How to convert a load cell 1mV/V to labjack U12 "readability" (10V)?

Discussion in 'The Projects Forum' started by cndg, Mar 31, 2013.

  1. cndg

    Thread Starter New Member

    Mar 31, 2013
    3
    0
    Hi all - I bought a load cell for part of a propeller design rig I'm building, and I'm stuck trying to work out how to hook it up. The doc reads like this (eBay item 350651562527):-

    Rated Output 1.0946mV/V
    Input Impedance 410±10Ω
    Output Impedance 350±3Ω
    Excitation Voltage 5-10V DC

    However, my LabJack U12 is rated like this:-
    8 Single-Ended, 4 Differential 12-Bit Analog Inputs
    ±10 Volt Analog Input Range

    So I'm needing some way to convert the mV load cell output into the ±10 V range my instrument expects to read.

    Can someone help, or point me at "nubie compatible" instructions for this kind of thing? I'm good with the software stuff (I even wrote the open-source perl drivers for the U12), but I'm out of my expertise zone on this voltage-conversion problem.

    Alternatively - if anyone knows a different instrument that will let me read a bunch of these things at once into a PC via USB or RS-232 - that too would suffice.
     
  2. JohnInTX

    Moderator

    Jun 26, 2012
    2,614
    1,185
    You will have to amplify the load cell signal with an instrumentation (true differential with high common mode rejection ratio) amplifier to get the signal up to something you can read. The analog section will need a relatively high degree of precision. Your software will have to scale the mV/V/weight to some useful value (lbs, Kg etc). Your software will need to have some smoothing to compensate for physical vibrations coupling to the cell and take into account drift over time (auto-tare).

    Here's a list of appnotes/info from Analog Devices that explain the process with reference circuits.

    Omega also has ready built modules.
    National Instruments has lots of stuff also. Here's a tutorial.

    Many off the shelf scales have RS232 outputs. If you just need an instrument to weigh things, consider going that way.
     
  3. cndg

    Thread Starter New Member

    Mar 31, 2013
    3
    0
    Hi JohnInTX - thanks!!

    Looking at those diagrams, I'm guessing my gauge has an inbuilt Wheatstone bridge... do you think it would be possible to isolate only the wires connected to the gauge itself (e.g. to bypass all those resistors), and to incorporate that into a different Wheatstone bridge to bring the whole thing up to the voltage level I might be able to read directly? I don't need temperature compensation or calibration etc - I can do all that stuff in software (I'm already reading temperature).

    I also found this USB thing for 35 Euros:
    http://digital-measure.com/html/voltmeter.htm
    Range: -0,2 V to + 0,2 V
    Resolution: 0,000 015 V
    Error: max 0,015 mV

    Would I be correct to guess that "1.0946 mV/V" refers to the maximum output at the full load-cell weight (300g), so with a 5V (USB powered) input, I'll be wanting to read from 0c to 0.005473V (5v*1.0946mV/V) ?
    If so - that resolution of 0.000015V would give me a scale of 0 to 364 over the full 300grams - or slightly better than 1gram accuracy - which is fine - however - what do you think "Error: max 0.015mV" means? That's 275% higher than the entire range of my load cell already - do you think that's a "constant" error (which I can compensate for in software), or some kind of show-stopping problem that will render that gadget useless for my needs?

    Final question - the load cell says "Excitation Voltage 5-10V DC" ... what would happen if I pushed around 180volts (super low current DC of course) into that? Will that give me an output of just under 0.2 volts at 300 grams, or are load cell "excitation" voltages things never-to-be-exceeded?
     
  4. JohnInTX

    Moderator

    Jun 26, 2012
    2,614
    1,185
    I don't think so. 'All those wires' go to two strain gauges on the frame working in opposite directions to get the specified output. I'm not one to trade one problem with a known solution (conventional IA-ADC) for another without one.
    The mv/V rating is (if I recall) the full scale output (mV) per volt of excitation volts. So if you have 5V excitation, you get 5*the 1.0946mv at 300g.

    The Error is the voltage offset at zero load (its not 0mV - your calibration routine has to take it out). The slope as you add weight conforms to the mv/V line.

    Sorry. 10V would be the limit. The strain gauges are film-deposited onto substrate and would likely be zapped. Remember, too, that the output voltage is a differential voltage + common mode voltage. For example, at 5V excitation, 0 grams and 0 error both output wires will be 2.500V (a differential voltage of 0.000V). A 2mv 'output' means 2.499V on one leg and 2.501V on the other. Exciting the thing at 180V makes the output 90V on each wire +/- the 2mv differential. Imagine how happy your USB voltmeter will be to see that. You need a front end.

    Maybe. Its one way to get a voltage to the PC. Only 6 conversions / sec, though. Is that enough? Cute, though.

    Some definitions to wade through.
    EDIT: And a good one from Omega
     
    Last edited: Apr 1, 2013
  5. cndg

    Thread Starter New Member

    Mar 31, 2013
    3
    0
    Hi JohnInTX - you are awesome. That was way loads more than I could ever have hoped for (and more than I could figure out from google!), and everything I need to get going. You should stick your bitcoin wallet ID into your sig so folks can tip you :)

    In case you're interested, photies of my 3d setup and links to some of the particle-swarm source code I've done are in this post: http://www.homebuiltairplanes.com/f...rting-xfoil-how-account-scale.html#post159892

    This gauge setup will be mounted to a "test bench" for measuring "actual" propeller performance (in real time) - instead of a wind tunnel though, my plan is to tie the whole rig onto the roof of my car, run the cables in through my sunroof, and then measure thrust/torque/rpm/current/windspeeds(various)/temp/alt/humidity at a selection of driving speeds (upto 110km/h - the max I'm allowed)... The PC will automate almost everything (using a servo-controller, variety of measuring attachments, and custom software - not my car speed though of course).
     
  6. JohnInTX

    Moderator

    Jun 26, 2012
    2,614
    1,185
    Good luck with your research and let us know how its going..
     
Loading...