Incorrect readings from NI 9237 with full-bridge pressure transducer

[Matt_Holmes]

I am using a NI 9237 bridge module in a cDAQ-9171 USB chassis to take pressure measurements from an Omega model PX1009L0-100AV pressure transducer (PX1009 series).  The pressure transducer is a full-bridge strain gauge style.  It is an absolute pressure gauge, with range 0-100psia.  Sitting on my desk, it should be reading something slightly over 11psi (at my altitude).  It is reading approx. -2.5 psi.  I cannot figure out the problem.

I am attempting to collect the data using the daqmx Create Channel vi, with the "AI Pressure Bridge (Two-Point Linear)" choice (reached by selecting analog input; pressure; bridge (two-point linear scale).  I have attached the simplest version of the vi that I can which isolates the problem.  I have also tried the same setup within MAX (created a task) and I get the same incorrect reading of about -2.5psi.

I have attached the data sheet and the calibration record (EDIT: can't attach more than three files, so the calibration will have to wait) for the pressure transducer.  The data sheet states that it is nominal 400ohm resistance.  The transducer has only 4 connections (for the excitation and the signal).  I have the small terminal block (NI 9949) for making my connections; I have placed jumpers in the appropriate positions in order to make the RS-, RS+, and two SC connections.  I have very carefully checked and re-checked these dozens of times.  I have made the connections according to the documentation in the manual; I have attached a pdf which shows the two relevant diagrams needed.

My cal sheet tells me that, when used with an excitation of 10.000V, a pressure of 0psia resulted in a voltage of 0.758mV, and a pressure of 100psia resulted in a voltage of 31.018mV.  I divide the voltage reading by the excitation value, and I obtain for my two points of linear scaling 0psia = 0.0758mV/V and 100psia = 3.1018mV/V.  These are the numbers I enter into the scaling settings on the VI.  I choose units of mV/V and psi. 

I run the VI.  I see approx. -2.5 psi.  If I push on the pressure inlet of the transducer with my finger, I can watch this signal rise slightly, indicating that it is indeed reading something real (and moving in the right direction--positive).  Regardless whether I perform the offset nulling and/or shunt calibration, I get readings in the same range.

I have an Omega strain gauge panel meter, the DP25, and when I configure that according to the (cryptic and stupid) instructions, I get a readout from the transducer of ~13psi (not very accurate, but at least in the right range).  FYI I am not connecting both measurement devices at the same time, just the 9237 OR the DP25, in case anyone was concerned.

So now I'm at a loss as to what the problem might be.  I would be tempted to say that I have the leads mixed up on the pressure transducer, except that it reads roughly correct when plugged into the omega gauge. 

Any ideas? 

thanks,

Matt

[ChaisePotato]

Hey Matt_Holmes,

What voltage are you reading from the transducer using a basic DMM on the signal leads when it is at rest? Do these values correspond to what you are reading in on your NI 9237?

[Matt_Holmes]

Okay, I figured it all out. There was nothing wrong with anything—the behavior I was witnessing was expected behavior.  It was a scaling mistake on my part.  I maintain that the scaling situation can be confusing, as evidenced by how long it took me to sort it out.

Here’s the deal:
The omega pressure transducer has listed on the calibration sheet:
0psi = 0.758mV
100psi = 31.018mV
at 10V excitation

So, I was laboring under the mistaken impression that I simply needed to take the above numbers, divide by the excitation voltage to get mV/V, and then these would be “two-point scaling” numbers that I should enter into the labview vi:
0psi = 0.0758mV/V
100psi = 3.1018mV/V

But if I do that, what happens is that the software first performs a null offset to make the initial reading 0mV/V internally, then uses my scaling—which says that 0mV/V should equate to roughly -2.5psi.  This happens to be the case because they performed the calibration at a factory that was nearer to 14psi ambient.

All of this seems obvious in retrospect, but I was confusing myself because this is an absolute pressure transducer that we’re dealing with, and I was expecting it to read ~11.4psi after the null offset and scaling were performed.  But the thing is, since this is a bridge-style gauge, if you want to zero it, it will be zeroed to whatever the ambient pressure is.  It has no way of “knowing” exactly what imbalance should initially be present in order to equate to the ambient pressure—it just measures imbalance and zeroes it.  So, in effect, if you perform a null-offset in order to function most accurately with the bridge style (pretty much essential), you end up using it as a gauge pressure sensor rather than an absolute pressure sensor.  There is no way to make it function as an absolute sensor without dialing in--manually into the scaling--exactly what offset one should use in order to get the ambient pressure.

So my solution is a two-step process:
First, check the system using the following scaling (which is a “gauge” scaling):
0psi = 0mV/V
100psi = 3.026mV/V
I run this and confirm that the initial reading is 0 and stays at zero.

Then, I use a completely separate, and trustworthy pressure gauge in the lab that gives me an accurate atmospheric measurement of 11.3psi.  

The I change my scaling to be:
11.3 psi = 0mV/V
111.3 psi = 3.1018mV/V

I other words, I manually dial in the appropriate offset by referencing a separate pressure transducer.

Then, finally, my recorded measurements from the VI are accurate and psiA rather than G.

Matt

[Matt_Holmes]

Thank you for responding and trying to help, I do appreciate it even though in the end I discovered that I was in fact the problem...