Discrepancy Between Sensor Reading and Thermometer


Thread Starter


I have a customer making liquid food products (sauces, etc), in a steam jacketed kettle. They're measuring temp with a 3-wire RTD, mounted in a flush mount sensor similar to Pyormation's TTA. The sensor is installed near the bottom of the tank, in the steam jacketed area. The top 2 feet or so of the tank are not steam jacketed.


The RTD signal goes to a Prosense XTD transmitter, and the 4-20 signal is looped to a PLC, chart recorder, and panel meter. All of those readings are the same value.

They sometimes have an issue where the value from the transmitter is about 50 deg F higher than the temperature measured by sticking a mechanical thermometer into the top 6 inches or so of the product.

We're not sure if it's an issue with the electronics, or simply a matter of a temperature gradient within the tank.

Any suggestions on ways to narrow down the issue?
Almost certainly a temperature gradient issue.


Walt Boyes, Life Fellow, ISA; Fellow InstMC
Chartered Measurement and Control Technologist,
Member Assoc. of Professional Futurists
Spitzer and Boyes LLC

**Spitzer and Boyes LLC publishes the Industrial Automation and Process Control INSIDER (www.iainsider.com)**
Disconnect RTD from transmitter and try to check your RTD resistance (ohm). Compare it with RTD table (ohm & degC/degF). Check whether the measrument of temperature is the same as output of transmitter. Sometimes transmitter 4-20 mA need calibration.
What instrument was used for the topside measurement?

Vapor actuated thermometer? Liquid-in-glass? bimetallic thermometer?
To what depth was it immersed?
How long was it immersed?
Has the mechanical thermometer ever been cal'd or checked against a standard?

Those mechanical instruments all have a different minimum immersion sensing length and each has a relatively lengthy thermal constant (Tau). Bimets typically have the bimet element welded to the tip of the sheath, and that's the only thermal conduction point. Other heat transfer is radiant through the sheath.

The 4-20mA receivers are clearly all ranged the same, but does that range match the transmitter's output range? If the receivers are 0-200 and the transmitter is 0-150, there's a scaling issue.
That is a design issue involving the cooking process and product testing. Hardly a simple matter.

In the simplest cases (typically non-food) a mixer is used.

In your case, with food products you have a high temperature limit to avoid damaging the product, and a minimum temperature limit to insure proper cooking in addition to duration of the cook.

Even if you resort to multiple sensors, you will still need a mixer as any cook will tell you.
Also the used thermowell can make a difference. If it is too short the influence of the steam jacket can be significant. The RVS material of the thermowell conducts heat better than the product.

Turn off the steam for half an hour and measure again.