Calibrating in the right range? Ovens - 600 +/-10deg Centigrade

[JohnLMorris]

One of our sister companies have some ovens which they need to have calibrated to say +/-10degC and the critical range is at approx 600deg C. I have noticed that the firm who do the calibration are only looking at it up to about 250degC. Thier reasoning was that 'the range where the temperature controllers are calibrated is to prove the controller is within tolerance, and that there is little deviation from its linear path'.
This does not sound like a very convincing answer to me. I would be grateful of other peoples opinion.

 

[DICKIE]

If your QS 9000 don't forget to make sure you calibration company is ISO 17025 accredited and the calibration cert has stated uncertainty.

 

[Ryan Wilde]

John,
It does not sound correct from your description of the calibration. Generally, a linearization would encompass bottom of range, top of range, and some points in between to prove linearity. Saying that a device is linear while only checking the lower 25% of range does not sound kosher BUT...

Before you accuse them of improper calibration, there are other factors to take into account are:

a) your calibration provider may be following the controller manufacturer's procedures;

b) your calibration provider may be using a calibrator to check the controller without the need to warm the oven at all, and only warms the oven to verify their findings and check the thermocouple/thermister/PRT/whatever.

c) The controller may be a multi-range controller, and they performed linearization in a different range band, then did a bias check in the range that you actually use.

I would contact the manufacturer of the controller or oven and ask them for their recommended calibration procedure and compare their suggested calibration points and method against what your provider is giving you. At a bare minimum, I would tag along and ask questions when they are in the next time. It annoys the technician (trust me, it does!), but it is well within your rights as a concerned and responsible customer.

Another thing that really helps the calibration provider is complete documentation of what you require. Customer requirements are the most important thing on the P.O. to a cal lab (other than the Bill To address), but the standard paperwork that I saw said "Calibrate IAW ISO/QS9000", which really just means that your calibration must be traceable and according to whatever procedures that the provider has documented as satisfactory. Put your specific requirements on the P.O.(ex: Calibrate oven at 600ºC ±10°C and perform linearity calibration IAW...), in bold print with flashing LED's. If your 'contract' requests calibration at 600°C, and they don't do it, they are in the wrong, and you have documentation to back it up. Then you can request a corrective action and that is NEVER a good thing to receive from a customer!

Ryan

[This message has been edited by Ryan Wilde (edited 30 May 2001).]

 

[Jerry Eldred]

There's two halves to my answer, as there are two halves to fully calibrating an oven. The controller is calibrated, and the chamber is verified for uniformity.

If all that is being done is calibrating the controller, there is no way of knowing how well the chamber is functioning. Depending on chamber size, there can be quite a bit of deviation from one part of the chamber to another. A long used method for chamber verification is with a 9-point profile. This originated from back in the MIL-STD-883 days, and is, I believe, also included in some other standardized test methods.

As for calibration of the controller, if it is a digital controller, a zero check for DAC offset, and two points along the scale are satisfactory for assuring the temperature controller is within it's specs. So I wouldn't fault a vendor for verifying accuracy of a controller at less than upper end of scale. HOWEVER, this applies ONLY to simulation method (applying equivalent millivolts via a cold junction from a temperature calibrator: could be a DC calibrator, or a dedicated temp calibrator that outputs millivolts). If they are measuring your control thermocouple that is installed in the chamber in a drywell or other temp bath, I highly recommend it be measured at the normal operating temperature of the chamber for the reason that the chamber is used at a pretty high temp, and so thermal breakdowns that may occur at that temp MAY not show up at lower test temps.

As to the necessity of the chamber profiling, that depends on chamber size. I would reserve judgment as to whether they do or do not need to do that without that information.

------------------