Sciencemadness Discussion Board

Thermocouples

Arrhenius - 30-5-2009 at 12:28

I'm running a PID controller with a K-type thermocouple. When I ran the device up to 400ºC, the PID read "EEEE" indicating the circuit was open. I thought I had burned up the thermocouple, but now it's working okay again today.

Any idea what's going on? I'm not too familiar with thermocouples, but this seems abnormal. A K-type ought to read well above 400ºC, right?

watson.fawkes - 30-5-2009 at 13:31

Quote: Originally posted by Arrhenius  
When I ran the device up to 400ºC, the PID read "EEEE" indicating the circuit was open.
It's possible you have a temperature-dependent open connection. Isolated heating with a hot air gun, plus close visual inspection, might find this out, if this is it.

Swede - 30-5-2009 at 13:59

I think WF is probably correct. There are teo flavors of thermocouple, grounded, and ungrounded. I forget the specifics, but essentially, any protective sheath becomes part of the circuit with the grounded TC. Ungrounded has the junction isolated from any protective sheathing. Expansion of the TC may e causing a short or grounding.

Did the error happen at EXACTLY 400K? If it did, and it is consistent, it sounds more like the controller... some internal software limit, maybe. If it happens at 390, then 412, etc, it's the probe.

Probe info galore at omega.com, lots of decent reading there. Probes are not expensive, it might be worth investing in another if it is the probe.

Arrhenius - 30-5-2009 at 22:30

Well.. my heat gun only got it up to 250ºC; I guess there's just a lot of air flow. It has a steel sheath around the wires, and a nut on the thermocouple, so I can't actually see much.

I had set the PID to 400ºC, and it cut out just around when the temp got up to the set point.

I do intend to buy one from them, but there are so many i'm not sure what's appropriate.


merrlin - 1-6-2009 at 19:35

You might want to look for sources of electrical interference. If you are using alternating current to drive an electric heater, the AC can couple to the thermocouple and confuse the controller if the power wiring is in proximity to the thermocouple. At RF frequencies (e.g. induction heating), radiation can be a problem. Some PID controllers allow for averaging time base adjustment to compensate for 60Hz noise.

Arrhenius - 1-6-2009 at 19:38

Thanks Merrlin! I really don't have much info on the PID (eBay special... super cheap). I had the thermocouple tucked under the heating tape. The tape doesn't have a ground, and the thermocouple isn't grounded. I'm trying to think how I can isolate the thermocouple from the heating tape while still maintaining good thermal conductivity.

merrlin - 2-6-2009 at 10:42

Quote: Originally posted by Arrhenius  
Thanks Merrlin! I really don't have much info on the PID (eBay special... super cheap). I had the thermocouple tucked under the heating tape. The tape doesn't have a ground, and the thermocouple isn't grounded. I'm trying to think how I can isolate the thermocouple from the heating tape while still maintaining good thermal conductivity.


What time and temperature profiles are you running? I usually like to have a thermocouple closer to the process and at some distance from the heat source. Although the response lag makes tuning the PID controller more difficult, the temperature of the process can be measured with more accuracy.

Arrhenius - 2-6-2009 at 19:31

I'm trying to get a constant reactor temp of 400-450ºC (with accuracy). The reaction involves gaseous HNO3, so I don't think I can actually get a thermocouple in the reaction. As for a time profile, I'm not sure what you're referring to.

DJF90 - 2-6-2009 at 20:18

By time profile I think merrlin is referring to the ramping speed, eg. 2C/min. Otherwise I'm not too sure either.

merrlin - 2-6-2009 at 20:25

Quote: Originally posted by Arrhenius  
I'm trying to get a constant reactor temp of 400-450ºC (with accuracy). The reaction involves gaseous HNO3, so I don't think I can actually get a thermocouple in the reaction. As for a time profile, I'm not sure what you're referring to.


