LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

k-type thermocouples bugging out at high temperature

Solved!
Go to solution

Hi there,

For our reactor experiments, we have a PID controlled furnace and several heated feed lines which are controlled manually by adjusting a VARIAC. We have 5 k-type thermocouples which are connected to the OMEGA PID, which controls current to the furnace heating elements through an OMEGA solid state relay. These thermocouples are functioning properly, and the furnace can be heated to 1000 K or more.

 

There are 8 k-type thermocouples connected to the various heated feed lines. The first four are connected to an NI 9211 DAQ, and the next four are connected to a second 9211 card. The temperatures of most of the feed lines are maintained around 500K. One of the 8 thermocouples reads temperature near the entrance of the furnace, so will reach temperatures near 1000K.  Interestingly, all of these thermocouples function properly as long as the temperatures never get too high. However, when the thermocouple that reads inside the furnace approaches 1000K, all of the thermocouples connected to the same DAQ card begin to fluctuate rapidly.

 

I have attatached  a bunch of screenshots which I will explain in order to capture the issue I am seeing.

 

capture1.png

 

Shows thermocouples 1-4 working properly.  Thermocouples 5-8 are working properly until thermocouple 7 reaches ~950K, and then all the thermocouples on the same DAQ card start to bug out. Thermocouple 6 is not connected at this point, so you can ignore that plot except to notice that it begins to fluctuate at the same time as all the others.

 

capture2.png

 

Thermocouples 5 and 8 begin to work correctly at the exact moment that Tc 7 is disconnected. Later, Tc 6 is connected and also works properly as long as Tc 7 remains disconnected.

 

capture3.png

 

Things start getting a bit interesting here. This picture shows that I have cut the power to all heating elements, and temperatures are increasing. This also shows that when I reconnect Tc 7, it shows the proper temperature of 965K and all Tc's are functioning properly.

 

capture4.png

 

At this point I realized that it is not that the Tc is malfunctioning at high temperature alone. It is malfunctioning at high temperature *when* the power to the furnace is also connected. Capture 4 shows that the readings on Tc's 5-8 only fluctuate when I plug in the power to the furnace.

 

------------------------------------------------------------------------

So, to recap:

 

(1.) All Tc readings are steady when the furnace is on, but at low temperatures.

(2.) All Tc readings are steady when the furnace is off, but at high temperatures.

(3.) Tc's 5-8 fluctuate when the power to the furnace is on, and Tc 7 is at high temperature.

 

And, as you recall from the beginning of the post, the furnace temperature control is done through an OMEGA PID system. The Tc's 1-8 that are connected to the labview DAQ chassis are not at all connected with the furnace circuit.

 

One final point: if I connect that pesky Tc 7 to a voltmeter with a k-type thermocouple adapter module, it reads the correct temperature and Tc's 5, 6, and 8 work correctly because 7 is not connected to the DAQ card in that case.

 

I read on this other thread (http://forums.ni.com/t5/LabVIEW/Thermocouple-k-type-fluctuation-reading/m-p/1599228#M581857) that grounding can be an issue. But I have "grounded the hell out of everything" and have yet to resolve this issue.

 

I'm really confused by this. Does anyone have any thoughts?

 

 

0 Kudos
Message 1 of 19
(6,339 Views)

It sounds like the DAQ card is starting to pick up 60 Hz noise.

 

What happens if you swapped the DAQ cards?  Does the noise stay with t/c's 5-8, or does it move to t/c's 1-4?

0 Kudos
Message 2 of 19
(6,327 Views)

Possibly your thermocouple weld is failing at high temperatures?

 

Open thermocouple readings can fluctuate all over the map.

========================
=== Engineer Ambiguously ===
========================
0 Kudos
Message 3 of 19
(6,311 Views)

I appreciate your help. The problem changes from 5-8 to 1-4 when I switch the position of the cards in the chassis. The Tc's that are in the same card as the high temp Tc begin to fluctuate.

