Load Cell Signal Conditioning

Thread Starter

elixirnova

Joined Jul 18, 2010
4
Howdy,

I am a mechanical engineering student with limited circuitry knowledge to start...

Anyhow I have a [-500,+500]lbf load cell that is essentially a wheatstone bridge strain gauge. I want to give it either 9Vdc or 12Vdc excitation voltage since I have a regulated power supply that can output those voltages. The load cell is rated for 10V excitation with a maximum of 15V though.

Anyhow here is my problem. I am using a NI USB 6009 DAQ to record the voltage output from the load cell. At zero load I tested the voltage outputs for both 9V and 12V input excitation voltages.
9V input -> 4.19V output
12V input -> 5.65V output

The Full Scale range of operation I hope to use this load cell in would result in a maximum/minimum increase in output voltage to
9V input -> 4.19V +/- 0.027V output
12V input -> 5.65V +/- 0.036V output

The other device I have is the USB 6009 DAQ that can ready +/- 10Vdc but it has limited resolution, if I hook the system up to the DAQ without conditioning the output signal I only get +/-5lbf output accuracy and that is when sampling at 1hz. I would like to maximize how fast I can sample and greatly improve the output accuracy. So I figure I need to "zero" the output voltage with a circuit and I suppose scale the output up to +/- 10V.

I hope I am not asking too complex of a question. I just don't really know where to begin. All I have come up with so far is:
-Low Noise Op-Amp to apply a GAIN to the output signal and power it with the DC voltage source
-Use a voltage divider circuit in parallel with the excitation source voltage to "zero" the output voltage.

I must not be doing any of it right though since the circuits I have thrown together from guides on the net have not done anything that I am wanting them to do :(.

Any help on where I need to look now?

Here is some other info on the load cell
Nominal Input Resistance: 394ohms
Nominal Output Resistance: 350ohms
 

beenthere

Joined Apr 20, 2004
15,819
A load cell is a resistive bridge whose output is take from the difference in output from either path. An instrumentation amplifier is a usual device to receive and amplify this differential signal.

The 10 volt excitation level is what is recommended for sensitivity, and to prevent the strain gauges from self heating. The usual output for the bridge is on the order of 3mv/volt.

A really comprehensive guide to these things is at www.omega.com
 

Thread Starter

elixirnova

Joined Jul 18, 2010
4
I managed to read through most of omega's info on load cells as well as some national instrument information on wheatstone bridges, noise, and other such articles.

I do remember seeing many guides that refer to instrument amplifiers like INAxxx chips I assume? Where should I look to learn about using an instrument amplifier and which one to purchase?

The other issue is the output voltage at zero load not being zero. Is this something instrument amplifiers can handle or do I need to make a circuit to apply a voltage drop to the output?

Edit: I'll look through the omega website further later today and get back to ya.
 
Last edited:

beenthere

Joined Apr 20, 2004
15,819
The bridge signal may not be balanced with no load applied. That is usually due to extraneous stuff like the fixture adding some weight. The offset is converted and saved as a tare, which is subtracted from subsequent readings to remove the offset.
 

Thread Starter

elixirnova

Joined Jul 18, 2010
4
Performing a software TARE wouldn't be a problem if my USB 6009 DAQ had better accuracy at reading analog input voltage.

The only way I could possibly get useful accuracy out of my DAQ would be to go from the 5.65V output I am currently getting down to say 0.5V by having a voltage drop 5.15V or however I would perform a mathematic subtraction) and then scale the "zeroed" voltage output up so that it spans 0-10V.

The labview program I have right now allows me to TARE the voltage output of my load cell (5.65V) so that it reads 0lbf and performs the conversion equation based on the load cell sensitivity (3.002mv/V FS). That works fine, but without a circuit to perform a voltage "subtraction" and subsequent "amplification" to have the output span 0-10V instead of 5.65-5.68V that I currently get.

I hope that clarifies.

I have been messing with this circuit simulator to try and make something that will perform the mathematical routines to the output signal I need performed... Maybe if I figure something out I can post a pic of the circuit.
 

Thread Starter

elixirnova

Joined Jul 18, 2010
4
Well here is a cool guide I found that looks pretty legit.
http://www.nakka-rocketry.net/strainlc.html

Some things I notice about his circuit.
-He uses two separate power sources to power the voltage regulators and they look to be 9V batteries.
-He does not need to balance his output voltage.

Although there is a link to balance the wheatstone bridge I have no idea what the resistances are for R1,R2,R3,R4. Just a general input resistance 394ohms and output resistance 350ohms. I suppose looking at his diagram of a load cell wheatstone bridge I can just hook a multimeter up to each combination of terminals to the wheatstone bridge and figure out the resistances with the multimeter. Then maybe I can use the following calculator and add some zeroing resistors.
http://www.nakka-rocketry.net/vyvazeni_mustku.html

Although I am not sure why, my load cell seems to be VERY far off of zero and I do not know why... hopefully these balancing resistors can fix the huge offset as it seems like good method.

I wonder why he could not have used one 9v battery and hooked it up to both the +5Vdc and -5Vdc regulator... I suppose I'll find out if I give it a go.

Maybe since my output voltage is so far offset from 0 I need to balance the output a different way... perhaps with a second instrumentation amplifier?
 
Top