TLV320AIC24K: Microphone input ADC conversion failure

Hello Archana, 

The AIC24k is an older device, but the ADC architecture in that device was leveraged in many other devices in the TLV320AIC3xxx devices.  We have not seen this problem before.  All of our devices undergo a strenuous testing to ensure functionality in the rated conditions, this includes ESD and latch up.  

Without any additional information it is hard to imagine what may have happened here,  can you give more information? 

best regards

-Steve Wilson

Hi Steve,

Thank you for your reply.

This part is used in a product in flight.  The failure was observed in just one unit, out of thousands of the product that are successfully flying.

When we received the unit from the field, we were still able to see the failure where the ADC was not working.

The unit was working fine for quite some time in flight before the actual failure was observed and the unit was returned back.

We removed the faulty codec IC and replaced the same IC back, only to find it working back again.

We did multiple tests in lab, but could not recreate the failure.

We are clueless as to why the latchup happened.  Do you suspect any design flaws (such as ESD protection or interference) that could have caused temporary latchup in codec IC rendering ADC failure?

We are not sure how to proceed on this.  Need your guidance in this regard.

Regards,

Archana Rao

Hi Steve,

Any update on above request?

Regards,

Archana Rao

Hi Steve,

Please help to list required additional information to take this forward.

Hello Steve,

My name is Vinod working with Archana. Some more details captured during my debugging are added. Hope this will help.  

Here is additional information which I collected during the debugging.

1. Note this is in production and this is the first one we noticed over 2000+ we shipped. We do extensive testing of every product including ESS to make sure there are no soldering issues.
2. MIC interface pin 34,35 and Out2 38,39 is used. Rest of the inputs are grounded and outputs left open.
3. Only MIC is not working but the OUT2 is working as usual.
4. The codec is interfaced to TMS320VC5470 through McBSP.
5. We plotted the data received by the Controller which is shown below.

1. Replaced the IC with new one problem solved. Put the IC back failed again .This is how we confirmed the IC is failure
2. The failed IC put on another card it worked and then we put it back on the original card still worked and after that never failed
3. Stressed the card over temperature from -40 deg to +70deg. Still we donot find any failure.
4. We are totally inconclusive as what would be the reason of this intermittent failure.
5. We verified the solder thermal profile we used and reflow profile of the chip ,but didn’t find any issues there as well. The suspected pins 34,35 are connected to signal traces and not to any kind of copper planes to sink temperature and make soldering inefficient.

Here is my thoughts

It looks like the MIC is sending garbage which happens when the interface is open or unable to receive mic signal. Yes , I do consider dry soldering as one of the cause but I would like to know if there is any other internal causes in the chip can cause such failures like MIC interface open internally for a short time and then restores. Can an excess voltage at MIC interface create some kind of temporary latch or any other corruptions in side the chip which could recover with a thermal shock (soldering)?.

 Please provide your response as we are unable to conclude on this.

Hello Vinod, 

I will inquire with the design team to get their insight,  but what I can say is that the AIC24k was tremendously popular for mobile phones back when it was released, and many customers continue to use them in public safety radios, and other communications devices.  What you are describing is not something I've seen with this device before.  

best regards,

-Steve wilson

Vinod, 

I've sent you a connection request.  I would like to know if you can share the schematics for this project.  the Codec section at least. 

-Steve Wilson

Steve

I sent you a private message with schematics. Hope it helps.

Vinod, 

The unused inputs should not be terminated with resistors,  but rather with capacitors.  I don't believe this would cause this issue, but just a note. 

There are no known latch-up states for this device, and we do test them thoroughly before they are released.  

Are you able to reproduce the issue after desoldering the device and it working again? 

best regards,

-Steve Wilson

Hi Steve,

No we are unable to reproduce the failure. Our protection TVS on this interface is of 30V . So what happens if 30V spike comes on the MIC interface with a low pass filter of 30KHz (500ohm and 10uf) +another series resistance of 4.7K? The current should be very feeble but i am not denying the fact of appearing 30V at the MIC input? 

Vinod, 

There are internal ESD diodes on all input pins, but these are not intended to protect against in circuit ESD events,  mainly for handling and manufacturing. 

If you anticipate voltage levels of 30Vs on the MIC inputs,  there needs to be some kind of protection circuitry. either in line TVS Diodes, and definitely some kind of series resistance between the microphone and the input terminal of the device.  a 50-100pf capacitor to gnd on those input pins might make some sense too. 

best regards,

-Steve wilson

I sent you the input circuit what we use. We do have TVS diode at the input but that is of 30V device.