Thermocouple Response Errors and Overshoot

[sammorris95]

Setup: K-type ungrounded thermocouples in 0.125" steel-sheathed probes -> NI-9219 Analog input module -> cDAQ-9174 -> lab computer running FlexLogger 2025 Q1.

I've noticed two problems:

-When these thermocouples are inserted into sand and then placed in a high-temperature (550*C) furnace, the initial measured response is a slow (minutes) downward temperature transient (contrary to what should happen, a period (minutes) of very slight temperature rise), followed by a gradual rise to furnace temperature. The thermocouples eventually read the correct temperature as the sand equalizes with the furnace, but the initial downward transient is concerning.

-When removing these thermocouples from a high-temperature environment (air or sand) to room temperature air, the measured response overshoots below room temperature by at least 10*C, and takes several minutes to equalize with room temperature. This overshoot and long equalization time is concerning.

The thermocouple probes are new, have proper polarity and read boiling water (the single calibration check point I have access to) accurately. The thermocouple cables are correctly attached to the NI-9219 with no severe kinks or bends. The thermocouples are equalized at room temperature prior to being inserted into the sand, and then equalized with the sand's internal temperature before being placed in the furnace. There are multiple opportunities during the handling of the sand for any static electricity to be discharged. The furnace uses resisitive heaters and inserting the bare probes in the furnace does not cause the downward transient (just in case it was magnetic induction from the heating coils).

I initially used an NI-9213 thermocouple module with the same response, and switched to the NI-9219 so that I could simultaneously run RTDs. I've tried using grounded thermocouple probes with the same response. RTDs have no issues.

Any help in understanding either of these problems would be welcome.

[Darylld]

Hi,

I was able to narrow down the solution to your issue into two points, a primary issue and a secondary issue:

1. Primary issue can be related to the Themal Mass of the sheathed K-type probe:

• The TC can experience a delayed heat transfer to the junction during rapid transitions, due to the presence of the Sheath.
• The non-uniform heating of the sheath creates a transient voltage artifact which leads to overshotting/undershotting. The propagation of this transient voltage could occur when pulling the probe out of the furnace -> the outer sheath cools faster than the insulated junction inside -> EMF mismatch -> leads to undershoot below the ambient temperature

2. Secondary issue (less likely issue) can be related to the Cold Junciton Compensation(CJC):

• The NI-9219 has a built-in CJC which uses internal compensation and filtering to satbilize the CJC from the radiant heat of the furnace. However, if the cDAQ module is very near to the furnace, or near airflow, there can be fast ambient temperature shift at the module itself which can cause the issue you are facing.

So, following are the solutions that I would recommend you to try and let me if it works:

1. Try using a smaller diameter or fine-wire TC for faster response.
2. Shield the NO-9219 and the cDAQ module from the furnace heat or airflow if it is near to the furnace since even a small thermal drift at the module can affect the CJC accuracy especially during transitions.

Please revert back if you need any further assistance.

Thank you