• State
  • Locked Locked
  • Replies 12 replies
  • Subscribers 82 subscribers
  • Views 3272 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.

TMS320F28035: Interference in ADCxx Channels

Needhu Gunasegaran
Intellectual 2890 points
Part Number: TMS320F28035
Other Parts Discussed in Thread: SN74LVC1G08

Hello Team ,

Could you please help us answer the question from my customer.

we found that there is interference in the AD measurement when we use the pin AIO14 as a digital output. Even if we do not configure the pin AIO14 we get errors.

The device is configured as follows,

           AdcRegs.ADCSOC0CTL.bit.CHSEL = 0x0003; // ADCINA3

           AdcRegs.ADCSOC1CTL.bit.CHSEL = 0x0006; // ADCINA6

           AdcRegs.ADCSOC2CTL.bit.CHSEL = 0x000B; // ADCINB3

 

And pin ADCINB6 / COMP3B / AIO14 as a digital output, which has been toggled in certain periods.
In this case, faults can be seen in the signal waveform of ADCINB3 when AIO14 is high. If ADCINB3 is configured to ADCINB7 instead of ADCINB3, then the errors are at B7.

For a comparison, we tried a different hardware configuration:

 

           AdcRegs.ADCSOC0CTL.bit.CHSEL = 0x0003; // ADCINA3

           AdcRegs.ADCSOC1CTL.bit.CHSEL = 0x0002; // ADCINA2

           AdcRegs.ADCSOC2CTL.bit.CHSEL = 0x000B; // ADCINB3

 

And the pin ADCINB2 / COMP1B / AIO10 has been configured and toggled as a digital output. The disturbances occur at B3, which comes to A2.

 

Summary:

 

ADC

Toggled pin

noice  at

A3

A6

B3

ADCINB6/COMP1B/AIO14

B3

A3

A6

B7

ADCINB6/COMP1B/AIO14

B7

A3

A2

B3

ADCINB2/COMP1B/AIO10

B3

A3

A2

B7

ADCINB2/COMP1B/AIO10

B7

 

For the purpose of checking, we have dispensed with the use of the pin 23 and therefore, via another means, communicates with the microcontroller.
The following points were observed:
• There were no more disturbances to see.
• When the pin 23 has been externally set to high / 3.3V, the fault has reappeared, that is, As long as the pin 23 is high, disturbances can be seen on B3 and B7, respectively.
• The width of the interference depends on the duration of pin 23 on high.
• The pin 13 / ADCINA2 has also been observed when we pinged the pin 20 (ADCINB2 / COMP1B / AIO10) (see "Tabular summary").

The problem always occurs in the cases described above, but during the operation of the current control, etc., we have no problems at the moment.
We are looking for the cause of the disturbance to avoid possible problems in the future. We have tested our software (the same hardware with TI-035 but different projects) and could not find something striking.

 

Thank you very much in advance.

Best Regards,

Needhu

  • 0
    TI__Genius 17370 points

    Hello,
    Thanks for the detailed description. A C2000 team member has been assigned to this post and should be answering shortly.

    Regards
    Baskaran

  • 0 in reply to Baskaran Chidambaram
    TI__Intellectual 2890 points
    Hi Baskaran,

    Thank you very much. Looking forward for the support.

    Thanks and Best Regards,
    Needhu
  • 0 in reply to Needhu Gunasegaran
    TI__Guru 62975 points
    Needhu,

    What is the nature of the load being driven by the AIO pins?

    Generally speaking, digital activity (both input and output) on the AIO pins can interfere with ADC conversions. The AIO circuitry draws power from the VDDA/VSSA analog supplies so there is a direct path for noise coupling.

    -Tommy
  • 0 in reply to tlee
    TI__Intellectual 2890 points
    Hi Tommy,

    Please find the loads as follows.

    Pin 20 / AIO10: LED
    Pin 23 / AIO14: for RS485


    Pin 13 / ADCINA2: DC link current
    Pin 12 / ADCINA3: Current U (from Current Sensor)
    Pin 10 / ADCINA6: Current V (from Current Sensor)
    Pin 21 / ADCINB3: Current W (from Current Sensor)

    Thank you.

