• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 34 subscribers
  • Views 156 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TRS3232E: Reliability issues with the TRS3232

Joan Fainé
Prodigy 10 points
Part Number: TRS3232E


Tool/software:

we have problems with a batch of boards where the RS232 driver fails. Sometimes it fails from the beginning, others work for a while and then fail during the factory test and others (the worst) fail at the customer's home.
The marking  on the chip is the following: 61DC93K G4 TRS3232EC.

Can you tell us if there is any known error in this batch. We may have found about 20 defective chips out of 500.

Thanks

  • 0
    TI__Guru* 75061 points

    Joan,

    Do you have a schematic we can review?

    Do you have more information on what the failure is?

    Are you suspecting the devices are damaged?

    -Bobby

  • 0 in reply to BOBBY
    Prodigy 10 points

    Hello Boby,

    The problem is not always the same. In some cases, the signal that arrives through the RS232 does not go through the TTL part to the microcontroller and in other cases the reverse has happened. The scheme is the same as many years ago and it had never happened to us. In the last batch of boards received from the assembler we have not had the problem until now.

    Schematic:

  • 0 in reply to Joan Fainé
    TI__Guru 53666 points

    Hi Joan,

    Today is a holiday. Please give some time until Monday for feedback, thanks.

    Best Regards,

    Michael.

  • 0 in reply to Joan Fainé
    TI__Guru* 75061 points

    Hi Joan,

    Do you have any scopeshots of either of the cases you described? I want to verify for the RS232 signal arriving, if the device is seeing a valid RS232 signal.

    For the case where the TTL signal is applied but no RS232 output, can you verify the V+ and V- signals are at the expected voltages? 

    Can you also verify the charge pump cap's rated voltage in the schematic? 

    Joan Fainé said:
    The marking  on the chip is the following: 61DC93K G4 TRS3232EC.

    Do you have a picture of this top marking? I'm not entirely sure the 6 in this is correct unless its an older unit from 2016. (Not impossible though).

    -Bobby

  • 0 in reply to BOBBY
    Prodigy 10 points

    Hi Bobby,

    I'm sorry but I don't have any oscilloscope captures but the RS232 signal is correct because the error has happened on different equipment in different places.

    When I was looking at it with the oscilloscope there was simply no signal coming out on the other side of the driver where it was supposed to come out.

    In the case that fail the RS232 side, the V+ and V-  are ok (If I remember correctly) but I did notice that charge pump that generates the positive/negative voltage wasn't working normally. It wasn't oscillating like it does on a good chip. I checked the capacitors, and they were correct. These capacitors are X7R 100nF/50V.

    I'm sending you a photo of two chips that I keep:

    Joan

  • 0 in reply to Joan Fainé
    TI__Guru* 75061 points

    Hi Joan,

    I think it might be due to damage on the charge pumps though it's a little hard to say without actually seeing the scope plots.

    Any idea if the issue occur after some use out in the field is it something that immediately happens after powering up?

    My guess is maybe some sort or short or transient event could be occurring on the RS232 pins during the initial plug in. Maybe you can probe the pins that tend to break on a good unit and set a trigger to above +/-13V and then try to do a couple insertions to see if the voltage during insertion exceeds 13V. Though I'm not sure if this suggestion is something you can do in practice..... 

    Joan Fainé said:

    I'm sending you a photo of two chips that I keep:

    Yeah, I'm pretty sure these are from 2016. I cannot locate them in our system (I think our system only tracks them for 5 years). 

    -Bobby

  • +1 in reply to BOBBY
    Prodigy 10 points

    Hi Bobby,

     Sometimes the failure appears from the beginning, when starting the equipment for the first time, others fail after being running all night at 40 degrees Celsius (factory test) and in some cases, the least, they have failed the customer after a while.

    The equipment has been manufactured for years without problems with the same schematic.

    From what you tell me about the date of manufacture of the driver, I suspect that the assembler has used a batch of chips of suspicious origin, because it is not normal for a device assembled in 2024 to have chips from 2016.

    Do you have any information about a defective batch of chips with these manufacturing dates?

    Thanks for the help. We will also consult with the board assembler.

    Joan

  • 0 in reply to Joan Fainé
    TI__Guru* 75061 points
    Joan Fainé said:
     Sometimes the failure appears from the beginning, when starting the equipment for the first time, others fail after being running all night at 40 degrees Celsius (factory test) and in some cases, the least, they have failed the customer after a while.

    Since they are breaking at different points and without human interaction in some cases, it likely isn't a transient event unless one was done during the factory testing. 

    Joan Fainé said:
    From what you tell me about the date of manufacture of the driver, I suspect that the assembler has used a batch of chips of suspicious origin, because it is not normal for a device assembled in 2024 to have chips from 2016.

    I agree with that, longest I've seen is about a 3 year gap for a commercial device like this.

    Joan Fainé said:
    Do you have any information about a defective batch of chips with these manufacturing dates?

    I've asked one of my contacts to see if they could look up the device in their system and verify if that specific lot trace code has ever been returned for quality issues. That lot trace code did not show up in our system (this system goes back further than the 5 years I alluded to in an earlier post). 

    It could potentially be related to how the devices were stored (assume they are legitimate devices). If they were stored in a room with humidity or the seals are broken, the device itself could become defective (I think this is related to the metal/bond pads/adhesive element weakening with moisture, may explain why some pins don't see outputs). Generally well sealed devices can be stored for 10 years. 

    -Bobby