You may have damaged your thermocouple when you were connecting voltages across it.
Incorrect !
I test it now in my device
and is showing the ambient temp fine, exactly as the other one. (I have 2 of them) And also responsive to my body heat, no problem.
Hmmmm..... we are missing clearly something. But what?
In my mind, an opamp input(s) expect difference of voltages.
So how I see it, in the current circuit, Im not giving it enough voltage to read properly and do its opamp job.
Do you have the + side of the thermocouple connected to the + of the opamp?
How were you making the connection?
As I mentioned before you need to do cold junction compensation. If the thermocouple and the connector at the same temperature then the output will be zero volts.
Exactly as in my circuit diagram:
And HOW TO DO THAT ? This part is a bit obscure to me. It looks like a reference voltage of some sort maybe... or another thermal probe to measure constant the ambient temp maybe? Hmmm You tell me and also come with the cct for it, if you can.
"If the thermocouple and the connector at the same temperature "
-what connector?
The easiest way is to use a thermocouple amplifier that automatically does cold junction compensation.
What you are doing with the breadboard and jumper wires is OK for experimenting and learning how a thermocouple works but you will never achieve any kind of accuracy with that set-up.
To do cold junction compensation you need another temperature sensor to read the temperature of the cold junction. In your set-up there are three. The jucction inside the yellow connector, where the wires connect to the breadboard and where the IC plugs into the breadboard.
1 Like
I dont believe you. Prove it with a cct diagram. I want to see it with my 2 remaining eyes.
bloody hell, you are right. I watched some shorts on youtube.
I do have some glass thermistors ... if they are any good, but from my experience with them, they are changing very slowly.
So basically the [cold junction] is actually the ambient temperature !!! And the uC is comparing the ambient with the [hot junction]. Right? hmmm
-I did not know this stuff !
-Alright then, how to build this [cold junction] module and then how to compare it with the [hot junction] ?
-Now I start to understand why that cable is so long ! hahaha. oh boy. Very interesting.
I also find this:
https://www.youtube.com/watch?v=-BsDLBI166U&ab_channel=EngineeringProf.
also this 'dude' deserves a quote here as well, very good explanations.
https://www.youtube.com/watch?v=AYblSfpKRUk&ab_channel=EEVblog
I start to see and realize even that yellow plug is part of the sensor, because that is another metal on the pins of the plug that are participating in the "temperature difference" problem. Wow, definitely not what I signed for !
No it the temperature of the cold junction which could be at ambient or higher or lower.
Every place you make a connection to the thermocouple, you create another cold junction. You have three that could be at different temperatures.
1 Like
yes, I realize it now. Well, this was a boot camp lesson for me. I didnt know all these things and details until today. Let's close our eyes and make abstraction of this little detail I made. Is working fine ! haha.
- Now, after I watched some YT videos, I put the links there, mister Dave is mentioning at the end of it's video that the "cheap ass digital thermometers" are actually using a constant of 41uV/ºC !!!! So as he said id, not me, it is possible to use this constant and "cheat" the [cold junction], basically bypassing it. We will have a slightly less precision but "Who cares?" !!! Do correct me if I didnt get it right. I hope I did.
So...can we actually implement his idea into our project here?
The 41uV is true if the cold junction is at zero degrees and even then it an approximation since the response is not linear. Like I said if the cold junction and the thermocouple are at the same temperature, then the output will be zero volts. So you can't measure absolute temperature, only temperature relative to the cold junction.
1 Like
I need solutions !
Adafruit has several
1 Like
Hi,
Your thermocouple cable has two wires in it, what colour are they?
What colour are you connecting to gnd?
What colour are you connecting to the amplifier input?
Tom.... 
You mean the pos and neg ends. The yellow plug has them marked on it. I followed the markings as well when I added my small wire to plug it into breadboard, even if I was not completely aware of its importance. So as it is connected now, I have the pos end of the wire, or pin, connected to pos input of the opamp and the neg end of the wire to the gnd, as shown in my earlier schematic.
In all my tests, I did inversed a couple of times my small dupont pins from this wire probe, to check if is doing something different. It didn't. I always put them back in the correct polarity. I believe, the effect will inverse if the wire is not put correct. So if I apply heat to the sensor, I will obtain a reverse reading, counting down, instead of up as it is normal when the temp is increasing. This is after all a piece of wire, not a fragile silicon element where polarity matters and can destroy the thing.
I look under the loupe at the exposed end of the wire, where the welded blob is, and both wires color are the same, its the color of metal, gray and shiny.
My orange plastic plug probes are marked with a K on them so I know they are the right thing. I also find some lists with different probes marking and what their characteristics means. Mine is this:
What colour is the insulation.....
It should be RED = Negative and Yellow = Positive.
Tom....
The isolation is only a white fiber glass mesh tube and the wires are twisted together. Exactly as in the commercial picture.
Thank you @jim-p for all the discuttion yesterday !
I got a MAX6675 K type thermocouple module + it's probe for 3.33$ from Aliexpress.
I could have taken only the naked chips, 4 of them as SOP8, for $4.67, a bit less than 5$, but this project is so specific, I opted for now for that module with a board included.
Good decision
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.