ADS1112: Intermittent Failure Mode where All Ports Read Zero

Hello,

Please let me know what happens with the ABA swap you performed, this was a good strategy. 

As for possibilities of device registers being reset, the power level would be the most relevant factor. If the device browns out, where a significant dip in the power occurs but not a total zero of the power supply. For this device, when it powers up, it automatically performs a reset. and sets all of the bits in the configuration register to their default settings, a brown out effect could initialize a reset process.

Regards

Cynthia

Thanks.  I did not connect a scope or I2C sniffer, so I was trusting what the software told me.  The software is supposed to be setting the config register with each read and then checking the Ready bit before reading the data register, but I cannot confirm that's happening.  

Since I replaced the IC, the problem has not occurred.  Is there any way TI test it or do some analysis if I sent it to a TI lab?

That is great to hear. 

I would suggest to run the new device for sometime before concluding that the original device was the source of the issue. If the original device was damaged somehow that resulted in this behavior, this new device could result in the same. 

Confirming communications is also a good idea, the methods you mentioned would help to confirm that the firmware is in fact communicating as expected. 

You can reach out to your field/sales TI representative. If you do not have one, I can put you in contact without someone in your area. They will be able to help you file the correct information to initiate a return. Although, I believe that if the device was damaged during your use, this will not really be much of help, but your representative would know best. 

Regards

Cynthia

Cynthia said:

I would suggest to run the new device for sometime before concluding that the original device was the source of the issue. If the original device was damaged somehow that resulted in this behavior, this new device could result in the same. 

I swapped the part with a part from my supply in the lab, making no other changes.   -->  We have not seen this issue since.

The problem has occurred on another system.  As before, power cycling resolves the issue. 

Hypothesis: This board build used a bad reel of ADS1112.  They bought some parts from brokers.  The parts could be counterfeit.

My plan is to swap the ADC chip with another one from this same board build.  Then, if it happens again, I will swap the part with an ADC chip from my supply in the lab, which I've never had a problem with.  If the problem does not occur, this points to the recent build having used a bad batch of chips.  

Could I sent the parts to TI for testing/analysis?  

Cynthia said:

You can reach out to your field/sales TI representative. If you do not have one, I can put you in contact without someone in your area. They will be able to help you file the correct information to initiate a return. Although, I believe that if the device was damaged during your use

I used to, but I'm not sure who it is anymore.  Could you look it up?  

I could provide the part that's showing the issue and an never-powered part from the same build.  I would rather not provide a good part from my supply because they are hard to come by now. 

Charles, 

That is very unfortunate, if they do turn out to be counterfeit parts. Given the current state, it is possible that they are. 

Your plan sounds very good, this would help to confirm performance. looking at the parts, do they show similar package markings? Would you please send a picture of the top markings of the possible counterfeit device?

I have contacted a field engineer about your situation nd decide what best steps to move forward. 

Regards

Cynthia 

Cynthia said:

Your plan sounds very good, this would help to confirm performance. looking at the parts, do they show similar package markings? Would you please send a picture of the top markings of the possible counterfeit device?

Here is the part.

It turns out the CM bought this part 2.5 years ago from Digikey, not brokers.  Date code = Aug-2018.  Lot code = 8857351ML8.

I think it's a longshot that it's counterfeit or that there's any problem with the lot, but I appreciate your investigating.

Cynthia said:

I have contacted a field engineer about your situation nd decide what best steps to move forward. 

Is this the support engineer for our geographic area or for this product line?  Can you send me their contact info?  It would be good for me to know who our field sales rep is.