Why does the shunt calibration simulated strain change in MAX using cdaq with NI-9236?

[ARBaker]

Hello - Looking for an explanation or clarification. I'm using a NI-9172 cdaq chassis and used a NI-9178 cdaq chassis with a NI-9236 quarter bridge module. When I first create the channels in a task in MAX and perform the first offset null and shunt calibration the simulated strain within the dialog box is the same for all channels. This is how it should be I believe since I am using a shunt of 100Kohms. See two images below that show an initial calibration:

When I perform the calibration I measure what is simulated and then get a close to 0% error so this makes sense:

When I finish this out and then come back into the calibration screen the simulated strains are all different and therefore the error % are also way off, but it still calibrates successfully and the measured strain does show that it measures what the simulated strain should be:

This happens with multiple modules, chassis and computers. I have used this hardware for quite some time and have not noticed this problem before. I was out in the field and had to take data with this issue and I don't think the data is invalid since it does show the correct measured value according to the shunt and the channels seem to respond according to what I expect but I'm still a little leary. Is this some kind of software bug or am I interpreting the screen wrong?

Thanks for any info,

[fromm8]

Hello Adamm

With simulated, you mean that you created a task and you ran the calibration for the device without a real strain connected to it?

Regards

Frank R.

[ARBaker]

Hello Frank - Thanks for the reply and helping me make sense of this.

The simulated strain that is shown in MAX as part of the calibration process is from the shunt resistor that is applied to the wheatstone bridge to simulate a known strain and therefore measure an expected voltage. The gain is then automatically adjusted to scale the voltage correctly. The strain gages have to be connected for the wheatstone bridge to be completed and the shunting to work. There is initial apparent strain from the wheatstone bridge being out of balance to start but that should be accounted for from the offset nulling that happens before the shunting takes place.

What I'm confused about is why the simulated strain changes from the first calibration attempt to the second attempt. The same shunt resistor is being used for shunting and therefore should be simulating the same strain value each time. 

[fromm8]

So you are simulating the strain and you created the strain task for the real 9236, the 9236 is not simulated?

Regards

Frank R.

[ARBaker]

I spoke with an NI application engineer and he was able to reproduce the same issue on multiple strain cards, chassis and computers. Data that is being acquired should be okay since the system is not re-calibrating to the wrong simulated strain values thankfully. So, it seems that it is just a user interface type of bug. NI applications engineer mentioned he will work with the development group to to see what can be done to document and explain this issue or fix in the future.