU comes with 80 TOPS AI accelerator for robotics and autonomous applications appeared first on CNX Software - Embedded Systems News.

SB Components’ 2×2 Quad Display Board is an MCU development board fitted with either a Raspberry Pi Pico W board or an ESP32-S3-WROOM-1 module used to drive four small color displays in square or round shapes.

The board specifically features either four 1.54-inch square TFT displays or four 1.28-inch round displays, a microSD card, an RTC with coin-cell battery holder, and a USB-C port for power and programming, plus a few buttons. It may feel like it’s coming out of the but-why-because-we-can department, but the company expects it to be used for signage, interactive displays, art projects, portable devices, data loggers, education, and more.

2×2 Quad Display Board specifications:

• Main control (one or the other)
  • Raspberry Pi Pico W
    • MCU – Raspberry Pi RP2040 dual-core Cortex-M0+ microcontroller @ 133 MHz with 264KB SRAM
    • Storage – 2MB QSPI flash
    • Wireless – WiFi 4 and Bluetooth LE 5.2
    • USB – 1x Micro USB 1.1 port used for power and programming
  • ESP32-S3-WROOM-1
    • MCU – ESP32-S3 dual-core Tensilica LX7 up to 240 MHz with 512KB SRAM, up to 8MB PSRAM
    • Storage – TBD
    • Wireless – WiFi 4 and Bluetooth LE 5
• Storage – MicroSD card
• Display configurations (one or the other)
  • 4x 1.28-inch round displays, 240×240 resolution, 65K/262K colors, GC9A01 display driver
  • 4x 1.54-inch square TFT displays, 240×240 resolution, 65K/262K colors, ST7789 display driver
• USB
  • Pico W – 1x micro USB 1.1 port on Raspberry Pi Pico board
  • ESP32-C3 – 1x USB-C 1.1 OTG port on carrier board
• Sensor – BME280 temperature, pressure, and humidity sensor
• Expansion – 4-pin GPIO connector
• Misc
  • 4x user button
  • DS3231 RTC + coin-cell battery holder
  • Buzzer
• Power Supply
  • 5V DC via USB port
  • LiPo Battery support via connector and charging circuit
• Dimensions – TBD

It seems to be a variant of the company’s Dual Roundy and Dual Squary displays introduced last January, but they upped the ante with four displays on the same side of the board.  Some may claim you could just achieve the same result with a larger display, although I reckon the 2×2 Quad Display Board may look neater than the alternative once it’s housed in an enclosure.

As usual, SB Components will not provide any code samples or deep technical details while the crowdfunding is in progress, and they usually release the software part before shipping the rewards, so MicroPython, CircuitPython, and Arduino samples should become available in due time. In the meantime, we can have a look at some of the demos in the video embedded below.

SB Components has launched the 2×2 Quad Display boards on Kickstarter with a 500 GBP target (620 USD). Rewards start at $57 US for the ESP32-S3 variants of the 2×2 Quad Display board, and $63 for the Raspberry Pi Pico models. Shipping adds 7 GBP to the UK, and about $19 to the rest of the world. As a comparison, an ESP32 board with a 3.5-inch touchscreen display like the Elecrow ESP32 3.5-inch HMI display costs around $20 plus shipping bearing in mind it lacks some of the features of the SB Components board. The 2×2 Quad Display boards are schedule to ship in June 2023 if everything goes according to plans…

The post 2×2 Quad Display Board uses Raspberry Pi Pico W or ESP32-S3-WROOM-1 module to drive four displays (Crowdfunding) appeared first on CNX Software - Embedded Systems News.

ADLINK MXA-200 is an NXP i.MX8M Plus-based 5G IIoT gateway running Debian or a Yocto-based Linux distribution with two RS-232/422/485 isolated serial ports, two gigabit Ethernet ports, two USB 3.0 port, and an operating temperature range of -20 to 70°C.

The fanless gateway also features two M.2 slots for integrating Wi-Fi and/or 4G LTE or 5G modules, and ADLINK expected the MXA-200 to find its way into renewable energy applications, EV chargers, smart city and smart factory applications that require big data collection, cloud-based applications, and video related monitoring solutions.

ADLINK MXA-200 specifications:

• SoC – NXP i.MX8M Plus
  • CPU – Quad-core Arm Cortex-A53 processor @ 1.6/1.8GHz
  • MCU – Arm Cortex-M7 real-time core @ 400 MHz
  • GPU – Vivante GC520L 2D GPU, Vivante GC7000UL 3D GPU
  • VPU – 1080p60 hardware decoder (HEVC, H.264, VP9, and VP8) and encode (H.265/H.264)
  • AI accelerator – 2.3 TOPS NPU
• System Memory – 2GB or 4GB LPDDR4 @ 4,000 MT/s
• Storage – 32GB or 64GB eMMC flash and microSD card slot
• Video Output – HDMI output up to 4K @ 30 Hz or 1080p60
• Networking
  • 2x Gigabit Ethernet RJ45 ports
  • M.2 2230 E-key socket for WiFi/BT plus two external antennas
  • M.2 2242 B-key (USB 3.0) socket for 5G or 4G/LTE plus two (huge) external antennas
• USB – 2x USB 3.0 host ports
• Serial – 2x isolated RS-232/422/485 4-wire
• Security – TPM 2.0
• Misc – RTC, Watchdog timer
• Power Supply – 12 to 24VDC
• Dimensions – 131 x 110.5 x 40 mm
• Temperature Range
  • Operating – -20°C to 70°C
  • Storage – -25°C to 85°C
• Vibration resistance
  • IEC/EN 61131-2 compliant
  • 5 to 9 Hz Single amplitude 3.5 mm (0.14 in)
  • 9 to 150 Hz Fixed acceleration: 9.8 m/s2
  • X, Y, Z directions for 10 cycles (approximately 100 minutes)
• Shock resistance
  • IEC/EN 61131-2 compliant
  • 147 m/s2, X, Y, Z directions for 3 times
• Electrical fast transient/burst immunity
  • IEC 61000-4-4
  • 2 kV: Power port,1 kV: Signal ports
• ESD immunity – IEC/EN 61000-4-2 Level 3; Contact discharge method: 6 kV, Air discharge method: 8 kV
• Certifications
  • UL – cULus 61010 for IT equipment
  • EMC – CE / FCC class A/ UKCA / RCM / KC / EAC

The ADLINK MXA-200 Arm-based 5G IIoT gateway ships with either Debian 11 or Yocto Linux, and Windows 10 IoT will also be supported. It’s also compatible with the EdgeGO device management platform for remote monitoring and configuration. The gateway supports mounting options such as DIN rail and wall mounting to facilitate installation into machines and control cabinets

ADLINK did not provide availability and pricing information for the MXA-200. You may find additional details on the product page and in the press release.

The post ADLINK MXA-200 5G IIoT gateway ships with Debian or Yocto Linux appeared first on CNX Software - Embedded Systems News.

TwoTrees SK1 is an FDM 3D printer with a working area of 256 x 256 x 256 mm and a CoreXY system that can print 3 to 5 times faster than a typical 3D printer. The movements of a CoreXY 3D printer are achieved by moving two motors simultaneously. The TwoTrees SK1 CoreXY 3D printer can print at a maximum speed of 700 mm/s and features automatic Z-Tilt base leveling with a 36-point mesh bed. The print head uses a dual-gear direct extruder to feed the filament directly to the nozzle and the nozzle supports up to 300 degrees Celsius. The printer also comes preloaded with the popular Klipper firmware.

TwoTrees SK1 specifications

• Printing size – 256 x 256 x 256 mm
• Dimensions – 400 x 400 x 530 mm
• Printing speed – Up to 700mm/s
• Acceleration – Up to 20,000 mm/s2
• Nozzle temperature – Up to 300ºC
• Guide rail type – Bail screw guide rail
• Display – 4.3-inch touchscreen display
• Leveling – Z-Tilt included leveling
• Hotbed panel – PEI
• Optional support for Webcam (“AI camera”)
• Optional support for WiFi
• Klipper Firmware

TwoTress SK1 3D printer unboxing and assembly

The 3D printer ships with a range of accessories including tools such as a hex screwdriver, cutting pliers, a brass brush, a microSD card, filament, screw nuts, and so on.

Slide Rail and Belt System

The SK1 relies on a linear guide rail for every axis (X/Y/Z) that supports high-speed movement. And has been used in this 3D printer. The X and Y axis each feature a belt and two stepping motors to move quickly and accurately in their respective direction.

The Z-axis is controlled by three lead screws and motors.

TwoTress SK1 3D printer control screen

A 4.3-inch color touchscreen display is attached to the top of the printer.

Extruder

A dual-gear direct extruder system feeds the filament directly to the 0.4mm nozzle which can be heated up to 300 degrees. One fan is used to cool the workpiece, and there’s also another high-powered ventilation fan.

Inductive sensor on the SK1 3D printer

An inductive sensor is used to check the distance between the print head and the hotbed for printing accuracy and also prevents the print head from colliding with the plate.

Unlocking the axis

In order to prevent damage during transportation, each axis of the 3D printer will be locked. So we need to remove the locks before using the machine. We show how to do it in the video embedded at the end of the review.

110V/230V voltage selection

Before turning on the machine for the first time, make sure the correct voltage is selected. There is a switch to adjust to 110V and 220V on the left side of the SK1 3D printer.

If the voltage is not set correctly, this happens (110V set when the mains are actually 230V):

The power supply blew up but it could be replaced by purchasing a MEANWELL LRS-350-24 power supply available on Amazon, Aliexpress, and other shops for under $30.

WiFi antenna installation

The WiFi antenna can easily be attached to the SMA connector on the side of the machine using the provided hex key.

Printing plate

A PEI magnetic sheet will make sure the workpiece sticks firmly in place and does not fall off during printing. It can be used with a range of plastic materials and is suitable for printing with small surface areas or workpieces that require a smooth bottom surface. Once the print is done, the workpiece will come off easily without having to use tools to take it out.

TwoTrees 3D printer filament

Our kit came with a 1kg HS-PLA filament spool from TwoTrees that is supposed to adhere better to the hotbed than standard PLA. The temperature used for printing is between 200 to 230°C and the diameter of the filament is 1.75 mm.

The filament spool can be installed in the steel hanger installed on the side of the SK1 3D printer and secured with two screws.

Preparing the TwoTrees SK1 CoreXY 3D printer

Ports

The mainboard comes with four ports as follows:

1. Ethernet EJ45 jack
2. USB 2.0 port is used to connect a flash Drive or a PC to send data to the printer
3. USB 3.0 port is used to connect a flash Drive or a PC to send data to the printer.
4. RJ11 connector to connect an additional display

TwoTress SK1 user interface

The Home page displays the temperature of the nozzle and hotbed, as well as the WiFi connection status.

It can also be used to control the LED strip included with the printer…

… by tapping the light icon on the bottom left (White: off; yellow: on).

Tapping the nozzle icon allows the user to set the temperature of the nozzle by entering the required value between 0 and 300°C.

We can do the same for the heated bed between 0 and 100°C.

The second item in the bottom bar enters the Move/Temp mode allowing the user to manually control the motor X/Y/Z and set the default movement distance increment to be 1mm, 10mm, or 30mm. It’s also possible to adjust the temperature and/or stop all motors in this window.

The fan mode provides options to independently control the model, auxiliary, and case fans from 0% (off) to 100% (full speed).

Tapping the gear icon in the bottom bar enters the System menu to set screen brightness, automatically off the display after a period of time, select the language, and perform a self-test.

This menu can also be used to configure the network (WiFi or Ethernet) to access the 3D printer via the web-based user interface provided by Klipper.

Leveling the bed of the SK1 CoreXY 3D printer

The same menu can be used to level the printing bed as the SK1 3D printer has a function for automatically calculating the tilt of the print bed by probing 36 points on the bed and ensuring the plate as perfectly horizontal as possible even though the chassis of the printer may not be. The calibration step is showcased in the video at the end of this article.

Printing head height calibration

The height of the printing head of the SK1 3D printer also needs to be calibrated. We can do this manually by adjusting the height in the user interface and once we get the correct value we can tap on the OK button.

How do we determine the correct Z value? By using an A4 sheet of paper as we did in our Creality Ender-3 S1 Pro 3D printer review. We must adjust the height of the nozzle by making sure it touches the paper but still allows us to move it with some friction.

Loading/unloading the filament

We need to insert the filament into the nylon tube that comes with the 3D printer and push it in as far as it will go.

Once this is done, we can go to the Load/Unload section of the user interface and tap on the “Load” button with the desired temperature. The 3D printer will then handle this for us automatically.

Using the PuraSlicer program with the TwoTress SK1 3D printer

PuraSlider installation and SK1 configuration

3D models must be converted into G-Code before printing. This is done by a slicer program, and we will use the PrusaSlicer program (version provided by TwoTrees) in this review.

The first thing to do is to import a configuration file for the 3D printer through File >>> Import >>> Import Config Bundle in the top menu.

Select the file named PrusaSlicer_config_bundle.ini then click “Open”.  You’ll find the file on the microSD card provided with the 3D printer or on MediaFire.

The SK1 3D printer is successfully configured in the PrusaSlicer program.

Converting 3D files to G-Code

We can now import a 3D model file by clicking on the Add icon as shown in the screenshot below.

Select the 3D model file with extension STL, 3MF, STEP, OBJ, AMF, or PRUSA, then click on “Open”. We selected “EiffelTower.stl” for our test.

