Strain Guage Measurement - Drift Problem

Thread Starter

FlyingSwissman

Joined Apr 6, 2009
6
Hi all,

I intend to measure (small) strains on a metal plate. I have run into a problem which I can not get rid of.

Problem:


I am using a BA 660 amplifier to amplify the signal of a half-bridge setup (see setup.png) consisting of two general purpose strain gage (350 Ohm) of Vishay and two resistors of 270 Ohm. The amplified reading seems to drift and is not very stable. Additionally, it has come to my attention that the excitation voltage of the bridge is not constant and drifts initially.



Could someone please tell me:


  • what could be causing this or a procedure to find the orgin of the problem
  • how I could remove this drift/instability i.e. how I may obtain a stable readin

I have attached:


  • Test setup diagram: setup.png
  • Graph of initial drift of setup: graph-drift.png
  • Graph of Voltage development after initial drift and compensation: graph-drift-after-offset.png
  • Graph of initial excitation drift: graph-drift-excitation.jpg


Additional information:


  • Power Supply (amp):
    12 [V], 42 (sometimes also 39, 40 or 43) [mAmp]
  • Strain Gage:
    Currently temporary attachment (double-sided tape), untouched/moved/altered during data acquisition





Any help would be greatly appreciated,
Thanking you in advance,

-FlyingSwissman

--------------------------------------------------------------------------------------
Edited:
Change made: gaph-excitation-drift.jpg changed
Reason: wrong graph initially uploaded.
 

Attachments

Last edited:

t_n_k

Joined Mar 6, 2009
5,455
Well if the bridge excitation drifts, then this won't help your cause. Although I think your excitation variation graph indicates only about a 30mV change over a couple of hours - which isn't that bad.

What resistors are you using for the 270Ω - are they a high temperature stability type? "Cheap" resistors will not be much use where you are looking at resolving small bridge output changes. It's possible there are thermal currents in the closed plastic box causing problems - is it getting warm inside the box? Perhaps due to the amplifier being in there as well ....

Also, what are the specs for the Vishay strain elements?
 

t_n_k

Joined Mar 6, 2009
5,455
Also, if you have access to a good mV output range voltage calibrator (to provide a stable input signal) you could check that the BA 660 amplifier is behaving "nicely".
 

Thread Starter

FlyingSwissman

Joined Apr 6, 2009
6
Thank you for your reply t n k!

Excitation Drift

Firstly, you made me realize that I had uploaded the wrong excitation drift graph (original post is modified, graph is now correct). The actual drift of the excitation Voltage is not 30 [mV] but rather only about 2 [mV] (in the first hour).

Strain Gage specs

I have the C2A-XX-125LW-350 General purpose strain gage.
Datasheet: http://www.vishay.com/docs/11200/125lw.pdf

Resistor Specs

The resistors are not highly stability type. They are 1%, 0.6W, 270 Ohm resistors. I thought a half bridge configuration should be relatively temperature stable, even with "cheap" resistors. Could this be causing the drift? If so should I invest in a ready made bridge module?

Temperature distribution in box

The amplifier module does not seem to heat up much (subjective judgement) and I think it should remain temperature stable after an hour.

Amplifier stability test

I will log the amplified signal whilst supplying the amp with 1 [mV] from a second, stable D.C. power supply to see if the drift is caused by the amplifier electronics and report the results.



Thanks again for the help, very much appreciated. I am very keen on solving this issue as I have been unable to resolve it for some time now.

-FlyingSwissman
 

Thread Starter

FlyingSwissman

Joined Apr 6, 2009
6
I supplied the amp with a voltage from a second power supply @ 2[mV]. Here is the development of the voltage over 30 minutes. And a linear trendline of the data. (Note: Drift orientation has changed, this is I believe, due to the fact that the input signal is inverse)
 

