I am using a cRIO-9012 controller, 9112 chassis, and modules as follows:
1) NI 9237
2) NI 9214
3) NI 9263
4) NI 9265
5) NI 9477
6) blank
7) blank
8) blank
I managed to get my temperature measurements / control working on the NI 9214 using the Wait on Interrupt / Acknowledge Interrupt nodes similarly to what is shown in the shipping example for that module. When I programmed the NI 9237, I followed a similar structure, since I don't require the high data rate capability of the module. Thus, I am performing a single sample, asserting an interrupt, waiting for it to be read and acknowledged on the LabVIEW RT controller, and then proceeding with another iteration in the FPGA loop. I understand that this is not efficient, but as I mentioned, I don't require a high rate of acquisition. The problem that I am seeing is that the measured signals appear to be very noisy. I am using two pressure transducers (50,000 psi Honeywell TJE) connected to channels 0 and 1 of the NI 9237, with the other two channels open. I have the RJ50 version of the module, with TEDS cables supplied by Honeywell / Hoskin Scientific, and I am able to successfully read and parse TEDS information. I don't suspect a cable problem, as step changes in pressure (tested using shop air <200 psi) are reflected correctly in the displayed signal; however, the signal appears to be randomly noisy, far outside both the 24 bit resolution of the NI 9237 and the specs of the pressure transducer (352 ohm full bridge, 2.84 mV/V) Scaled data shows +/- 30 psi random noise, which I have never seen before with similar transducers. I wonder if it is somehow due to my programming implementation, as the shipping FPGA example for the NI 9237 entails the use of buffered acquisition using a FIFO, and I am merely polling the FPGA I/O node once per iteration of my while loop, with an interrupt asserted each time. I also thought it might be related to the excitation voltage, since I am using internal excitation, and the module may not support both bridges at 10 V, but setting this down to 3.3 V has no apparent effect on the output. I'm stumped. Thoughts?
Sean