AD5422 incorrect output issues.


We have been using the AD5422 DAC in a production board for over 4 years and recently have been experiencing a fault with around 10% of the boards. The fault affects Iout and Vout, for Iout it has a negative 4mA offset so the full range goes from -4 to 20mA instead of 0 to 24mA, while for Vout it is stuck at –22Volts regardless of which Voltage mode it is currently in.

We have this problem occur on different batch numbers and both packages (TSSOP & LFCSP) of the AD5422 chip. We have no issues communicating to the chip and have read back the registers and found that that they have been set correctly. We have also tested the surrounding parts which connect to the chip but have found them to be working correctly.

We have checked the voltage rails, Refout and Refin and all of them are at the correct voltages.

Any ideas to what could be causing this problem or solutions to fix it will be greatly appreciated.

Thanks for the help.

adding circuit diagram
[edited by: LukeAW at 3:55 PM (GMT 0) on 2 Sep 2019]
  • 0
    •  Analog Employees 
    on Sep 4, 2019 8:51 AM


    Someone is currently looking into your query and respond to you as soon as possible.



  • 0
    •  Analog Employees 
    on Sep 5, 2019 12:14 AM

    Hi ,

    The schematic diagram you attached is too small to get a proper view, can you re-attach again with bigger image please?

    Can you also provide the following details:

    1) Power supply voltages and power up sequence

    2) Can the problem easily replicated? What are the steps required to replicate the problem?

    3) When the problem started to occur? Have you noticed any change on the test conditions? Any part of the circuit changed?



  • Hi Ivan

    The image is larger when clicked on. Does this not work for you?

    1) the AD5422 is powered by +/-24V with the internal 3.3V LDO enabled.  The sequencing of the +/-24V looks identical between the ones with and without the issue.  see below - first image with without the issue, second is with the issue.
    2) With the ones that have the issue, it happens every time.  The ones without the issue never seem to have this issue.  No special steps are required to make this occur.
    3) This problem started a few months ago. There hasn't been any changes to the design for about 4 years.
  • 0
    •  Analog Employees 
    on Sep 18, 2019 6:58 AM in reply to LukeAW

    Apologies for the delayed response.

    1) can you check the /FAULT pin if it asserts (both during IOUT and VOUT mode)? this pin is in open-drain configuration which means you need pull-up resistor to digital supply to get valid measurements.

    2) if it does assert, please do a readback on the status register:

    3) have you tried removing the +vsense buffer (ADA4522), operate AD5422 in IOUT mode then re-measure the IOUT if the -4mA offset still exists?

    4) what is the load (and the value) used in your test?



  • Hello,

    Sorry for the very long delay on this but the problem seems to go away then comes back quite randomly, we have started having this issue again.

    From further investigation we have found that the Iout is fine its just that the voltage output is always outputting around -22.3 volts and since the outputs are joined it causes the offset (tested this by separating the outputs).

    I did try adding the external resistor as from the data sheet I should use this when the outputs are joined and I enabled that in the code too but this did not fix the bad voltage output. 

    I have checked the sense buffering and this works correctly too.

    I read back the control and status buffers they are all correct too, with the fault only outputting on a open circuit when in current output mode. (control reg: 0011 0000 0000 0011 status reg: 0)  

    The unit is quite hot when in this broken state but the status reg never indicates that it is overheating. 

    So its mostly seems like the voltage output is just always outputting -22.3V regardless off what mode so any idea of how to fix this or what could of caused this to break?

    Thanks for the help, apologies for the long delay.