After a while getting wrong measurements from SCXI-1102!

[Ricbace]

Hallo guys!! 

     So, I have some weird problem here, I'm using Labview 8.5.1 and NI SCXI-1000 and SCXI1600 with cards SCXI-1102 with block conector SCXI-1303 and SCXI-1161.

     I build a program to read 3 thermocouples, voltagedrop, and control some funs, everything is running fine. I also have a Power supply that is controled by RS-232 running fine also.

     The whole system is working fine but, I have a random problem. The problem is that after some hours or days running the program I got wrong number from the NI module, but I received the right number from the power supply from the RS-232 and the program keep running without any error message.

      First I thought that is some problem with the drives so I deleted all tasks and I reinstall everything and check all configurations, like the block connector etc.

      The first time that the problem hapens it was 10 hours after the program start to run... and now... after 5 days running I got the same problem... so the measurements that I received from the thermocouples and voltage are like that:

                                         Voltage    PoweSupply   TC1             TC2               TC3               TC4

3/14/2009 3:31:03 AM      0.143463    2.970000    28.362855    30.588772    25.907197    30.114149
3/14/2009 3:31:03 AM      0.162388    3.310000    28.319368    30.566905    25.893820    30.108349
3/14/2009 3:31:03 AM      0.172766    3.310000    28.249820    30.550180    25.895867    30.114149
3/14/2009 3:31:04 AM      1.071393    3.710000    355.329016    154.680682    142.246452    307.193015
3/14/2009 3:31:04 AM     -0.364152    3.870000    -467.043430    -76.405986    -19953.389723    -2466.626749
3/14/2009 3:31:04 AM     -1.430661    3.880000    -2734481.533474    -857940.095837    -19953.389723    -4888367.277856
3/14/2009 3:31:04 AM     -2.159269    3.890000   -62691484.055380    -21968422.561662    -19953.389723    -82647514.967648
3/14/2009 3:31:05 AM     -2.658947    3.930000   -313587207.471697    -132362824.563333    -19953.389723    -366381634.715659
3/14/2009 3:31:05 AM    -2.998069     3.950000   -796328401.927470    -401481382.453773    -19953.389723    -875074649.761019
3/14/2009 3:31:05 AM    -3.227610     3.990000   -1415933250.424770    -831170646.007739    -19953.389723    -1503203536.040207

So as you can see ... suddunly.. I got strange (bigger and after negative) numbers from the NI hardware... but my power supply (collunn 3) still works fine... and the program doesn't get any message... it go on running without any error message....

The weird thing is that if I stop the program and run again... it works fine again.... 

I put the initialization of the tasks outside of my loop so I just use  DAQ-mx read.vi   inside of the loop...

Can someone guess what's going on?

Thanks in advance!!

Ricardo

[CoastalMaineBird]

Why did your power supply jump from 2.99 V to 3.71 V in less than a second, right where your TC data started going wonky?

[Ricbace]

The current is controled by a PI code... so its change fast that's is normal situation...  and they are independent equipment, so nothing related with NI hardware, I've just saved the data at the same file....  I just want to show that the problem is not something related with the program... since the measurements from the power supply are in the right range.... and these changes are normal condition....

 The problem is that the thermocouples start to receive crazy numbers.... and that it will be impossible...

Thanks!

[CoastalMaineBird]

I was thinking maybe the power supply controlled the temp-measuring stuff.

Are you using CJC (cold junction compensation)? 

The weird thing is that if I stop the program and run again... it works fine again.... 

That says it's definitely a software problem. 

Has it happened more than once?  Can you make it happen? 

[Ricbace]

yeap.. I'm using CJC and I also checked the block connector if it was the right one..... so anyway... I chose CJC at MAX like.. build in....

no I can't make the error happens ....   I really don't know why after some time ... (unknow time).... the measuring... get that wrong numbers... it happened twice until now... first it was running for 10 hours without problems... then start to receive that negative big numbers  ... but the program was running without stop after get the wrong numbers....... I mean... there is no error message .... so I have to stop the program and see why I was receiving that wrong values.... so I just stopped the program and start again..... and then ... it was reading correctly.....   it's like... when you have a termocoulpe that is not connected... so you get this big numbers..... but my situation is for all inputs....  TCs and the voltage drop....   so it's like I got all inputs open at the same time.....

one detail is when I got the wrong values.... it keeps... I mean... doesn't go back to the right measurement....  so it's necessary to stop the program...

so.... I couldn't figure out any other reason for that.... 

[CoastalMaineBird]

The fact that a program restart fixes things for a time suggests that it's a software problem.

Can you monitor the temperature from the CJC channel? 

If that's going wonky too, then that might tell us something.

Do you have any open channels? By that I mean unwired ones?

If you have open, unwired channels in a differential setup, then maybe they are floating up and up and up in voltage until the mux is overloaded, thus polluting all later channels.

Configurinig the board (at restart time) allows that to drain off.

Tie any unused inputs to analog gnd.  And try again.

[Ricbace]

Hi Steve!

  Thanks for the answer... and sorry for the delay...

  So l'll check if I can monitor the CJC channel.... that it will be reasonable...

  The board has more free channels (28) but I havent' create or configurate then... just the channels that I need.... so I wont have problem with floating up right?

  Where I can find this configuration (at restart time)?  

thanks for your reply!!!

Ricardo 

[CoastalMaineBird]

The board has more free channels (28) but I havent' create or configurate then... just the channels that I need.... so I wont have problem with floating up right?

Yes you might, that's what I'm saying. I'm speculating (I've seen is happen before, but not on your board) that open (unwired) channels float to a voltage that is higher than the mux can keep isolated, so subsequent channels get polluted because they can't be read correctly.

It doesn't matter if they're used or not.

To test this, tie them all to analog ground. 

Where I can find this configuration (at restart time)?

Wherever you initialize the task.  If you create channels in code, that's the place.  If you let MAX do it, then it's done for you when you run the program.

Either way, my speculation is that the act of configuration discharges this electrostatic build-up, so you read OK for a while. When when it builds up to a certain point, the mux cannot handle it.

[Ricbace]

Hallo Steve!

So thanks for the explanation!

It's going to take some time now to answer you because I have started to run again two days before, so I need to wait to finish it or crash 😛

So as soon as I stop, or get some answers I'll let you know!

Thanks!!

[Ricbace]

  Hallo Steve!

   So I have some news, the system now have broken 6 times. Because of the same problem that floating numbers.... I have tied them to the ground, the other channels, like you said but still getting the same problem.

  Now I can see that the problem happens always at the same point, I have a thermo cycling here.... and the system crash when It's almost finish to cooling down .... I have some funs controled by the card scxi-1161 .... do you think it can be something related with this card? I have also changed the slots of the cards, this last time but still crash.

 Thanks!

 Ricardo