DIY Force Sensor: Craft or Buy for High School Physics Labs?

[mishima]

I am looking for a way to make my own "force sensor" similar to this Vernier dual-range force sensor for high school physics labs.(Vernier) How is this different from a digital spring scale, like for fish? Can digital spring scales also measure pushes? Can the components be purchased from electronics suppliers or crafted? I'm hoping I can just modify a cheap fish scale to read voltage changes on my computer.

 

[MrSparkle]

a fish scale would be pulling only. They've got no need for pushing, so that'd be a waste of money for them to put it in. You could probably combine a fish scale with an electric kitchen scale.

 

[mishima]

Well, I found a Quarrow fish scale at the local fish shop for 14 dollars, so decided that was cheap enough to risk. Inside looks like actually a small strain gauge with 4 leads.\

It gives a negative reading on the LCD display for pushes so I think it might work, but its only accurate to 50 grams. Reading this awesome page I assume the relevant electronics are a Wheatstone bridge and amplifier, so if I can't just splice a wire somewhere I could build my own circuit fairly easy. I'm not sure if the more expensive Vernier sensor uses a similar strain gauge...I was thinking there was some other technology used for some reason.

 

[mishima]

So now I am guessing under the gunky white stuff on the load cell there is already a wheatstone bridge. I supplied 3.3V to the red wire while reading voltage across the white and yellow wires. Just sitting on my desk it read 0.1 mV. When I squeezed it (to the point of pain in my fingers) it climbed up to 0.4 mV. Letting go the voltage dropped back to 0.1 mV.

I just need an amplifier that can change that range into something that looks nice graphically on a computer. I use my soundcard mic input and a program called RadioSkyPipe2 for other similar projects. Any suggestions?

 

[jim hardy]

Some reading material for you...

http://www.analog.com/static/imported-files/seminars_webcasts/49470200sscsect2.PDF

http://www.analog.com/static/imported-files/seminars_webcasts/927412250sscsect4.PDF

http://www.vernier.com/engineering/stem/sensors/strain-gage-measurement/

http://www.maximintegrated.com/en/app-notes/index.mvp/id/3396

http://www.intersil.com/en/products/amplifiers-and-buffers/all-amplifiers/amplifiers/ISL28134.html

 

[mishima]

Thank you, I had never even heard of instrumentation amplifiers. I had been bumbling with 741s and BJT amplifiers...

The IC you posted looks like surface mount (beyond my workshop capabilities), do you think this INA126 would work?

http://www.ti.com/lit/ds/sbos062a/sbos062a.pdf

Apparently there is also INA125 with a built in reference voltage for double the cost...fairly sure I do not care about that since the values will be seen graphically.

Are there "standard" in-ops that everyone uses?

 

[AlephZero]

Thank you, I had never even heard of instrumentation amplifiers. I had been bumbling with 741s and BJT amplifiers...

If you don't want to go all the way to an instrumentation amp in one leap, try replacing your 741s with TL081/82/84s. They are pin-compatible but have much higher impedance inputs, lower noise, etc.

http://www.ti.com.cn/cn/lit/ds/symlink/tl084.pdf

 

[jim hardy]

I think the INA will do well for you.

It saves having to hand pick matching resistors.

Aleph is right, 741 is a sort of primitive opamp. LM324 is another handy one for single supply applications, and cheap. But you want some precision around your bridge.

 

[jim hardy]

Here's a nice intro to bridge sensors. It goes off on a computer tangent, but first part you may find useful.
http://www.ti.com/lit/ml/slyp163/slyp163.pdf

 

[mishima]

Finally got a few INA126 in the mail. Very happy to share with you some success with this project. I used a 6V 300mA wall wart for power and a 100 ohm resistor to set the gain near 1000. This took my tenths of a millivolt readings up to tenths of a volt as desired (getting a range of around 0.6 to 1 V for pushes). I fed the output through a chopper (this insanely handy circuit) and the chopper into my laptop microphone jack. Using RadioSkyPipe2 I plotted the following:

Here the blue is me pushing slowly as hard as I can. The red is me pulling as hard as I can (its harder to pull since its just a little rod at the moment). The green is just sitting on my workbench. I'm ecstatic! I likely did not wire the INA126 in an optimal fashion, to be honest I am confused about the 5 pin. Here is what I did (the Y is the yellow line coming from the bridge, W is the white line):

Any suggestions on that front are most welcome, although as is it the circuit seems to be doing everything I had hoped for. Finally here is a pic of the super sloppy breadboarded circuit and bridge sensor, to be cleaned up later of course:

Anyways, a BIG thanks for the help.

 

[jim hardy]

What is voltage on your Y and W wires?

Reason i ask is - opamps like to have a little "headroom" between their input pins and their power supply pins.
Look at datasheet here
http://courses.cs.tamu.edu/rgutier/ceg499_s02/sbos062.pdf
(the one from TI isn't displaying characters correctly in my version of acroobat)
page 4 bottom right chart labelled "Input common mode voltage range vs output voltage, Vs=+/- 5"

What it is showing graphically is this -

This IC likes to see its input pins at least a volt away from both V+ at p 7 and V- at pin 4.
Pin 5 would best be set about halfway between pins 7 and 4,
hence that little dashed rectangle showing locus of "desirable" inputs for single supply of +5volts on pin 7 and 0 volts on 4, with pin 5 halfway between them at 2.5.
Wow that's a mouthful - does it parse okay ?

