PPK2 stops measuring on long sessions

[torkelatgenet]

Hi Nordic,

I've tried to do some long duration measurements first 5 days, but also just 24 hours. Both times the Power Profiler desktop app was still running when i came back before the end of the session, and the DATA LOGGER view was scrolling past, but there were no measurements, no blue line. Stopping the session and scrolling back showed that it had only logged for about 16hours. When the software is in this state, starting a new session still does not measure anything, only the graph area scrolls by. Changing sampling rate or duration changes nothing, except setting the sampling rate to 100,000 where it works. My guess would be that there is a problem with the down-sampling of the signal. Are you aware of this bug? Is there a fix? Am I doing something wrong?

The ppk2 is connected directly to my laptop.

Running with downsampling:

Running with 100khz sampling:

Fullscreen nrfconnect-system-report-2021-05-25T07-39-38.349Z.txt Download

# nRFConnect System Report - 2021-05-25T07-39-38.349Z

- System:     Dell Inc. Latitude 5400
- BIOS:       Dell Inc. 1.10.1
- CPU:        1 x Intel® Core™ i7-8665U 1.90 GHz 8 cores (4 physical)
- Memory:     2.3 GB free of 15.5 GB total
- Filesystem: /dev/mapper/ubuntu--vg-root (ext4) 466.3 GB 44.9% used

- OS:         Ubuntu (20.04.2 LTS) linux x64

- Versions
    - kernel: 5.4.0-73-generic
    - git: 2.25.1
    - node: 12.13.0
    - python: 2.7.18
    - python3: 3.8.5

- Connected devices:
    - /dev/ttyACM1: FD158A69D96D 
    - /dev/ttyUSB0: DM1MWYTB 
    - /dev/ttyUSB1: FT9JAR75 
    - /dev/ttyACM0: 000683454330 PCA10056

Also, I have a few questions/suggestions about the ppk2:

1. Would it be possible to let the user chose the size of the downsampled buffer or is there a performance limitation?

2. Does the hardware have a way to measure the input voltage to the ppk2 in ampere-meter mode? If so will it be added to the desktop software or do I need to look into scripting? This would be very useful for battery-powered use-cases where the battery's dynamic characteristics are significant.

Thank you for your help!

Mvh
Torkel

[Øyvind]

Hi Torkel, 

I will need to ask our PPK2 team. Will get back to you within tomorrow.

Kind regards,
Øyvind

[torkelatgenet]

Hi again Øyvind,

I tried with the developer console open. It lasted a while but it stopped after a day or two.

Best regards,
Torkel

[Øyvind]

Hi Torkel, 

The system report that you provided, is this right after it stopped? Are there any indication in your system that the memory is full or similar, when the issue appears?

Working with the team to replicate this, however, they have not been successful yet.

Kind regards,
Øyvind

[torkelatgenet]

Hi,

The report is generated at another time. I will run another test and monitor the system memory. Is the memory estimation provided not necessarily accurate?

I do notice that nrfconnect has quite a lot of processes with some RAM usage combined:

I'll get back to you with the result in a few days.

-Torkel

[Øyvind]

Great, thanks! Please update when you have more information. 

-Øyvind

[torkelatgenet]

Hei Øyvind!

It took 30+ hours, but it did happen again. I've just pasted two large screenshots for you to look at. The screenshots were taken at the same time. The bug happened ~30 minutes ago or 2040 seconds. The system monitor is scaled to 6000 seconds. So you can see that there was still RAM left, though some swap-usage anyway. I also ran 'free - m' command around the time of the screenshots.

# nRFConnect System Report - 2021-06-23T12-01-21.570Z

- System:     Dell Inc. Latitude 5400
- BIOS:       Dell Inc. 1.10.1
- CPU:        1 x Intel® Core™ i7-8665U 1.90 GHz 8 cores (4 physical)
- Memory:     206.3 MB free of 15.5 GB total
- Filesystem: /dev/mapper/ubuntu--vg-root (ext4) 466.3 GB 45.3% used

- OS:         Ubuntu (20.04.2 LTS) linux x64

- Versions
    - kernel: 5.4.0-74-generic
    - git: 2.25.1
    - node: 12.13.0
    - python: 2.7.18
    - python3: 3.8.5

- Connected devices:
    - /dev/ttyACM1: FD158A69D96D 
    - /dev/ttyUSB0: DM1MWYTB 
    - /dev/ttyUSB1: FT9JAR75 
    - /dev/ttyACM0: 000683454330 PCA10056

➜  ~ free -h     
              total        used        free      shared  buff/cache   available
Mem:           15Gi       9,5Gi       297Mi       2,3Gi       5,6Gi       3,3Gi
Swap:         979Mi       682Mi       297Mi
➜  ~ 

I'm sorry for the messy post, but I think you will find your way around.

Regards,

Torkel

[Øyvind]

Torkel, 

Thanks for providing this information. Our SW team has been able to reproduce the issue, and are working on the issue. 

Kind regards,
Øyvind

[Oette]

Hello,

I'm having the same issue. Is there any updates on it?

I have tried it on both a Windows 10 and a Linux Ubuntu machine. On the Linux machine it happens after several hours and on the Windows machine it sometimes happens after hours and sometimes almost immediately after I start sampling.

Thanks

[Øyvind]

Hi, 

yes, I did get an update yesterday. The team responds:
"...working on releasing a potential fix for this problem with PPK 3.1.3 as we speak. So it would be great if you could ask the customer to try to reproduce with 3.1.3 when they are able to update."

Not sure when this update will be available, please keep your eyes open and test as soon as available to you. 

Thanks!

Kind regards,
Øyvind

[Øyvind]

Hello, 

v3.1.3 is now released. 

Kind regards,
Øyvind

[Oette]

Hello,

I updated to v3.1.3 and now it doesn't stop measuring, but there seems to be lots of corrupted samples which results in gaps in the graph:

It seems to occur when the current is varying between <50mA and >50mA. I assume this could be related to PPK2 switching the measurement range frequently.