OT, 1 thermocouple to 2 devices

Captain Haddock

Senior Member
Not picaxe related (yet) but does anyone have any idea how to split the output of a thermocouple to 2 different reading devices without messing up the millivolt reading? One of the devices used a pre-made harness that does not use proper thermocouple wire (type k btw) and requires calibration on a pc to correct (poor design realy).
As you are a clever lot I thought this would be a good place to ask.
 

g6ejd

Senior Member
Here's my thoughts.

A thermocouple produces a voltage for a given temperature at the junction of the two wires. So it's a voltage source of quite low impedance. Your type-k is 41uV/Deg-C so at room temperature will produce ~820uV.

You could connect two PICAXE ADC's inputs together as their input impedance is more than the thermocouple and thus read the temperature independently.

For a 20 Deg-C reading = mid-range of the supply (5V) , you should use an Op-Amp with a gain of 2.5V / 820uV = 3048. Or scale the output as you desire for your temperature.

Using an op-Amp, woudl isolate the reading devices from the voltage source and prevent loading and errorenous/non-linear results.
 

nick12ab

Senior Member
I'd have to agree with g6ejd on the op-amp.

You'll also need to ensure that noise does not interfere with the readings and cause the reading to be different at both devices or the reading to jump about all the place. Decoupling capacitors on the power supply rails are important. I can't think of anything better than just putting a small capacitor on the ADC input to each device to help stabilize it right now.
 

premelec

Senior Member
Your real issue might be if the two devices need to be isolated from each other and ground loops... give it a try per suggestions above... hope it works!
 

Goeytex

Senior Member
Attaching a thermocouple to a standard op-amp could be an exercise in time and patience if not futility. It's not that easy. This is why there are dedicated thermocouple amplifiers such as the AD595.

You did not say what the other two devices are that the thermocouple needs to connects to.
Are they looking for microvolts direct from a thermocouple or millivolts from an amplifier ?
 

fernando_g

Senior Member
I'm with Geoytex on this one. Condition first the TC's output with a proper, dedicated thermocouple amplifier IC. Then you can route the signal anyway you like.
 

mrburnette

Senior Member
Not picaxe related (yet) but does anyone have any idea how to split the output of a thermocouple to 2 different reading devices without messing up the millivolt reading? One of the devices used a pre-made harness that does not use proper thermocouple wire (type k btw) and requires calibration on a pc to correct (poor design realy). <...>
I going to take the path less taken on this one and suggest that depending on your specific need for monitoring, splitting beyond use of a simple resistor may not be necessary:
http://www.picaxeforum.co.uk/entry.php?100-Test-Results-On-AD-Using-Supply-Rail-Reference
ALSO
decoupling the lowZ thermocouple from the PICAXE may not be necessary, therefore, an OpAmp may not be necessary:
http://www.picaxeforum.co.uk/entry.php?88-Thought-on-thermocouples

IT is all about requirements... these will translate into the resolution required and the scaling of the thermocouple voltage. There is an amazing amount of simple work that can be done solely by the PICAXE, however.

- Ray

PS: However, there are times when use of an instrumentation amplifier for the thermocouple is necessary; I just finished the prototype of an 8-channel ATmel 2560 based that has 2 channels dedicated to exhaust gas temperature, senders are K-thermocouples. I used the MAX31855 for each sender and then coupled the output to the 2560 via SPI. Other channels are used for oil temperature, water temperature, cylinder head temperature x2, etc. The other senders are thermistors, so an array is used to convert their voltage-divider outputs into readings on the OLED.
 
Last edited:

fernando_g

Senior Member
Ray, I agree with you, there will be instances where one only needs to know whether the temperature is "Cold", "OK" or "Hot". In this case, the KISS-principle fully applies.

But I wanted to make aware to readers -those who may be unfamiliar with the finer details of reading temperature with thermocouples, that all that is required for high accuracy applications is something as trivial as a general purpose opamp.
I also believe that was Geoytex concern.

My other concern, is that with long or multiple wiring runs, noise pickup is a serious possibility. Therefore, it makes sense to amplify the microvolt signals to millivolt levels before splitting them, making less immune to the noise. Lastly the bane of all thermocouple applications: an open thermocouple....which to an unsuspecting user of a wrongly-designed circuit will appear to be a cold temperature, regardless whether the UUT is undergoing a China Syndrome meltdown.
 

Captain Haddock

Senior Member
Thanks for all the input guys, a few things to think about now.
The AD595 looks promising, I will need 8 of them but my boss will be paying :) I have a feeling the thermocouple inputs on the pre-made harness bit (dynamite engine readout) can be calibrated to accept 0-5v analogue input on the pc.
I have a feeling the AD595 can be piggybacked onto the existing readout input without messing up the readings so that would work well.
Thanks again.
 

Captain Haddock

Senior Member
Thanks for all the good info, AD595 was definately the way to go, had a test out with one channel and got the second device reading nicely without messing up the reading on the original and very simple too, as always you lot were spot on.
:)
 

John West

Senior Member
I got worried when I read the phrase "dynamite engine." But I looked it up and am now somewhat less concerned. :)
 
Top