    -Needhu
  • 0 in reply to Needhu Gunasegaran
    TI__Guru 62975 points
    Needhu,

    Would you happen to know the current loads for the LED and RS485? Is there isolation between AIO and loads greater than a few uA?

    The internal VDDA power distribution circuitry was only sized to support the internal ADC and comparator activity -- it was not meant to drive external loads.

    -Tommy
  • 0 in reply to tlee
    TI__Intellectual 2890 points
    Hello Tommy,

    Please find the answers to your question below.

    To Current
    a. There is a high-resistance logic component (SN74LVC1G08) between the LED and the the controller. According to data sheet, it has a 1M Ohm resistance at 3,3V. This corresponds to 3,3uA.

    b. For RS485, ADM2481 is used and according to data sheet max 10uA.


    There is no isolation.

    Thank you.

    -Needhu
  • 0 in reply to Needhu Gunasegaran
    TI__Guru 62975 points
    Needhu,

    This generally sounds ok to me. How are SN74LVC1G08 and ADM2481 powered? Are they sharing the VDDIO supply? Is VDDA also tied to VDDIO or is it isolated?

    Can you describe the waveform sampling? Is the signal buffered to the ADC pin? What ACQPS setting are they using? It may be possible that the ADC-B S/H cap is charging up between conversions when the AIO is high, and the ACQPS window is too short for the S/H cap to settle.

    -Tommy
  • 0 in reply to tlee
    TI__Intellectual 2890 points

    Tommy,

    Please find the answer to your questions as below.

    •        How are SN74LVC1G08 and ADM2481 powered?

    o   Both are powered from +3V3

    •        Are they sharing the VDDIO supply?

    o   Yes, they have common voltage supply.

    •        Is VDDA also tied to VDDIO or is it isolated?

    o   Yes VDDA is tied to VDDIO and it is not isolated.

    •        Can you describe the waveform sampling?

    o   The blue signal waveform is recorded by MATLAB/Simulink and sampling the ADCINB3.

    •        Is the signal buffered to the ADC pin?

    o   Low pass filter between current sense out and ADC pin

    •        What ACQPS setting are they using?

    o   ACQPS is 6, but the problem exists with different ACQPs (6 to 25), too.

    Thank you.

    Best,

    Needhu

  • 0 in reply to Needhu Gunasegaran
    TI__Guru 62975 points

    Needhu,

    Does the magnitude of the ADCINB3 error change with ACQPS settings?  What if the maximum ACQPS is used?

    Another experiment (if possible) is to disable the AIO function and drive the ADC pin directly using an external voltage source.  If the error is still there with AIO disabled, we can eliminate the AIO circuitry from the picture.  Another good data point from this would be to see if the magnitude of the error changes with the forced voltage.

    Do you have the schematic for the ADCINB3 low-pass filter?  What are the component values?

    -Tommy

  • 0 in reply to tlee
    TI__Guru 62975 points
    Needhu,

    When an ADC-A channel is sampling, its complementary ADC-B channel is also sampling even when configured for sequential sampling mode. This is a carryover behavior from implementing the simultaneous sampling feature. For example, when A6 is sampled, B6 is also sampled. In the case where B6 has a 3.3V signal, the ADC-B S/H capacitor will be charged up to 3.3V.

    This behavior should have no effect on the system as long as the ACQPS values are all sized correctly. In this instance, it looks like the passive RC filter on ADC-B3 may be preventing the ADC-B S/H from charging (and discharging) fully when it samples. Reducing the R and increasing the C should help. A good experiment would be to increase the C from 1nF to 100nF or larger.

    -Tommy
  • 0 in reply to tlee
    TI__Intellectual 2890 points
    Tommy,

    Thank you.
    I will keep you updated on the outcome of the test.

    Thanks,
    Needhu
  • 0 in reply to Needhu Gunasegaran
    TI__Intellectual 2890 points
    Tommy,

    Unfortunately, the results are not as expected, the interference is still the same.

    Thanks,
    Needhu