PrusaSlicer will then display a 3D representation of the model on the printing plate.

We can set the print resolution to be 0.08, 0.12, 0.16, 0.20, 0.24, or 0.28mm which will impact the quality of the sample and printing time.

Once this is done, we can click on “Slice now” to begin processing the file.

The program will also calculate the printing time and amount of material required. We can then download the G-Code file to a computer or put it in a USB flash Drive by clicking on “Export G-Code” to complete the file conversion process.

Klipper firmware web interface (Fluidd)

We can log in to the Klipper dashboard (Fluidd) after connecting the 3D printer to a WiFi network and using the IP address shown on the display.

Uploading G-Code to the SK1 3D printer over WiFi (or Ethernet).

Let’s go to the Jobs section by clicking on the relevant icon on the left banner and then clicking on the + sign to add a new job.

Select the .gcode file we have just exported (EiffelTower.gcode), then click on the “Upload” button.

The file will be uploaded to the 3D printer, and after that, you can select it and print the workpiece.

Selecting a G-Code file from a microSD card or USB drive

Alternatively, the TwoTress SK1 3D printer can also print 3D parts from a microSD card or a USB flash drive through the touchscreen display by going to the USB mode of the user interface.

Printing samples with the TwoTrees SK1 CoreXY 3D printer

Eiffel Tower model

Material: PLA; layer: 0.28 mm; infill: 15%; speed: 200 mm/s; travel: 500 mm/s

Print time: 3 hours 10 minutes

Calibration cube with different resolutions

An XYZ 20mm calibration cube was printed with the following parameters – material: PLA; layer: 0.08 to 0.28 mm; infill: 20%; speed: 200 mm/s; travel: 700 mm/s

A visual inspection does not reveal any obvious differences between different layer thicknesses.

Benchy model printed at maximum speed

We also printed a Benchy at maximum speed (700mm/s).

Material: PLA; layer: 0.20 mm; infill: 25%; speed: 700 mm/s; travel: 700 mm/s

Print time: 18 minutes.

The resulting print is not too bad and much faster than Karl’s speed printing experiment with the Longer LK5 Pro 3D printer where he could print the Benchy in 37 minutes after some optimization and some “cheating” (0.3mm thickness), while his 0.2mm print would take well over 45 minutes. So CoreXY 3D printers clearly have the upper hand when it comes to speed printing without sacrificing the quality too much if at all.

Flexible dragon model

Material: PLA; layer: 0.28 mm; infill: 15%; speed: 300 mm/s; travel: 700 mm/s

Print time: 3 hours 32 minutes.

Lion model

Material: PLA; layer: 0.20 mm; infill: 15%; speed: 400 mm/s; travel: 700 mm/s

Print time: 2 hours 11 minutes

Video review of the TwoTrees SK1 CoreXY 3D printer

Conclusion

The TwoTrees SK1 3D printer works well and enables impressive speeds thanks to its CoreXY design that enables speeds up to 700mms that are several times faster than traditional FDM 3D printer, but for instance, a Benchy could be printed in 18 minutes using a 0.2mm layer.

The Z-Tilt auto-calibration feature is convenient to use and lowers the risk of issues during printing. The SK1 3D printer has a good build quality with a solid structure that supports the movement of the X/Y axis. The Klipper firmware also helps with delivering good quality prints and a familiar user interface enabling the upload of G-Code files over WiFi. The 3D printer is suitable for small businesses, enthusiasts, or even the classroom.

We would like to thank TwoTrees for sending us the SK1 CoreXY 3D printer for review. The 3D printer can be purchased for $499 on TwoTrees’ online store, on Aliexpress, and they also have an Amazon store, but the printer is not listed there just yet.

CNXSoft: This article is a translation – with some additional insights – of the review on CNX Software Thailand by Kajornsak Janjam, edited by Suthinee Kerdkaew.

The post TwoTrees SK1 review – A CoreXY 3D printer with high printing speeds appeared first on CNX Software - Embedded Systems News.

After going through an unboxing and teardown of the Beelink SEi12 i7-12650H mini PC, and testing the 12th gen Intel Core i7-12650H mini PC in Windows 11 Pro, we’ll now report our experience with Ubuntu 22.04 in the third part of the review to show how the Beelink SEi12 performs in Linux.

We will perform feature testing, run some benchmarks to evaluate performance, test Ethernet, WiFi 6, and Bluetooth, submit the system to a stress test to check its cooling performance and measure fan noise and power consumption. We’ll also compare it to the GEEKOM Mini IT12 mini PC with the same Intel Core i7-12650H CPU.

Ubuntu 22.04 installation

We shrank the Windows partition is about half to install Ubuntu 22.04 in dual boot configuration along with Windows 11 Pro.

We then installed Ubuntu 22.04.3 ISO from a USB flash without any issues.

Ubuntu 22.04 system information

After a system update, going to Settings->About shows we have Ubuntu 22.04.4 running on an AZW SEi computer equipped with a 12th Gen Intel Core i7-12650H processor with 16 threads, 32GB RAM, and 626.8 GB of storage (the preinstalled 500GB M.2 SSD plus our own 128GB SATA drive).

We can find more details in a terminal:

aey@SEi12-cnx:~$ cat /etc/lsb-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=22.04
DISTRIB_CODENAME=jammy
DISTRIB_DESCRIPTION="Ubuntu 22.04.4 LTS"
aey@SEi12-cnx:~$ uname -a
Linux SEi12-cnx 6.5.0-21-generic #21~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Fri Feb  9 13:32:52 UTC 2 x86_64 x86_64 x86_64 GNU/Linux
aey@SEi12-cnx:~$ free -mh
               total        used        free      shared  buff/cache   available
Mem:            31Gi       1.6Gi        27Gi       539Mi       2.2Gi        28Gi
Swap:          2.0Gi          0B       2.0Gi
aey@SEi12-cnx:~$ df -mh
Filesystem      Size  Used Avail Use% Mounted on
tmpfs           3.2G  2.4M  3.2G   1% /run
/dev/nvme0n1p5  227G   15G  202G   7% /
tmpfs            16G   46M   16G   1% /dev/shm
tmpfs           5.0M  4.0K  5.0M   1% /run/lock
efivarfs        192K  110K   78K  59% /sys/firmware/efi/efivars
/dev/nvme0n1p1   96M   33M   64M  34% /boot/efi
tmpfs           3.2G  128K  3.2G   1% /run/user/1000

The inxi command provides more details:

aey@SEi12-cnx:~$ inxi -Fc0
System:
  Host: SEi12-cnx Kernel: 6.5.0-21-generic x86_64 bits: 64
    Desktop: GNOME 42.9 Distro: Ubuntu 22.04.4 LTS (Jammy Jellyfish)
Machine:
  Type: Desktop Mobo: AZW model: SEi serial: <superuser required>
    UEFI: American Megatrends LLC. v: ALDER112 date: 06/09/2023
CPU:
  Info: 10-core (6-mt/4-st) model: 12th Gen Intel Core i7-12650H bits: 64
    type: MST AMCP cache: L2: 9.5 MiB
  Speed (MHz): avg: 445 min/max: 400/4600:4700:3500 cores: 1: 400 2: 400
    3: 657 4: 400 5: 534 6: 400 7: 400 8: 400 9: 400 10: 521 11: 400 12: 609
    13: 400 14: 400 15: 400 16: 400
Graphics:
  Device-1: Intel Alder Lake-P GT1 [UHD Graphics] driver: i915 v: kernel
  Display: wayland server: X.Org v: 1.22.1.1 with: Xwayland v: 22.1.1
    compositor: gnome-shell driver: gpu: i915 resolution: 1600x900~60Hz
  OpenGL: renderer: Mesa Intel Graphics (ADL GT2)
    v: 4.6 Mesa 23.2.1-1ubuntu3.1~22.04.2
Audio:
  Device-1: Intel Alder Lake PCH-P High Definition Audio
    driver: snd_hda_intel
  Sound Server-1: ALSA v: k6.5.0-21-generic running: yes
  Sound Server-2: PulseAudio v: 15.99.1 running: yes
  Sound Server-3: PipeWire v: 0.3.48 running: yes
Network:
  Device-1: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet
    driver: r8169
  IF: enp3s0 state: down mac: 7c:83:34:bf:01:c2
  Device-2: Intel Wi-Fi 6 AX200 driver: iwlwifi
  IF: wlp4s0 state: up mac: 70:d8:c2:14:a4:4c
Bluetooth:
  Device-1: Intel AX200 Bluetooth type: USB driver: btusb
  Report: hciconfig ID: hci0 state: up address: 70:D8:C2:14:A4:50 bt-v: 3.0
Drives:
  Local Storage: total: 583.76 GiB used: 13.39 GiB (2.3%)
  ID-1: /dev/nvme0n1 vendor: Crucial model: CT500P3PSSD8 size: 465.76 GiB
  ID-2: /dev/sda model: CJ225128TC size: 118 GiB
Partition:
  ID-1: / size: 226.75 GiB used: 13.36 GiB (5.9%) fs: ext4
    dev: /dev/nvme0n1p5
  ID-2: /boot/efi size: 96 MiB used: 32.6 MiB (33.9%) fs: vfat
    dev: /dev/nvme0n1p1
Swap:
  ID-1: swap-1 type: file size: 2 GiB used: 0 KiB (0.0%) file: /swapfile
Sensors:
  System Temperatures: cpu: 43.0 C mobo: N/A
  Fan Speeds (RPM): N/A
Info:
  Processes: 360 Uptime: 20m Memory: 31.11 GiB used: 2.94 GiB (9.5%)
  Shell: Bash inxi: 3.3.13

We have an Intel Core i7-12650H processor with 10-core (6-mt/4-st) and 16 threads closed up to 4700 MHz, a Crucial CT500P3PSSD8 with size 465.76 GB capacity, a CJ225128TC SATA SSD that we installed for testing, and an AX200 module for WiFi and Bluetooth. The CPU temperature is reported to be 43°C at idle which seems about right, but we’ll test that again.

Ubuntu 22.04 benchmarks on Beelink SEi12 i7-12650H mini PC

Let’s start with sbc-bench.sh:

aey@SEi12-cnx:~/Downloads/sbc-bench-master$ sudo ./sbc-bench.sh -r
Starting to examine hardware/software for review purposes...

sbc-bench v0.9.64

Installing needed tools: apt-get -f -qq -y install gcc make build-essential powercap-utils curl git sysstat links mmc-utils smartmontools stress-ng p7zip, tinymembench, ramlat, mhz, cpufetch, cpuminer. Done.
Checking cpufreq OPP. Done.
Executing tinymembench. Done.
Executing RAM latency tester. Done.
Executing OpenSSL benchmark. Done.
Executing 7-zip benchmark. Done.
Throttling test: heating up the device, 5 more minutes to wait. Done.
Checking cpufreq OPP again. Done (14 minutes elapsed).

Results validation:

  * Measured clockspeed not lower than advertised max CPU clockspeed
  * No swapping
  * Background activity (%system) OK
  * Powercap detected. Details: "sudo powercap-info -p intel-rapl" -> https://tinyurl.com/4jh9nevj


Full results uploaded to http://sprunge.us/KNs4cT




# AZW SEi  / i7-12650H

Tested with sbc-bench v0.9.64 on Wed, 28 Feb 2024 15:41:49 +0700. Full info: [http://sprunge.us/KNs4cT](http://sprunge.us/KNs4cT)

### General information:

    Information courtesy of cpufetch:
    
    Name:                12th Gen Intel(R) Core(TM) i7-12650H
    Microarchitecture:   Alder Lake
    Technology:          10nm
    P-cores:
      Max Frequency:     4.700 GHz
      Cores:             6 cores (12 threads)
      AVX:               AVX,AVX2
      FMA:               FMA3
      L1i Size:          32KB (192KB Total)
      L1d Size:          48KB (288KB Total)
      L2 Size:           1.25MB (7.5MB Total)
    E-cores:
      Max Frequency:     3.500 GHz
      Cores:             4 cores
      AVX:               AVX,AVX2
      FMA:               FMA3
      L1i Size:          64KB (256KB Total)
      L1d Size:          32KB (128KB Total)
      L2 Size:           2MB
    L3 Size:             24MB
    Peak Performance:    1.35 TFLOP/s
    
The CPU features 2 clusters of different core types:

    i7-12650H, Kernel: x86_64, Userland: amd64
    
    CPU sysfs topology (clusters, cpufreq members, clockspeeds)
                     cpufreq   min    max
     CPU    cluster  policy   speed  speed   core type
      0        0        0      400    4600   Golden Cove
      1        0        1      400    4600   Golden Cove
      2        0        2      400    4600   Golden Cove
      3        0        3      400    4600   Golden Cove
      4        0        4      400    4700   Golden Cove
      5        0        5      400    4700   Golden Cove
      6        0        6      400    4700   Golden Cove
      7        0        7      400    4700   Golden Cove
      8        0        8      400    4600   Golden Cove
      9        0        9      400    4600   Golden Cove
     10        0       10      400    4600   Golden Cove
     11        0       11      400    4600   Golden Cove
     12        0       12      400    3500   Gracemont
     13        0       13      400    3500   Gracemont
     14        0       14      400    3500   Gracemont
     15        0       15      400    3500   Gracemont

31855 KB available RAM

### Clockspeeds (idle vs. heated up):

Before at 51.0°C:

    cpu0-cpu11 (Golden Cove): OPP: 4600, Measured: 4585 
    cpu12-cpu15 (Gracemont): OPP: 3500, Measured: 3488 

