Type K thermocouple inaccurate at high temps?

Breakout boards, sensors, other Adafruit kits, etc.

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
allears
 
Posts: 26
Joined: Mon Sep 05, 2011 5:13 pm

Type K thermocouple inaccurate at high temps?

Post by allears »

I'm using a Type K thermocouple with an AD8495 breakout board connected to an Arduino. I calibrated it as per the instructions, using an ice bath and boiling water. It was off by a little bit at either end, but pretty close. I've incorporated the calibration formula in my sketch.
I'm trying to measure the temperature of the stovepipe on my wood stove. There's a cheap analog stove thermometer magnetically attached to the pipe. When the fire gets going, the analog thermometer reaches a bit over 500F (a bit less than 300C) at the max.
I've wedged the end of the thermocouple between the analog thermometer and the stovepipe, so that they should be as close in temperature as possible. When I fire up the stove, the thermometer and the thermocouple track each other pretty closely, up to about 2-300F. Then the thermocouple levels off, as the thermometer keeps increasing. The thermocouple never gets higher than 300-350F.
I tried powering the AD8495 from the Vin port on the Arduino (with a 12v power supply) instead of the 5v port I had been using. The thermocouple definitely registered a higher max than before, but still 75-100F less than the thermometer.
I don't have any reason to believe that the analog thermometer is super accurate, nor am I looking for that -- within a few percent would be fine. But this is on the order of a 25% difference, and the gap increases as the temperature goes up -- it seems like the thermocouple tops out at a certain point.
Is there some special consideration when measuring higher temps? Is it possible the thermocouple is faulty?
Any advice appreciated.

User avatar
adafruit_support_bill
 
Posts: 88093
Joined: Sat Feb 07, 2009 10:11 am

Re: Type K thermocouple inaccurate at high temps?

Post by adafruit_support_bill »

Like all thermocouples, K-type thermocouples are not linear over their entire temperature range. This app-note for the device discusses the non-linearity and how the AD8495 deals with it:
http://www.analog.com/media/en/technica ... N-1087.PDF

This guide discusses a linearization algorithm for the MAX31855 using the NIST thermocouple linearization coefficients:
https://learn.adafruit.com/calibrating- ... earization

allears
 
Posts: 26
Joined: Mon Sep 05, 2011 5:13 pm

Re: Type K thermocouple inaccurate at high temps?

Post by allears »

Much thanks for your tips.
To the best of my understanding (which is being challenged here), these specs and performance graphs don't really account for such a large variance as I'm seeing. Do you think that might explain a 25% difference at 300C?

User avatar
adafruit_support_bill
 
Posts: 88093
Joined: Sat Feb 07, 2009 10:11 am

Re: Type K thermocouple inaccurate at high temps?

Post by adafruit_support_bill »

You should be closer than that at 300C. Where is the thermocouple amp relative to the heat source? A photo of your setup may be helpful.

allears
 
Posts: 26
Joined: Mon Sep 05, 2011 5:13 pm

Re: Type K thermocouple inaccurate at high temps?

Post by allears »

Thanks again. Amp (and Arduino) is about 18" from the stove, probably at 80F or so. I'll post some photos when I get a chance, might be a couple of days.

User avatar
adafruit_support_bill
 
Posts: 88093
Joined: Sat Feb 07, 2009 10:11 am

Re: Type K thermocouple inaccurate at high temps?

Post by adafruit_support_bill »

One potential source of inaccuracy is if there is a temperature gradient between the amplifier chip and the 'cold junction' (where the
probe attaches to the terminal blocks). This is a common problem if the probe is short and/or the board is mounted near to the heat source.

It is a short distance from the terminal block to the chip. But if you have heat being conducted through the probe wires, the temperature at the point of contact with the terminal block can be higher than the core of the chip.

allears
 
Posts: 26
Joined: Mon Sep 05, 2011 5:13 pm

Re: Type K thermocouple inaccurate at high temps?

Post by allears »

I don't think that there would be a significant gradient between the terminals and the chip -- the amplifier board is a couple of feet from the stove, the full length of the thermocouple lead.
I apologize for not posting pics immediately. It got unseasonably warm, I took down the wires around the living room stove, I got busy with some other stuff... I hope you get the picture. But I will definitely follow up in a few days, as soon as I can. Thanks again for your advice.

allears
 
Posts: 26
Joined: Mon Sep 05, 2011 5:13 pm

Re: Type K thermocouple inaccurate at high temps?

Post by allears »

OK! Sorry I got distracted for a little while, I finally found time to take some pictures.

Here's the overall setup. Arduino and thermocouple amp are on small wood pedestal on the stairs.
Image

Here's a closeup of the thermocouple wedged behind the stove thermometer.
Image

Another overall view, showing the laptop.
Image

A view of the output screen, showing thermocouple output plateauing at about 375F.
Image

Stove thermometer, taken a few seconds later, showing about 450F.
Image

Seems to me the cold junction on the thermocouple amp should be at about room temperature. That's a bit warm near the stove, but probably no higher than 80F.

