• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 566 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.

ADS1235: Problem with unstable readings on ADC.

João Almeida
Prodigy 30 points
Part Number: ADS1235

Hello. I am designing a PCB for a Strain Gauge measurement system using 2 AD8557 as buffers for the ADS1235 controlled by an STM32. I assembled 20 pcbs, but on 10 of them I get an instability error, the measurements after some time are completely meaningless and replicated in both channels. Since this occured only in 10 PCBs, I assumed soldering issues were involved, but I also suspect that the problem might be in the design/layout. This is my schematic:

The ADS is configured to work in continuous conversion mode, and I get the readings through DRDY interrupt at 4800 SPS.

I use the SerialPlot software to visualize the data, where this is what a working system looks like:

(with and without strain)

  

and this is the problem that appears in some pcbs (without strain applied):
 

 (snapshot of the ECAD software. ADS is in top Layer (L1) while the MCU is in bottom layer (L4), separated by 1 GND and VDD layers)

I hope you can see in the image that the magnitude of this instability is way higher that of the measurements. I concluded that the problem is in the ADS1235 since the error occurs on both channels, and it is driven by 2 separate AD8557s. Sometimes the errors are symmetrical and sometimes not.

I hand-soldered the components on the PCB using hot air pistol at 260C as indicated by the datasheet. Can this problem result from poor layout? The analog traces are separated from the higher speed digital traces, away from the crystal and the PCB has 4 layers with alternating VCC/GND planes, and the component thermal pad is connected to GND. 

I would be glad if anyone could provide me with testing procedures to make sure that the ADS is working fine or if there is any way to identify the source of this problem. If there is any additional information I can supply to help, I'm willing to share. Thanks  

  • 0
    Prodigy 30 points

    Another comment I forgot to mention, when I heat the part, it sometimes work normally for some time then the problem occurs again (which again leads me to believe the problem may be with soldering?)

  • 0 in reply to João Almeida
    TI__Guru 51846 points

    Hi João Almeida,

    Due to the New Year, our response to your question will be a bit delayed. Please expect a response in the next 1-2 days. Thanks for your patience!

    If you determine any new information about your issue in the next few days, please let us know so we can try to resolve this issue as quickly as possible

    -Bryan

  • 0 in reply to João Almeida
    TI__Mastermind 25020 points

    Hello,

    Seeing how the yield is at 50%, I also agree that there seems to be a system level issue that is at the edge of a marginal issue, and not an issue with the device functionality. 

    I am not too sure I understand the the plots you shared. But I suggest we focus on segments to isolate and eliminate sources of error. 

    I would suggest removing the input sensor and using known input voltages to confirm that the ADC measurements are correct, and communications to the host is correct. 

    Then look at the input sensor and circuit (where i suspect the issue lies) without the ADC sampling. confirm all the voltages are as expected with a multimeter. 

    Comparing the measurements to a known good system can also be helpful here to confirm expected nodal voltages. 

    Now, layout can be a source of error in this case. Taking a look at the layout you provided, I do not see a solid ground connection to the thermal pad of the ADC. The thermal pad should be grounded, and usually has a ground pour on top layer connected with multiple vias. hand soldering the thermal pad is rather difficult, thus the device may not be grounded properly. 

    The boards also has analog connections running along and crossing paths (top layer vs bottom) with digital connections. This is not good layout practice. It is recommended to have digital connections and analog connections separate from each other. 

    Regards

    Cynthia