Incorrect reading from Thermcouples

[rchaoua]

I sure hope someone can help me with this.

My application: I have wired six Omega Type T theromocouples into a 40' long refrigerator with three compartments. The refridgerator is set at -30 degrees Celsius. Each room contatins two thermocouples. I am using a field point module to read the thermocouples and then send this data via ethernet (xover cable) to a PC. I use the temperatures in each room to make a series of decisions ( do I turn on a set of heaters to make the room warmer or turn them off to make them colder). I also store the temperatures into an excel file.

My problem: My program seems to run fine (ok there are a few small problems). However, my thermocouples are giving me weird readings. In each room, I read two completely different temperatures. For example. in room 1 one thermocouple reads -13 C while the otehr reads -29 C. This is the same for room 2 and room 3. When I warm up the rooms by turning on the heaters the thermocouple that reads the lower temperature begins to show an increase in temperature while the thermocouple that reads the higher temperature shows a slight decrease in temperature (never more than -31C). Additionally my room 3 does not register a temperature change at all (even though I know the heaters are on and have went in the room to check).  When I was running the application prior to installing my thermocouples into the refridgerator they were running fine. Of course now my thermocouples must travel at most 35' to reach the FP module when before they only traveled 2'.

Does anyone have any expereince or advice for m? This is all very new and I can not figure out what the problem is. I have checked to make sure they were wired up correctly. My thermocouples are "butt spliced" to thermocouple wire wheich runs (at max 35') to the FP module. I have also refered to the Knowledge base article on calibrating my CJC Sensor (under Analog INput #2) in MAX. That did not seem to help any.  Any help at all please...I am getting desperate!

Thanks!

[Matthew_Williams]

HI rchaoua.  I have a couple of suggestions.

1) Make sure that you have the proper TC wires connected to the proper side all the way through;
been there, done that, felt silly.  Have another pair of eyes confirm that 'this color' goes all the way
out to the TC on one side, and 'that color' comes all the way back to here.

2) When you say butt splice, is there another piece of hardware involved between the connected
wires, or do the wires physically touch each other?  Wire to wire only is best, no junction is better.
If you must use a junction, use a TC junction block.

3) Can you use a CB-3 terminal block (with built in CJ)?  I haven't used cFP with an external CJ,
but we have had good results with the one built into the CB-3, and have tested up to several hundred
feet of TC wire.

4) As a last resort, pull all the wires out, spool them up and make sure they work on the bench configured as they
will be installed.

<edit> 5) Are you getting good temps in MAX but not your application?  Do MAX and your application agree?

Given that the sensors are behaving uniquely, it sounds like wiring per sensor vs. a systemic (CJ) issue.

Matt


Message Edited by Matthew Williams on 12-03-2007 01:37 PM

[rchaoua]

Matt,

Thanks for the reply.

1) I have double checked that the TC wires are the same throughout but may check again to be sure and will get someone to double check.

2) The wires are physically touching.

3) I am using a CFP-CB-3 I must of omitted that peice of information. Is there anything special I need to do withthe internation CJ I did the calibration mentioned in the KB article I found?

4) Unforutnately, this requires uninstalling most of my hardware from the box and is very undesirable but if I do not see any results I may have to result to this eventually.

5) No the temps in MAX  and application do not agree iether. The funny thing is when I initially installed everything I tested all of my devices and the thermocouples read very similar readings. When I came back the next day and the refridgerator had been turned on (thus the rooms where colder) I got this weird readings.

I spoke with the company I purchased the thermocouples from and was told to check the voltage readings from the thermocouples manually and compare with what is read by the application. I figure I will do that today. I figure while I am manually checking each TC I will also double check (again) that each wire is connected proerply. I hope it is a wiring isssue. I figure if I cannot get anything working after today I am going to call NI and see if they can help.

Any other thoughts/suggestions.

[Matthew_Williams]

MAX and your application not agreeing is interesting.  I would focus first on
getting good solid readings in MAX, then work on the application.

Does MAX 'find' all of your modules properly?  Have you confirmed in MAX that
it thinks all of the TC inputs are type T?  Deadband set to '0' (for now, at least)?
Saved the configuration in MAX?

Does MAX at least give a proper temperature for the CJ (should be 273.15K +
room temperature (C) of the CB-3; result should be in K).  Don't bother measureing
in the CB-3 yet, just observe whether the CJ in MAX reads about 290-300K,
assuming the CB-3 is at roughly room temperature.  If it does, the CJ and its
calibration are probably ok.