After at 84.0°C:

    cpu0-cpu11 (Golden Cove): OPP: 4600, Measured: 4587 
    cpu12-cpu15 (Gracemont): OPP: 3500, Measured: 3490 

### Performance baseline

  * cpu0 (Golden Cove): memcpy: 22040.2 MB/s, memchr: 34825.5 MB/s, memset: 27417.4 MB/s
  * cpu12 (Gracemont): memcpy: 8546.0 MB/s, memchr: 19112.5 MB/s, memset: 14312.9 MB/s
  * cpu0 (Golden Cove) 16M latency: 29.04 19.67 21.91 19.10 20.29 19.43 19.30 20.31 
  * cpu12 (Gracemont) 16M latency: 41.20 32.55 37.93 33.07 34.33 30.04 30.17 36.84 
  * cpu0 (Golden Cove) 128M latency: 104.9 104.3 105.0 100.9 102.9 103.7 90.12 95.98 
  * cpu12 (Gracemont) 128M latency: 126.8 121.7 126.9 123.3 128.1 123.0 122.3 122.5 
  * 7-zip MIPS (3 consecutive runs): 45080, 41688, 41058 (42610 avg), single-threaded: 5420
  * `aes-256-cbc    1275788.40k  1575155.20k  1603204.35k  1607265.62k  1606284.63k  1587265.54k (Golden Cove)`
  * `aes-256-cbc     973922.28k  1213948.46k  1254906.45k  1265038.68k  1255445.85k  1265090.56k (Gracemont)`

### PCIe and storage devices:

  * Realtek RTL8111/8168/8211/8411 PCI Express Gigabit Ethernet: Speed 2.5GT/s (ok), Width x1 (ok), driver in use: r8169
  * Intel Wi-Fi 6 AX200: Speed 5GT/s (ok), Width x1 (ok), driver in use: iwlwifi
  * 465.8GB "Crucial CT500P3PSSD8" SSD as /dev/nvme0: Speed 16GT/s (ok), Width x4 (ok), 0% worn out, 213 error log entries, unhealthy drive temp: 68°C
  * 118GB "CJ225128TC" SSD as /dev/sda: SATA 2.6, 3.0 Gb/s (current: 3.0 Gb/s), drive temp: 40°C
  * Winbond W25Q128 16MB SPI NOR flash, drivers in use: spi-nor/intel-spi

"nvme error-log /dev/nvme0 ; smartctl -x /dev/nvme0" could be used to get further information about the reported issues.

### Challenging filesystems:

The following partitions are NTFS: nvme0n1p3,nvme0n1p4 -> https://tinyurl.com/mv7wvzct

### Swap configuration:

  * /swapfile on /dev/nvme0n1p5: 2.0G (0K used)

### Software versions:

  * Ubuntu 22.04.4 LTS (jammy)
  * Compiler: /usr/bin/gcc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 / x86_64-linux-gnu
  * OpenSSL 3.0.2, built on 15 Mar 2022 (Library: OpenSSL 3.0.2 15 Mar 2022)    

### Kernel info:

  * `/proc/cmdline: BOOT_IMAGE=/boot/vmlinuz-6.5.0-21-generic root=UUID=18e952a3-4ef5-4c4d-8004-84f53b2e1f0d ro quiet splash vt.handoff=7`
  * Vulnerability Spec store bypass:    Mitigation; Speculative Store Bypass disabled via prctl
  * Vulnerability Spectre v1:           Mitigation; usercopy/swapgs barriers and __user pointer sanitization
  * Vulnerability Spectre v2:           Mitigation; Enhanced / Automatic IBRS, IBPB conditional, RSB filling, PBRSB-eIBRS SW sequence
  * Kernel 6.5.0-21-generic / CONFIG_HZ=250

Waiting for the device to cool down................................... 53.0°C^C

No thermal CPU throttling was detected, and the 7-zip benchmark score was 42,610 MIPS on average, with the first run as high as 45,080 MIPS (due to the initial burst in performance) and it then stabilizes at 41,688 and  41,058 MIPS in the second and third runs. That’s much higher than the score in the GEEKOM Mini IT12 (35,730 MIPS) and we initially thought this could be due to the default PL1 and PL2 power limits:

aey@SEi12-cnx:~$ sudo powercap-info -p intel-rapl
enabled: 1
Zone 0
  name: package-0
  enabled: 1
  max_energy_range_uj: 262143328850
  energy_uj: 28853949632
  Constraint 0
    name: long_term
    power_limit_uw: 35000000
    time_window_us: 55967744
    max_power_uw: 45000000
  Constraint 1
    name: short_term
    power_limit_uw: 55000000
    time_window_us: 2440
    max_power_uw: 0
  Constraint 2
    name: peak_power
    power_limit_uw: 80000000
    max_power_uw: 0
  Zone 0:0
    name: core
    enabled: 0
    max_energy_range_uj: 262143328850
    energy_uj: 14946156195
    Constraint 0
      name: long_term
      power_limit_uw: 0
      time_window_us: 976
  Zone 0:1
    name: uncore
    enabled: 0
    max_energy_range_uj: 262143328850
    energy_uj: 57907322
    Constraint 0
      name: long_term
      power_limit_uw: 0
      time_window_us: 976

PL1 is set to 35W, and PL2 to 55W in the Beelink mini PC, while the Mini IT12 had those set to 35W and 80W respectively.  So the power limits should not be involved here, and I’m guessing the thermal design is better in the Beelink mini PC so multi-core performance is better than in the GEEKOM Mini IT12, as we had already reported in the Windows 11 Pro review.

sbc-bench.sh also reports errors for the NVMe SSD drive as well as a high drive temperature:

aey@SEi12-cnx:~$ sudo nvme error-log /dev/nvme0
Error Log Entries for device:nvme0 entries:16
.................
 Entry[ 0]   
.................
error_count	: 215
sqid		: 0
cmdid		: 0x8
status_field	: 0x2002(INVALID_FIELD: A reserved coded value or an unsupported value in a defined field)
phase_tag	: 0x1
parm_err_loc	: 0x28
lba		: 0
nsid		: 0
vs		: 0
trtype		: The transport type is not indicated or the error is not transport related.
cs		: 0
trtype_spec_info: 0
.................

aey@SEi12-cnx:~$ sudo smartctl -x /dev/nvme0
smartctl 7.2 2020-12-30 r5155 [x86_64-linux-6.5.0-21-generic] (local build)
Copyright (C) 2002-20, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Model Number:                       CT500P3PSSD8
Serial Number:                      2330E8647A1B
Firmware Version:                   P9CR40A
PCI Vendor/Subsystem ID:            0xc0a9
IEEE OUI Identifier:                0x00a075
Controller ID:                      1
NVMe Version:                       1.4
Number of Namespaces:               1
Namespace 1 Size/Capacity:          500,107,862,016 [500 GB]
Namespace 1 Formatted LBA Size:     512
Namespace 1 IEEE EUI-64:            6479a7 7f3000005a
Local Time is:                      Sat Mar  2 16:12:08 2024 +07
Firmware Updates (0x12):            1 Slot, no Reset required
Optional Admin Commands (0x0017):   Security Format Frmw_DL Self_Test
Optional NVM Commands (0x005e):     Wr_Unc DS_Mngmt Wr_Zero Sav/Sel_Feat Timestmp
Log Page Attributes (0x06):         Cmd_Eff_Lg Ext_Get_Lg
Maximum Data Transfer Size:         64 Pages
Warning  Comp. Temp. Threshold:     85 Celsius
Critical Comp. Temp. Threshold:     95 Celsius

Supported Power States
St Op     Max   Active     Idle   RL RT WL WT  Ent_Lat  Ex_Lat
 0 +     6.00W  0.0000W       -    0  0  0  0        0       0
 1 +     3.00W  0.0000W       -    0  0  0  0        0       0
 2 +     1.50W  0.0000W       -    0  0  0  0        0       0
 3 -   0.0250W  0.0000W       -    3  3  3  3     5000    1900
 4 -   0.0030W       -        -    4  4  4  4    13000  100000

Supported LBA Sizes (NSID 0x1)
Id Fmt  Data  Metadt  Rel_Perf
 0 +     512       0         1
 1 -    4096       0         0

=== START OF SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

SMART/Health Information (NVMe Log 0x02)
Critical Warning:                   0x00
Temperature:                        48 Celsius
Available Spare:                    100%
Available Spare Threshold:          5%
Percentage Used:                    0%
Data Units Read:                    4,937,505 [2.52 TB]
Data Units Written:                 2,694,686 [1.37 TB]
Host Read Commands:                 46,223,348
Host Write Commands:                38,475,534
Controller Busy Time:               87
Power Cycles:                       27
Power On Hours:                     164
Unsafe Shutdowns:                   2
Media and Data Integrity Errors:    0
Error Information Log Entries:      215
Warning  Comp. Temperature Time:    0
Critical Comp. Temperature Time:    0
Temperature Sensor 1:               48 Celsius
Temperature Sensor 2:               55 Celsius
Temperature Sensor 8:               48 Celsius

Error Information (NVMe Log 0x01, 16 of 16 entries)
Num   ErrCount  SQId   CmdId  Status  PELoc          LBA  NSID    VS
  0        215     0  0x0008  0x4005  0x028            0     0

I’m not 100% sure what the error is, but it looks like the controller may just not support some commands, and it’s not a big issue.

We then tested the CPU single-core and multi-core performance with Geekbench 6.2.2.

That would be 2,589 points for the single-core benchmark and 10,208 points for the multi-core one. The single-core result is similar to the one in the GEEKOM Mini IT12 (2,575 points), but the multi-core result is better in the Beelink SEi12 i7-12650H mini PC (Mini IT12: 9,874 points) although not with a difference as great as in the 7-zip benchmark.

We’ll evaluate 3D graphics performance with Unigine Heaven Benchmark 4.0.

The Beelink SEi12 i7-12650H mini PC achieved 38.5 FPS on average and got a score of 969 points at 1920×1080 resolution.

Next up is YouTube video streaming at 4K and 8K resolution in the Chrome web browser.

8K 30 FPS played great with no frame dropped after watching the video for 9 minutes.

Same thing for 8K 30 FPS (4320p).

As usual, 60 FPS becomes more challenging, but the Beelink SEi12 mini PC handles that relatively well at 4K 60 FPS with only 92 frames dropped out of 54,226 after playing the video for 15 minutes.

Sadly, 8K 60 FPS is another story, and the video is unwatchable with 2045 frames dropped out of 5825 in a short test, and we can also see the loading wheel from time to time although the buffer has enough data for 20 seconds of playback.

We tried to play the video in Firefox just in case, but the result is even worse with 2,285 frames dropped out of 2,689. This issue did not occur in the Beelink SEi12 i7-12650H Windows 11 Pro review, and more surprisingly, the GEEKOM Mini IT12 could play 8K 60 FPS YouTube videos with AV1 codec just fine in Firefox… Not sure what happened here.

Speedometer 2.0 was used to estimate the web browsing performance in the latest version of Firefox.

The system averages 291 runs per minute with the results varying between 277.7 and 295.4 runs per minute.

Beelink SEi12 i7-12650H Ubuntu benchmarks against other mini PCs

Now that we have Ubuntu 22.04 benchmark results for the Beelink SEi12 i7-12650H, let’s compare its performance against the GEEKOM Mini IT12 (with the same Intel Core i7-12650H), GEEKOM Mini IT13 (Intel Core i9-13900H), GEEKOM AS 6 (AMD Ryzen 9 6900HX), and Khadas Mind Premium (Intel Core i7-1360P).

Before looking a the results, we should probably check out the main specifications of the five mini PCs.

	Beelink SEi12	GEEKOM Mini IT12	GEEKOM Mini IT13	GEEKOM AS 6	Khadas Mind Premium
SoC	Intel Core i7-12650H	Intel Core i7-12650H	Intel Core i9-13900H	AMD Ryzen 9 6900HX	Intel Core i7-1360P
CPU	10-cores/16-thread processor up to 4.70 GHz	10-cores/16-thread processor up to 4.70 GHz	14-core/20-core processor up to 5.4 GHz,	8-core/16-thread processor up to 4.9 GHz	12-core/16-core processor up to 5.0 GHz
GPU	64 EU Intel UHD Graphics (Alder Lake-P GT2)	64 EU Intel UHD Graphics (Alder Lake-P GT2)	96 EU Intel Iris Xe Graphics	AMD Radeon Graphics 680M	96 EU Intel Iris Xe
Memory	32GB DDR4-3200	32GB DDR4-3200	32GB DDR4-3200	32GB DDR5-4800	32GB LPDDR5-5200
Storage	500 GB NVMe SSD	1TB NVMe SSD	2TB NVMe SSD	1TB NVMe SSD	51TB NVMe SSD
Default OS	Windows 11 Pro	Windows 11 Pro	Windows 11 Pro	Windows 11 Pro	Windows 11 Home

Ubuntu 22.04 benchmark results comparison.

