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.

TPS7A33: Why is it so senstive and easy to damage?

Part Number: TPS7A33
Other Parts Discussed in Thread: TPS7A88, , TPS7A47

Hello,

I have a design with few analog supply rails that use TPS7A88, TPS7A47, and - unfortunately - TPS7A33. While the first two work great without problems, the negative rail is giving a few mV of output and tens of mV on the Vref pin.

My application is simple datasheet-based with ceramic caps 10 uF on input, 3x10 uF on output, 68 nF Cnr, 10 nF Cff. Resistor sets are 1,2M & 100k and 1M & 160k, PCB layout as in datasheet. Once I got it running after two rounds of resistor and IC matching, third ICs got up and running, but after a few minutes they are gone (maybe I touched them?).

What am I doing wrong? are these ICs really sensitive to some conditions, which I am not aware of?


  • Hi, 

    To help with your questions, will you please provide some startup waveforms for the 3rd unit that shows Vin, Vout and Vref? If the part is damaged, most likely some pins may get stress over its absolute maximum voltage. A touching could potentially create an ESD event that may result in a damage on the ESD structures on the output pins and this could lead to a output shorting to ground failure. 

    Regards, 
    Jason Song

  • Hello,

    thank you for your answer.

    Unfortunately, I saw reference voltage only on the NR pin, FB pin shows a few tens of millivolts. In attachment, you will find my waveforms of broken IC.
    Board is supplied from bench supply (R&S HMP2030), but as I read other posts on this forum, TPS7A33 is sensitive to any positive voltage on FB pin. Someone even had positive voltage during startup. Could it be the case? I cannot understand the connection, why FB could get positive when there was totally no load on both rails, but what is interesting, output goes positive in the broken part.

    Best regards,
    Bartosz

  • Hi Bartosz, 

    Thanks for providing the scope-shot. I can see that unit 3 has been damaged. When using TPS7A33, you need to ensure that the absolute max voltage ratings on all pins not only FB pins are not violated. 

    For the FB pin, here are the absolute maximum ratings. Have you noticed any voltage violations during your application? If you could share your schematic, I could help to review to make sure it follows the datasheet recommendation. 

    Regards, 
    Jason Song

  • Hi,

    sorry for the late response, I was out of the lab last days.
    Attached you will find my design - both schematic and board. PCB has 4 layers with one internal ground layer and one supply layer (after the TPS, TPS is supplied by the trace on the top layer), one capacitor is on the same layer as TPS - bottom, close to the TPS, other capacitors are on the top layer directly above them.

    I have soldered my last piece of TPS, but this time - with a switch on EN pin and after a few power-up cycles with a sequence: bench supply on, EN switch on, it is still alive. I haven't observed any voltage violation during startup, any positive voltage on any rail. 

    Could you explain why there is a positive voltage on the output of damaged TPS, even though it is supplied only from the negative?

    Best regards,
    Bartosz

  • Hi Bartosz, 

    I reviewed the schematic and it looks okay. For the damaged IC, the output pin could be floating which means the voltage could be positive or negative. It seems to me that the failure you have observed could be an isolated case, and I would suggest you run more tests and see if the failure could be duplicated. During the tests, please make sure to monitor important signals and look for any voltage violation and please also take scope-shoots if you anticipate any possible voltage violations. 

    Regards, 

    Jason Song

  • Hi Jason,

    thank you for the review. What do you mean saying 'isolated case'?
    Is it safe to have an R-C network on EN pin to make a small delay? The datasheet does not say anything about the possible inputs.

    Best regards,
    Bartosz

  • Hi Bartosz, 

    I was using an isolated case to refer to a single case that could be caused by an accident or ESD event. You could have RC network to create a delay on the EN pin. 

    Regards, 
    Jason Song