strain gauge guidance

It's a very long time since I did this at college I remember you only need one - just checked http://www.allaboutcircuits.com/vol_1/chpt_9/7.html

yes you do but according to this article four is best. Without seeing the physical layout it's difficult to see if they are wired in the correct sequence with respect to each other. Have you considered calibrating it with known weights it may be correct. What will happen next is someone far wiser will be more specific and give a better post. I had never heard of this use of strain gauges before, it's a great application of the technology. Good luck.

Actually that post of mine was worse than useless, and it's too late to remove it. I remember on a project to weigh a vehicle a steel bar had to be thinned to get it to flex more so the strain gauge would pick it up. In the info on the net was there a recommended thickness for aluminium? Gut feeling, you are at the top end but still ok.

Elastic Strain [epsilon] = Extension [deltaL] / Original Gage Length [L]

can you use this formula, measure gauge, the linear deflection and use pythagoras?

I am assuming your perch is cantilevered. Without knowing the exact dimensions of your setup, I cannot calculate your maximum allowable displacement at the end of the perch at 3% strain. I suspect that you're fine. If you do damage the gauges, with the circuit you're using, you'll know right away when the load is removed. One option you have is to rigidly support the end of the perch (via angle brace) and presto you've built a load cell. This will retain the functionality of your system while drastically reducing the deflection. No matter what, you'll have to calibrate the circuit using known weights. Sounds like a novel setup - if you can get it working reliably and can package a self contained unit, patent it! Good luck.

Sounds like you're doing O. K. The only thing I would suggest is to put a mechanical stop on the perch to prevent overstraining the aluminum and strain gauges from something like a marauding squirrel or large bird.

Yes, that is way too much deflection. I would be worried about the bond between the gauge and the aluminum breaking due to the amount of deflection. Also, the aluminum will most likely not return to its original position (gets bent) over time. Also, when the bird leaves the perch it will probably push off the perch and bend it even more.

I would think you want to size the rod of perch so that you need to amplify the bridge output at about 100X. I use a Texas Instrument IC part number INA326 to amplify my bridge. See:

http://focus.ti.com/lit/ds/symlink/ina326.pdf

You need to design a circuit and PC board for the amp, but that is not as bad as it seems.

Another thing to consider, how will your weight reading vary depending on where the bird stands on the perch? Is it designed such that it is independent of the bird's position? A cantilever perch will give you different readings at the end than it does in the middle. I don't know how you are mechanically doing this, but there are a lot of things to consider. It sounds like a great idea and I hope you succeed!

I don't know how long you are running this experiment, but there are thermal drift problems with strain gauges and aluminum has a pretty high coefficient of thermal expansion. Both will result in errors as the weather changes...

A load cell might be a better approach.... google load cells for more info.

The full-bridge setup he's using self-compensates for thermal expansion/contraction.

I would agree with the others comment to put stops on it. all so include side guides and stop limiting the upward travel. The wind could place loads on it in many different directions.

You could reduce the amount of force on the strain gauge with levers

Hi all,

I'll try to resume the situation:

Everybody assumed the cantilever model. If this is the case, of course, the position of birds along the beam results in a different bending moment and changes the bridge output. One solution could be to limit the position available to the bird and to consider the weight as a uniform distibuted load over that length.

No mention is done about the adhesive used to bond the gages: not all are suitable for all services. For example, cyanoacrilates are used because its fast curing (a couple of minutes) but are not suitable for large strain.

I suppose the full bridge is formed by 4 active gages located longitudinally and 2 in the upper side of the sheet and 2 just in the opposite side which results in an "amplification" factor of 4 (the bridge output is 4 times the output that will give only one gage)

With this configuration, the bridge is compensated for temperature changes along the measuring time, and compensated too for axial loads or torsional torques, measuring thus only the vertical load (the bird weight).

If one or several gages are misbonded, the output will be erratic. This can be checked before installing the sheet, by rubbing softly the gages and looking if after rubbing the output is the same than before.

With the bridge configuration already commented, the bridge relative output is:

Delta V/V = K*epsilon (K is the gage factor indicated in the gages package)

Once determined strain (epsilon), the bird weight is:

Weight = epsilon*E *W/L where E is the Young's modulus of the aluminum sheet (about 71 GPa), W is the section modulus of the sheet section and L is the length beteween the center of the gages to the point of load application.

