Filtering and multiplexing low voltage thermocouple signals

Discussion in 'Analog & Mixed-Signal Design' started by hunterage2000, Oct 8, 2018.

  1. hunterage2000

    Thread Starter Senior Member

    May 2, 2010
    471
    1
    Morning all,

    I want to multiplex 8 thermocouple signals with the output going to a thermocouple amplifier IC AD595.

    As the thermocouple signal is in the mV range I was thinking to filter it with a LPF of 3Hz before the multiplexer.

    Is this way fine? or should I be amplifying the signal before the multiplexer stage? The AD595 is expensive at £10 and wondered if a cheap differential amplifer would be ok.
     
  2. danadak

    Well-Known Member

    Mar 10, 2018
    1,887
    394
    You could always consider PSOC 5LP. Shown is one channel, trivial to
    drag and drop the mux component (internal to PSOC) and set up for
    16 differential channels. Additionally use the DFB component and
    DMA (both internal as well) to handle the filtering. DFB is the digital
    filter block.

    upload_2018-10-8_6-15-58.png

    Regards, Dana.
     
  3. danadak

    Well-Known Member

    Mar 10, 2018
    1,887
    394
    Here is a typical use of the DFB component inside PSOC. Note this is
    single channel single ended, ignore that, can use diff 16 channel mux
    at the input of DelSig. All on chip, including precision Vref.

    [​IMG]

    Note the DFB block has a wizard to set it up, you just pick the
    type, IIR, FIR, sampling rate, and it will help you by showing
    freq and time domain response plots. Quite easy in fact. The
    DMA a little more challenging, there is 2 very good ap notes
    covering that. It also has a wizard that makes it less difficult.


    Regards, Dana.
     
  4. ebeowulf17

    Distinguished Member

    Aug 12, 2014
    2,560
    474
    My concern with multiplexing thermocouples would be cold junction compensation.

    You need accurate temperature readings at each thermocouple termination, which means another set of signals through another multiplexer, or you need to make sure they're all at the same temperature as each other and you can use just one compensation temperature.

    With one chip per input, AD chip handles that all for you, assuming the TC junction is close enough to the chip's input pins. If you go with your own cheap amp, then you need to measure junction temp and add the corresponding voltage to your TC voltage.

    I'm not aware of any reason the multiplexing arrangement couldn't work, but it won't be as simple or straightforward as it might initially appear.
     
  5. danadak

    Well-Known Member

    Mar 10, 2018
    1,887
    394
  6. hunterage2000

    Thread Starter Senior Member

    May 2, 2010
    471
    1
    Thanks all for your comments. I have found an app note for the AD595 which shows how to setup it up with a multiplexer.
     
  7. ebeowulf17

    Distinguished Member

    Aug 12, 2014
    2,560
    474
    That's great news! Can't do any better than a good app note. I'm guessing you're referring to AN369:
    http://www.analog.com/media/en/technical-documentation/application-notes/AN-369.pdf

    That's the very first app note I read, and it got me through my first electronics project. I only understood a small fraction of it, but I learned a lot from it, and was able to simply follow directions to make things work for the parts I didn't completely understand. Good stuff - brings back fond memories.
     
  8. hunterage2000

    Thread Starter Senior Member

    May 2, 2010
    471
    1
    Hi ebeowulf17, you didn't do the multiplexed thermocouples did you? I'm assuming that V2 is the measured temperature of IC that you use?
     
  9. ebeowulf17

    Distinguished Member

    Aug 12, 2014
    2,560
    474
    I didn't do the multiplexing. I only needed two TC circuits, and the multiplexing was way over my head, so I just built two complete circuits with two AD595 chips. More recently I've needed just one or two at a time in different projects and have used a pre-made breakout board that makes it even easier (lazy, but effective for what I've needed them for.)

    Yes, it looks to me like you need to run one more thermocouple, beyond the ones you're reading, and attach it to read the temperature of the AD595 itself, hooking its wires up at the location on the isothermal block marked V2. Pay very close attention to the signal routing there. I misread it the first few times I looked at it. The extra thermocouple is in series with one side of the output of the multiplexer, before the input to the AD595. Very clever arrangement to cancel out temperature differences between the isothermal block and the AD595 - it makes my head spin thinking of the great minds who figure this stuff out for us in the first place!
     
Loading...