Exponential amplifier using LM741

Ron H

Joined Apr 14, 2005
7,063
On dc, a ground is the return path to the power supply.
AC it is the steak in the ground

GROUND is used as a 0v reference, but you could use a different reference if you were so inclined.

(And I would use a rod longer than 6 foot) ;)
Steak is beef. Stake is what you dispatch a vampire with.:D
 

retched

Joined Dec 5, 2009
5,207
You realize, now that you called me out in open forum, I cannot edit my 4th grade miss-steak. ;)

While were at it, what do you call a cow with no legs?

Ground beef. :D
 

Thread Starter

summersab

Joined Apr 8, 2010
161
Sigh. At long last, it works. I don't really get why, but at this point, I don't much care :)

Thank you all SO much for your help. I might post later if I have trouble getting the bridge and everything else to work, but right now, I'm happy with this.
 

retched

Joined Dec 5, 2009
5,207
Well you started this yesterday afternoon, and your done now.

You could have only done better if you lost the 741 on your way home from the store, and when you returned, they were out, and you bought the 339 for kicks.

At least you kept going through it all till the end. Very admirable.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
Thanks, retched.

Any thoughts on taring the Wheatstone bridge setup? I am using two load cells, and instead of using two bridges, I'm considering using one bridge and putting the load cells opposite one another (having run the math in an Excel spreadsheet, it seems to give the same effect as combining the outputs of two separate bridges).

I have considered using a potentiometer in series with one of the other resistor legs of the bridge to tare the setup, but the output seems VERY sensitive, so I'm not sure if that's a good way to go.

Attached is a (really disorganized) schematic of what I'm thinking. Any better thoughts for tarring?
 

Attachments

retched

Joined Dec 5, 2009
5,207
A pot would be the way to go. Actually a trim-pot. These come in very small sizes and are easier to "dial-in".

If you really want to loose your hair, try to incorporate a digital trim-pot into the mix.

You will also need a display of some sort to know where you are at in respect to the tare value.

Without a display, you wont know which way to go. You could be over, or under.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
Do I still need to ground the negative input to my comparator for it to work? That grounds one of the legs of my bridge, and I've been discovering that it messes up the output of the bridge (I have yet to test the bridge in conjunction with the comparator since I don't have a trimpot to tare everything - I've just been taking outputs from the bridge, and when I tie one leg to ground, my readings are bizarre). In the comparator setup that you edited, you grounded the negative input, so I wondered if it was necessary. Maybe I've just breadboarded it wrong.
 

Attachments

Ron H

Joined Apr 14, 2005
7,063
Do I still need to ground the negative input to my comparator for it to work? That grounds one of the legs of my bridge, and I've been discovering that it messes up the output of the bridge (I have yet to test the bridge in conjunction with the comparator since I don't have a trimpot to tare everything - I've just been taking outputs from the bridge, and when I tie one leg to ground, my readings are bizarre). In the comparator setup that you edited, you grounded the negative input, so I wondered if it was necessary. Maybe I've just breadboarded it wrong.
No, you don't have to ground either input. Look at the circuit I posted in post #22.
One problem with the dual load cell bridge is that both passive legs have to have the same resistance as the LOADED load cells (assuming they are identically loaded).
Isn't that a problem?
With a single load cell, the resistances of the other side of the bridge (R3 and R4) can be identical, but don't have to be the same value as the load cell. This saves power consumption.
There are several ways to tare the circuit. Below is one way.
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,230
Ron_H,
Couple of things...

1) I didn't know this, but 78xxL regulators seem to have gone the way of the dinosaur. I can't seem to find a major authorized distributor who stocks them.

2) A 7805 regulator has a ~5.5mA current to GND to begin to establish the internal reference voltage; however it requires another 5mA load in order to provide guaranteed regulation of the output. Unless the output of the 339 is sinking current from the 1.5k resistor, the minimum current requirement for guaranteed regulation won't be met.

3) I have a feeling that the plan is to power it via a 9v PP3 battery. Our OP may not be aware of how wimpy they are.

I'm wondering if the only need to regulate voltage here, is to provide the 0v-5v output for the TTL signal? In that case, why not make the ground common between the module that needs the input signal, and simply use a pull-up resistor (say, 3.3k to 4.7k) to that module's 5v rail?
 

Ron H

Joined Apr 14, 2005
7,063
Ron_H,
Couple of things...

1) I didn't know this, but 78xxL regulators seem to have gone the way of the dinosaur. I can't seem to find a major authorized distributor who stocks them.

2) A 7805 regulator has a ~5.5mA current to GND to begin to establish the internal reference voltage; however it requires another 5mA load in order to provide guaranteed regulation of the output. Unless the output of the 339 is sinking current from the 1.5k resistor, the minimum current requirement for guaranteed regulation won't be met.

3) I have a feeling that the plan is to power it via a 9v PP3 battery. Our OP may not be aware of how wimpy they are.

I'm wondering if the only need to regulate voltage here, is to provide the 0v-5v output for the TTL signal? In that case, why not make the ground common between the module that needs the input signal, and simply use a pull-up resistor (say, 3.3k to 4.7k) to that module's 5v rail?
7805L was a typo. I thought I changed that to 78L05, which is available. The minimum current on them is only 1mA.
Having said that, I like your idea of using the vcc of the load module for the pullup.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
I return once again with more. Problems. I had a feeling this would happen.

