• State TI Thinks Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 2 answers
  • Subscribers 47 subscribers
  • Views 353 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

ADS131M08: Need clarification on input shorting feature

Afzaal Ameer1
Prodigy 140 points
Part Number: ADS131M08

I am trying to get the offset by shorting the inputs using the CHx_CFG registers (Page 64: https://www.ti.com/lit/ds/symlink/ads131m08.pdf)  but it's not behaving as I expect it.
Below is table of the mean value over 4K samples at 4K sampling rate at different GAIN settings.
When it is not physically shorted, the offset keeps changing with the gain values, but once I physically short it, the value remains consistent across gain settings.

I am perhaps misunderstanding this feature and there is an explanation for this. But I am in the dark right now. So if someone can please throw some light on this, I'd be grateful.

Gain Internal short using CHx_CFG Register External Short External + Internal Short
0 261 581 664
1 587 589 672
2 1119 584 677
3 4408 573 672
4 6394 567 671
5 35111 557 664
6 70748 558 652
7 141995 556 657

It's running off 8MHz clock, all channels enabled, in continuous conversion mode. I've only changed GAIN and CHx_CFG Registers.

  • 0
    TI__Guru 57346 points

    Hi Afzaal,

    Thank you for your post.

    The internal short is applied at the input of the PGA. Both inputs are internally connected to AGND during this test. 

    Your results look fairly consistent across gain setting with only about 25-30 codes difference, but some variation in input-short offset is expected across gain. See Figure 6-7 in the data sheet.

    In addition, Gain >= 8 uses high-impedance pre-charge buffers on the inputs. Once these pre-charge buffers are enabled, we expect there to be less offset variation for the higher gain settings. 

    Regards,

    Ryan

  • 0 in reply to Ryan Andrews
    Prodigy 140 points

    I think there has been some misunderstanding there. In case of the last two columns there is a physical short. The first column only has the internal short, and the offset goes from 261 to 141995.

  • 0 in reply to Afzaal Ameer1
    TI__Guru 57346 points

    Ah my apologies - my eyes jumped to only the last two columns. Yes this is much more obvious. These results are the mean raw ADC output codes for each gain setting?

    When MUXn[1:0] = 01b, the switches which connect the input pins to the PGA are open, so there should be no difference between Internal Short (only) and Internal + External short. Something certainly seems off.

    Have you observed the same trend on multiple channels?

    Regards,

    Ryan

  • 0 in reply to Afzaal Ameer1
    TI__Mastermind 18900 points

    Hello Afzaal,

    To piggyback off Ryan, I just recreated your data collection on the EVM. It supports Ryan's claim (which is expected). You might have been changing the gain on the wrong channel

    See data below.

    Config Gain (V/V) Mean (codes) Sigma (codes) min code max code code spread
    Shorted internally and externally 1 1122.59 40.78 989 1278 290
    Shorted internally and externally 2 2367.48 69.84 2140 2616 477
    Shorted internally and externally 4 4920.6 129.71 4468 5412 945
    Shorted internally and externally 8 5112.9 158.32 4510 5781 1272
    Shorted externally 1 1632.15 41.21 1469 1795 327
    Shorted externally 2 3395.05 71.01 3137 3657 521
    Shorted externally 4 6989.06 132.79 6503 7567 1065
    Shorted externally 8 10586.22 160.58 9864 11110 1247
    Shorted Internally 1 1116.25 41.79 958 1269 312
    Shorted Internally 2 2355.59 71.23 2080 2585 506
    Shorted Internally 4 4919.61 132.18 4462 5402 941
    Shorted Internally 8 5089.11 156.61 4566 5831 1266

    Best,

    -Cole

  • 0 in reply to Cole Macias
    Prodigy 140 points

    I am in fact setting the GAIN and MUX on all channels. I'll do more diagnosis and add more details on Friday. Thanks !

  • 0 in reply to Ryan Andrews
    Prodigy 140 points

    I did everything from scratch but with a slightly different approach. It's a new unit so different offsets.
    Below I have mean for the three MUX settings. Shorted, positive and negative test signals. And I have estimated the offset using mean of the positive and negative signals.

    I can see that the estimated offset follows the pattern from the datasheet. It drops at Gain 8 and goes up and flattens out. But mean raw value of shorted channel is not accurate for gain 64 and 128. For now I am planning to use the test signals to get the offset, but it would be nice to know why the short doesn't work as expected.


  • 0 in reply to Afzaal Ameer1
    TI__Guru 57346 points

    Hi Afzaal,

    Thanks for the additional data. Cole and I will look at this again on Tuesday (Monday is a holiday for us) and report back to you what we think may be the root cause.

    Regards,

    Ryan

  • 0 in reply to Afzaal Ameer1
    TI__Mastermind 18900 points

    Hello Afzaal,

    I converted your codes to voltage so we can compare with the datasheet [using n_code*(2*VREF/2^resolution)/G)]. If anything, the negative and positive test signals look incorrect, the "short" column has a flat relationship like the datasheet. 

    I wondering the accuracy of the test signals to hold 1mV which is needed in the G=128, I'll do some confirmation when I run your tests on my EVM. Note, my plot is just the tabularized data above converted to voltage so we have them as a comparison. I should be able to check it out within the next few days.

    Your data:

    Mine:

    Datasheet: 

    Best,

    -Cole

    Edit: I forgot the 2*VREF because it is a bipolar contribution to the FSR. Updated all the figures, looks like they match much more with the datasheet