Trying to get quasi-static load data from 9234 module with IEPE load cells

[JohnnyF]

I'm trying to acquire load data for what should be essentially a square wave loading function. The current setup I have is cDAQ module 9234 on LV 8.5 I initially was using a 9233 module I had but after searching here and elsewhere it seemed that the 9234, with DC coupling, would allow me to acquire the quasi-static load (the load is applied for 2.5 seconds) and not just the change in load each time the load was applied or removed. So, now I've got the 9234 connected and I get nearly no change in output regardless of whether I have DC or AC coupling selected in MAX. Additionally, it still seems to be only register the initial change in load, as before-this is the larger concern I have.

I had previously configured it as an AI voltage signal and just scaled the input, which at least got me reasonable changes in load when I used the 9233. I had at least expected that I would see the same/similar values with the 9234 configured in AC mode. If anyone can shed some light on this I would greatly appreciate it.

Here is the load cell spec sheet: http://www.dytran.com/img/products/1203V.pdf

I've also attached my code, in the event it might shed some light on the issue (please be gentle-I'm sure there is a better way to code this but the various cleaner methods I tried did not cycle properly).

Thanks in advance for any help!

Message Edited by JohnnyF on 11-13-2008 04:21 PM

[Seth_B.]

Johnny,

Lets first see if we can isolate the problem to either a hardware or a software issue.  Is it possible for you to run the test and just watch the voltage readings in Measurement & Automation Explorer (MAX) Test Panels?  If you open MAX and right-click your 9234, you can select "Test Panels...".  Set the mode to "Continuous" and apply a load to the load cell.  This will let us determine if the trouble is due to hardware limitations or the software programming.

[JohnnyF]

Seth,

Thanks for the response. I ran MAX and manually applied the load but did not see any changes in voltage in the test panels. Just for kicks, I ran the program and captured the data. The itneresting thing is that the magnitudes of the data are reduced when a load is applied when they should be increasing, per the manufacturer's data sheet. I'm at a bit of a loss and realize this info probably isn't much help.

Thanks again.

John

[Seth_B.]

John,

The inverted reading is likely due to the differential wires being reversed.  Try reversing them and that should give you an increasing reading.

Which model of the 1203V load cell are you using and what load are you applying?  Using the sensitivity of the specific model and the load, we should be able to calculate what voltage we should be seeing versus what we are actually seeing.

[JohnnyF]

Hi Seth,

Thanks for the response. I'm not quite sure what you mean on the wiring-it is a 10-32/BNC cable so I'm not understanding what could be reversed? The load cell is the 1203V2, with sensitivity at 10 mV/lbf.

[Seth_B.]

John,

Are you doing any kind of scaling in your DAQ tasks? That might explain the decrease versus the increase then.  It appears that in the data you posted, the baseline reading for the caliper is about 50 mV, thus about 5 lbf.  Then, it is loaded to about 250 mV, so about 25 lbf of compression three times.  The cable force seems to be loaded in tension three times (negative voltage reading).   What was the actual loading during the test?

[JohnnyF]

I was doing scaling within MAX. I took the sensitivity of each sensor (mV/lbf) and created a table scale from there. For one sensor having sensitivity 9.2 mV/LBF, I got:

9.2             mV/LBF

.0092          V/LBF

108.6957    LBF/V

483.4783    N/V

So, from there and using the ±5 V range for this input I got ±2417.391 Newtons corresponding to each end of the range. These are the two values I used in creating the table.

You're correct that the loading on the cable is in tension but I'm actually measuring the force the cable exerts on the housing, which should be in compression. The cable force should have been approximately 25 lbf and in this case the caliper force should have been virtually zero (the wheel was not rotating so actuating the brake caliper wasn't exerting the tangential force that we are using to measure braking force).

Thanks again for helping me try to sort this out.

[Seth_B.]

John,

Generally, the table scale is designed for sensors that have a variable sensitivity, depending on the reading they take.  Instead, your load cell should for the most part be linear.  Thus, we will want to apply a linear scale.  We will calculate the slope based on the sensitivity, as you calculated below.  The offset will be the negative of whatever your typical unloaded reading will be.

Let's set up a task in MAX to test this.  Right-click on your device in MAX and select "Create Task" and create a Voltage task.  When this is done, take a few readings with the cells completely unloaded.  This will give you an idea of the offset inherent in your load cell. (Load cells often have an unloaded offset).  Create a custom scale for your task and use the negative of the unloaded value as the y-intercept and use the sensitivity of your device (in units of Force/Volts) as the slope.  Then, apply a known load to the cell, to test your output.  Usually setting something heavy (like a known 10 kg weight) on the cell works.

Message Edited by Seth B. on 11-21-2008 11:26 AM

[JohnnyF]

Hi Seth,

Thanks for the follow-up. I kind of stumbled upon the linear scale when the table yielded the offset you mentioned. I've got that taken care of, and in talking with Brooks realized that the ideal setup for this sensor would be to provide IEPE excitation wtih DC coupling enabled in the software. I've gotten that set up using some external current sources I had laying around and started to check calibration using some weights. The only issue I have now is that the signal decays quickly. I need to be measuring average force applied for a duration of 2.5 seconds but as you can see from the screen shot attached, the signal decays quickly (this was static application of a 10 lb weight). Is there a way to programatically work around this or somehow compensate for the TC of the sensor? Thanks again!

Also, is there anything that can be done about the sinusoidal component of this signal? Is that somehow related to an AC component or am I just way off base?

Message Edited by JohnnyF on 11-24-2008 07:09 AM

[Seth_B. · Accepted solution]

John,

Both the sinusoidal portion and the exponential decay are indications of either providing AC excitation or using a sensor that uses AC Coupling.  From looking at the spec sheet that you linked and Dytran's website, I cannot determine if the load cell has any sort of internal AC Coupling.  I would recommend you contact Dytran and explain the behavior you are seeing at a constant load to see if this is expected behavior or not.  It's possible that the load cell is primarily meant for dynamic or transient loading and wasn't meant to measure static loads.  Let me know what you hear back from Dytran and we can continue to troubleshoot this.