Let me know if I can supply more info.

User avatar
PaulRowntree
 
Posts: 391
Joined: Sun Apr 03, 2016 12:41 am

Re: Type K thermocouple inaccurate at high temps?

Post by PaulRowntree »

When working with high or low temperatures, I always have a bolted-on connection between the TC and the object. It is hard to see the 'wedged' connection in the picture, but you may want to try to harden that up.

The dial indicator is either welded or hard-bolted to a welded contact.

allears
 
Posts: 26
Joined: Mon Sep 05, 2011 5:13 pm

Re: Type K thermocouple inaccurate at high temps?

Post by allears »

The dial indicator is held to the stovepipe by a magnet. The thermocouple tip is wedged between the dial indicator magnet and the stovepipe.

I'm assuming the thermocouple is making as good contact with the stovepipe as the dial indicator is. However, I realize that this is not a well-controlled setup to say the least, and I wasn't expecting accuracy in any absolute sense.

I was just surprised at the magnitude of the discrepancy between the dial indicator and the thermocouple. Also that they both seemed to track well with each other up to a certain point, and then the thermocouple leveled off.

If you feel this is adequately explained by the less-than-perfect mechanical contact, even though the dial indicator seems to show a higher temperature with a similar contact pressure, then I don't have an easy way to modify the setup to improve the contact.

I very much appreciate the feedback.

Much thanks!

User avatar
adafruit_support_bill
 
Posts: 88093
Joined: Sat Feb 07, 2009 10:11 am

Re: Type K thermocouple inaccurate at high temps?

Post by adafruit_support_bill »

Doesn't look like a setup that would be prone to a cold-junction temperature gradient problem.

What type of magnets are you using to hold the probe to the pipe? The curie point of most rare-earth magnets is in the 300F-350F range. If the magnet temperature is reaching above that, I'd expect them to have lost most of their magnetism.

User avatar
PaulRowntree
 
Posts: 391
Joined: Sun Apr 03, 2016 12:41 am

Re: Type K thermocouple inaccurate at high temps?

Post by PaulRowntree »

Thanks for the correction regarding the magnet.

How confident are you in the accuracy of the stuck-on 'analog' temperature reading?

allears
 
Posts: 26
Joined: Mon Sep 05, 2011 5:13 pm

Re: Type K thermocouple inaccurate at high temps?

Post by allears »

I don't know what kind of magnet it is -- it's just a cheap wood stove thermometer, bought at a hardware store quite a few years ago, that's got a magnet on the back for sticking onto a stovepipe. Here's a photo of the back of the thermometer. The tip of the thermocouple is wedged behind one of the ridges on the back, where it contacts the stovepipe. So at that one point of contact, the thermocouple tip is between the thermometer and the stovepipe. The thermometer is still strongly attached to the stovepipe, magnetically, and the thermocouple tip is pinched and held next to the stovepipe by that force.

Image

I wondered if the magnetic field could affect the thermocouple, but what I've found so far on the subject seems to say that a static field shouldn't have any effect.

And no, I don't have any reason to be confident in the accuracy of the stove thermometer. It could be way off for all I know, although it seems to show a full range of temperatures in a consistent way.

What seemed like an anomaly to me is that the thermometer and the thermocouple agreed very closely up to about 300F, but then began to diverge, with the thermometer registering a smooth rise to 475F and higher, while the thermocouple leveled off at about 375F. It's that discrepancy I'm curious about.

If it's just that these are funky conditions, the thermocouple is poorly connected to the stovepipe, or the like, I can certainly understand that. This is far from a controlled experiment, I was just curious to try to chart the stove temperature variations.

I just wanted to be sure I wasn't overlooking some technical point about thermocouple usage that was skewing the results. It may be that a thermocouple just isn't suitable in this situation.

allears

Posts: 20
Joined: 05 Sep 2011 21:13

User avatar
adafruit_support_bill
 
Posts: 88093
Joined: Sat Feb 07, 2009 10:11 am

Re: Type K thermocouple inaccurate at high temps?

Post by adafruit_support_bill »

That looks like a ceramic magnet. Those are typically good to 475 F or more. That would make sense for a thermometer designed to measure temperatures in that range. And your research is correct: A static magnetic field will not affect the thermocouple reading.

I don't see any other possible explanations for the divergence in readings. We can try replacing the thermocouple amp to see if thaat makes a difference. Please contact [email protected] with a link to this thread.

allears
 
Posts: 26
Joined: Mon Sep 05, 2011 5:13 pm

Re: Type K thermocouple inaccurate at high temps?

Post by allears »

Thanks much. I got immediate response to my email, and placed the order for the new breakout board.

I'll definitely post my results when I get it.

Sure is a pleasure dealing with Adafruit! Great products, great service, and an education too. What's not to like?

Locked
Please be positive and constructive with your questions and comments.

Return to “Other Products from Adafruit”