interfacing thermocouples (revisited)

Go To Last Post
9 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I've looking to find the *cheapest* solution for interfacing J-type thermocouples to a MCU. The idea being to create a multi-therocouple input block that will be attached to a PC or DRO to display readings. The block will be split into about 10 input pairs or channels. What is not so important here is the actual temperature reading in degrees but the delta shift between each of the input pairs over time, although any reasonable approximation of the actual temp would be a nice.

From what I have found trawling the boards & google it looks like I have 2 or maybe 3 possible options, instrumentation op-amps hooked to ADC pins or a bunch of those fancy max66xx spi chip's @ 6-15 GPB ea. :shock:

I did however uncover this thread: https://www.avrfreaks.net/index.php?name=PNphpBB2&file=viewtopic&t=32138&highlight=thermocouple

I am really interested in giving it try myself but an a little unsure of the input filtering, what is best RC, LC, both? Could a transistor be used as an active filter/pre-amp here?

Really just looking for some general direction here, am I just wasting my time with that idea? If so, any recommendation on suitable op-amps ~1.00 GBP? If not then are there any other possibilities I might have missed?

Note it does have to be for thermocouples, not thermistors or PT100/1000 etc.

Thanks for reading.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I had trouble getting a max6675 to read 212 deg F boiling water until I realized the pins I had crimped onto the tc leads were like little batteries generating spurious temperature sensitive microvolts. I since then I got the 'Steinhart Equations' to read a Negative Temperature Coefficient sensor, and its about 99% like the Thermocouple eqn and coefficients (well known). I'd try an opamp instrumentation amp on the tc, read the temp of the cold junction with an lm35 or a ds12b80, and just solve the tc eqn using fp. Worth a try I think. Cheaper than a max. Read the maxim-ic appnote on cold junction compensation.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I'd try an opamp instrumentation amp on the tc, read the temp of the cold junction with an lm35 or a ds12b80, and just solve the tc eqn using fp. Worth a try I think. Cheaper than a max.

Ditto to what Bob says. I like the INA122, a good performing and inexpensive programmable gain instrumentation amp. A simple rc filter at the input will work fine. Using a lookup table is also a simple/good way to do the TC linearization, and not very many table values are even needed.

Tom Pappano
Tulsa, Oklahoma

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You didn't need anything more than AD7708 ADC, and a free ATmega ADC input, in order to do exactly whan you have suggested: up to ten TC readings at once, and with cold temperature compensation. Been there, done that. Simple RC filtering, or even C filtering would be enough for such ADC.

http://www.analog.com/en/analog-...

You perhaps would need also a VRef IC, like REF025, and a 'virtual' ground if you plan to measure negative or low temperatures, but this is easy to do.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks for all your replies, looks like good op-amps are going to be the way forward then :)

I am wondering though, since i'll actually need nearer 20 adc channels, is there a way I can simply measure the delta between pairs using a single adc pin?
Seems like something could be done with the op-amp configuration, just not quite sure what. Time to read up again on op-amp theory me thinks, but any pointers would be greatly appreciated.

Thank you all again for your time.

edit:

Guillem Planisi wrote:
You didn't need anything more than AD7708 ADC

Sorry, I just realised this was not another op-amp suggestion. I will have look at that in more detail and check availability in the UK.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Be warned: delta temp may be is no proportional to delta voltage.

With two of those ADC's you can read 20 TC's without any problem, ten each one. Believe me, it was pretty straight forward. I had achieved <1ºC of accuracy between -150 and 1200ºC (TC K, IIRC), measured with a NIST traceable TC calibrator.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have managed to source AD7708 in the UK from RS @ 8.35 GBP ea. Half that price from Farnell but they're currently not stocking the SOIC package. (About the limit of my soldering skills)

Anyway it looks ideal for my needs, thanks once again for your advice.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Also explore http://www.discovercircuits.com/... and similar sites.

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Alternate possibility: https://www.avrfreaks.net/index.p...

But probably not any cheaper, and it would take quite a while to read 10 channels with one max6675.