Thermocouple Type 'K'

Out of linear range. I'm a mechanical engineer so that is a guess.

That is well within the type K temp range. Does it have a sheath, or is it a bare TC? If it is 304 or 316 stainless steel, it won't stand up to those temperatures. Inconel is the MOC (material of choice) for that temp range.

Just for fun, I'd stick the TC onto a multimeter to see what you get there. 500° C is getting up around 10 mV if I remember right; you may be hitting some kind of limit in a defective analog card.

http://www.omega.com/temperature/z/pdf/z204-206.pdf

So much for memory. And that was the last organ still working (sigh).

I checked the Tc with temperature controller its measures correctly upto 900 dec

When i connect with the analog card it could not read correctly

Is there any memory problem

Has this problem just arisen on a system that was working ok, or has something been changed? Before getting in too deep check the readout instrument.It is possible that the readout instrument is not scaled to read above 500DegC or the scaling has been changed (many of these instruments can be programmed). If it is just a simple LCD / LED readout and the only thing between the Tc and the readout is the card, then it looks like the card, it may be the scaling on that and there may be some means of trimming the card. You really need a Tc calibrator, or a mV source with an equivalents chart to check the system properly.

Only my memory. I remembered the tc voltage incorrectly. Check the ps voltages on the analog card.

Answer: drift

Any thermocouple is a junction of two dissimilar metals. The interpretation of a temperature value from a millivolt signal comes from millivolt versus temperature reference tables. (yes, it can be a polynomial function or a look-up, but the basis is the standardized reference tables)

Those reference tables are valid only when the junction is made from 'pure' alloys, in the case of type K, chromel and alumel alloys.

High temperatures promote ion transfer between the junction and the protection tube, protection sheath or even the atmosphere. The hotter the junction, the more actively ion transfer takes place. The results of ion transfer is chemical pollution; the junction is no longer made up of its original 'pure' alloys, and most importantly, the standard reference tables do not apply.

The thermocouple in question has a polluted junction and hence it is no longer a type K thermocouple, it's a thermocouple of some unique chemical composition with its own unique and unpublished millivolt versus temperature table. The millivolt signal does not correspond to the published standard thermocouple reference tables, so the temperature value is incorrect.

Drift can be in either direction from the true value (up or down) and of varying magnitude.

There is no cure for a drifting thermocouple, except replacement with a new thermocouple. There is no correction factor for a drifting thermocouple because the chemical make-up is an unknown, hence there's no basis for making a correction.

Industries with critical temperature measurement requirements, like heat treating of aircraft components, require periodic replacement of thermocouples to avoid the worst cases of thermocouple drift.

Since the thermocuple reads correctly on a temperature control you must look at your analog board. Thermocouples must have a cold junction either real or a compensating electronic artifical junction. Even though the thermocouple is rated for a highter temperature, I have cheap multimeter that reads type K TCs but only to 1000 degrees F. So the analog board may not have the correct range. Or there may be a bad connection on the board - when you are dealing with millivolts it does not take much to give a false reading.

You may have to re-calibrate your analog circuit's sensitivity, range and linearity so that it will track and match linearly the thermocouple's temperature linear range.

Thermocouples are linear! One of the problems when designing a calibrator or recorder.

Thermocouples are not linear. There are devices designed to linearize the signal though. Also, these devices have a cold junction reference to account for varying temperature of the instrument itsself. Please see http://www.icpdas-usa.com/products.php?PID=3246 as an example of device of converting TC readings to on demand computer readable digital signals. It's type K range is up to 1372C.

Drink an extra glass of wine and you leave out the "NOT"!

Yes, I wondered if it might be a typo.

I'm giving you a GA for that. TCs are indeed not linear:

Calibrate the temperature measurement system in a calibration laboratory.