Strain gage question

Self contradicting sentence as you don't include the permissible errors of readings.

Every part of the system has effect on readings so all of them have to be accurate and stable during the "life time" cycle. Or include a calibrating known deflection of reading (on strain gauge bridge a stable resistor in parralel with one leg of the bridge) to give you the starting point.

Been there, done it!

HTH

Stanislaw

I should have mentioned that, my current method is probably no more than 5% accurate. So anything less than that would be a plus.

Well I have a good bit, but the fine points are beyond me. I understand strain gages change very little in resistance throughout there usable range, and usually require bridges and very high quality power supplies for ultimate use, but then they yeild results as accurate as .03% in many cases on a 5000 lb load cell. That's why I'm posing the question here, and framing it as I have. One percent accuracy should be more than adaquate for my needs. So will most ohm meters costing less than $100 have the ability to measure say 600 steps in the useful range of a strain gage?

Permissiable error of one percent would be very nice. As for calabration, we can just load the scale with a know quantity most any time. That's what we do now. But now we are using a 1.125 dia, oil filled cylinder with a 1000 lb. pressure gage. But I don't think we are getting closer than 5%, and it seems especially suseptable to temperature change.

A typical full-bridge load cell will give 5 millivolts of output per volt of excitation at full load. So 9 volts of excitation gives 45 millivolts of signal. Some DVMs have enough sensitivity to be useful with such signals.

The output will droop as the battery dies, so a regulated power supply would be a better idea.

Buy a load cell that's rated at or just a bit over your expected load.

You can buy a complete load cell excitation/display box fairly cheap; they're a commodity.

You will generally have to adjust for zero offset ("tare") and full-scale (using a known test weight) calibration, as load cells aren't always very accurate. They usually are extremely linear and stable.

John

Now you _are_ talking sense!

5% is NOT in my vocabulary as measuring range. In my lifetime in testing anything greater than 1% was a failure static and/or dynamic. So now the temperature creeped in, next will come something else, so do yourself a favor, do it once but do it as perfect as possible to eliminate insecurities, it will pay in the long run. Are you measuring now by reading pressure on circular gauge or having pressure transducer and getting the readings from that? And just a stupid querry, how time consuming is this activity, if it is daily routine for hours it costs in people's time plus errors in results. As you carry the calibrating quantity initially why not to use it on a lever as counterweight and watch for microswitch opening as indication of equality?

Just talking

Stanislaw

Isn't a load cell and excitation/display box about $700? If not, how cheap and where.

--
The viability of the whole thing depends on the granularity of the
measurement and what kind of repeatability/stability you\'re looking
for.

I suggest you Google "Strain gage tutorial" to find out what you\'re
up against.

Measuring now by mechanical gage, but suspect 5% errro is primarily caused by friction in the cylinder.

Measuring operation is done at least 30 times an 8 hour day, and increasing.

Space constraints eliminate the possibility of a counter balance.

Omega has indicator boxes for around $250, and I'd expect there is cheaper stuff around.

If you want really cheap, try ebay.

Have you heard of google?

John

Don't you _?know?_. So what improvement you want to apply?

As mentioned before, do it once and do it properly. If the hydraulic system has some faults replace it with load cell, most reasonable quality load cells are self temperature compensated (one of your complains), many include precise calibrating resistor inside to give indication of part scale reading of the full range. So a three point mechanical suport with one point beying a load cell gives you a weghing possibility. One point attachment can be used by fully hanging the load on a load cell. Not electronics, mechanics;^)

Have fun

Stanislaw.

When given the possible 5% error you can either gain or cheat one and a half units per day ( maximum taken as sample), so make some calculations how much it will cost you over the continuation of this activity, against some investment in reducing the errors.

Have fun

Stanislaw.

Thanks for you help.

If you'll be satisfied with 1% accuracy use a mechanical balance.

Bye. Jasen