This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DAC80004: Multiple occurrences of channel B defect

Part Number: DAC80004

Hello, in one of our recent designs we use this Quad-DAC and found a defect to channel B for multiple PCBs. Currently, we've seen 4 devices failing after several months of operation. All devices show a defect to channel B which is stuck at approx 800mV, while all other channels remain functional. Is this a known issue and how can we prevent it? We evaluated the input signals upon power-up and found that we already raise the CLR# input pin to 3V3 while the supply voltage (5V) is not yet present or was ramping up simultaneously. Can this cause this kind of failure? As we are preparing the design for mass production we need to know if this failure can be prevented by delaying the input signals until the supply voltage is present, or if we need to take additional measures on the PCB. Channel B is connected to another device (at the same voltage level) via a 40kOhm resistor, so large currents in the output are not expected. After replacing the devices all channels are working fine. Thanks..

  • Hi Robin,

    We have not seen this failure before.  Certainly biasing the CLR pin before the VDD supply violates our maximum operating conditions, but I would not expect that to impact a single channel's operation.  I think we should look at the basics:

    1. Can you share your schematic, showing the immediate circuitry connected to the DAC outputs?

    2. Does an A-B-A swap show the device recover? You stated that replacing the DAC resolved the issue, but did soldering the failing DAC on a otherwise good system allow the device to recover? This would help determine if there is a manufacturing concern.

    3. Can you find some kind of transients on the output or the supply pin that occurs in normal operation?

    Thanks,

    Paul

  • Hello Paul,

    We did not try the A-B-A swap and currently we are not able to do this as all available PCBs are in use. We however collected the faulty devices and can ship them to TI for analysis, is that possible?

    The affected channel is connected to an instrumentation amplifier as shown below. I cannot provide more detailed schematics.

    Some information that may be helpful is that we've seen the channel "hanging" at the initial voltage setting that we do and then not responding anymore to any new settings. This occurred before the actual defect and the situation could be maintained for several days. This issue was solved by preventing to raise the CLR# pin before the supply voltage is present. The CLR# pin was basically supplying the device via the internal ESD diodes which probably caused a latch-up effect in the device (causing the channel to hang and perhaps also causing a high shoot through current in the device?).

    Thanks,

    Robin

  • Hi Robin,

    The image did not load, as you need to use the "Insert/Edit Media" button to add pictures.  Are all the channels connected to the same circuit? Or is channel B unique?

    It is possible that the CLR pin is causing some kind of latch up behavior, but it would surprise me.  It would be interesting to mount a failing device and see if preventing CLR from biasing the device would resolve the issue.  

    If you prevent CLR from biasing the device, is there any failures?

    We would ask you to complete the A-B-A testing before applying to return the units, as it would help eliminate PCB assembly issues from being the culprit.  Returns would be initiated on TI.com if the units were purchased directly, or through your distributor if you ordered indirectly.

    Thanks,

    Paul

  • Hi Paul,

    After solving the issue with the CLR signal, we actually tried the QUAD-DACs before replacing them, but it did not solve the problem (they were really defect). PCB assembly issues are not expected as we actually have about 50 QUAD-DACs powered all day for several months and they are running fine. The issue arose on a board that is powered-on and off each day (in contrary to the other units), so I expect the power-on or power-off sequence triggers the defect. There are no transients on the outputs and the ch B output is connected to an AD8276 which has 40kOhm on its inputs.

    We actually have an evaluation board available (DAC80004EVM). Will try to use that to see if repetitively raising the CLR pin before VDD can cause this kind of situation.

    Robin