• State TI Thinks Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 4 answers
  • Subscribers 48 subscribers
  • Views 696 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.

AMC1200: Output noise frequency component dependent on input DC value

Gabriel Brunheira
Intellectual 440 points
Part Number: AMC1200
Other Parts Discussed in Thread: OPA2227,

Hi!

As shown in the schematics below, we're using a AMC1200 to isolate a voltage signal, an instrumentation amplifier (AD8421) to convert the output to single-ended and an OPA2227 (U201B) to buffer the signal to a BNC coaxial connector (J201), named here as OUTPUT. VDD1 is provided by an isolated DC/DC converter, with stitching capacitors between its inputs and outputs. The relation between the OUTPUT signal on J201 and the INPUT port (VOLTAGE-TAP-IN's on left) is approximately 0.5 V/V.

With a 0V INPUT, there's an offset of 18 mV (INPUT related) and a frequency component of approximately 4 kHz on the OUTPUT:

Applying -15 mV INPUT, the DC value in the OUTPUT approximates to 0 mV and the frequency component shifts to near 1 kHz:

Approaching a -18 mV INPUT, the frequency decreases until it disapears:

 

Varying the INPUT with a sinusoidal signal, we see once more the frequency decreasing when the INPUT is near 0V, and increasing when it moves away. We also see a phase inversion of the component when moving from positive to negative inputs:

All scope images above were captured in the OUTPUT BNC connector J201, with the proper 0.5 V/V gain setting to provide INPUT related measurements.

We traced back this frequency component up to the AMC1200 outputs. We don't observe it on its inputs. We tried changing the stitching capacitors to see whether it was an EMC issue, but didn't have any effect.

Since near 0V the frequencies are too low, there's no way to filter them out.

What could be the explanation for this behavior? Do you have any suggestion on how to deal with it?

There's no information on AMC1200 regarding this, but looks like it could be explained by the delta-sigma modulator, since it operates varying it pulse density, hence "pulse frequencies". 

Best regards,

Gabriel

  • 0
    TI__Guru 55342 points

    Hi Gabriel,

    The outputs of the AMC1200 are driven and I see there is no resistor in series before the connection to the AD device which may be causing the outputs to ring. 

    I recommend adding a resistor in series with VOUTP and VOUTN. You could remove the AD device and see if the outputs become stable to verify. 

  • 0 in reply to Alexander Smith
    Intellectual 440 points

    Hi Alexander!

    Thank you for your response! I must admit I didn't understand your suggestion.

    First, what do you mean by "the outputs are driven"? You mean they're providing a significant amount of current?

    If so, I don't see why, because the AD8421 connected to its outputs is an instrumentation amplifier, thus has very high input impedance (30 GOhm) and low capacitance (3 pF). Also, the traces between them are very short (in green below). So I don't see how such ringing would occur, neither how would the DC value affect the frequency of this oscillation.

    Does it have anything to do you the internal workings of the AMC1200 outputs? In this case, do you have a minimum value recommended for the resistor?

    Lastly, not only there's no information on datasheet, but also it states these output could directly drive ADC inputs, and even show a circuit example:

    Best Regards,

    Gabriel

  • 0 in reply to Gabriel Brunheira
    TI__Guru 55342 points

    Hi Gabriel,

    The output of the AMC1200 are driven by buffered opamps. 

    As AD8421 is not a TI product, I'm not familiar with it's design nor can I support any issues it may have. 

    I agree that this is an odd behavior and perhaps the AD8421 is not part of the issue, but removing the AD part to isolate the issue and verify if it is an issue with the AMC1200 itself may help with the debug process. 

    If it is an issue with the AMC1200, you could try replacing the AMC1200 to verify if that unit was damaged. 

    If resistance is necessary, a value of 100ohm-1k should be fine. 

  • 0 in reply to Alexander Smith
    Intellectual 440 points

    Hi Alexander,

    If there's no solid explanation for the behavior, I agree we should proceed with this '"perturb and observe" approach.

    Just to confirm, the "delta-sigma modulator hypotheses" should then be discarded?

    Best regards,

    Gabriel

  • 0 in reply to Gabriel Brunheira
    TI__Genius 9076 points

    Hi Gabriel,

    Thanks for your response! Alex is currently out of office.

    It is hard to tell if the delta-sigma modulator could be the cause of this issue. Trying Alex's suggestions of trying series resistors at the VOUTP and VOUTN, replacing the AMC1200 device with another AMC1200, or taking the measurements without the AD device would help verify if the issue is internal to the AMC1200 device or not. I hope this helps!

    Best,

    Samiha

  • 0 in reply to Samiha Sharif
    Intellectual 440 points

    Hi Samiha! Ok, we'll try it!

    thank you again!