• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 359 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.

AWR1843AOP: AWR1843AOP : GPADC operation related

DongHyun Lee
Prodigy 110 points
Part Number: AWR1843AOP

Tool/software:

Hi.

We are monitoring the input voltage through channel 2 of the GPADC on the AWR1843AOP, as shown in the diagram below.

However, as indicated in the table within the same diagram, there is a difference of about 4 bits in the ADC readings between the sensor stop state and the sensor start state.

To address this issue, we modified the circuit.

As shown in the diagram below, we removed the 10nF capacitor (C86) or replaced it with a capacitor of 120pF or lower, which resolved the issue.

Our question is: Why does the GPADC value differ between the sensor start and sensor stop states when the capacitor (C86) value is 10nF?

  • 0
    TI__Expert 4190 points

    Hi,

    Thanks for posting. Could you let us know what market and end application you are targeting?

    Thanks again,
    Deexith.

  • 0 in reply to Voona Deexith
    Prodigy 110 points

    We are developing automotive radar, and the circuit in question is part of the radar system used for collision avoidance.

    Could we receive an answer to our inquiry?

  • 0 in reply to DongHyun Lee
    TI__Expert 5461 points

    Hi DongHyun, 

    From the ADC spec in the datasheet for the GPADC, sampling time is 400nS and there is an internal cap 10pF. Adding an external 10nF cap will significantly increase the RC time constant potentially causing the input voltage to not settle fully within the 400nS sampling window which could result in the discrepancy. However, using a smaller cap (~pF) for filtering purposes should be fine still allowing the input to settle within the timeframe as stated in the datasheet. 

    Regards,

    Aydin 

  • 0 in reply to Aydin Kiasat
    Prodigy 110 points

    Hello,

    The time constant (RC) remains the same regardless of whether the AWR1843AOP is in sensor on or sensor off state, so it cannot be the cause of the ADC value discrepancy between these states.

    To restate the issue: when C86 is 10nF, the ADC value in the sensor off state drops by about 4 bits when the sensor is turned on. It seems that C86 is influencing the AWR1843AOP’s ADC when the sensor is turned on, leading to this decrease in value.

    However, without detailed knowledge of the internal circuitry of the chip, it’s difficult to pinpoint the exact cause.

    Could you provide an explanation of what might be causing this phenomenon?

    Thank you.

  • 0 in reply to DongHyun Lee
    TI__Expert 5461 points

    Hello DongHyun,

    From your experimentation above, the only discrepancy appears to be during the sensor start condition when the 10nF cap is present while the sensor stop values remain consistent with or without the cap. This points to a settling time issue. 

    Regards,

    Aydin 

  • 0 in reply to Aydin Kiasat
    Prodigy 110 points

    When I mention "sensor stop/start," I am not referring to the power on/off of the AWR1843AOP. Instead, I mean starting and stopping the transmission and reception of RF signals while the chip is already booted. Since ADC values are still output in the sensor stop state, I believe that both the sensor stop and start states should have the same settling time. Therefore, I believe the difference in ADC values between the sensor stop/start states when C86 is 10nF is not due to a settling time issue.

    If I have misunderstood something, could you please explain in more detail why or how a settling time issue could occur only during the sensor start state? I'd like to better understand the potential reasons or causes behind this.

  • 0 in reply to DongHyun Lee
    TI__Expert 5461 points

    Hello DongHyun,

    Just to verify, how long do you wait after starting and stopping the chirping before taking a GPADC reading? 

    The reason I'm asking is that when the Radar starts chirping, there will likely be a transient at the supply so the 10nF cap (low pass) at the input helps to smooth this out. However, it might introduce a lag in the response of the voltage as it settles. PMIC datasheet should reflect the Load Transient response of the PMIC you are using. The RC time constant you have is already 10nF *10k = 100uS (Without accounting the trace losses). This means, if any transient is present at the input, it'll take about 5*100uS = 500uS for the voltage to settle to 99% of the final value.

    To verify, I recommend measuring the voltage at the GPADC input on the scope during start/stop with and without the cap. This will help determine if there's any transient behavior or slow settling due to the cap. 

    Since the12V being monitored directly through the divider, it'll burn constant power, and it is also recommended to have a buffer between the supply and the ADC pin.

    Regards,

    Aydin