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.

ADS1263 32-bit A/D converter noise floor

Other Parts Discussed in Thread: ADS1263, ADCPRO, ADS1262

Hi,

We are at the late stages of a design using the ADS1263 32-bit A/D converter.  We need to get our product to the field but are stuck on a noise issue.

We have both the ADS126XEVM Evaluation Board and three custom prototypes.  Everything is working fine except for our noise is very high compare to what the datasheets say it should be. 

Both the Evaluation Board and our prototypes have 20uV of noise that we cannot get rid of it.  We have spent close to two weeks on this issue and are totally out of ideas so hopefully someone can help.  If it was just our prototypes I would keep looking, but we are stumped how we can’t even get the ADS126x Evaluation Board to meet specifications either.

According to the ADS1263 datasheet and other documentation like the TIDUAC1 reference design we should be able to get noise <1uV and both our design and the evaluation board are 20uV noise.

Here is our setup:

- We are trying to measure temperature with 0.000X Celsius precision. Right now we are 0.00X with our 20uV noise.

- We are using the 4 wire configuration as in page15 of the ADS126xEVM data sheet.

- 1k ohm Platinum RTD

- 1k Precision reference resistor.

- ADC Pro is setup to match the configuration and it measures properly (not consider the noise floor).  We are not using any gain.

We have tried everything from a Faraday Cage, every option in the software, different reference values, different RTD values, shorting the inputs together, etc.  Always 20uV noise on both the Evaluation board and our prototype.

Any help would be appreciated since we are not sure what to do at this point.

Thanks,

Aaron

  • Hi Aaron,

    First, welcome to the TI E2E Forums!

    What data are you using for the ADS1263?
    My initial thought is that you might be observing power-supply noise. I have noticed noisier than expected results with the ADS126xEVM, and looking at an FFT of the data revealed a significant 60 Hz signal to the noise. Selecting one of the data rates that provides a filter notch at the line cycle frequency usually removed this noise and provided results much closer to data sheet values.

    Since you've already tried shorting the ADC inputs (effectively removing the input signal and reference noise from the result), I would try looking at an FFT of the data to see if there is an interference signal, such as 60 Hz line cycle noise.

    Best Regards,
    Chris

  • Hi Chris,

    Thanks for the reply.

    I just tried the Eval Board on a battery and with the power supply. In both cases the FFT shows just a steady noise floor around 80dBc. There are no 60Hz or other signals in the spectrum from 0Hz to 200Hz. I can get the noise floor to move a little from 80dBc as I change the filters but not much.

    I am not 100% sure what you mean by “what data am I using for the AD1263”. We are using ADC number 1 only. We are saving the data to a file and plotting it in Excel or just viewing the Multiscope output. I am concerned we are just not setting up something correct since the Evaluation Board has the same noise as our prototype. However, we have spent weeks trying all the options and the evaluation software is quite easy to use and understand, we don’t really have any non understood settings.

    Is a flat 80dBc a high noise floor for this part?

    Thanks,

    Aaron
  • Hi Aaron,

    Sorry, I was trying to ask what "data rate" are you currently using?
    Also, you mentioned PGA gain of 1 V/V, but which filter mode (SINC1/2/34 or FIR)?

    ...Knowing these settings will help me figure out what you're expecting for the noise floor.

    Thanks,
    Chris
  • Hi Chris,

    We are using 400SPS and the SINC 1 filter.

    Thanks,

    Aaron
  • Hi Aaron,

    OK - Would you be able to share your data (with inputs shorted), and perhaps the register settings as well, so I can try to replicate the issue with my ADS1263EVM?

    Something seems way off if your observing 20 uVrms, when the expected performance is 1.4 uVrms (for 400SPS, SINC1, PGA = 1V/V).