	Beelink SEi12	GEEKOM Mini IT12	GEEKOM Mini IT13	GEEKOM AS 6	Khadas Mind Premium
sbc-bench.sh
- memcpy	22,040.2 MB/s (P-Core)	24,180.5 MB/s (P-Core)	24,014.4 MB/s (P-core)	19,131.7 MB/s	25,389.5 MB/s (P-core)
- memset	27,417.4 MB/s (P-Core)	27,539.6 MB/s P-Core)	26,647.9 MB/s (P-Core)	16,781.4 MB/s	24,731.8MB/s (P-core)
- 7-zip (average)	42,610	35,730	56,540	54,592	44,430
- 7-zip (top result)	45,080	41,360	60,981	56,251	50,396
- OpenSSL AES-256 16K	1,587,265.54k (P-Core)	1,630,005.93k (P-Core)	1,844,401.49k (P-Core)	1,249,203.54k	1,771,334.31k (P-Core)
Geekbench 6 Single	2,589	2,575	2,745	1,992	2,093
Geekbench 6 Multi	10,208	9,874	11,974	9,535	8,891
Unigine Heaven score	969	1,002	1,333	1,553	1,349
Speedometer (Firefox)	291	302	273	202	242

Just like the GEEKOM Mini IT12, the Beelink SEi12 i7-12650H mini PC has excellent single-core performance as illustrated in Geekbench 6 (single core) and Speedometer results, while the 64EU Intel iGPU is a weak point of the Core i7-12650H processor. Multi-core performance is however quite better in the Beelink mini PC than in the GEEKOM device as we can see from the 7-zip results, and to a much lesser extent in the GeekBench 6.2 multi-core benchmark.

Storage performance

iozone3 command line utility can help us evaluate the performance of the 500GB NVMe SSD that comes with the mini PC:

aey@SEi12-cnx:~$ sudo iozone -e -I -a -s 1000M -r 4k -r 16k -r 512k -r 1024k -r 16384k -i 0 -i 1 -i 2
	
                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
         1024000       4   494627   650667   178133   177569    72708   638548                                                                
         1024000      16  1386403  1692866   198590   187623   249485  1649280                                                                
         1024000     512  1850221  1847238  2039320  2047148  2032015  1837493                                                                
         1024000    1024  1825982  1884587  2737001  2759891  2754761  1845733                                                                
         1024000   16384  1820277  1843006  4431412  4468037  4464029  1837319                                                                

iozone test complete.

That’s a sequential read speed of about 4,431 MB/s and a sequential write speed of around 1820 MB/s which compares to 4,836 MB/s and 1,905 MB/s in Windows using CrystalDiskMark.

We also tested the SATA interface with an entry-level 128GB 2.5-inch SATA SSD using iozone3:

aey@SEi12-cnx:/media/aey/New Volume$ sudo iozone -e -I -a -s 100M -r 16384k -i 0 -i 1

                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
          102400   16384   144165   128322   252595   255083                                                                                  

iozone test complete.

That’s about 252 MB/s for reads and 144Mb/s for writes, or about the same as in Windows, and expected for this specific SATA SSD.

USB ports testing

The speed of the USB ports was tested with an ORICO M234C3-U4 M.2 NVMe SSD enclosure together with lsusb and iozone3 command line utilities. A Seagate USB HDD will be used for the USB 2.0 ports since the ORICO enclosure is not backward compatible.

For reference, here’s the output for the front left USB port (10 Gbps)…

aey@SEi12-cnx:~$ lsusb -t | grep uas
    |__ Port 2: Dev 2, If 0, Class=Mass Storage, Driver=uas, 10000M

aey@SEi12-cnx:/media/aey/EXT4-REVIEW$ sudo iozone -e -I -a -s 1000M -r 16384k -i 0 -i 1

                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
         1024000   16384   991006   992899   870106   872809                                                                                  

iozone test complete.

… and the results for the top USB 2.0 port on the rear panel:

aey@SEi12-cnx:~$ lsusb -t | grep uas
    |__ Port 5: Dev 10, If 0, Class=Mass Storage, Driver=uas, 480M
aey@SEi12-cnx:/media/aey/USB3_EXT4$ sudo iozone -e -I -a -s 100M -r 16384k -i 0 -i 1
 
                                                              random    random     bkwd    record    stride                                    
              kB  reclen    write  rewrite    read    reread    read     write     read   rewrite      read   fwrite frewrite    fread  freread
          102400   16384    40168    42192    39006    41072                                                                                  
 
iozone test complete.

Summary of the results for the five ports:

• Front panel (left to right)
  • USB-A #1 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – Read: 870 MB/s; write: 991 MB/s
  • USB-A #2 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – Read: 854 MB/s; write: 958 MB/s
  • USB-C #1 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – Read: 853 MB/s; write: 944 MB/s
• Rear panel
  • USB-A #1 (top) – USB 2.0  – USB 2.0 Hight-Speed (480 Mbps) – Read: 39 MB/s; write: 40 MB/s
  • USB-A #2 (bottom) – USB 2.0 – USB 2.0 Hight-Speed (480 Mbps) – Read: 40 MB/s; write: 31 MB/s

Everything works as expected here.

Networking performance (Gigabit Ethernet and WiFi 6)

The Gigabit Ethernet port was tested with iperf3 using AAEON UP Xtreme 11 mini PC (192.168.31.12) on the other side.

• Download

aey@SEi12-cnx:~$ iperf3 -t 60 -c 192.168.31.12 -i 10 -R
Connecting to host 192.168.31.12, port 5201
Reverse mode, remote host 192.168.31.12 is sending
[  5] local 192.168.31.74 port 58720 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate
[  5]   0.00-10.00  sec  1.10 GBytes   942 Mbits/sec                  
[  5]  10.00-20.00  sec  1.10 GBytes   942 Mbits/sec                  
[  5]  20.00-30.00  sec  1.10 GBytes   942 Mbits/sec                  
[  5]  30.00-40.00  sec  1.10 GBytes   942 Mbits/sec                  
[  5]  40.00-50.00  sec  1.10 GBytes   942 Mbits/sec                  
[  5]  50.00-60.00  sec  1.10 GBytes   942 Mbits/sec                  
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.04  sec  6.58 GBytes   941 Mbits/sec    0             sender
[  5]   0.00-60.00  sec  6.58 GBytes   942 Mbits/sec                  receiver

iperf Done.

• Upload

aey@SEi12-cnx:~$ iperf3 -t 60 -c 192.168.31.12 -i 10
Connecting to host 192.168.31.12, port 5201
[  5] local 192.168.31.74 port 56696 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-10.00  sec  1.10 GBytes   943 Mbits/sec    0    363 KBytes       
[  5]  10.00-20.00  sec  1.10 GBytes   942 Mbits/sec    0    795 KBytes       
[  5]  20.00-30.00  sec  1.10 GBytes   942 Mbits/sec    0    795 KBytes       
[  5]  30.00-40.00  sec  1.10 GBytes   942 Mbits/sec    0    795 KBytes       
[  5]  40.00-50.00  sec  1.10 GBytes   942 Mbits/sec    0    795 KBytes       
[  5]  50.00-60.00  sec  1.10 GBytes   942 Mbits/sec    0    795 KBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.00  sec  6.58 GBytes   942 Mbits/sec    0             sender
[  5]   0.00-60.05  sec  6.58 GBytes   941 Mbits/sec                  receiver

iperf Done.

• Full duplex

aey@SEi12-cnx:~$ iperf3 -t 60 -c 192.168.31.12 -i 10 -bidir
Connecting to host 192.168.31.12, port 5201
[  5] local 192.168.31.74 port 42466 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-10.00  sec  1.10 GBytes   943 Mbits/sec    0    468 KBytes       
[  5]  10.00-20.00  sec  1.10 GBytes   941 Mbits/sec    0    468 KBytes       
[  5]  20.00-30.00  sec  1.10 GBytes   941 Mbits/sec    0    468 KBytes       
[  5]  30.00-40.00  sec  1.10 GBytes   942 Mbits/sec   36    454 KBytes       
[  5]  40.00-50.00  sec  1.10 GBytes   942 Mbits/sec    0    454 KBytes       
[  5]  50.00-60.00  sec  1.10 GBytes   941 Mbits/sec    0    454 KBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.00  sec  6.58 GBytes   942 Mbits/sec   36             sender
[  5]   0.00-60.04  sec  6.58 GBytes   941 Mbits/sec                  receiver

iperf Done.

All good here. It’s better than in Windows, where the data rate dropped to 883 Mbps in one direction (upload).

Time to test WiFi 6 through a Xiaomi Mi AX6000 router.

• Download

aey@SEi12-cnx:~$ iperf3 -t 60 -c 192.168.31.12 -i 10 -R
Connecting to host 192.168.31.12, port 5201
Reverse mode, remote host 192.168.31.12 is sending
[  5] local 192.168.31.127 port 40354 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate
[  5]   0.00-10.00  sec  1.08 GBytes   927 Mbits/sec                  
[  5]  10.00-20.00  sec  1.09 GBytes   939 Mbits/sec                  
[  5]  20.00-30.00  sec  1.10 GBytes   941 Mbits/sec                  
[  5]  30.00-40.00  sec  1.09 GBytes   933 Mbits/sec                  
[  5]  40.00-50.00  sec  1.10 GBytes   942 Mbits/sec                  
[  5]  50.00-60.00  sec  1.10 GBytes   941 Mbits/sec                  
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.04  sec  6.55 GBytes   937 Mbits/sec    0             sender
[  5]   0.00-60.00  sec  6.55 GBytes   937 Mbits/sec                  receiver

iperf Done.

• Upload

aey@SEi12-cnx:~$ iperf3 -t 60 -c 192.168.31.12 -i 10
Connecting to host 192.168.31.12, port 5201
[  5] local 192.168.31.127 port 34882 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-10.00  sec  1.08 GBytes   928 Mbits/sec   65   1.59 MBytes       
[  5]  10.00-20.00  sec  1.10 GBytes   942 Mbits/sec    7   1.76 MBytes       
[  5]  20.00-30.00  sec  1.10 GBytes   942 Mbits/sec    2   1.70 MBytes       
[  5]  30.00-40.00  sec  1.10 GBytes   942 Mbits/sec    8   1.27 MBytes       
[  5]  40.00-50.00  sec  1.10 GBytes   942 Mbits/sec    1   1.26 MBytes       
[  5]  50.00-60.00  sec  1.10 GBytes   942 Mbits/sec    4   1.44 MBytes       
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Retr
[  5]   0.00-60.00  sec  6.56 GBytes   939 Mbits/sec   87             sender
[  5]   0.00-60.05  sec  6.56 GBytes   938 Mbits/sec                  receiver

iperf Done.

The results – 937 Mbps download and 939 Mbps upload – are so similar to the Ethernet results that we had to double-check, and the IP address is indeed different. As usual, WiFi data transfer rates are better in Linux than in Windows 11, where the system achieved 768 Mbps (DL) and 778 Mbps (UL).

We also quickly tested Bluetooth 5.2 by successfully transferring a file from an Android smartphone.

Stress test and CPU temperature

We’ll now run a stress test on all 16 threads of the Intel Core i7-12650H processor while monitoring the package temperature with Psensor and the CPU frequency with the sbc-bench.sh script.

The screenshot above shows the CPU temperature stays around 87-89°C, with the P-cores running at around 2,800 – 2,900 MHz and the E-core at 2,400 – 2,500 MHz while running the same stress test on the GEEKOM Mini IT12 revealed a different temperature chart with the package temperature quickly jumping to 91°C for a few seconds before dropping and stabilizing at around 81°C over the long run as the Core i7-12650H cores were clocked at 2,600 MHz (P-Cores) and 2,400/2,500 MHz (E-Cores). See below.

These different behaviors should explain the 7-zip benchmark results differences between the two mini PCs, and that means the Beelink SEi12 i7-12650H performs better with the CPU operating at higher frequencies and temperatures under sustained heavy workloads. What we (the team at CNX Software) don’t know is how this might impact the expected lifetime of the processor and mini PC.

Fan noise

The mini PC’s fan is relatively quiet at idle and under light loads but becomes noisier under heavy loads without becoming too annoying. We measured the fan noise with a sound level meter placed approximately 5 cm from the top of the case:

• Idle and web surfing – 45 – 47 dBA
• Stress test (on 16 threads) –  50 – 53 dBA

Note that’s quite noisier than the Mini IT12 which we measured at 39.0 – 41.6 dBA at idle. The sound level meter measures 38 dBA in a quiet room.

Beelink SEi12 i7-12650H power consumption in Ubuntu 22.04

We measured power consumption with a wall power meter as follows:

• Power Off – 1.0 – 1.1 Watts
• Idle – 10.7 – 10.9 Watts
• Video playback – 57.0 – 63.5 Watts (YouTube 8K60fps in Firefox)
• CPU stressed test (stress -c 16)
  • During the first 30 seconds 73.0 – 86.6 Watts
  • After 30 seconds 60.0 – 63.0 Watts

Note: The mini PC was connected over WiFi 6, one 2.4 GHz RF dongle for a wireless keyboard and mouse combo, and a 14-inch CrowView laptop monitor connected via HDMI port and its own power supply.

The stress test power consumption is higher than on the GEEKOM Mini IT12, but that’s normal because the Beelink SEi12 operates at a higher frequency. The big shocker is the power consumption while playing an 8K YouTube video as it’s about three times higher and points to software video decoding which would explain why the 8K 60 FPS video can’t be playing smoothly. Firefox’s compositing is set to “WebRender (Software)” and we tried to change a few settings, but did not manage to enable hardware video decoding.

Conclusion

The Beelink SEi12 i7-12650H mini PC performed well in Ubuntu 22.04 with all features working as expected including video output, gigabit Ethernet, WiFi 6, Bluetooth, and USB ports. YouTube video playback is working fine up to 4Kp60 and 8Kp30, but not 8Kp60 in either Chrome or Firefox.

