Go Down

Topic: 3-Wire Strain Gauge (Read 9705 times) previous topic - next topic

nickvd

Hello!

I have decided to make by next project, and need some assistance...  I picked up a cheap digital scale from XScargo ($9.99).  After tearing into it, I discovered 4 strain gauges, one per corner.  During the course of my searches on how to interface them to the UC I could not come up with any explanations on how to wire up a strain gauge that has 3 wires, only 4 wires.  My electronics knowledge is entirely self taught, so there is a VERY large chance that I've just missed something simple.

Any ideas?  (i need roughly 70lbs and accuracy isn't all that vital; its to measure the level of propane in a heater)

Grumpy_Mike

Normally a 3 wire sensor is just wired up like a potentiometer, with a center tap and two ends. You wire the ends to a supply + & - and the sensor output is the center tap. You usually put this into one input of a differential amplifier with a simple potential divider into the other input.
Think of it as half a Wheatstone bridge.
http://en.wikipedia.org/wiki/Wheatstone_bridge

nickvd

#2
Nov 09, 2010, 12:41 am Last Edit: Nov 09, 2010, 12:46 am by nickvd Reason: 1
So then it is possible that my multimeter and (cheap diy scope) are just not sensitive enough to detect the changes... I knew about the amp requirement, but just assumed that I would at least be able to check via multimeter prior to building an amp... (i tested it with both 3 and 5v since the scale is run from a single CR2032).

I will likely have a few chips laying around that could do it... Any ideas which? :D (heh, I have no idea what I have).

That being said, there is a ready made pcb inside that ties all 4 of the gauges to a single 4 pin daughter board (using a single 1x4 header which is soldered onto the main board), so I'm anticipating a very easy hook up :)

nickvd

After further inspection, it seems as though the daughter board only connects the red wires (red/black/white) to the main board, the black and white wires are tied together in pairs of like colours (as seen below).  As I have said before, I am still pretty much a noob when it comes to electronics, but I cannot see how they would make that work in the scale, where is the power coming from? and if it's the red wire, how are they reading the gauges?

I also tested the multimeter again and did notice a predictable change in resistance when test the black/red wires, the resistance goes from 1001 to 1002 when pressed and 1003 when pressed as hard as my fingers would let me...




nickvd

#4
Nov 09, 2010, 03:49 am Last Edit: Nov 09, 2010, 03:52 am by nickvd Reason: 1
All the lines trace back to the big black blob :(

the bottom left is the connection point for the daughter board pictured above.

The resistances are as follows

white/red: 1002 ohm
black/red: 1001 ohm
white/black: 2000 ohm



nickvd

#5
Nov 09, 2010, 04:59 am Last Edit: Nov 09, 2010, 05:16 am by nickvd Reason: 1
from the opposite end of the '100K' marking (edit: heh, just noticed two 100k markings... 1 is on the top)

1 goes through a 33k res then blob-town
2 is ground
3 goes to a 51k res AND to blob-town, other end of res goes to another res/cap  (104 marking, to ground) AND blob-town
4 goes to a cap (marked 104, to ground) AND blob-town

heh... blob-town...

I drew it out to the best of my mspaint ability ;)

You'll want to load the larger version for sure....

nickvd

Tested the voltage between black and red while applying 3 and 5v across white/black.  Couldnt see a predictable change in reading when using 3 or 5v using my el-cheapo meter ($10), however when i cranked my power supply as low as it could go (read 100mV at red/black) it would change, but only slightly (as expected)... makes me wish i had an analog meter... or a better digi....

nickvd

As far as the amp is concerned, I came across a similar schematic that used an LM324.  Seeing as I don't have any, I found a couple LM358's  and well... since I know nothing, I am only assuming they will work the same...  Using his schematic, I more or less copied it, leaving out an led or two...  Will this work suitably?


nickvd

#8
Nov 10, 2010, 03:32 am Last Edit: Nov 10, 2010, 04:32 am by nickvd Reason: 1
Ok, I more or less understand that, (trying to rework it in my head :))... what opamp would you recommend, or are you saying the 358's would do... (5v preferably)..

That being said, I'll be trying the one posted since i have two sitting here, and if i release the smoke then, well, all in good fun I say ;)

edit: yeah.... nuthin... I'm still not entirely sure how to wire in the strain gauge.  I read .5v on the output of the amp circuit i posted, and it jumped to 3.06 after connecting the white/black to the input (black also tied to ground -- afterthought: possibly my mistake) with red going to +5v.  No change in the reading (via output/ground)..

Or am I just way off the mark? :)

nickvd

Ok, so I've been playing around with it with not much in the way of a result.

I re-soldered the 4 strain gauges back onto the daughter board pictured below:



and checked resistances between the 4 outputs (1-4 from left to right as pictured)

1/2 and 3/4 are both 1500ohm
1/4 and 2/3 are also both 1500ohm
2/4 and 1/3 are both 2000ohm

I read somewhere that the input voltage usually goes to the higher resistance of the pairs, so with +5v running into 1/3 (+/- respectively)
I did measure between 0.2mV with no load and 1.4-1.7mV with about 175lbs on it... I was expecting a bit more of a change...

I wired it into a opamp (lm3900n to be precise), and while it did amplify the voltage, I could no longer detect a change when under load.  Not knowing much about electronics I searched for some schematics and wound up using something similar to this:



Depending on how I re-jiggered things I would read either ~100mV or ~4.5V with no change under load.

Here's the current setup:



I have a bunch of random IC's I picked up for free, I searched through my inventory spreadsheet and came up with the following (searched for description contains 'amp')

AZ4558CP      DUAL BIPOLAR OPERATIONAL AMPLIFIERS
LM301      SINGLE OPERATIONAL AMPLIFIER
LM386      Low Voltage Audio Power Amplifier
LM3900N      QUADRUPLE NORTON OPERATIONAL AMPLIFIERS
LM3900N      QUADRUPLE NORTON OPERATIONAL AMPLIFIERS
TL072CP      Dual Low-Noise JFET-Input General-Purpose Operational Amplifier
LF357N      Series Monolithic JFET Input Operational Amplifiers
LF398N      Sample-and-hold amplifiers
LM358N      Low Power Dual Operational Amplifier
MC1709CP1      MONOLITHIC OPERATIONAL AMPLIFIER
TL441CN      Logarithmic Amplifier 16-PDIP 0 to 70
LF356J      Series Monolithic JFET Input Operational Amplifiers
CA3130E      15MHz, BiMOS Operational Amplifier with MOSFET Input/CMOS Output
NE5534AP      Low-Noise Operational Amplifier
TL072CP      Dual Low-Noise JFET-Input General-Purpose Operational Amplifier
TL071CP      LOW-NOISE JFET-INPUT OPERATIONAL AMPLIFIERS
LF411CN      Low Offset, Low Drift JFET Input Operational Amplifier

Any thing here ring a bell?  I can only assuming (with my knowledge) that at least one of them would work for my purposes (trying to run at 5v, but I'll budge if I have to).

Thanks for the help!

nickvd

#10
Nov 14, 2010, 05:46 am Last Edit: Nov 14, 2010, 05:47 am by nickvd Reason: 1
There are no resistors on the board itself.  Just the solder pads for some surface mounts..

Here's a closeup:


nickvd

Also, I've switched out the lm3900 for a lm358 with the same (as in none) results (r1/rf = 100/10k [also tried 10k/1m])

Go Up