If you have available a negative supply of a few volts, try setting pin 4 to around negative 5volts.
If not, try setting pin 5 about +2.5.
In either case you'll want your input pins2&3 not very far from pin 5's voltage which ought to be ~midway between supply pins 4&7.

Your INA will thank you for that.
................

"Headroom" is my colloquial term for that allowable voltage difference between input and power supply pins.
It's given in datasheets as "common mode range" , an oblique term everybody hears but hardly anybody pays attention to until it causes them trouble.

...................

Congratulations on your most interesting experiment .
You prove once again the adage "We learn faster by doing than by reading about doing."
It's iterative - first it doesn't work at all, then you get it sort of working, then you get it working great, then you see how you should have started out in the first place !
Keep us posted... I'm vicariously enjoying your success.

old jim

 

[mishima]

My students also thank you, because now we can spend money on something cool like a slow motion camera instead of expensive Vernier sensors.

Question, does the graph on page 4 scale in a linear way for supply voltages other than $\pm$5? My wall wart is actually 8V (despite it being marked 6V). So, my common mode voltage on the Y and W is 4V. Raising pin 5 voltage to also be ~4V required the use of a 1Mohm resistor. This seemed to rescale my push pull graph in a way that clipped off some of the output, I must have hit some saturation range or something. It also raised the output voltage significantly, like a DC bias (is that the point of Vref?) I tried a few other resistors between 5 and ground.

Here the first is with 30kohm, second with 50kohm, and the last with 1Mohm. The push/pull/push at the very end is what I am saying looks clipped (but maybe this is a limitation of the mic input on my computer). The square forms are where I was adding/removing resistors.

I also tried using a few 3.3V battery packs to setup a dual power supply as mentioned (didnt have the stuff to make a dual 5V unfortunately), with pin 5 grounded. Looks good, but again there is some kind of scaling. From my other post above the first push jumps up about 400 units on the graph, but with this setup it jumps up only 200. Is it fair to call that a lower resolution? Is that because of the lower range of input voltage (6 compared to 8)?

 

[jim hardy]

So, my common mode voltage on the Y and W is 4V. Raising pin 5 voltage to also be ~4V required the use of a 1Mohm resistor.

4 volts common mode is too much for a 5 volt supply.
But you had eight...that's more headroom for you.

Now to pin 5... observe there's going to be some small current out of it that'll give a drop across any external resistance in series with pin5.
You'd do better to raise pin 5 by use of a lower ohm voltage divider - say a couple of 1K resistors in series across the supply, will give you Vsupply/2 at their junction. better yet would be to use an opamp buffer between the resistive divider and pin 5 for effectively zero ohms in series with pin 5. Here's what they say about resistance in series with pin 5...
page 7, "Applications information"

The output is referred to the output reference (Ref) terminal which is normally grounded. This must be a low-impedance connection to ensure good common-mode rejection. A resistance of 8 Ω in series with the Ref pin will cause a typical device to degrade to approximately 80dB CMR.

LM324 is a garden variety opamp that's cheap and useful for such tasks. Its common mode input voltage goes from zero to about 1.5 volts short of positive rail.

Question, does the graph on page 4 scale in a linear way for supply voltages other than ±5?

It is not intuitive what those graphs are showing you.
Best way to answer your question is:

Take a look at the adjacent graph on page 4, the one for for +/- 15 volt supply.
It shows that input common mode, on left vertical axis, can swing from at least -10 to +10, with slight shift positive as output goes more positive.

As i said earlier, the opamp wants a few volts between its input pins and the supply pins (supply voltage is often called "Supply Rails" - ie a hard limit you cant't go outside of).
The area enclosed by that parallelogram is the operating range of your INA, ie locus of desirable inputs.

Now back to your right hand graph
With +/- 5 volt rails and pin5 halfway between them at 0, your input common mode needs to be between -3.8 to +2.8 , that range shifting more positiove as output goes more positive.
Interesting ; 4 volts is too much common mode for even +/5 volt supply. But you had eight/zero , so it at least sort of worked for you.

Now - the little rectangle for +5/0 shows how severely limited is your input CMV range for that small supply.
There's a better explanation 'why' than i could compose at bottom of page 8, paragraph "input common mode range".

So your next step is to get supply voltages that are suitable for your input CMV from the Quarry scale,
and place pin5 ~halfway between them through low impedance.

Can you steal power from the Quarry scale? They solved this problem somehow. Can you locate its power supply rails on the circuit board?

here's a gizmo that'll turn a single supply into dual.
http://intronics.com/products/pdf/dc200.pdf
Wall-warts are plentiful in thrift shops
perhaps one of these little dual supplies and a wall wart to power it would do. That's how i power my breadboard.

From my other post above the first push jumps up about 400 units on the graph, but with this setup it jumps up only 200. Is it fair to call that a lower resolution? Is that because of the lower range of input voltage (6 compared to 8)?

I strongly suspect you are running into limits because of your small supply voltage combined with the 4 volts common mode from that Quarry scale.
Were your two bridge input wires nearer zero you'd have done well with the +/-3.3 and pin 5 to common i think.


You are progressing remarkably here - you've got it almost working great, keep on pushing !

old jim