The SEi12 performs similarly to the GEEKOM Mini IT12, except for multi-core workload where it is quite faster due to different behavior under loads where the Core i7-12650H operates at a higher frequency and temperature in the Beelink mini PC, but the Mini IT12 is configured in such a way to operate at a lower frequency and temperature under sustained multi-core loads. The main downsides compared to the Mini IT12 are the lack of USB4 ports, gigabit Ethernet networking instead of 2.5GbE, and a smaller and slower SSD for the models we tested. The latter also supported 8K 60FPS videos

We’d like to thank Shenzhen AZW Technology (aka Beelink) for sending a review sample of the Beelink SEi12 i7-12650H with 32GB DDR4 and a 500GB M.2 NVMe SSD. This model can be ordered for $439 on Amazon (after ticking on the $110 discount coupon), Aliexpress (some countries only), and on the company’s online store where you can get a $50 discount with the code 1265050 valid until March 31 (but they seem to extend it every month…). For reference, the GEEKOM Mini IT12 (32GB/1TB) typically sells for a little under $520, so the SEi12 model we tested is a cheaper device albeit with a smaller 500GB SSD and fewer ports.

CNXSoft: This review is a translation – with a few additional insights – of the original article on CNX Software Thailand by Suthinee Kerdkaew.

The post Beelink SEi12 i7-12650H mini PC review – Part 3: Ubuntu 22.04 Linux appeared first on CNX Software - Embedded Systems News.

French company Scaleway has just launched the “Elastic Metal RV1” bare metal servers which it claims to be the world’s first RISC-V servers available in the cloud with pricing at 0.042 Euros per hour, or 15.99 Euros a month excluding VAT.

Scaleway launched some Arm servers based on Marvell Armada 370/XP quad-core Cortex A9 processor in 2015 before phasing those out a few years ago, and they are now just offering AMD and Intel-based servers and hosted Apple Mac computers based on the M1 Arm chip. But the company has decided to try something new again with the EM-RV1 servers based on Alibaba T-Head TH1520 quad-core RISC-V processor, 16GB RAM, and 128GB eMMC flash and running Debian, Ubuntu, or Alpine.

EM-RV1-C4M16S128-A instance key features and specifications:

• SoC – Alibaba T-Head TH1520
  • CPU – Quad-core RISC-V Xuantie C910 (RV64GCV) processor @ 1.85 GHz
  • GPU – Imagination BXM-4-64 with support for OpenCL 1.1/1.2/2.0, OpenGL ES 3.0/3.1/3.2, Vulkan 1.1/1.2, Android NN HAL
  • VPU  – H.265/H.264/VP9 video encoding/decoding
  • NPU – 4 TOPS @ INT8 with support for TensorFlow, ONNX, Caffe
• System Memory – 16GB LPDDR4
• Storage – 128GB eMMC flash
• Networking – 100 Mbit/s Ethernet network card with public IPv4 and IPv6 addresses included
• Power Consumption – 0.96W to 1.9W per core @ ~1.8GHz; average: 1.3W per core
• Custom design with laser-cut chassis, 3D-printed blades
• Pricing  – 0,042 € per hour, 15,99 € per month

Scaleway also shared some benchmark results showing the performance of the EM-RV1 RISC-V server against the StarFive VisionFive 2 RISC-V SBC and some of their x86 instances. In Geekbench 6, it’s faster than a server based on an Intel C2350 dual-core processor (Dedibox Start-3-S), but still ways off the octa-core Intel C2750-based Dedibox Start-1-M.

Note the EM-RV1 instances are part of Scaleway Labs so it’s mostly for evaluation, but the company also says the RISC-V server can be useful for testing RISC-V applications, CI/CD, and AI applications thanks to the 4 TOPS NPU found in each TH1520 SoC. You can get started on the product page where you’ll also find additional information and extra benchmarks.

I didn’t try the Scaleway RISC-V server myself, but Bret Weber did and he reported his experience setting up an instance with Ubuntu 23.10 (GNU/Linux 5.10.113+ riscv64) and ran several benchmarks. Scaleway says the EM-RV1 servers have been designed in-house with “the soldering of electronic components, the development of specific firmware, and the manufacture of the enclosures using 3D printing”, but Bret also noted the arrangement of the ports on the first photo in this post looks very similar to the Sipeed Lichee Cluster 4A box.

So it looks like they used the Sipeed Cluster 4A box’s motherboard fitted with Sipeed LM4A modules, and customized the mechanical design so that they can fit several such boards into a rack.

The post Scaleway launches hosted RISC-V servers for 15.99 Euros per month appeared first on CNX Software - Embedded Systems News.

The Embedded Open Source Summit 2024 (EOSS 2024) will take place on April 16-18 and the Linux Foundation has already announced the schedule with conference sessions, lightning talks, and birds of a feather (BoF) sessions covering embedded Linux, Zephyr OS, and real-time (RT) Linux.

While I won’t be attending in person, I still find it interesting to check out the schedule as we may learn more about the current status of embedded Linux. So I’ve created my own little virtual schedule out of the available talks.

Tuesday, April 16 – Day 1, Embedded Open Source Summit 2024

• 9:05 – 9:45 – No, It’s (Still) Never Too Late to Upstream Your Legacy Linux-Based Platforms by Neil Armstrong, Linaro

Nearly 7 years ago, Neil already spoke about this subject in Berlin, and it’s still very true. Do you maintain or used to maintain a Linux-based board or SoC off-tree? Then there are plenty of reasons for you to push your changes to the mainline Linux. Some will say it’s too late, too complex, or too expensive but the long-term benefits of regular upstreaming truly outweigh these constraints especially if you have the right methods. In this presentation, Neil will elaborate on this question. Neil will then expose the various challenges of code upstreaming, like time constraints, copyright issues, and the community aspect of the work. For example, vendor GPL code is generally lying on an obscure GitHub repo, or in a hardly reachable tarball. In parallel, Neil will present practical tips to ease your day-to-day upstream work and this simple rule: the faster the maximum patches are upstreamed, the less work you’ll have to actually maintain the port in the future.

• 10:00 – 10:40 – Enabling Real-Time Secure Connectivity to the Industrial Edge with Single-Pair Ethernet and Zephyr by Jason Murphy, Analog Devices

Single-Pair Ethernet (10BASE-T1L SPE) is transforming low-speed, insecure fieldbus installations into flexible, higher data rate links with modern cybersecurity. SPE enables IP connectivity to field devices over twisted pair cabling, reducing installation costs while eliminating data islands that exist between OT and IT systems. Extending IP connectivity to the edge requires the integration of Ethernet connectivity into resource-constrained embedded devices around factories and buildings. This paper explores the use of Zephyr OS as an embedded software ecosystem for Ethernet-connected industrial devices. Zephyr provides a comprehensive networking software stack that supports protocols such as TCP/IP, MQTT, and TLS that are crucial in developing industrial IoT applications. We present an overview of SPE technology for IIoT applications, followed by a discussion on how Zephyr’s software ecosystem can be harnessed to drive rapid development of secure, real-time, Ethernet-connected industrial edge devices. The discussion aims to underscore the potential of SPE in combination with Zephyr OS, to drive smaller, smarter, and lower-cost intelligent nodes deeper into the built environment.

• 14:15 – 14:55 – Enabling Linux Support with Upstream Kernel on Snapdragon X1 Elite SoCs by Sibi Sankar & Rajendra Nayak, Qualcomm Innovation Center Inc

The session details how the upstream kernel was used during the Snapdragon X1 Elite SoC Linux bring-up, how it was used to obtain the reported ST and MT benchmark Geekbench scores, and the current upstream status of the patches. It also includes a demo booting the upstream kernel with a Debian/Ubuntu userspace on a Snapdragon X1 Elite QRD (Qualcomm Reference Device). The patches to enable basic console support along with a public branch with Display/GPU enabled were posted when the Snapdragon X1 Elite SoC was publicly announced. Boot-to-console support has already landed on kernel version v6.7 and is on track to have remaining kernel support land by the time the first commercial device with X1 Elite SoC comes out on the market.

• 15:25 – 16:05 – Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driver for a New GPU by Alejandro Piñeiro Iglesias, Igalia

The Raspberry Pi 5 was announced in October 2023. This new version of the popular embedded device comes with a new iteration of Broadcom’s VideoCore GPU platform and was released with a fully open-source driver stack, developed by Igalia. The presentation will discuss some of the major changes required to support this new Video Core iteration, the challenges we faced in the process, and the solutions we provided in order to deliver conformant OpenGL ES and Vulkan drivers. The talk will also cover the next steps for the open-source Raspberry Pi 5 graphics stack.

• 16:20 – 17:00 – Comparing Linux OS Image Update Models by Drew Moseley, Toradex

Today’s connected Embedded Linux devices increasingly rely on over-the-air updates to deliver security patches and new functionality. These updates can be challenging due to their substantial size. Delivering them can place a burden on mobile network connections. This session will compare several models for providing OS updates to remote devices focusing on their bandwidth reduction characteristics. We will dig into the architecture of each of the model’s approaches to size reduction including details such as storage formats, update generation, and implementation impact on target systems.

We will then benchmark them against a common set of input images to quantify the size reduction. Reducing the total amount of data transferred is obviously a cost savings, but it can also increase reliability; less data to transfer means there is less time for things to go wrong, resulting in fewer retries and an overall smoother experience. We will wrap up with a discussion of common OTA update systems focusing on the update model(s) supported by each. Armed with this information, attendees will be better equipped to decide which model, and ultimately which solution is appropriate.

Wednesday, April 17 – Day 2, Embedded Open Source Summit 2024

• 9:00 – 9:40 – Accelerated Porting of Linux, U-Boot and Yocto for Production Ready Embedded Systems by Vaishnav Mohandas Achath & Keerthy Jagadeesh, Texas Instruments

Embarking on the journey of bringing SPL/U-Bootc, Linux, and Yocto on a freshly minted System on Chip and a new hardware platform can be a daunting task. Based on the experience of porting multiple ARM-based SoCs to work with mainline Linux/U-Boot/Yocto and also the multiple custom hardware platform bring-up experiences, We discuss strategies to get Linux, U-Boot bootloader, and Yocto-based filesystem functional and product-ready on your system in a very short time, sharing insights, practical experiences, and tools to streamline the process.

While getting your system booting Linux and U-Boot quickly once the silicon/boards arrive is important, it is more important to have a clean production-ready system that can be maintained over the Long term, we also discuss best practices to ensure minimal effort long-term support for your systems as well in this session. By the end of this session, you will be equipped with the knowledge and resources to embark on your journey to bring up Linux, U-Boot, and Yocto porting adventure for your custom embedded platform whether you are a seasoned embedded developer or a Linux enthusiast ready for a challenge.

• 9:55 – 10:35 – Rethinking U-Boot Devicetree Story by Sumit Garg

The maintenance of device trees (DT) for embedded systems often appears to be fragmented with different DT sources appearing in different projects. The DT specification provides a base vocabulary to describe hardware but it is augmented by bindings documentation. Currently, the device bindings are maintained as part of the Linux source tree. This often leads to confusion in the embedded community, especially whether U-Boot should maintain its own DT. Many of the DT boot standards (EBBR, SystemReady, etc) require firmware to provide a DT that the bootloader and kernel can consume. Things get difficult when U-Boot and Linux contain different DT sources or disagree on preferred bindings… and doubly so when U-Boot is being used as part of the firmware implementation!

This session will focus on the journey to change the way we sync DT sources from Linux to U-Boot. In particular how we can switch away from ad-hoc syncs by board maintainers to regular full-tree alignment. This also includes bringing DT bindings checks into the U-Boot build system. We’ll wrap up by looking at how the DT contribution model can become more friendly for contributors/maintainers coming from different projects.

• 11:00 – 11:40 – Optimizing BLE for Throughput-Oriented Applications by Luis Ubieda, Croxel, Inc.

Even though BLE is known for being low power, its presence and accessibility in existing devices are ubiquitous; which enables it to be used for other applications where its throughput capabilities are critical for success. This session goes over the technical details of BLE with a focus on relevant aspects towards improving BLE data-throughput; including tools and tips for improving your BLE connection, and covers an example of optimizing a Zephyr-based application for BLE data-throughput.

This presentation is for you if:

1. You struggle to use BLE for anything more than a Low-Power & Low Data-Rate sensor.
2. Usually BLE either “just works“ or “is broken“. When is not working, you don’t know where to look.
3. You don’t know if BLE is fast enough for your application.

Content will include an overview of BLE from a throughput perspective, factors affecting its performance, throughput expectations, tools and tips, a live demo, and a comparison of BLE-throughput optimized devices vs non-optimized devices.

• 11:55 – 12:35 – Adding Support for Power Over Ethernet (PoE or IEEE Clause 33) to Linux Network Stack by Köry Maincent, Bootlin

Power over Ethernet (PoE) is a technology that combines electrical power and data transmission over a single Ethernet cable. It eliminates the need for separate power sources, simplifying installations for devices like IP cameras, and VoIP phones. In this presentation, we’ll initially delve into Power over Ethernet (PoE), It debuted in IEEE Clause 33 without explicit reference to the PoE nomenclature. We will look at what currently exists in the Linux kernel and user space to support PoE. We’ll continue our discussion with some details of the Linux implementation currently in development and the PSE core changes brought by this new support. The PSE framework core and bindings happen to need modification as it was not prepared for the PoE specificities. In parallel to these extensions of the PSE framework, we developed Linux kernel drivers for two distinct PoE controllers: the Microchip PD692x0 and the Texas Instruments TPS23881. Finally, we will look into the mainline status, the things that still need to be merged, and the future features that need development. This PoE Linux support is sponsored and funded by DENT Project.

