04-10-2019 07:04 AM
Because PXIe chassis do not allow for control of SCXI (which I discovered late), I am attempting to use the SCXI-1581 for excitation of multiple 100-ohm Pt RTDs, and a PXI-6254 for signal acquisition. I believe the measurement is basically accurate, but it's incredibly noisy. Digital filtering skews the reading, so that doesn't help me. Does anyone know why I would be getting 12-Hz noise from this setup? Is there any recommended special wiring that could help reduce the noise? I tried tying the negative and positive signal inputs to AIGND with balanced resistors of 12kOhm, but this did not help.
04-11-2019 02:15 PM
What other hardware is involved in the system? Can we verify that the 100 uA excitation from the SCXI is constant & stable? (perhaps with a multimeter or ammeter)
I don't see anything immediately suspect in the two devices listed. Other questions I'd have would relate to what "basically accurate, but incredibly noisy" translates to in terms of volts and converted degrees.
04-11-2019 02:19 PM
Thanks for the reply. Yes, I did use a multimeter to verify the SCXI-1581 current at 0.1001 mA. The RTD resistances all read good (~108-112 Ohm). Using a voltmeter and measuring across the channel pins, I read just over 10 mV, which makes perfect sense, but the signals read on the board are really noisy, +/- 20-30 Deg F.
04-12-2019 03:41 PM
So, to verify:
The SCXI module only provides the current excitation.
The RTD's are then read directly by the PXI DAQ card, correct? How is the conversion to degF being done? Custom scale?
What does the noise look like in terms of voltage?
04-12-2019 04:20 PM
The channels are setup for RTD input, external excitation, Pt100, alpha .003851. I don't know the voltage level of noise - I haven't configured them as voltage only.
04-16-2019 03:15 PM
What range do you have the 6254 configured for? A ~100 Ohm reading with 100 uA across it will be 10 mV. Per the specs, if you're in a large measurement range of the instrument (it supports up to +/- 10 V), you could see up to 280 uVrms noise, which, at 100 uA, corresponds to ~2.8 Ohms of RMS noise, which for a typical Pt100 RTD could be ~7 or so degF RMS. And that's just RMS noise, so you could definitely see peaks higher than +/- 7 degF from your target. 20-30 F seems totally reasonable if you're in the max range. If you choose a smaller input range, you could see that reduce drastically.
You're probably specifying the range in degrees F if DAQmx is returning Deg F as your data, so set a min and max that actually represents the temps you want to measure (-40 to 300 F, for example), and that could help reduce the noise.
04-16-2019 03:27 PM
Thanks, but that's not the issue - I set the input range down to 0-100 F, which should have auto-selected the +/-100 mV input range on the card, but the noise was unaffected, and significantly larger than 7 deg F.
I can only think that the SCXI-1581 is defective/noisy somehow. I've not had an opportunity to troubleshoot it; I'll most likely send it in for checkout/repair, or find an alternate solution. In the meantime, we've been using single channel signal conditioning stolen from another data rack. We don't have enough to accommodate the next phase of testing, so if we had an idea of how to fix the issue with the original method, that would be ideal. Or, if NI wanted to simply make it possible to control an SCXI with an M-series board in a PXIe chassis, I could just use an SCXI-1102 . . .
04-16-2019 03:45 PM
"If NI wanted to simply make it possible to control an SCXI with an M-series board in a PXIe chassis, I could just use an SCXI-1102"
You can do this with the SCXI-1349 adapter that goes on the back of your scxi chassis so that you can cable it to the 68-pin connector of your DAQ board (even PXIe X series boards). I'm not sure if NI will still sell them though since SCXI is an EOL'd product line. If the 1581 is noisy, you should be able to scope that noise with a current probe to confirm. A DMM will inherently use filtering and won't show high frequency noise.
Another option is to scrap the SCXI altogether and just use a PXIe-4357 in your PXIe chassis (which I'm assuming you have since you've mentioned PXIe a few time). It'll probably be ~$3,500 with the terminal block, but it'll take Pt100 RTD measurements natively and should be supported for a good long while.
04-17-2019 07:53 AM
Yep, I'm aware of the PXIe-4357; it's been a lead time and budgeting issue. I don't have any PXIe X-series boards, and didn't know about the PXI M-series issue until we attempted to control the SCXI with one. If you're interested, here's the issue:
We do have a PXI-6259, but being that it has the standard connector, it cannot be used in a PXIe chassis.
04-17-2019 08:30 AM - edited 04-17-2019 08:33 AM
Just for context, I worked at NI for 5 years and supported SCXI extensively. There are a couple of things that I think are misunderstandings. I'll attempt to clarify:
In a PXI/SCXI combo chassis like the PXI-1052 or PXI-1050, you must have a standard connector PXI DAQ board in the far right PXI slot if you want to control the SCXI portion of the chassis without using any extra cabling. In this circumstance, the DAQ board communicates along the PXI local bus to do all SCXI digital control AND to receive the analog signals from the SCXI chassis. However, this is a very small use case for how people use SCXI today since the combo chassis are so old.
For ALL PXI or PXIe chassis (including the combo chassis listed above), you can control an external SCXI chassis (or even the built-in SCXI chassis in the combo chassis!) from ANY PXI or PXIe M or X series DAQ board regardless of the backplane connector type. The backplane connector can be full PXI, hybrid PXI, or PXIe. You use a 68-pin cable connected to the front of the DAQ board and wire it to the SCXI-1349 that plugs in to the back of one of the SCXI modules in your SCXI chassis. ALL digital commands and analog signaling go through the cable and have absolutely no interaction with the backplane connector, which is why the backplane connector doesn't matter. The only thing you have to worry about is whether your DAQ board can fit into the PXI or PXIe chassis you have at hand.
Finally, just to be clear for any future readers, you stated, "We do have a PXI-6259, but being that it has the standard connector, it cannot be used in a PXIe chassis." This may be true of your PXIe chasis, but it's not true of all PXIe chasiss since there are several PXIe chassis that offer full PXI backplane connectors. The PXIe-1062Q and PXIe-1065 are the most popular of these variants. If your PXIe chassis has no full backplane PXI connectors, you can also send your PXI-6259 to NI for it to be hybridized. (The bottom part of the top connector will be removed.)
Again, I want to be clear. As long as you have a chassis slot that accommodates your hybrid PXI-6254 OR your full backplane PXI-6259, you absolutely CAN use either to control your SCXI chassis. You just need the standard 68-pin cable and the SCXI-1349 adapter.
Let me know if any of that was unclear or if I'm misunderstanding your core issue.