     

    I did a quick test with inputs shorted and observed about 1.6 uVrms noise (or 1364.5 codes StDev), as shown below:

    Here is my register setting text file, which can be imported on the "Register Map" tab of the ADS126XEVM plugin for ADCPro:

    ADS126x Register Map.txt
    Name	Addr	R/W	Init	Value
    ID	0x00	R	-	0x20
    POWER	0x01	R/W	0x11	0x13
    INTERFACE	0x02	R/W	0x05	0x05
    MODE0	0x03	R/W	0x00*	0x00
    MODE1	0x04	R/W	0x80*	0x00
    MODE2	0x05	R/W	0x04*	0x08
    INPMUX	0x06	R/W	0x01	0xAA
    OFCAL0	0x07	R/W	0x00	0x00
    OFCAL1	0x08	R/W	0x00	0x00
    OFCAL2	0x09	R/W	0x00	0x00
    FSCAL0	0x0A	R/W	0x00	0x00
    FSCAL1	0x0B	R/W	0x00	0x00
    FSCAL2	0x0C	R/W	0x40	0x40
    IDACMUX	0x0D	R/W	0xBB	0xBB
    IDACMAG	0x0E	R/W	0x00	0x00
    REFMUX	0x0F	R/W	0x00	0x00
    TDACP	0x10	R/W	0x00	0x00
    TDACN	0x11	R/W	0x00	0x00
    GPIOCON	0x12	R/W	0x00	0x00
    GPIODIR	0x13	R/W	0x00	0x00
    GPIODAT	0x14	**	0x00	0x00
    ADC2CFG	0x15	R/W	0x00*	0x40
    ADC2MUX	0x16	R/W	0x01	0x01
    ADC2OFC0	0x17	R/W	0x00	0x00
    ADC2OFC1	0x18	R/W	0x00	0x00
    ADC2FSC0	0x19	R/W	0x00	0x00
    ADC2FSC1	0x1A	R/W	0x40	0x40
    

    Best Regards,
    Chris

  • Hi Chris,

    It must be something about our hardware wiring on the AIN pins because I still get the same results even with your settings, that is the only dfference now. The noise looks nothing like what you have. I am not having any luck pasting screen shots into the forum yet so I will describe below.

    If you look at page 15 of ADS126XEVM we are wired the same:

    ANICOM to IN7
    IN7 to IN6 (usually a 1k RTD but tried a 1k precision and short here)
    IN6 to IN4
    IN3 to IN4
    1k precision resistor between IN4 and 1N5
    IN5 to ground

    Registers from ADCPro power up default are:
    INPUT MUX Tab: ADC1, AINP1=IN7, AINN1 = IN6
    REFERENCE TAB: ADC1 REF Source = Other, REFP = AIN4, REFN = AIN5
    IDAC\SENSOR BIAS TAB: IDAC1= AIN3, IDAC2= AINCOM, IDAC1 Magnitude=1000uA, IDAC2 Magnitude=1000uA

    Thanks,

    Aaron
  • Hi Aaron,

    First try testing the noise performance without the external circuitry. Configure the ADS1262 to measure an internal short and use the internal reference as the ADC's reference source.

    ...If you get good noise performance then the issue is likely related to the external circuitry.
    ...If you cannot get good noise performance in this configuration then there is probably something wrong with the hardware, in which case I can help get you a replacement EVM.

    Best Regards,
    Chris
  • Hi Chris,

    When I do an internal short of ADINP and ADINN on any signal input I get the same high noise results. Since we have our custom design (with the same noise issue) I just got a new ADS1263 part out of stock and changed it on the eval board. Same noise. There has got to be something we are missing. Despite the fact I changed the ADS1263 on our evaluation board can you send us a new one? I am willy to try anything at this point.

    Can you call me? This forum is great but I think some quick back and forth conversation might be more use full right now.

    Thanks,

    Aaron
  • Hi Aaron,

    I'm sending you an email on your E2E provided email address.

    Best Regards,
    Chris
  • Thanks Chris. I think we have all our issues resolved and are moving forward now.

    Aaron

  • Hi,

    dare I ask whether the noise was due to something particular to your setup or was there some general conclusion from which other systems might profit too in the future?

    Thank you very much

    Pio

  • Hi Pio,

    ADCPro was not displaying the data correctly and I think the noise performance was okay after all.

    Best Regards,
    Chris