• 14:00 – 14:40 – Beefy ML: Ultra-Low Power Algorithms on Cattle by Jordan Yates, CSIRO

Machine learning may conjure images of GPU clusters in a datacenter, but robust algorithms can actually be run in microseconds on a basic Cortex-M4F. Combined with thoughtful hardware design, this enables systems that can run ML algorithms indefinitely on the ear of an animal. In this session, Jordan will provide an overview of the journey to a real-world algorithm designed to monitor cattle feed intake from accelerometer data. The focus will be on the system design, algorithm implementation and validation, but data acquisition, annotation and training will also be touched on. The whole system is built upon multiple open-source components: Zephyr as the underlying OS, Zephyr Power Management for low-power operation, PyTorch for algorithm training, and CMSIS DSP for embedded implementation.

• 14:55 – 15:35 – Tuning RT Kernel to Improve Schedule Performance with Intel Platform by Junxiao Chang, Intel

This work focuses on Linux RT kernel thread scheduling latency performance improvement on the Intel ADL/RPL platform. Linux real-time performance is more and more important for industry, medical, and other domains. This work tunes RT kernel command line parameters, kernel options, and BIOS settings to get the best thread scheduling latency performance on Intel platforms. RT Linux kernel has a lot of kernel options which might impact thread scheduling. For example, adding “idle=poll” makes the CPU wake up faster. The latest generation Intel processors have a lot of features that might be related to RT kernel scheduling performance. Intel ADL and later generations processors have big/small cores. RT thread could have better performance if it is set to big core. With optimizations in BIOS, kernel options and kernel command line parameters, thread schedule latency is much better than it is in the original default kernel. Without these optimizations, scheduling latency might be > 1000us. With optimization, the latency is around 5 to 10 microseconds with the Intel ADL/RPL platform.

• 16:00 – 16:40 – Bluetooth on Embedded Linux Systems Deep Dive by Marcel Ziswiler, Toradex Inc.

Bluetooth is one of the most dominant wireless connection technologies. USB Bluetooth dongles are very common in the consumer world, but many modern embedded systems also contain directly designed-in Bluetooth solutions using interfaces like PCIe, SDIO, or UART. Often as part of a wireless solution with Wi-Fi and/or other wireless technologies like 5G, IEEE 802.15.4, etc. Marcel was tasked to evaluate various Bluetooth solutions and tested all major vendors that have upstream driver support.

This talk introduces the Bluetooth specification and discusses how it may be used on embedded systems from the Linux kernel, accompanying user space, BlueZ Bluetooth stack, and further application-level integration (e.g. home assistant). USB profiles are covered including audio interaction with PipeWire/WirePlumber. The last part concentrates on how to debug various Bluetooth-related issues one may encounter. The powerful BlueZ btmon and hcidump facilities are introduced which may be used to collect traces of Bluetooth communication and in combination with Wireshark allows easy visualization and Bluetooth protocol analysis. A live demonstration of some real-world Bluetooth use cases will also be presented.

• 16:55 – 17:35 – RISC-V and RISE Project BoF by Jeffrey Osier-Mixon, Red Hat

RISC-V is an open instruction set that is taking the world by storm, enabling new and creative hardware designs across the spectrum of computing devices – many of which are themselves open. This BoF is a meeting place at EOSS to discuss the current state of RISC-V as well as the RISE Project, an open-source initiative under LF Europe to support the RISC-V software ecosystem.

Thursday, April 18 – Day 3

• 9:00 – 9:40 – The State of Hardware Video Codecs in Linux by Andrzej Pietrasiewicz & Nicolas Dufresne, Collabora

The need for video codecs on embedded Linux is forever growing. Whether you make security cameras, in-flight entertainment, infotainment, digital signage, or even robots, it’s likely that you will need a video codec of some sort. While compression ratios have massively increased over the years, encoding and decoding complexity have exploded. Due to this expansion, some help from dedicated hardware is needed. In this talk, you will learn how the Linux Media subsystem have gained driver interfaces for various types of hardware codecs and numerous encoding formats such as H.264, HEVC, AV1, VP9, VP8, MPEG-2, and more. You’ll get an insight into the work that is currently underway and a plausible future plan.

• 9:55 – 10:35 – Rust for Linux – What Is Possible and What Is Still Work in Progress by Christina Quast, Independent

Rust has been the up-and-coming new programming language that will remove whole categories of bugs like memory leaks and race conditions from code forever. For a few years, step by step more Rust code has been added to mainline. This talk gives an overview of what kernel modules you can already write, and which functionality is still only to be found in the Rust for Linux project. Furthermore, it outlines how a person new to this topic could get started writing their first kernel module in Rust!

• 11:00 – 11:40 – SoC Development: From ROM to Application by Nadav Cohen Zukerman, Autotalks

This session will unfold a journey of developing a SoC from scratch using Zephyr. Zephyr is present all along our bootflow, from ROM to a bootloader to a fully operational application image. In the session, we will detail the development process, the challenges we had, and the lessons we learned throughout over 2 years of development with Zephyr, two tapeouts (and two upcoming), and two SoC bringups.

• 11:55 – 12:35 – Testing Rotation Sensor Drivers with LEGO Robots and Other Adventures in the Linux IIO Subsystem by David Lechner, BayLibre

Is it possible for a mostly remote team to develop Linux drivers for IIO hardware while not actually putting their hands on it? We’ve been giving it a go and would like to share why we are doing it, how it is going, what works, what doesn’t work, and hear from others who are doing the same. We will share how we’ve managed to update kernels, and use signal generators, logic analyzers, and even LEGO robots remotely in the course of development, testing & validation. We would also like to share some new developments in the IIO subsystem related to the work we are doing. There have been several efforts working towards enabling higher throughput and higher sample rates for IIO devices in several different areas.

• 14:00 – 14:40 – Maximizing SD Card Life, Performance, and Monitoring with KrillKounter by Andrew Murray, The Good Penguin

The underlying storage of an SD card is NAND flash, which is inherently unreliable, has a limited number of program/erase cycles, and has constraints on how data is written. Fortunately, the firmware in an SD card overcomes these challenges well enough to provide a cost-effective and reliable block-based storage medium; though, this is not without side effects that can impact performance and endurance. In this talk, we’ll delve into the ‘write to destruction’ testing that we’ve performed, which illustrates how access patterns and write amplification can significantly impact the lifespan of an SD card. We’ll show you what happens when an SD card fails and provide actionable steps to maximize the lifespan of an SD card in Linux. We will also examine the performance characteristics of SD cards and explore how access patterns can impact write performance. We’ll present methods for analyzing performance and provide practical steps for improving throughput. Finally, we will introduce Krill Kounter, an open-source daemon and library for embedded devices for monitoring SD card wear and indicators of write amplification over its lifetime.

• 15:15 – 15:55 – Using Picolibc in Embedded Systems by Keith Packard, Amazon

Picolibc is a C library designed for embedded environments. Providing a complete C17 library interface along with much of the POSIX 2008 additions, Picolibc offers standards conformance, broad architecture support, and integrated testing performed under emulation on the target architectures. This talk will start by providing an overview of Picolibc, including API support, standards conformance, and memory usage. Then, a description of the testing infrastructure, including the bare-metal test frameworks, emulator bugs fixed, and guidance on supporting new targets will be provided. Next, some specific examples of integrating Picolibc into various embedded RTOSes, including FreeRTOS, Zephyr, and RIOT will be presented. The talk will finish with the current status of Picolibc along with future plans.

• 16:10 – 16:50 – Compound Interest – Dealing with Two Decades of Technical Debt in Embedded Linux by Bartosz Golaszewski, Linaro

The GPIO subsystem is one of the oldest driver abstraction layers in the Linux kernel. First, somewhat unified GPIO interfaces appeared around 20 years ago. Over the years it has become one of the most ubiquitous subsystems in embedded Linux as GPIOs are used universally by all kinds of devices for a multitude of more, less, or not-at-all standardized functions. Over the years GPIOLIB has become its own, self-contained library, was integrated into the driver model, and acquired many new features (device-tree and ACPI support, GPIO irqchips, plug-and-play) but it came at a cost. The subsystem was hit especially hard by Arm fragmentation, the fallout of which is still visible in countless board files containing suboptimal code we need to maintain. We eventually ended up with two in-kernel APIs, two uAPI variants, complex glue code to pinctrl and interrupt subsystems as well as piles upon piles of quirks and corner cases. Efforts to improve the situation have been ongoing and have picked up in 2023/2024. This talk will cover how we try to improve serialization, hot-pluggability, and reduce the API abuse treewide while dancing carefully around existing legacy users.

---

You’ll find the complete schedule for the Embedded Open Source Summit 2024 on the Linux Foundation website. If you are interested in attending in person, you can register with the following rates:

• Standard – Feb 25 – Apr 7, 2024
  • Attendee – USD 799
  • Academic – USD 275
  • Government – USD 275
  • Hobbyist – USD 275
  • Small Business – USD 500
• Late – Apr 8–18, 2024
  • Attendee – USD 949
  • Academic – USD 275
  • Government – USD 275
  • Hobbyist – USD 275
  • Small Business – USD 500

A small business is a company that has less than $5 million in annual revenue and less than 30 employees.

The post Embedded Open Source Summit 2024 schedule – Embedded Linux, Zephyr OS, and Real-time Linux appeared first on CNX Software - Embedded Systems News.

WeAct STM32G4 is a tiny development board based on a 170 MHz STMicro STM32G4 Arm Cortex-M4F mixed-signal microcontroller with DSP instructions and suitable for applications such as motor control, building automation, lighting, digital power meters, and more.

Two versions of the board are offered one with an STM32G474CEU6 “Hi-resolution line” microcontroller equipped with 128KB RAM and 512KB flash, and the other with the lower-end STM32G431CBU6 “Access Line” MCU with just 32KB RAM and 128KB flash. The board also comes with a USB-C port for power and programming, three buttons, and two 24-pin headers.

WeAct STM32G4 specifications:

• Microcontroller (one or the other)
  • STMicro STM32G431CBU6 – Arm Cortex-M4F MCU @ 170 MHz with DSP instructions, 32KB RAM, 128KB flash, and math accelerator
  • STMicro STM32G474CEU6 – Arm Cortex-M4F MCU @ 170 MHz with DSP instructions, 128KB RAM, 512KB flash, and math accelerator; high-resolution timer and complex waveform builder plus event handler (HRTIM) for digital power conversion useful for the design of digital switched-mode power supplies, lighting, welding equipment, solar and wireless charging
• USB – 1x USB Type-C port for power and programming
• Expansion – 2x 24-pin headers with GPIOs, ADC, DAC, I2C, USART, LPUART, OAmp, CAN Bus, timer outputs, etc…
• Debugging – 4-pin SWD header
• Misc
  • Reset and Boot keys, 1x user key
  • Power LED, 1x user LED
• Power Supply
  • 3.3V to 20V DC input via USB-C port (Note: the STM32G4 MCU supports USB PD)
  • MicrOne ME6216A33XG voltage regulator with 3.3V output
• Dimensions – 36.28 x 28.14 mm

Getting started with the board may be somewhat challenging, and the documentation could be worked on. For instance, WeAct Studio did not provide any pinout diagram, so you’d have to look up the pin functions matching the pin names from the schematics or the silkscreen markings with the information from the datasheet or technical reference manual, the good ol’ way…

WeAct Studio does provide the PDF schematics, the STM32G4 datasheet and TRM, some code samples (Blink, ADC, RTC, MSC, SPI flash), and the WeAct Studio Download Tool (Windows only) to flash the firmware via USB or UART. You’ll find those resources on GitHub for both the value line board and the hi-resolution line board, but most people will need to check the tools and documentation on the STMicro website as well.

Besides their tiny size, the other key benefit of those boards is their price as the WeAct STM32G4 board goes for $2.92 and $5.79 for the STM32G431CBU6 and STM32G474CEU6 models respectively, including GPIO and SWD headers, but not shipping that adds about $1 in my case. For reference, the official STMicro NUCLEO-G491RE board is sold for $15.

The post WeAct STM32G4 is a tiny board based on STMicro STM32G4 mixed-signal microcontroller appeared first on CNX Software - Embedded Systems News.

CTL Chromebook NL73 Series based on Intel Processor N100 or N200 SoC will be offered with a Snapdragon X35 modem to support the new 5G RedCap (Reduced Capability) standard.

5G RedCap – also known as 5G NR-Light – keeps some 5G features such as low latency, low power consumption, enhanced security, and network slicing while limiting the bandwidth to around a few hundred Mbps. It was initially designed to target industrial IoT applications, but Qualcomm also mentioned its Snapdragon X35 modem could be used in smartwatches and XR glasses when it was first announced, and it might be used in other cost-sensitive devices such as Chromebooks.

