Getting odd readings from hacked digital scale.

Forum 2005-2010 (read only) Hardware Interfacing
1

As a follow up to my earlier post regarding a 3 wire strain gauge, I'd like to say that I got it "working"! The quotes are due to some odd readings I am getting from the other end of the amp circuit i am using (schematic below)

When I read directly from the sense lines i get a reading (using a cheap $10 meter) of about .8mV under no load, it goes upwards of 2.3-2.5mV with load (me pressing my upper body weight onto the scale).

After running it through the following instrumentation amp I created using 3 of the 4 amps in a LM324 I get a constant .6v under no load, but then it peaks out at 2.5v no matter what weight I put on it (I'm 175, my dad is 230, both read the same 2.5v).

I also noticed (using a graph in processing) that the reading dips down a little. About 4-6 units, directly from analogRead.

I also have to calibrate it since under no load it reads .6v, I understand that calibration should always be used, and I plan on it, however I am curious as to why it's reading .6v and not 0v.... (and on second thought, the wheatstone bridge reads .8mV under no load so it would be bumping that up to .6v, am I right?).

I plan on making a pcb for this so I can button it up inside the scales housing (already made one, but it didn't work reliably, so I scrapped it), should I taking anything into consideration when designing it? i.e. decoupling (2"x4" ss board), trace paths/sizes, etc...

If it matters (which it might). During construction, I broke one of the leads from a strain gauge. I went back and picked up the identical model scale for a replacement. After ripping it out and measuring, I found that it was 2-4 ohm off when compared to the others... It was also wired differently on the daughter board (the configuration of white/black was reversed).

Any help would be appreciated! :slight_smile:

Amp Schematic:

691×367 4.18 KB

Full Project Schematic (SV1 goes to a 10 led bar graph, IC4 is a LM2940):

2

Yeah, that's the primary purpose of this post... I am very confused why the limit is 2.5v... I am feeding everything from the 5v out on the arduino. I found a schematic in the datasheet that uses 100k's and a 2k pot for gain adjust... but by my calculations, I would need a gain of 1000 (.6mV -> .6v), so I figured that using the 22 and 10k would get me fairly close, but on further reading it seems like my calculations are possibly incorrect.

I am about to try again with the recommendations of an online calculator I found... 1K/1M...

As for the output, it seems to be fairly linear, I wrote a quick moving line graph in processing and it seemed to be linear as i applied body weight. It just maxes out at 2.5v. The scale originally ran with a single 3v cr2032 battery...

3

Well I used the following schematic, using a 1k pot for R2, and 1M for R1 and R5, the rest are as shown.

With no load I get a similar response to the previous experiments... with the pot turned down (reading about 25ohm) the response is very coarse (very little weight maxes the amp out at 2.52v) when turned up (reading 1025ohm) I get the response I'm looking at, however it is still capped at 2.52v... however with no load I get 39mV as opposed to .6V.

Is this a limitation of the chip? and if so, where in the data sheet is it noted..

4

It's a 324, but essentially the same thing, it's being powered directly by the arduino 5v out. According to the data sheet, it is designed to operate at lower voltages such as 5v.

5

You may be failing to appreciate that strain guages are dual action. That is, as well as pushing down on them you can also pull up on them. In effect their full range is from minus-weight to plus-weight. So in effect, in a body mass device you are spanning over only half the weight range.

Also because the op-amps can only span from 0 volts to +5 volts they need to be set to mid range at zero load to match the strain guage capability.

You have proven that you can vary gain by twiddling the adjustable pot. What you now need is a zero off-set adjustment so that you effectively blank out the -ve range capability of the strain guage and the mid-voltage offset of the op-amp.

There again I may be talking a load of bull.

jack

6

So how would I modify the schematic above to include offset adjustment in addition to gain adjustment?

Gain would be by replacing R2 with a pot (already done) and offset would be by replacing R7 with a pot? What values would be best?

7

I hate bumping threads without adding new information, however I got nothing :frowning: