thermocouople readings are jumpy

[MrDoc]

I have a program that reads up to 100 thermocouples.  It was written in rev 5.1.  I have it in rev 8.5 now, but it needs something.  The rev 5.1 gives a very smooth update to the display graph, and the temp readings are varying in a range of ±0.1°C.  The temp readings of the 5.1 program have been verified as accurate.

The rev 8.5 has the temp readings varying ±0.7, and the display graph gives a very choppy view of the temps. 

I am not a programmer, and only have limited experience with Labview.  The company that I work for has wanted to do the convertion from 5.1 to 8.5 for the past 2 years, but everyone is so busy.  I had some free time here and there and decided to make the upgrade myself, and it does run fine, except for the jumpy temp readings.  The thermocouples are running in a range of room temp to 50°C.  I have run the the calibration of the DAQmx read, but am unsure of if I did it correctly.  

Everyone thinks I did a great job with the conversion, and we would like to run this setup in rev 8.5, if only it was as solid as the rev 5.1 program. 

 I am looking for any input at this point.  Thanks

[Broken_Arrow]

Are you using cold-junction compensation?

Are you updating in the loop at the same speed as the 5.1 program?

Are you taking the same number of samples at the same rate?

Faster sampling, less averaging, and a faster loop speed will result in a choppier display, because you are simply seeing more reality. If you had 100 T/C's that read 0.1 degrees apart from each other - sounds fishy.

Can you attach the loop where you are reading the T/C's?

[MrDoc]

I've tried to add a delay into the loop, but that doesn't fix the problem.

Everything up to and including the SCXI chassis is the same as before.

Sorry I can't seem to get the copy / paste to work.  How do you get the block diagram on here?

[MrDoc]

I am using CJC as built in. 

The rev 5.1 had the provision to use a scan rate and number of scans.  DAQmx doesn't seem to have this.  I used millisecond wait to try and slow the loop down, but the readings were still not tight.

The rev 5.1 had little varations in the high and low readings, as well as each individual T/C only varied 0.1 to 0.2 °C.

The rev 8.5 has about 1.0°C varation between the high and low reading, as well as 0.3 to 0.7 for each individual T/C reading.

That is the change I see, and would like to make it not vary so much.

Thanks for replying.

[Broken_Arrow]

You can add the entire VI by using the Add Attachements feature, located below "Tags" when you are relpying. Just like the picture I attached.

Sometimes as screen-shot is easier. The old fashioned way is to hit Print Screen, then go to Paint and Paste, and save as a jpg or png.

Message Edited by Broken Arrow on 04-15-2010 01:28 PM

[MrDoc]

OK, here is the loop regarding doing the T/C read.

[Broken_Arrow]

ohhh , ouch , um , er , you might find that people are reluctant to open a doc file. Can't you post a VI, or a jpg?

[MrDoc]

Sorry about that.

Here are two block diagrams that I have tried.

I had a coworker sugest that I try using a create virtual channel and put a timing block to comntrol the speed.

The one with the DAQ assist gives me the ability to run a calibration and ALL the T/C come up at about the same temp at ambient. The problem with this one is my original one, that temp jumps around on each update / read and can change temp 0.5°C or more.  T/C should be give more stable output - not?  On rev 5.1 this output is very stable.

The one with the 4 DAQmx icons gives me the ability to control the speed.  The problem here, again running at ambient, is the T/C come out at 4 distinct levels, which are related to the four boards in the SCXI chassis.  Each SCXI 1100 will have it's T/C output at a different temp from the other boards. All T/C are sitting in the same room at the same temp.  The calibration from above lets me elimitate this problem.  But I like the speed control with this one.

Do you know the reason for the different levels?

Or can you explain the unstable ouput?

As my previous post - using the same T/C's and SCXI chassis, only running rev 5.1 on a different PC, the output is very stable.

Thanks

[MrDoc]

Oops forgot the files

[Broken_Arrow]

Hello again, thanks for the JPG's ! The setup looks fine in the DAQmx VI's. I can't comment on the DAQ Assistant.

Do you know the reason for the different levels? : You seemed to have answered this yourself! You have four SCXI-1100 cards, and the differences track the cards, so I suspect it is simply the slight errors in amplifier gain and resistor tolerances across each SCXI-1100.

Or can you explain the unstable output? : 1/2 degree differences per update may indicate a noisy interface.  You say the 5.1 software is on a different computer. Are you using the same DAQ and just transferring it from computer to computer? If not, have you tried swapping the DAQs?  

Also, regarding variances: Is the thermocouple wire being used throughout your comparisons from the same spool? Each manufacturer, gauge size, termination method, wire length, etc can contribute to slight variances. Even the same part number, from different lots, can produce differences.

As a reference for you, I currently have two J-type thermocouples of the same wire length, etc. etc., going into a single SCXI-1102, and they read 0.4 degrees C apart from each other. 

So, I would forget about the T/C to T/C variance and concentrate on the 0.5°C change per update.