NI 9237 on Scan Interface: Noisy Data

[ace.bot]

I am having trouble obtaining an acceptable signal on a 9237 strain gauge module. Application is a single axis servo reporting encoder position plotted against load. I would like to collect a data point every 10ms.

Loadcell: 1010AKC-500-B. Signal is 1mV/V.

Hardware:

cRIO-9075: Scan Interface programming mode.

9237: Full bridge configuration, 5V excitation. J-50 pin.

• I am using the EX +/-, AI +/-, and drain
• I have spliced together the supplied loadcell cable to as shielded cable from the J-50 terminal on the card. This includes the shield.
• A 10V excitation setting only gives 7.7V excitation when measured.

Loadcell: 1010AKC-500-B. Signal is 1mV/V.

9514 with soft motion module.

LabVIEW 2017 Robotics and the Scan Engine Module. I do not have the FPGA module, as I only need data every 10ms.

The scaled signal (500 lb loadcell) fluctuates +/- 3lbs while sitting idle. Applying filters helps, but I can do nothing to get the noise below +/- 1lb. Averaging enough samples to give a steady signal using a FIFO array gives an unacceptable lag in the data.

I have read the Getting Started with LabVIEW and the Real-Time Module, as well as the compact rio developer guide. From what I can tell, you an wire the real-time scan engine resource (AI0 on 9237) right into your data buffer.

I am at a standstill on this project until I figure this piece out.

• Is this a hardware issue or a software issue?
• What approach should I take to filtering data? Is there parameters within the card I can access? In MAX, it’s easy to set up tasks. Is there a RT equivalent?
• All of the Labview examples have the FPGA module. Is this possible without the FPGA module?

I am attaching a screenshot of scaled data and the VI for collecting it.

[AndyClegg]

Loadcells can be quite noisy, so +/-3lbs on a 500lbs range might be what the loadcell produces rather than any particular problem with noise, electrical interference or issue with the software. It sounds like you are fairly good with the interfacing, but make earthing correct with IO module, sheild and loadcell (just in case).

The 9237 has an internal sampling rate that is an integer of a clock frquency, so you might want to consider if the rate at which you are reading data in scan engine might be giving rise to some problems. In fact I just searched the forums thinking somebody may have asked question previously - only to find it was me posing a question about the 9237, in a cRIO in scan mode, sampling at 10msec (!) back in 2013.

Here it is - https://forums.ni.com/t5/Real-Time-Measurement-and/cRIO-with-NI-9237-Bridge-Module-in-Scan-Interface...

Just a coupe of other ideas:

• Do you get same noise when loaded, it may be noise is different in the working range.
• You could look at sampling faster and then the lag imposed by any filter will be less. You could also try a steeper higher-order filter

Good luck