There are ledgitimate reasons for checking furnace calibrations, but let me clarify what that is.
1. TC's are not necessarily calibrted. Not unless you need "absolute" accuracy. TC's operate against a (non-linear) standard +/- a tolerance. Most furnaces uses TC grade for temp chamber and EXTENSION grade TC wire for the connection to the a-d converter. If you wanted absolute accuracy each and every wire you have would have to be calibrated...and calibrated IN your machine AND every wire you use to profile boards with. 2. The analog-to-digital converters of the furnace are checked against NIST standards for accuracy. Example if 500 degrees Fahrenheit yields a 500mv output voltage for a type K TC, then a TC simulator can feed this to the A-D and the furnace displays what IT THINKS the temp is.
Why should I check this? Electronics fail. I have had Three furnaces fail. One failed 15 deg Celsius. Motorolla once asked me, "How do you know your furnace is working correctly?" It gets hot is not what they wanted!!!
I have run evaluations with "golden board" profiles to evaluate furnace stability. On one particular furnace, a upper zone had failed (heater open). The lower heater (directly below the failed heater) was compensating for the failed unit. Heat rises and the TC above was reporting all systems fine. They were not. The golden board never detected the failure within the UCL and LCL statistical limits derived for that furnace
Do you have your profiler calibrated? If yes, then how could you not have your ovens calibrated?
Profiling is not absolute. Just because your profiler says you reached a peak of 220C on channel 3, DOES NOT mean you actually reached that temp. This is only true if all junctions are cold-junction compensated, a-d converters are linearized and compensated AND your Thermocouple wires are individually calibrated (this means each wire has a calibration compensation table). You can even factor cross-zonal heating uniformity as another variable which WILL affect the resultant. Oh, and furnace TC feedback method also affects results of "golden board" profiling. The uncertainty factor for profiling is larger than you think.
Profiling is not absolute. I have only seen this in laboratories.
reply »