Chromebook NL73 “5G RedCap” key specifications:

• Alder Lake N-Series SoC (one or the other)
  • Intel Processor N100 quad-core processor up to 3.4 GHz (Turbo) with 6MB cache, 24EU Intel HD graphics; TDP: 6W
  • Intel Processor N200 quad-core processor up to 3.7 GHz (Turbo) with 6MB cache, 32EU Intel HD graphics; TDP: 6W
• System Memory – 4GB, 8GB, or 16 GB DDR5/5X RAM
• Storage – 128 GB UFS storage (I think it’s the first time I see UFS storage on Alder Lake-N hardware, but it’s indeed supported)
• Display – 11.6- HD touchscreen or non-touch display with 1366 x 768 resolution; 180 or 360° hinge depending on model
• Camera – 720p webcam
• Connectivity
  • WiFi 6E and Bluetooth 5.3 via Intel AX211 wireless module
  • Qualcomm Snapdragon X35 Modem-RF System with support for 5G-RedCap up to 220 Mbps download speeds, up to 110 Mbps upload speeds
• USB – 2x USB-A ports, 2x USB-C ports
• Durability – MIL-STD-810H certification, drop testing from 75 cm, hardened glass, and built-in carry handles

The ML73 Series runs Chrome OS with automatic updates guaranteed until June 2033.

The 5G RedCap Chromebooks from CTL are not available just yet, instead, the company demonstrated a prototype at MWC 2024. Besides the lower cost, using 5G RedCap instead of 4G LTE will also enable private cellular networks. Further information about the existing CTL Chromebook ML73 series can be found on the product page, and the press release has a few more details about the 5G RedCap demo.

Thanks to TLS for the tip.

The post CTL Chromebook NL73 Series to support 5G RedCap with a Snapdragon X35 modem appeared first on CNX Software - Embedded Systems News.

The MINIX Z100-AERO is an actively-cooled Intel Processor N100 mini PC with a form factor similar to the company’s Z100-0dB fanless mini PC but offered with a different choice of interfaces including dual Ethernet (1Gbps + 2.5 Gbps), and three video outputs through HDMI 2.1, DisplayPort, and USB-C connectors.

The mini PC supports up to 32GB DDR4 memory via a single SO-DIMM slot and M.2 NVMe SSD storage,  and features four USB 3.2 Gen 1 ports (5Gbps), plus one full-featured USB-C 3.2 port, and a 3.5mm headphone jack.

MINIX Z100-AERO specifications:

• SoC – Intel Processor N100 quad-core Alder Lake-N processor @ up to 3.4 GHz (Turbo) with 6MB cache, 24EU Intel HD graphics @ 750 MHz; TDP: 6W
• System Memory – 4GB, 8GB, or 16GB DDDR4-3200 via 1x SO-DIMM socket, upgradable to 32GB
• Storage
  • 128GB, 256GB, or 512GB M.2 SSD (PCIe Gen3 x1?), upgradable up to 4TB
  • MicroSD card slot
• Video Output
  • 1x HDMI 2.1 port up to 4Kp60
  • 1x DisplayPort up to 4Kp60
  • 1x DisplayPort via USB 3.2 Gen 2 port up to 4Kp60
  • Up to 3x independent displays supported
• Audio – 3.5mm headphone+mic jack, digital audio output via HDMI
• Connectivity
  • 2.5GbE RJ45 port via Realtek RTL8125BG-CG controller
  • Gigabit Ethernet RJ45 port via Realtek RTL8111H controller
  • Dual-band 802.11b/g/n/ac WiFi 5 and Bluetooth 5.1 via Intel Wireless-AC 9560 module and 2x external WiFi antennas
• USB – 4x USB 3.2 Gen 1 ports (5 Gbps), 1x USB 3.2 (5 Gbps) Type-C port with DisplayPort Alt mode, USB PD support
• Misc
  • Power Button
  • LEDs for Power, LAN (Active, Status)
  • CMOS reset pinhole
    
  • AMI EFI X64 BIOS
• Power Supply – 12V/3A power supply
• Power Consumption – 10-12W (Typical); 25-27W (Turbo)
• Dimensions – 12.7 x 12.7 x 4.3 cm
• Weight – 800 grams
• Certifications – CE, FCC, RCM, RoHS

The MINIX Z100-AERO comes preloaded with Windows 11 Pro, and ships with a 12V/3A power supply and accessories such as an HDMI cable, antennas, a VESA mount, and a user’s manual. As we’ve seen in the Windows 11 review of the MINIX Z100-0dB, a fanless design may not always extract the full performance of the processor, and it was a bit slower in benchmarks than actively cooled designs, so I’d expect the Z100-AERO to perform at a higher level.

Having said that, while the MINIX Z100-AERO provides some improvements with three independent displays and four USB 3.2 ports, there aren’t any 10 Gbps capable USB ports, only WiFi 5 is supported, and having two HDMI ports like on the fanless mini PC may have been more suitable to most people. I can also see the NVMe SSD is supposedly connected to a PCIe Gen 3.0 x1 interface which could negatively impact storage performance and as a result, overall system performance. Maybe it’s just an error in the specs.

The MINIX Z100-AERO mini PC is also cheaper than the Z100-0dB model with a starting price of $219 on the company’s store, and it should soon show up on the MINIX Amazon store as well.

Via AndroidPC.es

The post MINIX Z100-AERO is an Intel N100 mini PC with triple 4K display support, dual Ethernet appeared first on CNX Software - Embedded Systems News.

The Waveshare ESP32-H2-DEV-KIT-N4-M is a development board based on the ESP32-H2, available for only $6.65 on Aliexpress, but you’ll also find it on Amazon and Waveshare’s official store. This is a significant price drop compared to last year’s official Espressif ESP32-H2-DevKitM-1 board, which was priced at $10 without including shipping costs and with a similar design.

In 2021, Espressif Systems introduced the ESP32-H2 to the world. However, it wasn’t until 2023 that they released their first development board. Since then, there haven’t been many products built around this new module. Some exceptions include the Olimex ESP32-H2-DevKit-LiPo, LILYGO T-Panel, and the ESP Thread Border Router/Zigbee Gateway board, all of which feature the ESP32-H2 chip.

Waveshare ESP32-H2-DEV-KIT-N4-M specifications:

• Wireless module – ESP32-H2-MINI-1
  • MCU – Espressif Systems ESP32-H2 32-bit RISC-V microcontroller at up to 96 MHz with 320 KB SRAM, 128 KB ROM, 4 KB LP memory, Bluetooth 5.2 LE/Mesh, and 802.15.4 (Zigbee/Thread/Matter) radios.
  • Storage – 4MB flash storage
  • PCB antenna
  • Dimensions – 13.2×16.6×2.4 mm
• USB – 1x USB Type-C ports with CH334 USB HUB controller and CH343 USB to UART converter
• I/Os
  • 2x 15-pin headers with 19 programmable GPIOs, 2x UART, 3x SPI, I2C, I2S, Remote Control Peripheral, LED PWM, Full-speed USB Serial/JTAG Controller, GDMA, MCPWM…
  • Compatible with the pinout of the ESP32-H2-DevKitM-1 development board
• Misc – Power LED, RGB LED, Boot and Reset buttons
• Power Supply
  • 5V via USB Type-C port
  • 5V to 3.3V LDO
• Dimensions – 51.60 x 25.40 mm

The official ESP32-H2 devkit from Espressif features two USB-C ports: one port connects to the ESP’s full-speed USB 2.0 Host port, and the other connects to a USB to UART converter. Waveshare simplified this setup by including a USB HUB controller IC on the board, effectively combining the two USB connections into one.

You can check out Espressif’s documentation to find more hardware details and get started guide with the ESP-IDF framework.  Waveshare also provides its own Wiki page to get you started with the ESP-IDF using VS-Code and ESP32 Arduino Core 3.0.0 using the Arduino IDE.

The module comes in two variants one with soldered headers and another without soldered headers. As mentioned at the start of the article, the module can be purchased on Aliexpress, Amazon, and Waveshare for $6.65 and up.

The post Waveshare ESP32-H2-DEV-KIT-N4-M – A Low-cost ESP32-H2 development board going for $6.65 appeared first on CNX Software - Embedded Systems News.

The LuckFox LuckFox Pico Pro and Pico Max are two new Rockchip RV1106-powered development boards that offer a Linux-based development platform for IoT applications. The boards feature 10/100M Ethernet, up to 256MB DDR2 memory, and a 0.5TOPS NPU for AI tasks. With support for Buildroot and Ubuntu 22.04, this board can be used for smart homes, remote monitoring, and other AI-enhanced projects.

Last year, we covered the LuckFox Core3566, a Raspberry Pi Compute Module 4 alternative, and the LuckFox Pico with its RV1103 SoC which has a similar form factor and similar features to these new modules. But the new modules are built around the RV1106 SoC which features an Arm Cortex-A7 processor (up to 1.2GHz),  a RISC-V co-processor, a 0.5 TOPS NPU for AI tasks, and a 4M @ 30fps ISP for high-quality image processing.

LuckFox Pico Pro and Pico Max Specifications

• SoC – Rockchip RV1106G2
  • CPU – Arm Cortex A7 @ 1.2GHz with an integrated RISC-V co-processor.
  • NPU– 0.5 TOPS, supports INT4/INT8/INT16, TensorFlow/MXNet/PyTorch/Caffe/Onnx NN
  • ISP – 5MP high-performance, HDR, WDR, 3DNR, 2DNR, sharpening, defogging, fisheye and gamma correction, feature detection
  • VPU – 3072×1728 (5M) @ 30fps H.265/H.264 encoding, 16M @ 60FPS JPEG snapshot
• System Memory
  • LuckFox Pico Pro – 128MB DDR2
  • LuckFox Pico Max – 256MB DDR2
• Storage – 256MB SPI NAND flash, microSD card slot
• Connectivity – 10/100M Ethernet RJ45 port
• Camera Interface – MIPI CSI 2-lane support for camera modules.
• Default camera specifications
  • Type – Color Camera
  • Image Sensor– SC3336
  • CMOS Size– 1/2.8″
  • Resolution– 3 MP (2304×1296)
  • Aperture – F2.0
  • Field of View – 98.3°
  • Focal Length – 3.95mm
  • Distortion – < 33%
  • Max Frame Rate – 2304×1296 @ 30fps
  • Focus Mode – Manual focus
• USB – USB 2.0 Host/Device Type-C port
• GPIO – 2x 20-pin headers with 26x GPIO pins
• Misc – ACT-LED; BOOT and RESET buttons
• Power Supply – 5V via USB-C port
• Dimensions – 72 x 21 mm

LuckFox provides two versions of their device: the Pico Pro, equipped with 128MB DDR2 memory, and the Pico Max, which comes with 256MB of memory. Both models feature 256MB of SPI NAND flash for storage.

The board offers various interfaces such as MIPI CSI, up to 26 GPIOs, UART, SPI, I2C, and USB. It also includes a 10/100M Ethernet controller with MAC PHY, a MicroSD card slot, a USB Type-C connector, and a 2-lane MIPI CSI camera interface.

The company says that the LuckFox Pico Pro and Pico Max module supports the SC3336 3MP Camera, featuring a SmartSens SC3336 sensor. This 3M MIPI CSI module is said to deliver “superior low-light capabilities, thanks to its high sensitivity and excellent signal-to-noise ratio”. We have also seen this camera used in the Firefly CT36L AI Smart Camera which we wrote about just last month.

In terms of software, LuckFox Pico Pro and Pico Max support Buildroot and Ubuntu 22.04 (server), and you’ll find the instructions to get started on the Wiki.

The Luckfox Pico Pro and Pico Max can be purchased on Aliexpress starting at $12.71 and on Amazon for $22.90 (Pico Max only) The SC3336 3MP camera module can be purchased separately on Aliexpress or Amazon for about $10 and up.

The post LuckFox Pico Pro and Pico Max – Rockchip RV1106 powered boards with 10/100M Ethernet and camera support appeared first on CNX Software - Embedded Systems News.

Arduino Nano 33 BLE Rev2 is an update to the Arduino Nano 33 BLE board launched in 2019 that features two IMU sensors instead of one with the BMI270 6-axis accelerometer and gyroscope and the BMM150 3-axis magnetometer and also comes with a few changes made after feedback from users.

The new board is still powered by an nRF52840 Bluetooth LE module (u-Blox NINA B306) and remains Arduino Nano compatibility with two rows of 15-pin headers, but replaces the 9-axis IMU with the BMI270 and BMM150 chips, adds new pads and test points for USB, SWDIO, and SWCLK, a new VUSB soldering jumper, and brings changes to the power circuitry.

Arduino Nano 33 BLE Rev2 specifications:

• Wireless Module – U-blox NINA B306 module
  • SoC – Nordic Semi nRF52840
    • MCU Core – Arm Cortex-M4F microcontroller @ 64MHz
    • Memory and storage – 1MB Flash, 256KB RAM
    • Bluetooth 5.0 LE
      • Up to 2 Mbps link rate, long-range support
      • +8 dBm TX power
      • -95 dBm sensitivity
      • Power consumption (@ 3.3V?)
        • 4.8 mA in TX (0 dBm)
        • 4.6 mA in RX (1 Mbps)
    • 802.15.4 radio – Thread, Zigbee support
• Expansion
  • 2x 15-pin expansion headers with GPIO, SPI, I2C, USART, PWM, ADC, DAC, reset, and power signals (+3.3V, GND, Vin, AREF, …)
  • The board only supports 3.3 V I/Os and is NOT 5V tolerant
• Sensors
  • BMI270 6-axis accelerometer and gyroscope
  • BMM150 3-axis magnetometer with 0.3μT resolution
• Debugging
  • Via micro USB port, and 6-pin solder pads (SWD)
  • New pads and test points for USB, SWDIO, and SWCLK
• Misc
  • Reset push-button
  • Green power LED, SPI SCK LED, user RGB LED
• Power Supply
  • 5V via micro USB port or VUSB pin (with soldering jumper) on header
  • 5V to 18V via Vin pin
  • MP2322 DC-DC step-down converter – Regulates input voltage from up to 21V with a minimum of 65% efficiency @ minimum load; more than 85% efficiency @12V
• Dimensions – 45 x 18 mm (compatible with Arduino Nano)
• Weight – 5 grams

The Arduino Nano 33 BLE Rev2 remains compatible with the first generation of the board and you can program it with Arduino or MicroPython. Just make sure to update board support and libraries to the latest version. Further technical details, instructions to get started, and some tutorials can be found on the documentation website.