A time/temperature profile is a plot of the temperature vs. time. If you have to heat or cool quickly, the PID controller will typically have to use a larger proportional band and derivative action. If you do not have a requirement for rapid heating and cooling, and your process is more or less steady state, then a narrow proportional band and larger integral action can be used. In general, the more insulation that you can apply over the heating tape and vessel, the lower the temperature gradients will be. With lower temperature variations in the reactor, the thermocouple placement will be less critical. You could also use a low pass filter consisting of a series resistor and a shunt capacitor to reduce the 60Hz noise. The series resistor should small compared to the input impedance of the PID controller. Insertion and removal of the filter can be done to check for impact on accuracy. I'm not familiar with gaseous HNO3 chemistry, but if you are working with a neutral, or better yet, oxidizing environment, It might be possible to use a bare type K thermocouple in the reaction chamber.

Arrhenius - 2-6-2009 at 22:12

I don't believe I can set a ramp with the PID I've got. The heating rate doesn't really matter, and it takes ~1minute to get from r.t. to 400ºC. I'm going to try heating it up again once I get some Kaowool to wrap it in.

Filter on the thermocouple side?

Ya, metals are best left out of the reactor entirely. They interfere with the reaction itself, too.

merrlin - 3-6-2009 at 09:41

Quote: Originally posted by Arrhenius  
I don't believe I can set a ramp with the PID I've got. The heating rate doesn't really matter, and it takes ~1minute to get from r.t. to 400ºC. I'm going to try heating it up again once I get some Kaowool to wrap it in.

Filter on the thermocouple side?

Ya, metals are best left out of the reactor entirely. They interfere with the reaction itself, too.


Attached are two datasheets that have references to low pass filters. The two legs of your ungrounded thermocouple provide a differential input to the PID controller (+,- inputs). You would probably be able to use a cutoff frequency of 10Hz without any problem.

CIO-DAS160x-1x.pdf--On page 33 there is a schematic for an RC low pass filter for a data acquisition board that I have used.

AD8253_IN_amp.pdf-- Pages 19-20 shows another RC filter used at the input of an instrumentation amplifier that I am planning to use.

Attachment: AD8253_IN_amp.pdf (575kB)
This file has been downloaded 755 times

Attachment: CIO-DAS160x-1x.pdf (392kB)
This file has been downloaded 568 times

Arrhenius - 3-6-2009 at 19:57

Thank you.

Swede - 5-6-2009 at 09:01

Arrhenius, before you get really balled up with filters and such, which may very well work, I suggest you try another probe. That would be standard troubleshooting procedure. Look for the simple solution first. A type K probe with no sheathing, just two exposed TC wires welded into a junction, and hanging in air, no contact with anything, then slowly apply heat, maybe an alcohol lamp a few inches below the junction. Have the PID controller's output unconnected. You should see the process temperature (the temp of the probe) real time. In this case, the PID controller is simply acting as a TC thermometer.

I've had excellent luck with a cheapo PID TC/PT100 control, it works great. Here's a Blog Entry on how I adapted it to power the heater using a SS relay.

Arrhenius - 5-6-2009 at 12:49

I shall try that. Your PID looks quite a bit nicer than mine.

Arrhenius - 10-6-2009 at 22:27

Well... I tested the thermocouple above an alcohol lamp, and it cut out at exactly 400ºC.... Cool it off and it comes back at 399C. Is this likely to be the probe or the PID?

Arrhenius - 10-6-2009 at 23:14

:D Nevermind! I got it... would have been nice if it came with the instruction manual. I had the thermocouple type set incorrectly in the PID controller.

markgollum - 11-6-2009 at 19:51

I have a nice type K thermocouple probe in an a nickel superalloy sheath (not sure which super alloy) which I regularly take it up to 950C for up to 5 hours at a time. I have even kept it at 1100C for over an hour and it still looks and works great. I am posting this because when I first built my electric furnace I was under the impression that I would need a very expensive Pt RTD for those temperatures, upon finding out what those things cost, I decided to try to use a regular type K thermocouple but to treat it as disposable; I was pleasantly surprised that it held up.