What would suddenly cause low voltage readings on our SCXI 1102C / SCXI 1300?

[bluepeeple]

Have 6ma current source exciting 19 RTD's in series; differential voltage picked off for channels 11 - 29 and 100 ohm resistor on channel 30.  Other 4-20ma channels used to measure flow and pressure.  All voltages appear to be offset approximately 100mv.  Not sure if this could be caused by some component failure in our machine, the 1102C, or in the SCXI multiplexer.

[Logan_K]

That is a pretty large offset. Have you used a precision DMM to make sure that you are computing the expected voltage drop across your resistors correctly? A constant offset on all the channels like that makes me think that the current source may not be accurate, or maybe there is a bias resistor somewhere that is sinking current away from the RTDs?

If you are sure that the errors are not external, then it could be a settling time error. What are you measuring on channels 0 - 10? Do you have anything connected to those channels? The SCXI-1102/B/C modules have a non-intuitive scanning behaviour where they must start their scan at channel 0, and must scan all channels sequentially between 0 and the final channel in your scan list. So, if any of channels 0 - 10 are disconnected, then they could be floating which could result in settling time errors from the multiplexer.

So, try shorting channels 0 - 10 to ground to see if that helps. Also, try reducing your sample rate in order to give your channels more time to settle.

I hope this helps!

Logan Kunitz
National Instruments - Austin
.

[bluepeeple]

I did use a DMM and current loop calibrator to verify the current source and voltages.  The RTD current was confirmed to be 6ma.  The DMM read voltages that were consistently higher than the computer by approximately 100mv.  Channel 0 - 9 are 4-20ma output flow and pressure instruments that now read approximately 0.9 VDC instead of the 1.0 VDC tha I would expect (corresponds to negative flow with machine shut off/idle).  Ch 10 and 31 are spares.  All of the instruments appear to be functional & no open circuits.  I will try to determine status of the spare channels.

[Marni]

Dear Bluepeeple,

What happens when you short just one channel?

What happens when you short all of the channels?

Which terminal block are you using with the SCXI-1102C?

Please give me a description of the pins that you are using for a couple of channels.

Is your SCXI chassis configured under Devices and Interfaces>>NI-DAQmx Devices in the Measurement and Automation Explorer?

Have a great day!

Marni S.

[bluepeeple]

Dear Marni,

Please realize that this machine has run almost non-stop since 2001; suddenly, all of the channel voltages appear to read approximately 100mv low.

We are using a SCXI-1300 general purpose terminal block and have 29 channels of data.  We have a single chassis system; the 1001 SCXI chassis is configured as chassis 1, address 0 and connected to a PCI-6052E NI-DAQ device.  The SCXI-1102 is located in the first slot (module 1) and other cards in the chassis are a 1163, 1162, and three 1124's in slots 2 - 6 respectively; slots 7-12 are open.

Two of the channels are currently shorted, channel 10 and 31; they read approximately -0.100 vdc from the VB software or Measurement & Automation Explorer.  I have not tried disconnecting everything to short all the channels.  The configuration in Measurement and Automation Explorer shows two 1122 modules in slots 7 and 8, but they are empty.  The configuration shows 9-12 as empty.

Any ideas?

Jeff

[Logan_K]

I would like you to elaborate on "suddenly". If you didn't make any changes to the system (i.e. the software is the same, the hardware is the same, and you haven't touched anything) then I am leaning strongly toward a hardware fault or failure. If this is the case, then you really need to do some more debugging to isolate where the error is coming from. It sounds to me like there may be a hardware fault that is diverting current away from the RTDs which would result in a uniform voltage drop. 100mV is a peculiar number, because that is the maximum voltage level of the 1102 when it is set to a gain of 100 - what input range are you using for your inputs? If you have the 1102 set for a gain of 100, then your signals may all be railing at -100mV, which means that you may have a grounding issue, or like you mentioned a potential hardware issue.

So, if nothing changed, there are 2 possible options. The SCXI system has a failure of some sort, or the sensors that you are measuring have a failure somewhere (i.e. a short to ground, or a disconnected wire, etc...). Both of these require further debugging like what Marni was suggesting. I would recommend disconnecting your system and measuring the channels one at a time with one RTD connected at a time with a simple example program to see if you can isolate where the failure is happening.

-Logan K

[bluepeeple]

I cannot find where anything has changed in the system & it had been running basically non-stop for some time when the problem appeared. 

The system appears to be set for a gain of 1, no scaling and +/- 5V or 10V range, depending on whether you look at the SCXI-1102C analog input setup or PCI-6052E.  I did take a second SCXI terminal block, then shorted and sourced various channels, noting the indicated voltage in M&A Explorer and our custom VB diagnostic / calibration application; the same offset appeared to be present on all channels.  I tried disconnecting the other modules in the chassis to see if that had any effect on the voltage measured on slot 1 (the SCXI 1102C); it did not.  I next replaced the 1102C module in slot 1 with a new unit - no effect.

If the problem were limited to only this module, I could likely re-calibrate the system to eliminate the offset.  However, if there is a hardware problem, will it be stable and not affect the other I/O modules?

-Jeff

[bluepeeple]

I cannot find where anything has changed in the system & it had been running basically non-stop for some time when the problem appeared. 

The system appears to be set for a gain of 1, no scaling and +/- 5V or 10V range, depending on whether you look at the SCXI-1102C analog input setup or PCI-6052E.  I did take a second SCXI terminal block, then shorted and sourced various channels, noting the indicated voltage in M&A Explorer and our custom VB diagnostic / calibration application; the same offset appeared to be present on all channels.  I tried disconnecting the other modules in the chassis to see if that had any effect on the voltage measured on slot 1 (the SCXI 1102C); it did not.  I next replaced the 1102C module in slot 1 with a new unit - no effect.

If the problem were limited to only this module, I could likely re-calibrate the system to eliminate the offset.  However, if there is a hardware problem, will it be stable and not affect the other I/O modules?

-Jeff

[Marni]

Dear Jeff,

It sounds like you have one more more RTDs that are bad. Disconnect all of your signals. Connect an RTD. Do you still get the offset? If you do, try other RTDs. If you don't get the offset, keep adding RTDs until you get the offset. Have a great day!

Sincerely,

Marni S.

[bluepeeple]

I may not have been clear in the previous post.

I already tried disconnecting the original SCXI-1300 terminal block (with all of the RTD and other instruments connected) from the old SCXI-1102C module, then connected a second/new SCXI-1300 terminal block.  I then tried shorting and sourcing various channels in the new terminal block and read the corresponding voltages in the M & A Explorer as well as the machine's VB calibration and diagnostic application.  This should have eliminated any errors associated with a grounded signal, unexpected open channel, defective terminal block, etc.  The same offset still occurred.

Going a step further, I replaced the old SCXI-1102C with a new unit, still using the new SCXI-1300 terminal block with shorted and sourced channels, and still measured the same (-100mv) offset.  This suggests that the original 1102 module is not the source of the problem. 

Next, to eliminate the possibility that errors were originating from other modules connected to the SCXI-1001 chassis (slots 2 through 6) all of the other modules were unplugged to note any impact on the voltages read in slot 1; no difference.

Hopefully this is more clear.  I'm thinking that the problem is in the SCXI-1001 chassis (power supply, timing circuitry or multiplexer) or the PCI-6052E DAQ device.  Anyone agree?  Other ideas?

Regards,

Jeff