1794-IT8 (Thermocouple) Calibrating

[sacsac]

I have two 1794-IT8/A modules on ethernet. They are not reading the correct temperature. According to the user manual, modules on a Flex I/O sytem must be calibrated. I'm using the TB3T base. The plc is a CompactLogix 1769-L35E. I had them configured as type J thermocouple, Fahrenheit format. The user manual talks about calibrating the module, but over Remote I/O. Either way, I found the bits to use in my tags for the module. I followed the process described in the manual of enabling the channels I want to calibrate, toggling the CalibrationClock bit, and waiting for the CalibrationDone bit. It wouldn't do anything. So I wound up changing the configuration of the module from Type J to millivolt. When I did that, I got the CalibrationDone bit, but it wouldn't go off when I turned off the CalibrationClock bit.

Does anyone have any experience doing this? How does the module need to be configured for calibration, J type, millivolt, and what about the data format?

I'm actually coming in behind someone on this. The Cold Juntion Compensators were not put on until this afternoon. But, after they were put on, the modules still didn't read the right temperature. Do these need to be connected during calibration? I may just have to call tech support in the morning. There just isn't any recent documentation that I can find that explains this with very much detail.

Any help will be apreciated.

 

[robertmee]

I've never had to calibrate an IT8 module for flex I/O. I just commissioned the most recent one a few weeks ago, and all channels were dead on. If you don't install the CJ correctly you definitely won't read correctly. I'm wondering if in your attempts to calibrate before installing the CJ, you've messed up the factory calibration and now it won't read correctly either way. Check your wiring, check polarity, and make sure you have no termination points between T/C and your TB3T module. If you can, post a screen shot of one of your channel configurations.

 

[OZEE]

Yeah... I've never had to calibrate a module either. So that leads me to think that you have a hardware issue. And you really haven't told us HOW the readings are wrong (way high, way low, signal seems "inverted", does the reading change when you change the temp? etc...)

1) You say you configured your module for Type J. Are your thermocouples Type J?

2) Do you have any wire connections to anything other than Type J T/C extension wire? If you've connected to copper anyplace, you'll have cold junctions that will screw up the readings. Are you sure it IS a T/C? (I got called in on a repair job awhile back when the T/C wasn't working right -- They had installed an RTD instead....)

3) Do you perhaps have the leads reversed?

4) Is your T/C properly installed? I.e., is it really measuring what you think it's measuring?

5) Have you tried another T/C?

 

[sacsac]

The readings were about 130 degrees high, according to the other guy that was (still kind of is) working on this. This is the first time either of us has used these modules. What I didn't realize was that he didn't have the Cold Junction Compensators on until we had begun trying to calibrate. On the first module, I may have messed up the factory calibration as Robert said, because I got it to take the offset calibration once, at least I got the calibration done bit. I've got the second module that hasn't been messed with. I'm going to change the configurations back to J type and fahrenheit, wire a thermocouple to the second module with the CJC installed and see how it reads.

I believe everything hardware wise is correct. I'll verify this morning, though.

 

[sacsac]

I just wanted to update. Everything hardware wise was ok. There is a problem with the modules. The readings are off by alot. Today the modules were reading about 26 degrees F, and they should've been about 90. We hooked the thermocouples to a meter and it read correctly. The modules had not been reading high as I was told. The decimal was not taken into account. So 260 was actually 26 degrees.

We had a Rockwell support guru come in. He went back over the process the same as we had done and saw the modules were, in fact, reading incorrectly. He attempted to calibrate a module. I watched him toggle the bits and he did it the same way I had done. It didn't work for him either. So he's going to look into it next week. If he doesn't figure out a solution for these modules, we are going to go with two cards that plug directly into the rack.

 

[danw]

Are you sure that you're working in Deg F, not Deg C?

I had an experience with a measured temperature just above a m b i e n t and a combination of errors produced a situation with very similar values to yours. I doubt it's the same situation, since you stated that you're working in Deg F, and a 2nd person has seen the setup, but I'll bring it up, just in case . . . .

If your 26° reading were 26°C, not F, your 26° value is not that far removed your hot end temp of 90°F (32.2°C).

Are you sure the T/C is not connected backwards? Negative wire for a Type J (American color code) is red. T/C negative (red) attaches to the odd numbered terminals on row A: A1, A3, A5, etc.

Here's what should happen:

The CJC reads the temperature of the terminal block and that value is added to the temp reading calculated by the input/amp/A/D on the millivolts generated by the thermocouple gradient.

Suppose the terminal block were 29°C (84°F), a reasonable condition/situation.

The thermal gradient across the T/C with the hot end at 32.2°C (90°F) is 3.2°C [32.2°C - 29°C], so the thermcouple would create a millivolt potential equivalent = 3.2°C.

The analog input reads whatever mV value 3.2°C is and converts it to temperature.
The CJC value (the temp of the terminal block) is added to that mV conversion value.

3.2°C [T/C gradient] + 29°C [CJC] = 32.2°C = 90°F

Here's what I found when a T/C was wired backwards and the units were °C, not °F. My numbers were slightly different, but I'll use your numbers to illustrate.

When the thermocouple is wired backwards the analog input sees a reverse polarity so the resulting mV value is negative, not positive. The thermal gradient is the same, just the polarity is changed with respect to the analog input. So,

negative 3.2°C [T/C gradient] + 29°C [CJC] = 25.8°C = ~ your reading.

If the CJC temperature of the terminal block varies by a couple 10ths of a degree, the numbers match.

Not terribly likely that you've got compound errors, but it's happened to others.

Dan

 

[sacsac]

Yeah, the module was set for F. And it was wired correctly. The Rockwell guy that came went back over the whole procedure doing the same things we did. He finally admitted that there was a problem with the modules.