At first, I was a little confused, as it’s not the first “Nano 33 Rev2” board we’ve covered, and last year, the Arduino Nano 33 BLE Sense Rev2 was introduced with the same BMI270 and BMM150 chips as found in the board covered today, plus extra sensors namely the MP34DT06JTR microphone, APDS9960 gesture, light, proximity sensor, LPS22HB barometric pressure sensor, and Renesas HS3003 temperature & humidity sensor.

The new Arduino Nano 33 BLE Rev2 can be purchased for$24.17 / 22.80 Euros without headers or $26.92 / 25.40 Euros with headers on the Arduino store. A few additional details may also be found in the announcement.

The post Arduino Nano 33 BLE Rev2 board features BMI270 six-axis IMU and BMM150 magnetometer appeared first on CNX Software - Embedded Systems News.

The STMicro VL53L9 sensor is the latest addition to the company’s FlightSense product family. The direct Time-of-Flight (ToF) 3D LIDAR (light detection and ranging) sensor offers a resolution of up to 2,300 zones. The module is described as all-in-one and easy to integrate. It comes in a small, reflowable package that contains all the necessary components for sensing objects and processing images.

The sensor features an array of single photon avalanche diodes (SPADs) for photon detection, a post-processing SoC, and two vertical surface emitting lasers (VCSELs) powered by a dedicated bipolar-CMOS-DMOS (BCD) VCSEL. The VL53L9 is a multi-zone ToF sensor similar to the VL53L7CX and the VL53L8, meaning that it offers multi-zone distance measurements up to 54 x 42 zones with a wide 54° x 42° field-of-view.

Unlike most IR sensors, the VL53L9 sensor uses backside illumination direct ToF technology to ensure absolute distance measurement, regardless of the target color and reflectance. It captures 2D IR images and 3D depth map information at <5cm up to 10m using dual-scan flood illumination, a unique system according to the company. It can then stream the 2D image, depth information, and a confidence map at a 60Hz frame rate (which the company claims is the fastest on the market).

STMicro VL53L9 key features and specifications:

• Wide field of view
  • Emitter: 940 nm invisible light vertical cavity surface emitting laser (VCSEL) and integrated analog driver
  • 71° diagonal field-of-view (FoV) using metasurface optical elements (MOE) on both transmitter and receiver
  • Receiving array of single photon avalanche diodes (SPADs)
• Speed and accuracy
  • Resolution: up to 54 x 42 (2,268) separate zones and binning options
  • Post-processing SoC can stream a 2D infrared (IR) image with a depth and confidence map
  • Range: <5 cm up to 10 m
  • Frame rate: 60Hz
  • Scan by two vertical-cavity surface-emitting lasers (VCSEL) flood illumination
  • Histogram processing and algorithmic compensation
• All-in-one module with integrated SPAD sensor and VCSEL power management integrated circuit (PMIC)
• Power – Dual power supply operation: 1.2V and 3.3V
• Compatible with a wide range of cover glass materials
• Dimensions – 12.8 x 6.1 x 4.6 mm (miniature reflowable package)
• Safety – Class 1 certification for eye-safe operation and additional skin protection in normal use

The company expects the VL53L9 sensor to find applications in:

• Telephoto zoom camera assist.
• Augmented reality/virtual reality (AR/VR) enhancement.
• Laser-assisted autofocus (LAF).
• 3D room mapping and obstacle detection or SLAM (simultaneous localization and mapping) for robotics applications
• Content management (liquid level control, load in trucks, tanks, waste bins)
• Gesture recognition
• Smart buildings and smart lighting (user detection to wake up devices)
• Internet of Things (user and object detection)

In other news, ST has also announced a design win for the earlier-released VD551, a low-power, indirect Time-of-Flight (iToF) sensor. A design win means that a company has ordered 500 or more units of the sensor to integrate into their product. In this case, that company is Lanxin Technology, a Chinese company with a focus on mobile robot deep vision systems. MDDVS, a subsidiary of Lanxin, has selected the VD551 to bring high-accuracy depth-sensing to its 3D camera.

Mass production of the STMicro VL53L9 sensor is scheduled for early 2025 and samples are currently available for select customers. Pricing details and samples can be requested at local ST sales offices. You can find more information in the press release and the VL53L9 product page.

The post STMicro VL53L9 is a high-resolution, direct Time-of-Flight 3D LiDAR sensor module appeared first on CNX Software - Embedded Systems News.

Intel Core Ultra 5 134U and Core Ultra 7 164U are new 12-core Meteor Lake processors with 9W PBP (Processor Base Power) and 30W Max Turbo Power (MTP) that appear to have recently been added to Intel Ark.

Intel just announced Intel vPro Platform support for the Core Ultra Meteor Lake processors to make those more suitable to business customers with improved security through AI-powered Intel Threat Detection Technology, better manageability with the Intel Device Discovery comprised of cloud-based tools for managing PCs remotely, and enhanced stability thanks to Intel’s Stable IT Program aiming to validate and ensure Windows 10 and Windows 11 compatibility.

But Liliputing also notes that Intel has added some 9W parts besides the existing 15W parts such as the Intel Core Ultra 5 125U or Core Ultra 7 165U, namely the Core Ultra 5 134U and Core Ultra 7 164U. They are quite similar to the other 15W Meteor Lake-U parts with 12 cores in 2 P-Core+8 E-Core+2 LP-Core configuration, 14 threads, 12MB cache, and Intel UHD graphics with 4 Intel Xe GPU cores, but there are some differences as shown in the table below.

	Core Ultra 5 134U	Core Ultra 7 164U	Core Ultra 5 125U	Core Ultra 5 135U	Core Ultra 7 155U	Core Ultra 7 165U
PBP/MTP	9W/30W		15W/57W
Base frequency (P/E/LP)	700 MHz / 500 MHz / 400 MHz	1.1 GHz / 700 MHz / 400 MHz	1.3 GHz / 800 MHz / 700 MHz	1.6 GHz / 1.1 GHz / 700 MHz	1.7 GHz / 1.2 GHz / 700 MHz	1.7 GHz / 1.2 GHz / 700 MHz
Max Turbo Freq. (P/E/LP)	4.4 GHz / 3.6 GHz / 2.1 GHz	4.8 GHz / 3.8 GHz / 2.1 GHz	4.3 GHz / 3.6 GHz / 2.1 GHz	4.4 GHz / 3.6 GHz / 2.1 GHz	4.8 GHz / 3.8 GHz / 2.1 GHz	4.9 GHz / 3.8 GHz / 2.1 GHz
Max GPU frequency	1.75 GHz	1.80 GHz	1.85 GHz	1.9 GHz	1.95 GHz	2.00 GHz
RAM	64GB Up to LPDDR5/x 6400 MT/s		96GB Up to LPDDR5/x 7467 MT/s, up to DDR5 5600 MT/s
PCIe	12-lane PCIe 4.0 and 3.0 1 (x4) Gen4 + 4 (x1,x2,x4) Gen4 + 4 (x1,x2,x4) Gen3		20-lane PCIe 4.0 3 (x4) Gen4 + 8 (x1,x2,x4) Gen4
Package	FCBGA2551 (23 x 19 mm)		FCBGA2049 (50 x 25 mm)

The new 9W Meteor Lake U-series Type4 processors come in a smaller package, offer fewer PCIe 4.0/3.0 lanes, and support (only) up to 64GB LPDDR5 RAM, while the 15W SKUs can handle up to 96GB. They’ll likely be found in thinner laptops and mini PCs be it actively cooled or fanless. Other notable features include Intel AI Boost NPU for AI acceleration, Thunderbolt 4 interfaces, and support for WiFi 7 and Bluetooth LE audio which are common to all Meteor Lake processors.

The post Intel Core Ultra 5 134U and Core Ultra 7 164U are 9W Meteor Lake processors appeared first on CNX Software - Embedded Systems News.

A few days back, Inventia introduced the MT-058 cellular IoT telemetry module built around Nordic Semiconductor’s nRF9160 System in Package (SiP) module with support for LTE-M/NB-IoT connectivity with GNSS. The device is designed for low-power, battery-operated industrial applications such as water metering and environmental monitoring.

The module is IP68-rated and offers a combination of digital and analog inputs. It also comes with a battery that can last up to 5 years and there is also support for external sensor attachment for data logging applications.

MT-058 Cellular IoT telemetry module specification:

• Nordic nRF9160 SiP with
  • 64 MHz Arm Cortex-M33 application processor with 1 MB flash, and 256 KB RAM
  • Connectivity – LTE Cat M1 and NB-IoT
• SIM card options – Standard 2FF SIM card slot, with an option for a soldered Machine Identification Module (MIM).
• Antenna – SMA connector for attaching an external antenna.
• Inputs – 5 inputs for counters, up to 250 Hz, min 2ms pulse.
• Sensors
  • Temperature and humidity sensor mounted outside the housing
  • 1-wire interface for connecting an external temperature sensor
• Misc features
  • Alarm – Inputs can also be configured as alarm inputs to signal events such as the opening of a cabinet door.
  • Data Logging – Built-in Flash memory for data logging, allowing for effective energy management and data recording even without a radio signal.
  • Diagnostics and Configuration – Remote configuration, updates, and diagnostics; The device also includes diagnostic LEDs and a USB-C port for local setup.
• Battery – Up to 5 years of operation with the built-in battery pack depending on device configuration.
• Dimensions – 151 x 80 x 60 mm
• Durability – IP68-rated housing, making it waterproof down to 2 meters for 24 hours.
• Warranty – Comes with a 3-year warranty, ensuring reliability and customer satisfaction.

As we mentioned earlier the device has five binary inputs that can be configured as counter inputs, counting pulses with a frequency of up to 250 Hz, perfect for reading water and flow meters. These inputs can also be configured as alarm inputs to trigger alarms, like if a cabinet door opens. The device has one analog input which can be used to measure temperature and humidity if required.

On the software side of things, the module utilizes Nordic’s nRF Connect SDK, providing developers with pre-certified, precompiled downloads for rapid application development, and connects to the Dataportal.pl IoT platform that offers a cloud-based SCADA solution for data visualization and analysis, along with alarm notifications.

In our previous posts, we have seen the nRF9160 (part of the nRF91 series) used in various projects like the Actinius Icarus SoM DK, Conexio Stratus, nRF9160 Feather LTE IoT and GPS Board,  Ruuvi Node, and others. Feel free to check those out if interested.

At the time of writing the company has not revealed any pricing information. You can find more information about the device on inventia’s products page or in Nordic’s press release.

The post Inventia MT-058 cellular IoT telemetry module is powered by Nordic Semi nRF9160 SiP appeared first on CNX Software - Embedded Systems News.

If you could not care less about WiFi, the Waveshare RP2040-BLE module combines a Raspberry Pi RP2040 microcontroller with an unnamed Bluetooth 5.1 dual-mode chip, and supports a USB-C add-on board useful during development.

There are several Raspberry Pi RP2040 boards with a WiFi and Bluetooth module including the Raspberry Pi Pico W, but if you are only interested in using Bluetooth, the RP2040-BLE board from Waveshare may be more power-efficient and its design may be more suitable for integration into products.

Waveshare RP2040-BLE specifications:

• MCU – Raspberry Pi RP2040 dual-core Arm Cortex M0+ microcontroller up to 133 MHz with 264KB of SRAM
• Storage – 2MB of onboard Flash memory.
• Connectivity
  • Bluetooth 5.1 dual-mode function (BLE and classic)
  • Controlled by serial AT commands
  • Ceramic antenna
• Expansion
  • 24x through and castellated holes with 14x multi-function GPIO pins. 2x SPI, 2x I2C, 2x UART, 3x 12-bit ADC, 14 x PWM
  • FPC connector for board with USB-C port, Reset and Boot buttons
• Misc – Temperature sensor.
• Power Supply – 5V via VBUS pin or FPC connector
• Dimensions
  • RP2040-BLE board – 33.5 x 21 mm
  • USB-C board – 18 x18 mm

The RP2040-BLE module supports drag-and-drop programming using mass storage over USB and the MicroPython and C/C++ SDKs like the original Raspberry Pi Pico (W), as well as Arduino programming. You’ll find resources to get started in the wiki including details about the Bluetooth APIs as well as resources such as (partial) schematics since we’re not told which Bluetooth chip the module is using…  It’s unclear whether it’s compatible with the Raspberry Pi Pico Bluetooth implementation.

 

Waveshare RP2040-BLE module can be purchased for about $11 on Aliexpress or Amazon with free shipping, and the kit with the USB-C board and FPC cable goes for 90 cents extra on Aliexpress. So depending on your location, it may be more expensive than the Raspberry Pi Pico W, but it’s quite smaller. You’ll also find the module and kit on the Waveshare shop for $6.49 and $7.49, but those prices do not include shipping.

The post Waveshare RP2040-BLE is a Raspberry Pi RP2040 module with Bluetooth LE connectivity appeared first on CNX Software - Embedded Systems News.