• State
  • Replies 2 replies
  • Subscribers 46 subscribers
  • Views 35 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

Original question:

ADS1296R: CH4 as PACE input noisy

ADS1296R: CH4 as PACE input noisy

Raghavendra H S
Prodigy 25 points
Part Number: ADS1296R

Tool/software:

Hello Mr.Ryan,
 I would like to re-open this case.
After detailed study and many experiments the findings are below.
  • The pace input noise is not dependent on any channel of AFE. The noise is present in all the channels.
  • As you posted in the  E2E post the PACE input periodic spike correlates with the data speed. 
  • As you updated the oscilloscope image where the periodic spikes related to the data rate is further amplified in PACE detection circuit which mistriggers the pace pulses. 
  • We have reviewed the layout of our system and found that there is a clear separation between AGND and DGND.
  • We suspect this noise coupling is inside the AFE between AGND and DGND.
  • One more answer we are looking from you is , we have separated planes in our PCB between AGND and DGND of AFE. The Microcontroller (STM32) processing the data is in DGND plane which can have switching noise, do you think this separation of AGND and DGND plane can cause this coupling on the pace out ?
I have attached the latest schematic for reference. If needed we can also come for Video call to discuss this issue.
This has put us in a turmoil stage, please help at priority.SCH-3-86-220-0137-22-REV D_WIP.pdf
  • 0
    TI__Prodigy 310 points

    Hi Raghavendra,

    Thanks for reaching out on the forum! I understand this issue is critical, but expect a delay in response as Ryan is currently out of office.

    I can at least answer your last question for now. Your MCU should be on your DGND plane. Section 12.1 mentions the separation of DGND and AGND is for preventing return currents on the digital side from coupling to your analog circuitry. Because of the high impedance nature of the input signals, there are tons of ways noise can be introduced, especially involving the placement and routing on the board.

    For now, I would suggest decreasing the MCU's SPI signal drive strength in software if possible. This can improve EMI emission from those lines if noise is only an issue when communicating. If you've already been in contact with Ryan, I would also suggest sending him an email, and attach PCB layout screenshots for each layer.

    Regards,

    Shane

  • 0 in reply to Shane Aspenwall
    Prodigy 25 points

    Thanks Shane.

    We wait for Ryan to come back.

    We have mailed the gerber files to Ryan.

    If you look at the linked post for this, Ryan has tried in Eval board with him and he could notice almost 20mV spikes for the data rate at tha PACE out, after the PACE out we have slope detector with gain of almost 150 to detect small pace pulses. But due to this gain the spikes of data rate also amplifies.