Issue with AD74412R current & voltage outputs after long duration

Hi Canberk,

I will be happy to support you to help uncover the challenge in your system with AD74412R.

Thank you for your inputs, you did good job with initial debugging, let me ask couple of clarifying questions, that may be perhaps spark some ideas or lead you to more experiments. 

1# Let me double check, so there is no assumptions (based on your inputs): Is CCOMP capacitor connected, what is its value? Can you double check, if it is connected in schematic - between Rsense and BAV99.

2#

• In current output mode, when set to 10 mA, the output occasionally dips to about 8.3 mA for a very short time (~50–100 ms) and then recovers.

Is signal repeating with some period (for example every 1s or so?)? Is the experiment it repeatable with each board and individual channel? 

• In voltage output mode, we also observed similar irregular behavior.

Could you clarify what type of load is connected during this experiment? 

• This happens on multiple channels and multiple boards, but only after the system has been running for a long duration (≈6 hours or more).

Can you check the AD74412R chip temperature, once issue is observed. What is the temperature reading?

Please build in temperature sensor, which is measuring die temperature:

3# We also noticed that when a channel is set to current input and connected to another AD74412R board’s output set at 0 mA, the input reading shows short fluctuations (usually ~0.3 mA, sometimes up to 12 mA). If the input is left open, there are no fluctuations. Interestingly, a multimeter in series reports 0 mA while the AD74412R still sees spikes.

What was the ADC configuration of the monitoring (current input) chip AD74412R (sampling frequency). It seems to me, that this is occasional dynamic event (~ 50-100ms), multimeter is typically bandwidth limited and is measuring average value, for such reason you don't see any changes there. Could you monitor the same using the oscilloscope (triggering at the dip)? It would be handy if you can share oscilloscope capture of the dip.

Is there anything else connected to I/OP_x screw terminal beside recommended front end? (Any switches or DO circuit etc. or any other customer circuit?)

4# Additional notes:

• Board is 4-layer with a complete ground plane.

Could you comment please, how and where is AGND_SENSE connected - PCB layout?  

• Power supplies for the board and IC as well are stable, no drops observed.

Could you clarify, how did you monitored power supplies? Could you use oscillospe and perhaps synchronize with the dip event?

May I ask you also to monitor Voltage Reference (REFIN). Are you using internal reference of the AD74412R?

• We are monitoring the IC’s diagnostics as well. During the issue, two current-output channels reported “open circuit” errors even though there was no cable fault. It’s possible that all four channels experienced the error, but our firmware only raises a warning after seeing the same diagnostic flag a consecutive number of times.

It would be great to understand, if any other Alert was flagged in by reading: ALERT_STATUS register. 

Perhaps consider using only simple load as resistor and observe if same behavior takes place. 

• We followed the datasheet rules (<10 pF to GND at SENSEL/SENSEH/CCOMP, 2 kΩ resistors, fly cap close to CPUMP pins).

 Would it be possible to also monitor AD74412R voltages as ALDO1V8, ALDO5V, AVDD, AVSS? (Using Oscilloscope)

5# Board differences vs. ADI eval board:

• MOSFET: DMP6110SFDF-7 instead of FDC5614P

For now, it does not seem to be concern, however if we won't be able to narrow issue down, we can simply bypass the external cascode on the "faulty" channel and observe the difference. 

• Isolation diode: BAV99WT1G instead of BAV99LT3G

This seems to be no concern.

• CLOAD at screw terminal: 100 nF instead of 68 nF

If should not be a concern once CCOMP capacitor is connected. 

What could be the reason behind this issue? Any help or experience with similar AD74412R stability problems would be greatly appreciated.

So far, I am not certain of the reasoning and not similar issue comes to my mind. It is particularly tricky as it is related to "long time operation".

Please comment on my inputs, use oscilloscope to monitor REFIN, Power supplies and Dip event. Look for noise or any "suspicious behavior". 

Please share some oscilloscope, especially dip one, it would be greatly appreciated.  

Regards,

Arnost

Hi Arnost,

Thanks a lot for your detailed suggestions and support. After deeper analysis, we found that the issue was purely on the software side, not hardware. Once corrected, the AD74412R outputs have been stable without dips even during long runs.

Appreciate your help in narrowing this down!

Best,
Canberk