• State TI Thinks Resolved
  • Locked Locked
  • Replies 4 replies
  • Answers 1 answer
  • Subscribers 79 subscribers
  • Views 799 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.

Large errors are seen with channel A7 (for calibration) of SD16 in MSP430F4250

John J Wu
Genius 3870 points
The MSP430F4XXX User's Guide states the following (slau56h in page 29-6): "29.2.5 Analog Input Pair Selection The SD16_A can convert up to 8 differential input pairs multiplexed into the PGA. Up to five analog input pairs (A0-A4) are available externally on the device. A resistive divider to measure the supply voltage is available using the A5 multiplexer input. An internal temperature sensor is available using the A6 multiplexer input. *Input A7 is a shorted connection between the + and -- input pair and can be used to calibrate the offset of the SD16_A input stage." *However when we try to use input A7 to calibrate for input offset, it alway give us errors in the hundreds of uVolt and they are very inconsistence (varies significantly from one to the next calibration). We test the calibration by shorting another channel (i.e. A0), and it works just fine. We are using linear regulator to generate power for MSP430, and we use internal voltage reference. We are seeing problems with channel A7. Please help us. Thanks!
  • 0
    TI__Expert 7845 points

    A few data points and recommendations:
    1. If you use A7 to calibrate the input offset, then tie A0+/- to GND and measure A0 using the code example "msp430x42x0_sd16a_03" in the following zip file:
    http://www.ti.com/lit/zip/slac077

    what are the results you see? Please let me know if you said that this works fine from your post.

    2. To deal with the offset, you can connect two ADC channels together reversing the polarity of the second pair of pins. In software, then sample from both channels in quick succession and subtract the results, to take care of offset error.
     

    3. Since there are five channels on the MSP430F42x0 devices, there is one channel that can not use the switching method; for this channel use the offset calculated from another. This method can be used to give four channels in total, one of which has direct zeroing and will be the best correction. The other three channels should have good offset calibration.

    4. One thing should be noted with this method. It has the added advantage of doubling the signal resolution but this also can be a disadvantage as you must keep this in mind when dealing with the offsets of the other channels.

     

  • 0 in reply to Harman
    Genius 3870 points
    Thanks Harman for your detailed answers. When using for example A0, usually we will tie both inputs of the channel to Vref or to a voltage near it, to stay well within the common mode input range, and then measure the input, which usually will measure something in the hundreds of uVolt (this will be the input offset). Connecting both inputs to GND only works well when the input buffer is disable, on the other side if the buffer is enable it is better to be at least a few hundred millivolts above ground for reliable results. On the specific reading for channels A0 and A7, it varies with each device, temperature, buffer settings (on or off, buffer gain), SD16 settings, etc. Do you know if there is any significant difference between A0 and A7 in the die? It's so funny that customer has no issues with A0 but A7. Thanks!!!
  • 0 in reply to John J Wu
    Genius 3870 points
    Hi, anybody has any further update? Thanks! AV
  • 0 in reply to John J Wu
    TI__Expert 7845 points

    There's really no significant difference in A0 & A7 except the fact that A7 is internally shorted.

**Attention** This is a public forum