Hi, I have a nice, clean project about finished, but ran into a huge problem. I have 3 TC's connected to a custom PCB board that houses a Wemos D1 Pro, and send the data out over wifi. It's all working, until I am finishing up putting it in a nice enclosure. The problem is with the TC connection to the PCB. I'm using MAX31855K's to read the TCs. The TCs works fine when connected directly to my PCB terminal block. As seen in the first photo. I mounted a barrier block on the outside of my enclosure and wired it to the PCB, as seen in the second photo. Now my TC does not give a reading. I thought I read somewhere that you could do something like this. I can hardly wire the TC directly to my board and still have the project look neat and clean, as the TC would have to poke thru a hole in the enclosure. Plus, it would mean the end user having to wire the TC up to the delicate PCB, which will likely be disastrous. Is there a simple way to do what I'm trying to do here? Thanks.
Yeah, I think that's supposed to work. Any voltage form dissimilar metals should cancel-out as long as the additional junctions are the same temperature.
...I only have limited experience with them. All I know about thermocouples is that the voltages are very-low and the voltage changes/differences for small (say, 1 degree) temperature change is very small, so they can be finicky to use. Theoretically, they are ideal because the voltage-temperature relationship is defined by the laws of nature.
Now my TC does not give a reading.
It has to read something. If you are literally getting "nothing" then something else is wrong.
Thanks. finicky... that may be it. I actually only hooked up the first of three TCs when I discovered it not working. Now, after a lot of testing and fiddling, I see that the other two DO work. My PCB has 3 MAX31855K's on it. Two of them will read the temperature thru the extra connection to the Barrier block on the outside of my enclosure. The third, will not. I WILL read it ok if I wire it directly to the PCB. ODD. No idea how to fix this problem. I can make another custom PCB, but maybe a different one won't work???
For home use you should be fine connecting the TC to simple terminals.
Ideally, a thermocouple hot box is needed.
Comments. First, there are connectors and terminal blocks specifically for thermocouple connections.
Second, there are DVMs available with thermocouple connections and temperature reading ability.
Paul
You should get a reading as long as you have continuity. Whether you get the correct reading is another story. Most thermocouple chips and modules include compensation for one set of contacts that include the wrong material (typically called cold junction compensation).
When you go: PCB -> TCa -> TCb -> PCB there is one cold junction pair (PCB->TCa and TCb->PCB).
When you include the barrier block you know have
PCB->Wire->barrier block->TCa->TCb->barrier block ->Wire ->PCB which now includes 3 pairs of cold junctions.
You should get a reading but likely it is incorrect. The solution as Paul_KD7HB alluded is to use a barrier block of the TC materials and to connect the barrier block to the board via TC wire. The connections would now be
PCB-> TCa Wire-> TCa terminal ->TCa->TCb-> TCb terminal -> TCb Wire -> PCB
Thus reducing the number of cold junction pairs back to one.
P.S. my terminology may not be exactly correct but having just worked through this on a multi-million dollar piece of equipment, I know for a fact the theory is sound.
Thermocouples are really tricky things. If you can avoid using them your project will be much better. But you have it nearly finished and it won't be too difficult to save this one.
Every join in the thermocouple wire must be made with the same exact metal as the thermocouple wire. Except for the last join to the PCB with the MAX chip. Look a the special plugs used for thermocouples. The two pins are different colours: different metals.
My suggestion is to use proper thermocouple plugs and sockets on your box. It is probably acceptable to have inline sockets dangling on short leads.
It may take some searching to find a place to buy the proper components. Omega comes up in searches a lot but I have never been able to buy small quantities from them. Newark was the best source I found.
Thanks for all the help. I did some testing with a max31855k breakout board by Adafruit, and it is working fine with the extra jumper to the barrier block outside. I stuck the probe in ice water and got it as cold as I could and it went down to 2.5 C. I'm hoping I can just include an offset value in my code, unless the error from extra connections is not the same across the range of temperatures I'll be using. But it's for a meat smoker, so the max temp will be about 300 or 350 F. and a 5 degree error wouldn't matter if thats the worst it'll be. Mainly it needs to be reliable -- as in -- what ever it is, it doesn't fluctuate randomly, but I don't think it should. I looked up those TC connectors, and did find some panel mount ones, so if I run into problems, I'll get some of them.
SouthernAtHeart:
Thanks for all the help. I did some testing with a max31855k breakout board by Adafruit, and it is working fine with the extra jumper to the barrier block outside. I stuck the probe in ice water and got it as cold as I could and it went down to 2.5 C. I'm hoping I can just include an offset value in my code, unless the error from extra connections is not the same across the range of temperatures I'll be using. But it's for a meat smoker, so the max temp will be about 300 or 350 F. and a 5 degree error wouldn't matter if thats the worst it'll be. Mainly it needs to be reliable -- as in -- what ever it is, it doesn't fluctuate randomly, but I don't think it should. I looked up those TC connectors, and did find some panel mount ones, so if I run into problems, I'll get some of them.
The voltage differences due to use of differing metals in the connections can be either + or -, depending on the metals. Moreover, that voltage will depend entirely on the temperature of that particular junction. But if you are looking for a 5 degree error, none of those differences should matter.
Paul
Sounds promising! Now if only I hadn't fried my custom PCB. I accidentally plugged the wemos in backward AND while it was powered up. Now none of the MAX31855's work, or my 3.3v Mosfet that switches the small motor. The whole thing is pretty cheap, except those Max's... 
It still looks ok, but none of the MAX31855's give a reading, or the MosFet doesn't work. My Wemos does work, though, I've tested it pretty thoroughly on a breadboard. I'm going to make another PCB and be VERY careful with it.
And I'm sure your next PCB will include a polarity-protection diode or MOSFET. 