Interpreting Temperature Measurements

[D_Wood]

Despite the many standards we may all test to, I think the general intent of a temperature test is the same.

For the sake of clarity to my question, the intent of the temperature test is to verify the temperatures obtained in the product do not exceed the acceptable temperature ratings of each of the components used within the construction of the product. Temperatures recorded include such things as polymeric materials, switch bodies, wiring insulation, motor windings, etc.

One other assumption to be made: testing has stopped when steady state conditions / thermal stabilization has bene reached.

My question focuses on the the squiggly lines observed on a chart recorder, or the array of numbers on a spreadsheet collected by a data acquisition system. Do you record the maximum temperature observed - or - do you record the temperature associated with "steady state condition"?

For reference, I have been told (at different times) both methods are correct, and even given both answers from the same approval agency through different engineers.

Thank you,
-D_Wood-

[Jim Wynne]

What you record, or pay attention to, is a function of what you need to know. In other words, if you have a reason for monitoring temperature, or any other process variable, then you should already know what's important to keep track of, thus I don't understand the point of the question.

[Marc]

Maximum temperature since you're looking for the upper limit. A unit may heat up at startup and then cool a bit at stabilization.

A factor to consider - Is device load a factor?

[D_Wood]

It has always been my general practice to record the maximum temperature observed. An overshoot during the early stages of the test may be ignored if it does not repeat. An example may be a bearing. However, if there are "humps" or peaks that repeat over the duration of the test, lets say 10°C above what appears as "steady", then the peak temperature would have to be recorded (my opinion).

Jim,
The problem sometimes stems from the poor, general wording provided in a standard, leaving multiple interpretations to be made. Thus, a company may provide a written test procedure to take the "guess work" out of what interpretation is to be made. This usually coincides with the interpretation handed down from an approval agency. However, as times change, and you go to a different approval agency, a different interpretation may come into play.

Thank you both for your replies.
I recently asked 5 technicians and 3 engineers this same question. I got an even split as how they interpreted what value would be record.

-D_Wood-

[BradM]

D,

If I remember correctly, you are using thermocouples. Thermocouples (made from T/C wire) will yield virtually instant readings.

When I performed uniformity surveys, overshoot is overshoot, and has to be accounted for. The reason for this is your part is seeing that temperature; thus, it should not exceed the maximum allowed.

For maximum allowable overshoot, I recorded the max temperature. For uniformity/ system tolerance, I recorded steady state.

I hope that makes sense.

If your system it "too" sensitive, go with thicker T/C wire, or a recorder that don't record as often

[Marc]

Originally Posted by BradM

When I performed uniformity surveys, overshoot is overshoot, and has to be accounted for. The reason for this is your part is seeing that temperature; thus, it should not exceed the maximum allowed.

For maximum allowable overshoot, I recorded the max temperature. For uniformity/ system tolerance, I recorded steady state.

Same here, but again - An 'Under Load' requirement may exist. It depends upon the reason for the test and the definition of 'steady-state'.

[andygr]

The primary point on the graph is the highest point. Depending on your process you may need to be concerned with the spread of the hysteresis of the controller.
On preprinted charts there is one other thing to watch for that happens ever now and then. Verify that the scale of the chart is accurate. For process with tight tolerances this can make a differance.