MCC134 Giving Common Mode Error only above 1000C

[John Brock]

I have an ungrounded tip Type K TC connected to the Raspberry Pi hat MCC134. It works perfectly and accurately from room temp to 950C. It's flaky between 950C and 1100C returning intermittent Common Mode errors. Above 1100C, it returns only Common Mode errors.

I have replaced the TC with a known good TC's. I have changed the extension wire (which is also Type K of course) to shielded and unshielded and tried various methods for grounding the shield. 

Nothing changes this behavior. At 1100C and above it returns only Common Mode errors. When the temp drops below 950C, it continues to work again perfectly. Any ideas?

[Fausto]

Hello @John Brock.

Is the tip of the thermocouple exposed or in contact with a conductive surface?  If so, consider using a thin film tape (i.e. Kapton tape) or non-conductive epoxy.  Is there a fan or another piece of equipment that turns on when the temperature goes above 950C?  How long is the thermocouple wire and are there any additional connections besides the initial connection at the MCC 134 board?

Regards,

Fausto

[John Brock]

Thank you Fausto for your response. The TC tip is in air, not touching anything and fully insulated from the vessel being measured (electric ceramic kiln). I have tried multiple iterations of shorter wire (down to 4 feet, direct connected with no connectors and up to 15 feet with one set of mini-plugs). Exactly the same result in all cases.  No other equipment comes on during the firing so all is consistent there.

There are two other Type K TC's on the board running near room temp. I have also tried it with those disconnected. Any other ideas? Thanks again for responding.

JB

[Fausto]

Is the thermocouple wire heating up, especially at the screw terminals on the MCC 134?

[John Brock]

N,o the wire would not be hot at the MCC board. I have a monitoring TC there at the board and it's never over 78F.  At the measuring end, however it could get hotter. The TC sticks through the kiln and terminates on a ceramic block insulator. There are screw terminals for the TC wires. It's possible this is terminal block and wiring is get pretty hot. I have not been around when the kiln was at max temp but have found the terminals still hot sometime later.

At what temperature might that cause problems? Is there a way we can compensate for it? I used actual Type K wire rather than extension wire back to the MCC.

[Fausto]

Are the two wires at the end of the thermocouple welded to each other or are they individually connected to screw terminals on the ceramic block insulator?

Are you using one continuous thermocouple wire pair or are there junctions along the length of the thermocouple?

[John Brock]

The tip of the TC rods are welded. The other end of the thick rod leads are clamped in the terminal block with a screw. See photo. 

I have run both short continuous TC wire to the TC and am now using the mini connector show below - so 1 additional connection in the plug.  Same problem appears when I used the short wire with no connectors.

Thank you very much for debugging this with me!

[JRys]

I have my MCC 134 reading a stable 1198 degrees C. I'm thinking your issue has to do with the RFI/EMI electrical noise emitted by electric kilns. Electric ovens and furnaces experience this too. The common mode error is a false alarm caused by the noise disrupting the measurement. The solution is to use a thermocouple signal conditioner with a good analog low pass filter.  The Dataforth 8B47K-13 or 8B47K-14 signal conditioners utilize a 3Hz low pass filter and convert the thermocouple signal to a 0-5 volt signal (proportional to the input range). You would need to switch from the MCC 134 to a MCC 128 to use one. Reach out to www.dataforth.com for more information. 

 

[John Brock]

Thank you JRys for the idea. This may very well be the issue. It makes perfect sense with the phenomenon I'm seeing. I'd like to not loose my investment in the MCC134 board if possible. Is there any sort of conditioning that could be used with the the Type K and keep it in a format compatible with the MCC134? Maybe something in software?

[JRys]

I don't see any way to salvage the MCC 134. You need something with a better low pass filter. We did have a customer a few years ago doing roughly the same as you. They were using our USB-TC to measure kiln temperatures. Their solution was to switch to the USB-2001-TC. I don't know if it will solve your issue but it is not expensive and it can be used on Raspberry PI with our UL for Linux driver. 

https://digilent.com/shop/mcc-usb-2001-tc-single-channel-thermocouple-measurement-daq-device/

https://github.com/mccdaq/uldaq

best regards,

John

 

[John Brock]

Thank you for the information. I'd like to use the MCC 134 if possible because this is a test case for a decision. The other applications we have are less demanding so the MCC134 is economical and I'd like to standardize on those.

Do you think one of these might help?

https://lookingforsolutions.com/collections/thermocouple/products/thermocouple-connector-mv-v-with-noise-reduction-filter-board-model-lfs110-series

[JRys]

A thermocouple measurement consists of measuring its mV signal as well the temperature where it connects to the board. The mV measurement is the hot junction and where it connects to the board is the cold junction. The cold junction measurement is because merely connecting the thermocouple to the board forms another thermocouple that must be accounted for. Connecting a circuit board to the inputs will form another cold junction and that will throw the measurement off. In addition to the cold junction error, the filter module will add error to the hot junction measurement. 

Have you tried using shielded thermocouple wire from the probe back to the MCC 134? 

[John Brock]

I see your points on using the filter. I have tried shielded TC wire. Tried both grounding the shield only at the measurement site to the equipment and then only to the ground pins on the Raspberry Pi. No effect. I then switched to unshielded wire which actually goes to a little higher temp before common mode errors show up.

Thank you for continuing to work on this with me. Any other ideas?

[John Brock]

Just to close up this item: the filter did indeed cause a non-linear off-set as you predicted so I returned it. The good news is the USB-2001-TC has fixed the problems of EMF interference above 1000C. I finished testing last night and it worked beautifully to 1200C.  This is a good solution for us (and the other USB devices) since they can be used with any computer rather than just the Raspberry Pi hat products.  Thanks again for all the help on this issue!

[JRys]

Hey John,

Thank you for the feedback and I'm sure it will serve to  help other customer too. 

Happy holidays,

John