 

------------------------

capture5.png has a discontinuity when I reverse the positioning of the cards. Then you see heating go as normal for 5-8, and also that 1-4 beging to fluctuate when ~950K is reached.

------------------------

 

After that, I moved the high temperature Tc (previously #7), back to the other card. Then the Tc's on that card began to fluctuate.  So the malfunction follows around the high temperature Tc, it does not follow around the specific DAQ card. Is that consistent with what you would expect for the interference theory?

 

0 Kudos
Message 4 of 19
(6,310 Views)

@RTSLVU wrote:

Possibly your thermocouple weld is failing at high temperatures?

 

Open thermocouple readings can fluctuate all over the map.


It does kind of look like that, as I've seen similar things when connections are not good. But it does not seem to be the issue here, as when I connect the Tc to a voltmeter, it returns a steady signal.

0 Kudos
Message 5 of 19
(6,308 Views)

So the issues move with the thermocouple.  ??  It sounds like you are saying it is one t/c that has the problem but all other t/c's on the same card are being affected?

 

It does make me believe it is interference.  If it was just a bad t/c, I'd expect it not to go back to reading properly at lower temperatures.  But interference on 1 t/c could affect the others since the noise is being introduced into the card through that one t/c.

 

Try just disconnecting that one t/c and see if the other 3 on that card behave properly.  If they do, now you've narrowed it down to 1 t/c.  Try swapping that one out with another one.

 

I'm not sure why it would only see interference at high temperatures and not other temperatures.  But hopefully these tests help you narrow down where the problem originates.

0 Kudos
Message 6 of 19
(6,266 Views)
Solution
Accepted by Ben

Is the PID output constant at full on until the temperature reaches 950 K and then does it go to PWM or other on/off mode as it approaches the setpoint? What if you make the setpoint 800K? Do you see the fluctuations starting near 750 K?

 

I agree that interference is the most likely issue.

 

"grounded the hell out of everything" This is often not the solution and may actually make it harder to solve. What do you measure with your voltmeter (AC and DC) between the high temperature T/C and various grounds, particularly on the DAQ cards? What style T/C? Open, sheathed, grounded, ...?

 

What is the voltage and current supplied to the furnace? Are there fans, pumps, or other motorized equipment connected to the ssytem?

 

Lynn

Message 7 of 19
(6,259 Views)

If you connect a scope instead of a (handheld?) multimeter, what do you see?

(ideal device would be a TEK 7A22 (10µV/Div diff amplifier for 7k scopes :D), but a soundcard mic input with protection would also do (McGyver)  the job )

 

The hint with PWM near setpoint is a good one.   

 

 

Greetings from Germany
Henrik

LV since v3.1

“ground” is a convenient fantasy

'˙˙˙˙uıɐƃɐ lɐıp puɐ °06 ǝuoɥd ɹnoʎ uɹnʇ ǝsɐǝld 'ʎɹɐuıƃɐɯı sı pǝlɐıp ǝʌɐɥ noʎ ɹǝqɯnu ǝɥʇ'


0 Kudos
Message 8 of 19
(6,235 Views)

Some materials that are electrical insulators (like ceramics) at room temperature can become more electrically conductive as the temperature goes way up. Combine that with high-voltage electrical heating elements and you may start seeing stray power-line leakage around your furnace. You could try using shielded isolated thermocouples. Or, with the furnace up to temperature and on and with a digital multimeter set to a millivolts scale, with one meter probe grounded, touch the other meter probe to either of the the disconnected TC terminals/leads. If you see a voltage, then you know that you either need better TC isolation or better common-mode rejection in your readout device.

Message 9 of 19
(6,202 Views)

Common Mode Rejection ????

 

I hve seen situations were one channel exceeds the CMR of the device, all of hte other channel are affected.

 

Ben

Retired Senior Automation Systems Architect with Data Science Automation LabVIEW Champion Knight of NI and Prepper LinkedIn Profile YouTube Channel
Message 10 of 19
(6,182 Views)