Answering another of the questions, yes, there are strain gages for almost any range of strain. The limiting point usually is the adhesive more than the gage itself.

Obviously if "the voltage across the bridge was 0.4V at rest and went up to 1.4V when a weight of 70g" you are not using a signal conditioner, but if after removing the 70g the output is again 0,4V means that the sheet thickness is valid because no plastic strain occurs.

Of course there are load cells for this range of weights but trough the above paragraph I suppose you are trying to solve your problem in the cheaper way and I think you have selected a good one: the bridge output must be linear with weight and you can easily calibrate the system with two or three known weights.

Here you have a link which have a lot of free technical information about strain gages.

http://www.vishay.com/strain-gages/knowledge-base-list/

Wow, that is an impressive amount of feedback for a single day. It seems the majority opinion is that the setup is OK. I'm pretty sure my aluminum strip (which by the way is about 1.5 x 20mm in cross section) is stiff enough to retain its shape even after take off of a large bird. Also, I will use a mechanical stop to prevent extreme bending. Sorry for leaving out so many details about the bridge, but it looks as though most everyone guessed correctly about what I'm doing.

To address a few concerns, I will mention the following. The strip will be rigidly secured on one end as most folks guessed, and the bird will land on the other end. I plan to convert voltage across the bridge to weights by calibrating with 4 or 5 known weights (maybe more depending on what the data look like). Even if the output is not linear, I should be able to generate a conversion function. The actual perch for the bird to land on will project up from the end of the aluminum strip and will be perpendicular to the strip, so the position of the birds center of gravity will more or less be restricted to one location along the strip. There may be some issues with birds not landing in the middle of the perch and creating twisting forces on the strip, but I chose a broad strip with the hope of minimizing this. I'm breadboaring a simple op-amp that should boost the signal a bit more--like 3x or 4x. The circuit should be fully adjustable via a couple of potentiometers, so I should be able to get it going. This feeder is a prototype that will be tested in the next few weeks. I hope to eventually incorporate about 40 of them in a large scale research project. I checked out the load cell option, and the lowest price I could find was about $80 each. With strain gauges, the price is about $20 each and I can really fine tune the customization. Of course there is the issue of bonding the strain gauges. This is an art I have not yet mastered--but I think I am getting there. Word to the wise--don't use scotch tape!

One other detail of interest, the birds will be identified as they are weighed via an RFID tag, so I should have multiple weights for each bird. So long as this set up works some of the time (even as little as 30%) I should get plenty of weight data.

Thanks everyone for the helpful comments. Let me know if there is any interest and I'll post more details about the device.

Eli

Brid0030:

Don´t worry, The output shall be linear and with the full bridge configuration, the effects of twisting must be cancelled so the bridge output will be due only to the bird weight.

Regarding the comments about strain gages, forget them. Human beings use to be affraid from unknown matters. I've used some dozens of thousands of strain gages and if you know their behaviour, there should be no problems.

Anyway, apart from the calibration with known weights, you can try to make your own calculations before and them compare the results of calibration with the calculated output, then you can easily find if there are any gage damaged or bad bonded.

I think it´s a good idea to use some potentiometers to initially balance the bridge. You can use high resistance ones in parallel with gages or low resistance ones in series.

Some "gurus" offered you a "kitchen analogic item". I do not recommend this solution because in that case you would to instrument the item to get the signal or to be 24 h. a day watching (perhaps "gurus" do)

Here is a direct link to a html document with the formulae you need in the cantilever beam with a full bridge configuration. You will find in the same web page the cases for half and quarter bridge.

http://www.vishay.com/brands/measurements_group/guide/ta/ftm/ftmd.htm

Personally speaking, I despise strain gauges. We tried them in our installations and had problems getting accurate readings. They're not very forgiving when it comes to installation. Load cells are easier to use albeit more expensive.

That said, I suggest you construct your birdhouse so that the weight is concentrated at as small a point as possible. That way, the position of the bird won't matter.

Also, a thicker slab or bar of aluminum seems called for since, as you say, the metal bends under the weight of the birds. As someone pointed out, the aluminum may be bent permanently which will cause your zero point to drift over time.