If you have some more TC wire or sensors, try connecting them instead of the
installed sensors to confirm that you get valid readings with short sensors.  Once
this is confirmed, then the problems are most likely in the installed wiring.

I'll think a bit more.  Hopefully you will have an 'A HA!' moment in the mean time 😉

Matt

[rchaoua]

"I would focus first on getting good solid readings in MAX, then work on the application."

I agree.

Does MAX 'find' all of your modules properly?  Yes, I double checked this yesterday

Have you confirmed in MAX that it thinks all of the TC inputs are type T? Yes I double checked this yesterday too.

Deadband set to '0' (for now, at least)? Yes I double checked this yesterday as well.

Saved the configuration in MAX? Yes.

Does MAX at least give a proper temperature for the CJ (should be 273.15K +
room temperature (C) of the CB-3; result should be in K)? It did then I changed the range to Celsius. Of course it is non longer in room temperature so the readings would be different. I will check this again tonight and will return the CJ to the the Kelvin range and check. Though it will not be at room temperature ( it is very cold where the devices are located).

"If you have some more TC wire or sensors, try connecting them instead of the installed sensors to confirm that you get valid readings with short sensors. "

 This is a good idea I will give that a try tonight and see if it works.

Thanks for the help.  I will post my results.

[rchaoua]

Hi...Its been a few days of trying to figure things out. I have gotten no where. I still have the same problem. I checked the wiring and the thermocouples I think the problem is not on the NI side but the thermocouple side. I think my wires are being affected by the temperature in the refridgerator. Thus, they are what is giving me the volage drop and not th e thermocouple itself.

I have a question if the CJC temperature is vastly different the the temps read would that cause a problem?

[Matthew_Williams]

The CJ being at a very different temperature from the sensors shouldn't be a problem.

How cold is it where the FieldPoint unit is located?  The minimum operating temperature
specification for the controller is -25C.  Have you tried unplugging and re-plugging each
of the TC-120 and CB-3 several times since the unit was placed in the cold environment?

Are the CJ and sensors reading properly with MAX, or with the short, test sensors?

Matt

[rchaoua]

Matt,

How cold is it where the FieldPoint unit is located?  It is roughly -30C...It is located in an enclosure and other than the TC problem everything is running fine. I could try changing the temp in the reefer to a warmer temp and see if performance is affected.

Have you tried unplugging and re-plugging each of the TC-120 and CB-3 several times since the unit was placed in the cold environment?No do you think that would help...I will try that this afternoon.

Are the CJ and sensors reading properly with MAX, or with the short, test sensors? I am unsure if the CJC is reading proerply. I do not know the actual temperature in the compartment but from what I can tell it is in the ballpark of the refridgerator's set temperature. I have not tried connecting another TC I planned on doing that this afternoon (this is my second job). My thermocouples are attached to shielded & twisted thermocouple wire (from Omega). The wire runs any where from 10' to 40' to the FP module. In th e"control" room where it is -30C they terminate to the FP module. Last night I warmed up the wire in the control room and notice that my temperature readings had changed. Do you think the wire itself could be the problem? Now I plan on attaching a TC to one of the open channels and checking the readings. Then I was going to check the the temp of the TCs in each room (comparing the measurement with an independent temperature device). I really hope it is not the wire b/c that woudl be a real pain trying to run it again!

Thanks for all of your input Matt!

[Matthew_Williams]

While -30C is outside the specification range for the FieldPoint, I would
expect it to work, at least marginally.  You should, however, try to get
the FieldPoint unit to a temperature within its operating specifications.

With the plugging and unplugging, I was hoping to make sure that you had
good electrical contacts, thinking that if the unit had been warm, then taken
into the cold, there may have been condensation on the contacts that froze.
Pluggging and unplugging should wipe any ice away.

Where are the shields connected?  They should be connected only inside
the CB-3, and the CB-3 common should be connected to the ground lug
on the bottom of the backplane, which should be connected to earth ground.
http://www.ni.com/pdf/manuals/373344a.pdf , page 6.  The shields
should not be connected to anything at the sensor end.

I wouldn't think the wires themselves are bad yet, although that is certainly
possible.  If a channel works with a short sensor, then maybe run a new
sensor wire parallel (even just temporarily) with the installed one to confirm.

Matt

[rchaoua]

Where are the shields connected?  That may be one of the problems I did not connect the ground to the ground lug...I will do that for all of the sheilded wires.Do you think that could cause that big of a problem? (And all of the EE's shriek simulatanesouly) They are not connected to anything on the sensor end.

I doubt the wires are bad as well. That is a good idea about running a wire in parrallel.