Calibration Help: Nonlinear temperature offset?

[mjp]

Hello,

I have the JLD612 and the PT100 thermocouple (http://www.lightobject.com/6ft-PT-100-P ... -P218.aspx) set up with the two blue contacts up to 9 and 10 and the red contact set to 8. This is in a sous vide circuit with an SSR and heating coils. I've attached one of the blue leads to the shielding to stop the fluctuation problem, and the control circuit seems to be working well. However, the actual output of the temperature is inaccurate.

I several good-quality kitchen thermometers that I trust, and I have been using them to calibrate the JLD612, which gave me readings that were ~3 degrees C too low in the warm water. To fix this, I set the offset parameter based on the difference between the JLD612 and my thermometers, and now in ~60C water the readings are very near identical.

However, when I take the probes out and put them in colder tap water (and wait for everything to equilibrate) the JLD612 reads 17.4C and the kitchen thermometer reads 18.8C. At ~55C, they're about .5C off.

This means that either the error in the JLD612/PT100 setup is nonlinear, or I'm doing something wrong in the calibration or setup. Do I need an actual external resistor? As it stands now, I'm nervous about the accuracy of the JLD612/PT100, and don't want to have to calibrate for each new temperature I wish to reach. A few degrees in cooking means the difference between safety and dangerous bacterial growth

Any thoughts?

Best,
M

[mjp]

Also in boiling water it reads ~3c too high, and sometimes I hear a "click" sound and it will switch to reading ~40C. If it take the PT100 out of the water, it "click"s again and goes back to the correct range.

[Tech_Marco]

Did you insert the whole PT-100 probe inside water?? Note, a thermocouple or the pt-100 is not water-sealed and if you submersed it into water, it will casue a problem. Click sound could be coming from the relay . You need to look at the setting on AH and AL.

Marco

[mjp]

Thanks Marco!

I had at one point placed the whole probe in the water, but not during these tests. Is the water issue something that goes away with time (and evaporation) if the probe is kept dry or is the whole probe shot?

How do you recommend calibrating the PT100, or should a functional probe require little calibration?

Best,
Mike

[Tech_Marco]

You need to dry up the probe completely before using it again.
Also, water boiling water doesn't has to be 100'C, it varies according to the atitude of your location. Unless you have a high accuracy temperature reader available, it's hard to get you a prefect precise temperature.

See this: http://wiki.answers.com/Q/How_does_the_ ... t_of_water

Also, 100'C boiling point apply to pure water only. If water you used coming from well or river, temperature of boiling point varies. Simply, just a sodium chloride (salt), you can increate the boiling point way over 100'C!

[mjp]

Thanks, Marco! I'll dry it out and see what happens. So theoretically the offset should be fixed across all temperatures within the proper range if the probe is working properly?

Also, at what temperature is it best to run the auto parameter estimates, or does that not matter?

Thanks again!

Best,
Mike

P.S. I happen to be at sea level and have distilled water, so I'm pretty close to 100 depending on the pressure

[Tech_Marco]

For the Auto Tune to work, you need to setup the sytem in complete form. Means that you need to hook it up to a heaster, a SSR, and the probe, be a "closed loop system". Then hit the Auto Tune. It may take a couple hrs to 20hrs to finish the Auto Tune depends on the case. It varies.

Usually the offset is 1 or 2 degree in most case. Sometimes reverse the two blue may affect a little. It won't be normal if the offset is off by more than 5 degree. Then the probe must be defect I think.

You're at sea level and using pure water, then I think you should get perfect 100'C boiling point. If not, we need to talk to some scientist

Marco

[richiem]

I'll stick an oar in here and suggest that you buy another Pt100 probe -- I think it's possible that the one you have is not resurrectable after its innundation.

Also, if you don't already, use the probe type Pt10.0 setting on the 612 for higher accuracy -- it will really help you keep tight control.

My experience is that, in general, a Pt100 probe with a better than 1% specified accuracy (pretty typically, widely available Pt100 probes have 0.2% accuracy) connected to the JLD612 will give more accurate temperature readings than any other thermo you have. Pt100 probes have been preferred as standards for a very long time simply because they are so accurate. Of course, the accuracy of the reading in the 612's display depends crucially on the linearity computation done my the microcontroller in the 612, which converts the gently exponential rise of the platinum resistor's resistance in the probe into a linear change in temperature; but there's no reason for that to be hugely in error -- it's a well-known correction factor.

[mjp]

Thanks richiem!

I'll let it dry out for a day and see how well it does, and if not, swap it out for another one. I wish there were better manuals for these items...

Best,
Mike