0

K-Type Thermocouple resistances affecting readings
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

K-Type Thermocouple resistances affecting readings

by roherman00 on Wed Feb 13, 2019 9:46 pm

Hello,

I have a project where I need to measure temperatures at certain points on essentially a heated block of aluminum. Unknown to me at the time, the thermocouples I ordered were of the grounded type and using the MAX6675 boards I had, they worked great until I added them to the project. Which is when I realized the problem.

I then ordered ungrounded thermocouples, and now I am having issues with the getting any readings that make sense (the temperature reads 40-50°F whereas the grounded thermocouple reads correct 60-70°F).

I measured the unconnected resistance of the grounded/ungrounded thermocouples which are 5ohms/21ohms respectively, and their lengths are 1ft/3ft.

I browsed many forums trying to find if the MAX6675 handles ungrounded thermocouples and any connections/resistors needed if so, or if resistances affect readings (only one forum talked about this and it was a no), but I can't seem to figure out what is going on. I did parallel a 6.8ohm resistor with the 21ohm TC (to make 5ohms) and got a correct output but then nothing much else even on temperature changes.

I am also wondering if it is an option to cut the ungrounded TC wire shorter to make it the same length/resistance as the original grounded but there wouldn't be any turning back if it failed.

I am considering the MAX31855 but I would want to be positive that it will provide a correct output and that it works with ungrounded thermocouples.
The store page does state it is pin-incompatible with the MAX6675, but I noticed it only has an additional 3Vo between the 5V and GND. Could I cut the trace leading to the 3Vo and bridge the 3Vo/5V pads thereby making it pin compatible?

Thank you to anybody who has some insight into this.
Robert

roherman00
 
Posts: 19
Joined: Thu Apr 26, 2012 7:00 pm

Re: K-Type Thermocouple resistances affecting readings

by franklin97355 on Wed Feb 13, 2019 10:29 pm

thermocouples don't work on resistance but voltage. Are the two thermocouples the same type (K)?

franklin97355
 
Posts: 20118
Joined: Mon Apr 21, 2008 2:33 pm
Location: Lacomb, OR.

Re: K-Type Thermocouple resistances affecting readings

by roherman00 on Wed Feb 13, 2019 10:54 pm

There appear so. Both are specified as K-type thermocouples.
I couldn't find any specifications that might give the impression that they perform differently.

roherman00
 
Posts: 19
Joined: Thu Apr 26, 2012 7:00 pm

Re: K-Type Thermocouple resistances affecting readings

by roherman00 on Wed Feb 13, 2019 10:58 pm

On second thought if you would like to take a look. Would not be surprised if I missed something.
https://www.amazon.com/gp/product/B009H ... UTF8&psc=1

roherman00
 
Posts: 19
Joined: Thu Apr 26, 2012 7:00 pm

Re: K-Type Thermocouple resistances affecting readings

by adafruit_support_bill on Thu Feb 14, 2019 7:14 am

The color coding for that is non-standard for a K. In the US it is yellow(+) red(-). The international standard is green(+) white(-).

The MAX31855 requires an un-grounded K-type thermocouple.
The MAX31856 can work with any type of thermocouple probe: https://www.adafruit.com/product/3263

adafruit_support_bill
 
Posts: 71770
Joined: Sat Feb 07, 2009 10:11 am

Re: K-Type Thermocouple resistances affecting readings

by roherman00 on Thu Feb 14, 2019 9:19 am

Now I am more confused because the color-coding on the thermocouples all match (red/blue connectors).
I realize this would mean a T-type thermocouple but I am not sure if they are in fact (as the original ones work perfectly fine). And as this is supposed to be for measuring exhaust temperatures and T-type are for cryogenic applications.

Would there be a way to determine the type of thermocouple it is (K,J,T etc) if it is in fact different? I did measure there resistances at ambient temperature and they differ but from what I've read provided they are not RTDs (resistance would be much higher), it shouldn't affect how a controller uses them.

I've also tried various grounding options per the datasheet (jumper, 1MΩ resistor), and still nothing.

roherman00
 
Posts: 19
Joined: Thu Apr 26, 2012 7:00 pm

Re: K-Type Thermocouple resistances affecting readings

by adafruit_support_bill on Thu Feb 14, 2019 9:39 am

Thermocouples work using the Seebeck effect: https://en.wikipedia.org/wiki/Thermoelectric_effect
They generate a small voltage proportional to the temperature differential from one end to the other. Since the input impedance of the thermocouple amplifier is very high, very little current will flow, so the actual resistance across the probe should have a negligible effect on the reading.

