Go Down

Topic: Strain Gauges (Read 17516 times) previous topic - next topic

Dave_L

Hello,

I need to use strain gauges to measure the twist/torque strain of a hollow metal pipe. I would like to display the data in real time on a force vs time graph. Can anyone recommend a basic (cheap) model of Arduino that could do this?

Many Thanks

Dave   

retrolefty


Hello,

I need to use strain gauges to measure the twist/torque strain of a hollow metal pipe. I would like to display the data in real time on a force vs time graph. Can anyone recommend a basic (cheap) model of Arduino that could do this?

Many Thanks

Dave   



Any of the Arduino boards would handle this application equally well. The big hurdle is the electrical interfacing needed to attach a strain gauge to a arduino analog input pin. Strain gauges are resistive bridge circuits that only output millivolt range changes and typically require opamp amplification and bridge voltage excitation wiring. A link for the strain gauge datasheet would be useful. Some strain gauges include interfacing electronics that include internal amplification, so again linking to your specific gauge would be helpful.

Lefty


Dave_L

Thanks Lefty,

Im studying aerospace engineering so I am new to the electronics side of things. I was thinking of setting the strain gauges in a wheatstone bridge configuration and then connecting a voltmeter and amplifier chip(any suggestions?) and then connecting this to the arduino. Strain gauge wise would something like this work?
http://vishaypg.com/docs/11334/250us.pdf (I think this is already configured as a wheatstone bridge)

As for the wiring, could you suggest any reading on the voltage excitation wiring?

Any other advice would be great, at the moment Im in way over my head:)

Cheers

Dave


retrolefty

That datasheet is kind of generic to their offerings, a key missing spec is the millivolt change per volt of excitation voltage. That is needed to know how much amplification gain your amplifier is going to need. As far as excitation voltage the simplest method is to just use the regulated +5vdc board voltage. The best op-amp to use is whats called an instrumentation opamp. These have programmable gain set by one resistor and usually have very good low noise specifications. The need of course is that with the strain gauge under it's maximum rated load weight there should be a +5vdc output measurement voltage that can be wired directly to an Arduino analog input pin. This is a 10 bit A/D converter so a measurement between 0 and +5vdc will result in a digital value of 0-1023 'counts'. It's also possible with arduino software commands to change the measurement range of a analog input pin to 0-1.1vdc = 0-1023 counts, by utilizing an internal 1.1vdc bandgap reference built into the micro chip. This would lower the amount of gain needed by the opamp and in some cases might eliminate the need for any external amplifier, but that depends on the raw millivolt range that the sensor outputs through it's load range.

There is lots of information on the web, with example circuits, etc,  if you do a little searching around.

Lefty

liuzengqiang

Lefty,

The strain gauge is just the bare resistive wire, 4 wires in this case to increase sensitivity. I think if you connect 2 and 4 to VDD and 3 to ground, then read the differential voltage between 1 and 5, which scales linearly with the resistance/strain/stress change. Since it's for sheer and symmetrically made, a sheer to the left on the top will increase resistance on the top left and bottom right gauges, while decrease the resistance on the bottom left and top right gauges. This is a few times as sensitive as a single one.It also could compensate temperature variations.

Lots of force sensors have amplifiers that takes the differential voltage and amplifies it and spits out a ratiometric value so zero strain is 50% the VDD. This one is just harder to use without the amplifiers, and yes it needs an instrumentation amplifier and two single-ended ADC or a differential ADC channel(s).
Serial LCD keypad panel,phi_prompt user interface library,SDI-12 USB Adapter

retrolefty

Quote
Lefty,

The strain gauge is just the bare resistive wire,


I wasn't the originator of this post, and I don't think I asked any questions. And I have used strain gauge circuits before. So ?

Lefty

liuzengqiang

I mixed up who's asking what. Lefty, did you mean millivolt per Newton or Pound per excitation voltage?
Serial LCD keypad panel,phi_prompt user interface library,SDI-12 USB Adapter

retrolefty

#7
Feb 12, 2011, 07:29 am Last Edit: Feb 12, 2011, 07:37 am by retrolefty Reason: 1

I mixed up who's asking what. Lefty, did you mean millivolt per Newton or Pound per excitation voltage?


Passive devices like strain gauges that are internally wired like the classic resistance Wheatstone bridge typically have a full scale output voltage of some small millivolt output voltage, however the specific output voltage is dependent on what voltage you supply to excite the bridge, 10vdc, 5vdc, 2.5vdc, etc.

As bridge excitation voltage value is a user choice, how should a strain gauge manufacture state the full scale millivolt output for their device not knowing what voltage you will excite the bridge at? Rather then to list a number of standard output voltages Vs Bridge excitation voltage, they just rate it like this: 10mv/V. That's not division, it says it will develop 10mv per volt of bridge excitation voltage.

That means if your drive the bridge with one volt and load the cell to it's maximum rated strain it will develop 10mv output, and you can scale it to any voltage you want to actually use, so at 5vdc bridge drive will have a full scale output voltage of 50mv.

This is not the same spec as what amount force is required (in grams, ounces, tons, etc) to reach 100% full scale output, that is a separate specification independent of the bridge electrical sensitivity specification of xx mV/V.

Make sense?
Lefty

Dave_L

Hi Lefty,

Thanks for all the advice. I have a bit of reading to do but I think Ill manage to put this thing together. If not, Ill be back to pick your brain a bit more:)

Cheers

Dave

liuzengqiang


Make sense?
Lefty


Yes, quite a bit. So if you have a 10mV/V gauge, that means whatever the max measurable strain it does, it will output 10mV if excited with 1V? Then a separate number is max strain. Since the strain scales linearly with resistance change, you get a linear voltage output as strain gets higher.

So if you want strain, you glue this gauge on the object. But if you want to make a force sensor, possibly a sensitive one, you glue this gauge or other to a hollower object so it develops more strain under same stress. I hope I'm making sense. Lefty, could you give me a "That's right" or "You gotta be kidding?!" I guess I am co-asking, not really answering this question  :smiley-red: Thanks.
Serial LCD keypad panel,phi_prompt user interface library,SDI-12 USB Adapter

retrolefty

Quote
"That's right" or "You gotta be kidding?!" I guess I am co-asking, not really answering this question   Thanks.


That's basically correct. The manufacturer will generally package a strain gauge into a mechanical package that is designed to measure a specific kind of force, compression, sheer, expansion, etc and call them 'load cells'. With raw strain gauge like the link below, it's up to the user how to mechanically mount the printed resistor assembly to best measure the desired force.

Lefty

Go Up