measuring resistance with RC circuit

[engsolution2]

First off, to start with I am a mechanical by trade so my foray into electrical circuit design has been a little difficult for me.

What I am attempting to do is build a circuit used in a scale application for a consumer product measuring in the weight range of 0-200 lbs. (i.e. lower cost components)

I am attempting to utilize a thin-film load cell, but I am not for sure what is the best way to send the information back to a microcontroller.

It seems it would be simple enough to measure the varying resistance by measuring the time it takes a capacitor to discharge and correlating this to a weight. Is there a downfall to doing this? (such as not being accurate?) I am only looking to be accurate to the nearest lb.

Or would it be better to build a circuit that outputs a 0-5V output and utilize an A/D converter?

Or maybe there is another method that I am not aware of.

I appreciate any help you guys can offer as this kind of work strays a little bit from what I normally get into. Thanks!

 

[waross]

I often start at Omega.
How about one of these off the shelf devices?

http://www.omega.com/toc_asp/subsectionsc.asp?book=Pressure&subsection=C04&flag=1

Free strain gage literature.

http://www.omega.com/literature/strainNH25/

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Skogsgurra]

Definitely as Bill says. If you are building one or e few scales.

If you are building thousands of them, there are quite a few other techniques, like magnetoelasticity, resonance frequency of a spring or string, piezoresistivity and others. Whatever technique you choose, you need help from an EE to do this.

Back to the thin film strain gauge. It is possible to use an RC circuit to measure resistance changes. It is being used in lots of applications like joysticks and other low precision devices. The problem when it comes to a strain gauge is that instead of a 0-100 oercent resistance variation and low need for accuracy, you have at most a few percent resistance variation and also a rather high accuracy (1 lbs in a 200 lbs range is .5 %).

RC methods can be made with good accuracy, but needs a stable precision capacitor, highly accurate comparators and a low noise reference and overall design. I would not go that way.

A differential amplifier or instrument amplifier plus an (at least) ten bit A/D is the standard way of doing it.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[sreid]

Load cells have a small resistance change from zero to full load and amplification will be required. Stable amplification is what you need and I'd suggest a chopper stabalized op amp from, perhaps, Linear Technology. Microchip Technologies makes PIC processors which have A/D converters. The processor can also be used to crunch the numbers and drive a display, etc.

 

[engsolution2]

Thanks for the help! It is much appreciated.

I suppose it may depend on the particular strain gauge used, but does a circuit as described above (comparator - A/D) typically deliver a linear output throughout a certain loading range? Sreid, is this what you mean by chopper stabilized op amp???

Also, would I really require at least a 10-bit A/D converter if there I only have 200 possible levels of output? (meaning using an 8-bit in this application would be fine?) Also, I think I do not quite understand what exactly is outputted from an A/D converter. Is it some sort of on/off pulsing at varying frequency or something else?

Gunnar, I do agree I think I will have to get an EE involved at some point. I am at the point where I would like to build a few prototypes where it might make more sense to use off the shelf components. But, before I do talk to an EE I always like to be able to talk the language at least to some extent. I appreciate all of the replies so far! It has been very helpful.

 

[MacGyverS2000]

If you're new to the EE world, a high-quality op-amp circuit is probably not the easiest project to tackle. Not to mention learning how to program for the processor, the interface circuits required, etc.

If you want a prototype, pick a cheapy scale up at your local Borg store. If this isn't much more than a weight scale, you're going to have a VERY difficult time convincing the market (or investors) your idea is better... scales can be had for a few dollars these days, to the point they even make specialized bare dice that do all of the work mentioned above.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[sreid]

Chopper stabalized amplifier.

http://www.linear.com/pc/viewCategory.jsp?navId=H0,C1,C1154,C1009,C1027

 

[LionelHutz]

I didn't put much time into searching but look at this MAX1452. I know there are similar things from other manufactures too but this should give you a good push.

http://www.maxim-ic.com/quick_view2.cfm/qv_pk/2369

 

[Caledin]

...probably LTC1799 and similar chips have been overlooked
here. They are so small, inexpensive, nice things.
They convert resistance into frequency using their own
precise capacitor inside and the wonder is that the
resistor being measured is at constant voltage. The
temperature accuracy is better than 40 ppm/Cels'grade.
( Linear Technology offers valuable tips about such
solution. )

 

[Skogsgurra]

The LTC1799 is a good chip. But it will not work in a strain gauge application. And that is for several reasons. All cells are bridge circuits where the diagonal voltage is evaluated instead of the resistance change. This is done mostly because resistance changes a lot with temperature while the voltage dividing action in the two bridge halves is unaffected by temperature variations (within working range).

It is also done because the bridge resistance is rather low and the variation is quite small.

Finally, wiring may influence calibration more than you like when using an RC evaluation, which the LTC1799 does. You need to put the chip close to the transducer. Probably OK in your case, but often not practical.

A stable DC amplifier or, in extreme cases, a carrier frequency amplifier is what works.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Caledin]

LTC1799 actually evaluates current and voltage at its input and simulates RC evaluation. I used it for converting voltage into frequency ( temperature stability wasn't crucial then ).

 

[VEBill]

Gunnar beat me to it.

I've taken apart a couple of inexpensive digital bathroom scales and they both used a good old-fashioned resistive bridge as the first step.