If you have a good multimeter, you should be able to measure the voltage directly at the thermocouple leads to see if it agrees with the expected output of a K-type at a given temperature.
https://us.flukecal.com/Thermocouple-Te ... Calculator

adafruit_support_bill
 
Posts: 71770
Joined: Sat Feb 07, 2009 10:11 am

Re: K-Type Thermocouple resistances affecting readings

by roherman00 on Thu Feb 14, 2019 10:39 am

So I tried two different multimeters and both read 0mV when I connected them. I tried an oscilloscope and I got approximately 60mV from both (at ambient room temperature) and a 80-150Hz frequency for some reason.

From the link you provided it seems they are J-type thermocouples (because K and T have a range outside what was measured), but then I am still confused as to how one works but the other doesn't.

All thermocouples have been at the same temperature and place for more than a day, and I have tried every single one to see if maybe I happened to measure a bad one, but they all measure the same. I also swapped connections if for some reason positive and negative mattered, but nothing changed.

I don't know how much I would trust the oscilloscope reading though either since when I held the "working" thermocouple it rose to approximately 80-90mV which is outside the range of J-types (the link states −8.095 mV to 69.553 mV).

roherman00
 
Posts: 19
Joined: Thu Apr 26, 2012 7:00 pm

Re: K-Type Thermocouple resistances affecting readings

by adafruit_support_bill on Thu Feb 14, 2019 11:49 am

The oscillations in your oscilloscope readings are likely due to some external interference. Thermocouple probes can make fairly effective antennae for picking up EMI.

I'm not familiar with the manufacturer of the probe you linked. But we have seen mis-marked probes before (e.g. J types labeled as K and K-types with the color code polarity reversed).

adafruit_support_bill
 
Posts: 71770
Joined: Sat Feb 07, 2009 10:11 am

Re: K-Type Thermocouple resistances affecting readings

by roherman00 on Thu Feb 14, 2019 12:00 pm

That is what I thought since I was not even aware there was a standard for color-coding until today.

Is there anything particular that should be done when connecting thermocouples? I noticed on the MAX6675 datasheet it shows a ground connection from TC-. Is there anything required for ungrounded thermocouples? Or for the MAX31855?

From what I've read the grounded/ungrounded comparison is irrelevant for single thermocouples except in regards to response time or interference (as you suggested), and that they should operate identically otherwise.

roherman00
 
Posts: 19
Joined: Thu Apr 26, 2012 7:00 pm

Re: K-Type Thermocouple resistances affecting readings

by adafruit_support_bill on Thu Feb 14, 2019 12:19 pm

In most cases, you simply connect the two thermocouple leads to the thermocouple amplifier. Some amps - like the MAX31855 will not tolerate any connection between the probe and GND. So, for example, an un-insulated probe bolted to an engine block would be a problem for a MAX31855 - if the chip was also grounded to the chassis.

With shielded probes, the shield is typically connected to GND at the cold-junction end. The shield is generally not electrically connected to the actual probe.

adafruit_support_bill
 
Posts: 71770
Joined: Sat Feb 07, 2009 10:11 am

Re: K-Type Thermocouple resistances affecting readings

by roherman00 on Thu Feb 14, 2019 2:12 pm

So your comment about EMI causing the interference in my measurements got me thinking. I reattached both thermocouples to compare and I tried a variety of things to see if I could improve the problematic thermocouple.

I put it and the shielded cable inside a pipe and nothing happened. But when I touched the thermocouple to warm it, the temperature dropped to 0°C/32°F. So I moved the whole test away from my computer and the temperature approached the working thermocouple but still about 10-15°F under. Regardless of position, the grounded thermocouple was unaffected and performed correctly.

Finally, I ran a wire from the ground of the circuit and clamped it to the metal shielding of the problem thermocouple and voilà the temperature is reading "correctly", and unaffected by proximity to the computer.

I quickly tested the sensors in some snow and some boiled water to make sure they both read comparable values, and I'm happy to say everything looks to be in order.

As an aside I was expecting the grounded sensor to respond quicker than the ungrounded one (being physically attached to the metal housing), but I was surprised how much quicker the ungrounded one responded. It is likely just a quality aspect (having cost double the grounded one), but it was something of note.

After having spent days prior to posting this question trying to find an answer I'm glad to have found a solution. Thank you for getting me on the correct mindset. I didn't find anything in my searches pertaining to this particular problem (happy to be proven wrong), so hopefully this will help someone having the same issue.

roherman00
 
Posts: 19
Joined: Thu Apr 26, 2012 7:00 pm

Re: K-Type Thermocouple resistances affecting readings

by adafruit_support_bill on Thu Feb 14, 2019 2:26 pm

Good to hear you found a solution.

The response time of a thermocouple depends on a number of variables. There is the thermal mass of the junction plus the shroud and any insulation. There is also the thermal resistance of the shroud and insulating materials. All else being equal, insulated probes do tend to have slower response. But a little more money often buys better engineering and/or better materials.

adafruit_support_bill
 
Posts: 71770
Joined: Sat Feb 07, 2009 10:11 am

Please be positive and constructive with your questions and comments.