• State
  • Locked Locked
  • Replies 11 replies
  • Subscribers 32 subscribers
  • Views 1424 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Connecting strain gauge to LM3S8962

David Rosales23
Prodigy 30 points
Other Parts Discussed in Thread: EK-LM3S8962

Community,


I have a project where I want to connect a strain gauge to my Stellaris EK-LM3S8962 board. I intend to attach the gauge to a metal plate which will have forces acting on it. I want to be able to determine forces and deformation of the plate with the strain gauge and output this to the on-board led. I have experience using the on-board peripherals but nothing external. I would need to pass a voltage through one side of the strain gauge and read the voltage output on the other end. I see on the board there are pins with voltages next to them. Are these the only pins that would be able to send the voltage through the gauge? If this is the case, are there specific pins that would be able to read a voltage input as well? I will be using CCS as a development environment. Any help would be appreciated.


Thanks,

David

  • 0
    Guru 47900 points

    David Rosales23 said:
    I would need to pass a voltage through one side of the strain gauge and read the voltage output on the other end.

    You've chosen a task which may be a bit beyond your "present" capabilities.   KISS dictates that one build capabilities slowly, methodically - not in one "giant leap."

    That said - many/most such "strain gauges" implement a resistive bridge - which is "balanced" at idle - and becomes unbalanced under strain.  (or load)   When "balanced" the bridge output drives towards 0V - under strain that output deviates from 0V.   (and is likely too low in level for proper introduction to your MCU)

    Normally the following method would achieve your task:

    • Apply the proper voltage to your strain gauge's voltage input pins.
    • Identify and then properly connect to the strain gauge's (I'd expect 2, differential outputs)
    • That strain gauge output is surely too weak/low for your MCU - must be amplified via a proper Op-Amp stage
    • The output of that Op-Amp stage is then presented to the MCU's ADC input(s)

    Again - with facts in evidence - this response will lead to many more questions.   Not to be mean - but at this stage - w/your level of experience & understanding - I'm unsure if this project is fully/properly w/in your "wheel-house."

  • 0
    Guru 27665 points
    Giving your apparent level of experience I suggest you start with a pre-built conditioning module (often called a transmitter) such as this

    www.omega.com/.../DMD4059.html

    TI does have an appnote on measuring a strain gauge but it uses a delta sigma A/D.

    www.ti.com/.../slyp163.pdf

    Robert
  • 0 in reply to Robert Adsett
    Guru 47900 points
    Let the record show that (two) now note, "Mismatch between desire & capability."

    That's not bad nor evil - simple statement of fact. Best results occur when those two items "match."
  • 0 in reply to cb1
    Guru 27665 points

    Indeed, there are three  non-trivial tasks here

    • Reliably connecting the strain gauge to what it is measuring
    • Stain gauge signal conditioning , especially if it is a noisy environment
    • Converting, filtering and analysing the resulting signal

    Robert

  • 0 in reply to Robert Adsett
    Guru 47900 points

    I'd consider a 4th task - that of properly delivering the conditioned signal to the suitable ADC input pins.

  • 0 in reply to cb1
    Prodigy 30 points
    Thanks for the responses. When I proposed this project to my professor, he did in fact tell me that this was a very ambitious project and wished m luck. I will likely come up with another project other than this for the term. On that note, I would still like to know a little more on the topic.

    Here is some more information on the project for the next part of this post. The strain applied to the plate will be relatively small, under 10 pounds I figure. The duration of applied force would be a few seconds long.

    I gather that I would need to connect the input to the gauge to one of the voltage pins and the output to an ADC pin.

    For noise, would gathering a sample set and removing outliers, taking mean/mode, be suitable for obtaining a feasible reading? My original approach was going to be to plot the results of tests with different defined forces and determine an equation based upon this for the unknowns that I want to measure.

    CB1, can you explain a little more on what you mean by "properly delivering" the conditioned signal?
  • 0 in reply to David Rosales23
    Guru 27665 points
    Take a look at the TI application note I gave a link to, that will give you some idea of the requirements for interfacing to a strain gauge directly.

    Use a proper filter for removing noise rather than an ad-hoc approach. It will probably be faster and easier to confident the result is meaningful. The advantage of a transmitter is a good deal of the necessary filtering has already been done for you in the transmitter.

    Robert
  • 0 in reply to Robert Adsett
    Prodigy 30 points
    Thanks Robert, I'll have a go at that note.
  • 0 in reply to David Rosales23
    Guru 117855 points

    David Rosales23 said:
    properly delivering" the conditioned signal

    I concur w/Robert's "use a pre-built transmitter" suggestion.

    To "properly deliver" your strain gauge signal you should:

    • minimize connection distances... usually - but not always - this suggests NOT sending a low level analog signal anything beyond a few inches.   Amplify at/near the source!
    • avoid nearby noise sources
    • match the output impedance of your signal to the ADC's specified/desired input
    • place small cap. very close to ADC's input pin
    • amplify the signal to a level which "best overlaps" the input signal voltage range of your ADC

    Might this satisfy?   You've got a good general idea - yet the mechanics of the original project steer you from your MCU-centric course - and (Robert & I) both feel should be postponed...

  • 0 in reply to David Rosales23
    Guru 27665 points

    David Rosales23 said:
    For noise, would gathering a sample set and removing outliers, taking mean/mode, be suitable for obtaining a feasible reading?

    Apropos of this, a link to the following arrived in my e-mail today

    Create a filter circuit faster than ever before
    Filter Designer has always been a good, solid member of the WEBENCH® family of design software. Now it is not just good; its improvements make it easier than ever to use. You first start on the WEBENCH…
    By in Technical articles > Analog

    Robert

  • 0 in reply to Robert Adsett
    Guru 47900 points
    Blasphemy for sure - yet (almost) related - MAX7404CSA+ realizes "quick/easy" such filter. (we've used - recommend...)