Hi all,

I am having the (similar) drift issue but it's in different fashion. I have two strain gauges on my sample (one on the compression side and other on the Tension side). I am using two individual channels to measure two individual strains from each gauge and using quarter bridge connection with "dummmy" resistor.

When I use the gauge alone. The results are nice and steady; and am able to nullify the output at some point. But, as soon as I place (glue) the gauge on my sample the voltage started to drift again.

Assuming the possibility of the defective gauge, I tried using brand new gauge on my sample but the drift continued as soon as the sample comes in the picture. I also changed the sample with brand new gauges but couldn't get rid of the drift. Just to avoid any possible temperature affect, I changed the room where the room temperature was monitored, but the drift continued.

When, I interchanged the position of the gauges on the bridge I and II, the drift became negative and continued to -ve 5 volts in merely 10 mins.

As all these indicate the gauge on the sample may be the cause, I checked the resistance on the gauge (with sample) and it comes out to be 120 ohms as it should be.

Any hint what could be the trouble here.

Thanks in advance,

Tkp

 

I would review all of your connections. Remember that any dissimilar metal connections with impart a signal based on temperature. It is not uncommon as a procedure to allow such circuits a warm up time to stabilize before reading commences.

 

Thanks for your reply,

Yes, I did cross-check my circuit. Actually, I am using the NI SCXI-1321 module that connects to the SCXI 1121 (Isolation Amplifier). The only wires in my configuration is the two quarter bridge connections. One that connects to the Dummy resistor (Quarter Bridge connection I) and other connects to the strain gauge on the sample (Quarter Bridge Connection II). Interestingly everything works fine with simple gauge (prior the placing on the sample). The same is repeated for channel 2.

Thanks,

Tkp

 

You may well then be observing bimetalic action, even though your adhesive base isn't metalic. I believe the manufacturer can make some recommendations in regards to the expansion coefficient of the base material and the adhesives they supply.

 

YOu have a new thread now, as the original was pretty old, and not quite the same - http://forum.allaboutcircuits.com/showthread.php?t=22878&page=2

Is it possible that your gauges are characterized for, say, aluminum, and you have placed them on a steel fixture?

 

Thanks to all for your suggestions,

Just to give you brief information about the setup. Our sample is a bone. Actually, we have been using the same setup since past few years and never had any trouble until recently. Earlier we used to put two strain gauges on the bone (one on compression and other on tension) and measure each strain on individual channel with individual bridge for each. But recently, we added one more strain gauge (on the compression-tension surface) and tried to add one more bridge connection on the third axis (again individual bridge) and drift started to show-up.

I am not sure what could have gone wrong in adding one more channel but I also tried to use the two channel system (as we used to do earlier) but even that channels are drifting as well.

Thanks again,

Tkp

 

A schematic would be helpful. How certain are you of your excitation voltage source?

 

I'll try to create a schematic for that. I just cross-checked my excitation voltage and it's consistently 3.3 V.

-Tkp

 

It may be noise or cross-talk. What frequencies are you operating the read at?

 

Hi all,

Thanks again for your valuable information. I agree that it seems like the noise come in the picture as I place the gauge on the sample. By the way my read out sampling frequency is 4Hz and the operating frequency is 50 Hz.

Any hint on getting rid of the electromagnetic noise?

Thanks,

Tkp

 

Can you substitute a 120 ohm resistor for the strain gauge and see what the amplifier output looks like? Any difference with it in free space vice taped to the fixture?

 

What adhesive are you using?

Cyanoacrylate adhesives shrink as they cure, which will cause problems. Filled epoxy that does not change in volume during curing works much better.

 

Yes, I did try that option earlier and with the 120 Ω resistor (in place of the gauge) the output was smooth and steady, even the gauge was ,

I agree with the cyanoacrylate but the strain gauge was placed on the bone 4 days before and the drift was there immediately after I place the gauge on the sample and is still there. I think the glue would have shrunk enough in these day. Also, we have been using the same procedure since past few years but never had any issue with the drift before.

Thanks,

Tkp

 

So, ask yourself what changed?

Did you change the adhesive? Did the manufacturer change their raw materials?

Was something else installed nearby that could cause the interference, like an MRI scanner or X-ray machine?

Do you have fluorescent lamps and the ballast is going bad? Did you change from incandescent to fluorescent lamps?

Did you change power supplies? Is the power supply getting old, and the caps are going south?

Are you buying your components from factory-authorized distributors, or the manufacturer itself? Counterfeit ICs are getting to be a big problem recently.