Problem with ADA4004-4/ADG1436/AD7628

Dear All:

Recently I had a full board using, among others, the chips named above. As I got ADA4004 or ADG1436 burnt without any explanation, I did a prototype board using only the above chips to try to discover the problem and still have the same problem: without any apparent reason, suddenly one or both chips got burnt. I have checked the pinout and everything seems OK. I dont see any reason for this....could anyone please check the schematic to check if there is something wrong?

I attache the schematic.

I am using a power supply to check currents althoug the schmatic is thought to use a ATX power supply.

Thanks un advance,

Joaquin.

  • Thank you very much Harryh,

    - I attached the pictures for power up, power down and the last one with the full setup. In the last one you can se the protoboard I am using for test.

    The images are with any load becuase the board have burnt chip and is not usable anymore without producing a short

    - Regarding the unused parts and to be more precise, you mean, that both inputs should have a pull-down (10K) to ground and the same with the output, correct? Or do you mean another thing?

    Thanks again for you support,

    Joaquin

  • Hello Harry,

    No, the yellow arrow on the right side  set the trigger level, not zero volts, so +12V doesnt go below zero volts, but anyway I can add the shotkky diode as a protection.

    The spikes are without any load, so it is how the power supply works. On the schematic there are several decoupling capacitors.

    Supposing +12V does not go below 0V and -12V does not got up 0V, any other possible cause? Should add any other protection?

    I will replace ADA4004 and ADG1436 with new chips and also will add the protection shottky diodes and check again...and then will let you know the results.

    Thanks again.

    Joaquín.

  • JLC,

      --  Unused sections have to be tied off as buffers with the non-inverting input tied to a voltage within

    the input voltage range;  I would use a 5k-10k resistor to ground.

    --  You may have a power supply sequencing problem;  Can you provide scope pictures of +/-12V

    on power up and power down?

    Harry


  • J,

      In the first photo, does the yellow arrow head on the right side indicate zero volts?  If so,

    the positive rail is being pulled below ground, and that is probably your problem.  Put

    a Shottky diode from ground to +12V to clamp it at -0.3V.

    Also, the spikes on your supplies are weird.  Do you have any bypass caps on the rails?

    For unused sections of a dual or quad op amp, tie the output to the inverting input.

    Then put a 10k from the non-inverting input to some voltage that is within the input

    voltage range spec.  In your case, ground is convenient.

    Harry


  • 0
    •  Analog Employees 
    on Sep 29, 2015 3:08 PM

    Hi Joaquin,

    It sounds like a latchup issue if the devices are getting burnt out. Do you have any scope plots of the ADG1436 channels switching? It looks like you're switching between the supply rails of the device, so any glitches on the source pins that exceed the supply rails could trigger a latchup event. Or are there any ESD events that could be occurring?

    The ADG5436 is a latchup immune version of the ADG1436. It may be worth trying that part on the board to see if that resolves the issue, or at least rule out the switch as the source of the latchup.

    Rgds,

    Paul