Hi Jeremias,
We switch on the systems in the morning and turn them off in the evening. Yes, the occurrence rate is very high.
I believe rebooting the devices is not a sufficient test. If you want to try and reproduce this issue, I assume you should at least run one container using the ADC driver - just speculating.
We switch off and on our devices using a hardware switch.
I believe that we have to let the supplier of our boards investigate further, if you believe there can be an issue with the board.
Thank you for now
I believe rebooting the devices is not a sufficient test. If you want to try and reproduce this issue, I assume you should at least run one container using the ADC driver - just speculating.
Well I was thinking of this, but from what I can understand of your situation your container doesn’t even run properly because /dev/verdin-adc1 is missing. If I understood this right then that would mean your code doesn’t even have a chance to execute or do anything with ADC.
We switch off and on our devices using a hardware switch.
Okay that’s good to know.
I believe that we have to let the supplier of our boards investigate further, if you believe there can be an issue with the board.
Well I’m not 100% sure, I’m also speculating on this point. Though the carrier board is certainly one difference between your setup and mine.
On one of these devices could you get the contents of this file /sys/kernel/debug/devices_deferred. It could contain some logs showing why the ADC driver may not have been probed or initialized properly.
Best Regards,
Jeremias
Hi @jeremias.tx
What I meant is that using the ADC could affect the system during shutdown, with a problem that manifests itself upon the next boot.
I checked the /sys/kernel/debug/devices_deferred/ file and it’s empty - yes, I was root. I looked at it right after the problem occurred this morning, but I first shutdown the system to access the NIC port.
For now, we have disabled the ADC readings. I hope that will be sufficient. We will not be able to detect this error in the future - hopefully.
Thank you all for the help and support.
Hi @antonio.astorino ,
What I meant is that using the ADC could affect the system during shutdown, with a problem that manifests itself upon the next boot.
If the problem is related to hardware it may be possible, but I’m not sure the ADC would cause something that persists during a power cycle, at least related to software.
Do you have one of our carrier boards that you can test to see if this problem happens there?
For now, we have disabled the ADC readings. I hope that will be sufficient. We will not be able to detect this error in the future - hopefully.
How important is the ADC readings for your use case? Just to be sure, you’re not blocked by this problem, at least for now, is this correct?
Best regards,
Lucas Akira
Dear @lucas_a.tx,
Sorry for disappearing.
We are selling our first devices with the ADC feature disabled. Its purpose was to check the internally generated voltages and trigger alarms if those voltages were not within specs. It’s a feature I haven’t designed and I’m not even sure it makes sense since also the reference voltage was generated… Anyway, we are not blocked.
Nevertheless, it’s a feature for which we paid money and should work. It is therefore in our interest to know if the problem comes from the carrier board. We are busy with other projects at the moment and we cannot invest more time in investigating this issue. And it’s unfair to delegate you guys especially because I have the feeling the problem is caused by our carried board.
I believe when we power the board, the transient is too slow. The input voltage of the board is 24 V, internally stepped down to 5 V and 3.3 V using buck converters (I think cascaded). What I dislike is the LC filter placed next to the power connector. The inductance is very big and it has already lowered because (of course) it was creating overvoltages when the plug was disconnected, burning some silicon here and there. However, this inductor (I think) causes the initial transient to be slower than the Verdin module expects. My main supporting argument is that I noticed the first boot after a long time unpowered is the most likely to cause the ADC not to work. That is because (I think) all the capacitors (also that of the LC) are not charged and the voltage reaching the Verdin module increases too slowly.
Concluding, if you have the chance to power the module using a slowly increasing (even by hand) voltage, and verify that the ADC does not come up, that would give me the chance to open further investigations.
As you can see, I did disappear but I also thought a lot about this issue 
Thank you and everybody else for your kind support.
Best regards,
Antonio Astorino