Attachments

jans123

Joined Jan 30, 2010
20
I believe you have a temperature drift problem caused by the heat the bridge produce. The 270 Ohm resistor I guess is too low, try 470 Ohm instead, that produces less heat, then make sure that there will be no draft, that goes for both the resistors and the gages. Moving air can make your bridge a wind speed meter instead of a tension metering instrument. I’m not familiar with the amplifier you use. It is a good idea to provide the bridge with a stabilised voltage, some amplifiers can do that, otherwise you will have to ad a stab. The voltage over a 350 Ohm half bridge should probably not exceed 10V, but you can try to decrease it to 5V and see how it affects the timing of the drift. If it is different, then you have two possible problems: 1. If the amplifier you use have a regulator for the bridge, then it can produce heat that gives the electronics some headache, but that you can check, by connecting your bridge to the amplifier excitation the same way you show in the diagram, but the amplifiers input you can connect as you did when checking the drift against a fixed voltage. 2. Check your bridge by using four dummies in a full bridge, leaving the strain gages un-connected, this will probably give you an answer if the dummies are your main problem and if they are innocent, then I guess your problem is a heating problem in the strain gages and you must reduce excitation voltage. Remember: Half the voltage gives one fourth power dissipation.
 

jans123

Joined Jan 30, 2010
20
I believe you have a temperature drift problem caused by the heat the bridge produce. The 270 Ohm resistor I guess is too low, try 470 Ohm instead, that produces less heat, then make sure that there will be no draft, that goes for both the resistors and the gages. Moving air can make your bridge a wind speed meter instead of a tension metering instrument. I’m not familiar with the amplifier you use. It is a good idea to provide the bridge with a stabilised voltage, some amplifiers can do that, otherwise you will have to ad a stab. The voltage over a 350 Ohm half bridge should probably not exceed 10V, but you can try to decrease it to 5V and see how it affects the timing of the drift. If it is different, then you have two possible problems: 1. If the amplifier you use have a regulator for the bridge, then it can produce heat that gives the electronics some headache, but that you can check, by connecting your bridge to the amplifier excitation the same way you show in the diagram, but the amplifiers input you can connect as you did when checking the drift against a fixed voltage. 2. Check your bridge by using four dummies in a full bridge, leaving the strain gages un-connected, this will probably give you an answer if the dummies are your main problem and if they are innocent, then I guess your problem is a heating problem in the strain gages and you must reduce excitation voltage. Remember: Half the voltage gives one fourth power dissipation.
 

beenthere

Joined Apr 20, 2004
15,819
Besides the possibilities in post #7, having the dummy resistors track a different temperature than the bridge strain gauges can be another problem. Are the foils characterized for the material they are attached to?

By making the right connections, you can place a full strain gauge bridge on the one fixture and eliminate some of the problems associated with the half bridge (and double the sensitivity).
 

jans123

Joined Jan 30, 2010
20
Besides the possibilities in post #7, having the dummy resistors track a different temperature than the bridge strain gauges can be another problem. Are the foils characterized for the material they are attached to?

By making the right connections, you can place a full strain gauge bridge on the one fixture and eliminate some of the problems associated with the half bridge (and double the sensitivity).
-Yes, that is true… but when the two gages in a half bridge are close together on the same piece of metal they can be quite forgiving as they suffer together. The problem with temperature coefficients occurs mostly when using quarter bridges or if two gages in a half bridge are widely apart, then the biggest problems tend to be if two gages with correct temperature coefficients are on different sides of a structure and the wires connecting them are in an environment with a varying temperature. Then the wires can cause major problems. But of course, choosing gages with a temperature coefficient adapted for the material they are to be glued or welded avoids sometimes a lot of headache.

(Sorry for my limited vocabulary. Swedish is my native language)
 

eblc1388

Joined Nov 28, 2008
1,542
@FlyingSwissman,

Short circuit the bridge input signals of the BA660 amplifier at the junction of the two 270ohm and then log the performance of the circuit.

This will give you a baseline reference of how stable your setup is, unrelated to the strain gauge. Also this will let you know how much of the drifting is caused internally or via the strain gauges.
 
It has been my experience that all DC amplifiers are subject to drift problems. Maybe, there are some new linear devices that have licked this problem, but I have not heard of them. Why not consider the chopper stabilized amplifier circuit for these strain gage signals. Basically the chopper amplifier breaks up the DC signal into an AC signal that is then conditioned and converted back to a DC signal. See the www.Wikipedia dissertation on chopper amplifier specifics.

Cheers, DPW [ Everything has limitations...and I hate limitations.]
 
Top