Strain gauge signal conditioner

[Jaguarjoe]

Hi guys, it's been a little while since I've been here.

I was at an estate sale last week and bought 3 boxes full of electronica very cheap ($10.00)
Included in them was a very nice load cell, the kind that plugs into a ratchet wrench to measure the twisting torque on nuts, bolts, and other rotating objects. This a very high quality part, worth 100x the $3.33 I paid for the box of goodies.

Much of the other stuff is SMD which is useless to me so I will be giving that away here.

My problem is scaling the output of the load cell. It's output is 1.9965mv/v at 1500 in-lbs. I would like to be able to resolve 30 in-lbs. With 10vdc excitation, I'll get 150 in-lbs. If I build an amplifier with a gain of 5, I'll get 30.

Is that valid thinking? The last time I played with Strain gauges and conditioners was 1968. I forgot 95% of what I knew then and I think it was all discrete. Is there an IC solution?

Thanks,
Joe

 

[johansen]

If you would like a $6 solution, buy a 1 kilogram electronic scale off ebay, replace the load cell with yours.
You may get lucky and find that you can calibrate the scale with an arbitrary weight to compensate for the difference, and the digital readout will be 1 gram per in-lb.

however: My $6 1 kg scale runs everything off the same 3v battery. At 1 kilogram the load cell output is 2.4mv, and that's only 0.8mv/v on the load sensor. (perhaps its actually a 2 kg load cell, and my mv meter is wrong? idk)
So if you buy a cheap 1 kilogram scale you may lose the full range of your load cell, as 2mv/v makes 6mv at 1500in-lb, and the strain gauge amp may have too much gain for the cheap 16 bit adc that they probably used.

*checks ebay. Looks like inflation is a *****. the one i bought last year is now $9.49

here's a real expensive way to do it.
https://www.electro-tech-online.com/custompdfs/2012/03/Strain20Design.pdf

 

[Reloadron]

Cool, those are useful little units for sensing torque. With a stable 10 volts excitation and 1.9965mv/v the full scale output will be 19.965 mV. Actually, depending on the unit it should measure CW and CCW torque. So the voltage span should be -19.965 mV to +19.965 mV. Does the unit you have include a 4 pin bayonet type connector or just a cable out?

Depending on what you want to spend I like the AD-524 precision instrumentation amplifier, while not cheap it would be a good solution. That or the link johansen provided. Did you want a portable or mains powered solution? I may have a few AD524 chips lying around and could send you one. I know I had some somewhere.

Ron

 

[Jaguarjoe]

The load cell has a 6 pin "Lemo" plug on it. The inner space where the pins are is about 7mm dia. Being so small and recessed I'd never be able to solder leads to those pins or even get ohmmeter probes in there.
I can buy a 200g scale to experiment with for about $13 at a local Harbor Fright. If I'm careful, I can reassemble it and let my neighbor use it in his meth lab (JK!)
A 524 is about $24 at Digikey. For what you get it's pretty hard to beat (like a 120db CMRR). I like things like this because I should be able to figure out how to make it do what I expect.

 

[Reloadron]

Actually the AD524 I have lying around is the AD524CD the ceramic dip version. Just a better version that runs about $57 through Digikey. If you want one let me know and I will send it to you. Once you get the signal conditioned you can scale to read LbIn or LbFt whatever you want. I have a 100 LbFt unit lying around (1200 LbIn). Mine is a 4 pin. The six pin versions aren't unusual as they in most cases allow for the excitation voltage to use sense leads.

Ron

 

[Ubergeek63]

your calculation was correct, but you did not say how you wanted to read it out. there are excellent analog front ends for $10 from digikey... and SMT is not that bad to work with.... mind you, my positions are a bit slanted from the POV of a hobbyist.

Dan

 

[Reloadron]

Got your PM Joe. It should leave tomorrow from Cleveland, so give it a few days. Best of luck with it.

Ron

 

[Jaguarjoe]

I'm having a hard time opening PM's tonight. I'll try later or tomorrow.

Thank you very much!

Joe

 

[Jaguarjoe]

your calculation was correct, but you did not say how you wanted to read it out. there are excellent analog front ends for $10 from digikey... and SMT is not that bad to work with.... mind you, my positions are a bit slanted from the POV of a hobbyist.

Dan

Waaay back in the day, I would spend weeks peering through a microscope building tri-axial accelerometer signal conditioners on boards the size of a quarter, using SMT before most of the world new what it was. It was discouraging putting the whole thing together, potting it, sending it across the hall for shock testing and seeing it die in about a second. I'd dig out the potting, see what broke, redesign the fault it had. Repeat many times.

I was told then that I should not have went to Engineering school but should have become a surgeon.

Flash forward to today- because of my tremors I can easily solder 5 or 6 IC pins at once if I don't have a random spasm and launch the chip, or the pcb across the bench. It ain't pretty

 

[johansen]

these days i still use through hole, but i don't drill holes, just bend the leads out and solder them to the traces.

however, regarding that 57$ opamp...For that much money you can buy a buried zenar, a digital panel meter, a chopper stabilized opamp, a thermistor and a second opamp to keep it at a constant temp, and 0.5% resistors to keep it within spec.