Some background. My load cells are pulled from a very generic bathroom scale. Thy are simple beam-style, single deformation load cells. The strain gauge is already mounted and there are wires mounted securely to the terminals. Interestingly enough, there are three wires: yellow, orange, and brown. The resistance between orange and brown is 1.5k while the resistance between any of the remaining terminals is 750ohm. All resistances decrease proportionately as strain is applied in the same direction (meaning it's not a rosette or anything). This tells me simply thus: the strain gauge has three terminals, like this:

MM
***

Each asterisk is a terminal while the M's represent the strain gauge wiring. Between the first and second terminals is 750ohm, second and third is 750ohm, and first and third is 1.5k.

Background aside, this is the problem I'm dealing with. I've been using a pair of vice grips to test the strain gauges, and when I grip the gauge solidly, I can get the strain gauge to drop by about an ohm. When in a Wheatstone bridge of three 750ohm resistors and the 750ohm leg of the strain gauge, that only amounts to 1-3mV drop in output tops. That's not nearly sensitive enough to trip a comparator. So, my nagging question is this: the load cells worked just fine to measure body weight down to 0.5 lbs. How do I make my strain gauges more sensitive so I can get this darned circuit to work finally?

(Wow, this forum has NOTHING to do with the subject at this point . . . )
 

rjenkins

Joined Nov 6, 2005
1,013
I'd guess it's actually a half bridge (one side leg of a full bridge).

You need two more resistors for the other side leg of the bridge.

The absolute value is not critical as long as they are identical.
You could add a low value preset between the two for calibration.

It's a fairly common economy setup.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
I'm a little confused. Using this as reference (dang, isn't this a nice forum?):

http://www.allaboutcircuits.com/vol_1/chpt_9/7.html

I've got my gauge in a quarter bridge. Where is the half bridge bit coming in? If you're referring to the fact that the load cell package appears to have essentially a dual strain gauge in it, note that each terminal of the strain gauge is effected by a load proportionately (meaning terminal one to two changes as much as terminal two to three while terminal one to three changes twice as much. It's the same strain gauge, not something like picture 4 in the above page). Found a picture of what I believe is in the package:

http://www.dtbtest.com/IMAGES/Strain-Gauge-4.jpg


What is this "low value preset" you speak of? (Remember, I'm an idiot over here . . . )
 
Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
Take another look at how the strain gauges are attached to the beam. Isn't one on the top and another on the bottom?

The scale originally had an instrumentation amplifier in it.
 

rjenkins

Joined Nov 6, 2005
1,013
Unless you use exacly the same form of attachment and pressure angles on the load cell as it had in the original scale, it won't give the correct output.

I'd still expect it to be a half bridge.

Re the preset, rather than use just two fixed resistors for the reference half of the bridge, if you add a preset between the two resistors, the wiper of the preset becomes the center point and you can tweak the balance of the two resistors to set the zero point.

Depending on how much adjustment you need, the preset could be a tenth or even down to one hundredth of the resistor value. Lower values give less range but higher precison.
 

Ron H

Joined Apr 14, 2005
7,063
Re the preset, rather than use just two fixed resistors for the reference half of the bridge, if you add a preset between the two resistors, the wiper of the preset becomes the center point and you can tweak the balance of the two resistors to set the zero point.

Depending on how much adjustment you need, the preset could be a tenth or even down to one hundredth of the resistor value. Lower values give less range but higher precison.
That's what I posted in my last schematic.
 

Thread Starter

summersab

Joined Apr 8, 2010
161
Sorry to bump something so old, but I didn't want to do any more double-posting damage. I've had trouble with my outputs from the Wheatstone bridge being too small to detect unless massive forces are applied. I'm thinking of getting an instrumentation amplifier to see if I can get a read on my outputs.

My question, then, is thus: what is the difference between the multiple amplifiers? There's a slough of INAxxx models by TI, many more by other manufacturers, etc. Some are low power, and I take that to mean supply power, but does that make a difference? What's a general, single-amplifier IC package that is pretty cheap and most frequently used?

Once again, feel free to talk stupid to me :)
 

beenthere

Joined Apr 20, 2004
15,819
I've been using a pair of vice grips to test the strain gauges, and when I grip the gauge solidly, I can get the strain gauge to drop by about an ohm
Some of your problem may relate to this. A strain gauge is designed to change resistance due to a change in length (with a maximum of about 5%). Crushing the device with pliers is more likely to cause mechanical damage to the resistive element that to get a sensible output.

I've got my gauge in a quarter bridge.
A "quarter bridge" is a single resistive element. Without some additional fixed resistance to control current, you may destroy the gauge due to self heating. You also lose the ability to see the resistance change due to elongation to compression.

Take the battery voltage in the scale as the maximum excitation that can be applied to the gauges, plus the fixed resistors in series. The resistance of the gauge will probably be close to the fixed resistors for maximum sensitivity.

You can get literature from Omega Engineering that gives a great deal of information about strain gauges (they call then gages), along with how to use them.
 
Top