load cell and Analog Modules

dear all,
im currently doing a project where i need to integrate a load cell (1mV/V) to my analog module that has an input of 0-10V, 0-20mA. Can i just create my own amplifier? because i dont have a budget to buy a strain guage amplifier...is a fet transistor similar to an amplifier?
i will need to get the exact weight to the gram (1gram) because i will need to add them up later on, on the program.
please suggest.
thank you very much.

You surely are in need and yes, an FET can indeed be used in an amplifier circuit. Alone, it is just an FET. You really are going to have to get some kind of transducer to make your strain gauge signal compatible with your analog module. If you want to build your own amp, then good luck, but best to spend a buck or two to get a commercial grade one. Your requirements on precision (1gram) will make a homemade one tough to do.

would it be cheaper/ easier to connect the load cell to a weighing indicator and connect it to the plc via rs232? im using an omron plc CPM1A and i think it only allows 1 communications adapter, but i need two load cells/ weighers to attach to the plc...
please help

thank you very much for the reply randylud

student.in.need,

What is the make and model of the load cell?? What about looking into a used amplifier???

If this is a student project in which the lesson learned is suppose to be "signal conditioning and inferfacing", there are lots of stuff on the net to help guide you in building your amplifer. As per your needing to measure 1 gram, that will all depend on the resolution of the load cell.

Hope some of this helps.

Andrew Evenson

The signal conditioner for a strain guage does two things. First, it provides the excitation voltage to the strain guage, which in effect is a variable resistance device. Thats where the denominator in the mV/V comes from. If you supply the strain guage with 10 Volts excitation, your output will be 10 mV at full scale.

The second function is to convert the mV output to a high level signal like 4-20 mA. This not only allows you to connect to a standard PLC input, it also allows transmitting this singal more than a few feet before the losses and EMI/RFI noise make the signal inaccurate.

It is possible to get an analog input card for many PLCs that will accept the mV signal directly, but this doesn't deal with the effect of noise and signal integrity.

If you want, you can certainly roll your own amplifier, but it is quite possible that this task will become more time consuming than accomplishing the original objective of your system. I strongly recommend getting a signal conditioner. They are available for a few hundred dollars, and if you talk to your strain guage supplier about your project and explain that it is for educational use, they may give you a break.

We get these for $172...
Paul C.

The CPM1A does not support serial communications. The port is only for a programming device. The CPM2A/CPM2C does have instructions that support serial comm.

I also noticed that you are looking for 1 gram accuracy. Please note that the CPM1A-MAD01 is only an 8 bit unit. The CPM1A-MAD11 has a higher resolution (1/6000).

Hope this helps.

8 bit is pretty low resolution.

To determine what the resolution of your measurement is divide the full scale sensor output by 2^X where X is the A/D input resolution.

For example, if the load cell has a 1Kg range, then the resolution is 1000g/256 or 3.9 grams. You cannot divide the measurment any smaller than that.

If the load cell has a 100 gram range, then the resolution is .39 grams.

With a 1kg laod cell you could amplify a 20V range as long as you know you will never put more than .5kg on it and get 1.95 gram resolution. The smallest measurable voltage step is 39mV or , so bear that in mind when designing an amplifier.

We use a bipolar signal isolation card from Carotron Inc. it's easy to set up. look them up on the web. and if you use one I can help with set-up directions and a drawing as we use them.
Lynn

as a metrologist, I don't like the odds of attaining a measurement uncertainty level (at least 4:1) based on your application. Tim Kyptka, from Borg Warner, submitted a paper on this type of measurement at the 2001 NCSLI conference. Since the measurement is ratiometric you really need to monitor not only the load cell's output; but, also the power supply voltage. You need a signal conditioner. The FET theory is nearly obscene! Operation of a FET in the linear region at the levels you want is crazy. A bi-polar junction transistor would be better; but, you'll definitely require an op-amp at the least! The op-amp circuit would need to be characterized and calibrated prior to making any reasonable measurements. You will need calibration standards that correlate to at least 10% below the nominal you expect to see and up to the maximum of the load cell -- which should not exceed 90% of the load cell's full scale output. Linear transducers (even when ratiometric) are not known to be as accurate outside of the 10 to 90% range.


Looking at the AB PLCs -- even with 16 bit analog input from 0-10V. That equates to 1 bit for every 3.05mV. You cannot possibly get enough resolution to satisfy any type of uncertainty analysis with the PLC input.

Hardy

Check out Hardy Instruments.

They have a PLC plug in card. I have not used them but I would like to hear form somebody that has.

It is nice to say:
I need to measure within a gram.
I need a standard deviation of a gram.
I need a 95% confidence level of a gram.
and all this means something different.

A non-linear measurement can work nicely if you can calibrate it right but all that depends on the stats. What are you measuring and what is the standard deviation of the parts, and at what confidence level do you need to make the good/bad call.

I am not sure of what your doing so I'll stop.
If you do try Hardy Inst, let me know.

Thanks

Hey,
why not try an good old 741 op-amp, in your frequency range (0 hz) the 741 can acheive an open loop gain of 10^5, if you use 10v as excitation an a gain of 1000 you will be able to directly feed an 0-10v analogue input card, wouldn't hurt to throw in an opto-isolator for a little galvanic isolation between the card and the amp circut.

Lance.

Try

This reference: http://www.uoguelph.ca/~antoon/gadgets/741/741.html

Originally posted by unclebuck:

Check out Hardy Instruments.

Click to expand...

I haven't used these modules but I checked into them. You are looking at about a $2000 card. If student.in.need is thinking about building his own amplifier I think this is probably out of his price range. Unless they would loan him one for a while.

The Hardy card had some pretty cool noise rejection functions. I would have to guess they have a high pole count programmable phase filter built into the module. The demo video they have on their website is kind of wild.

Keith

Couldn't hurt to ask, and the tech guys might give him some points.
Talking to them would be a learning experience in itself.