Hello:

I am working on a strain guage project for our formula sae car. In order to measure suspension travel we want to hook up strain guages to each corner of the car so when the shock compresses, it sends a 0-5V signal to the mychron we have.

The question i have is this: i found this item online and i was wondering how we could implement this into our project. The strain guage will run through an amplifier to give us the 0-5v signal needed. We wish to run this off of the car battery (12V):

http://www.blh.de/pdfamp/ba660_e.pdf

looking at this since we wish to run 4 strain gauges (one per spring) can someone explain the pins in question?

screw terminal:
pin 1: is this the input from the strain gauge (+)?
pin 2: is this the input from the strain gauge (-)?
pin 3: what is this? excitation?
pin 4: what is this? excitation?
pin 5: self explanatory
pin 6: what is analog out?
pin 7: self explanatory
pin 8: what is this?

pin terminals:
pin 1: self explanatory
pin 2: self explanatory
pin 3-pin7: what are these for?
pin 8: what is this for?

finally why is there a power supply input for both sides? shouldnt there only be one since there is only one power source?

Thank You

 

You might want to look at www.omega.com for information of strain gauges and their placement. Also using full or half bridge arrangements.

Your amplifier board is only capable of handling one strain gauge bridge.

 

Thanks for pointing me to that link, that definitely looks useful for future reference. I was reading on the finer points on how strain gauge amplification works and i pieced together this schematic and i was wondering if this looked reasonable to you guys.

 

With resistors being as poorly-matched as they are (1% at best), you would get better results by the use of an instrumentation amplifier, although at a higher cost.

Here is a link to one IA - http://www.analog.com/en/other/militaryaerospace/ad620/products/product.html

TI bought Burr-Brown and so has a number of IA's available.

 

If you want this to work you will need a split supply which means some sort of precision reference supply. If you can avoid using the auto electrical ground then ±5 volts works really well from a 12 volt nominal auto electrical system. Don't use that op-amp. The drift will completely obscure the data. Use an ICL7650 or equivalent and many problems will vanish. Indeed with the 7650 the bridge is not needed as the system can be direct reading as the 7650 has no drift or offset for all practical purposes.

 

I don't know why a better power supply will help (much). I suppose you are concerned about supply ripple? it would be a common mode signal to the bridge / amplifier. so as long as the amplifier is truly balance (which it is not now), you should be largely immune to supply ripple.

 

If 350 ohm foil gauges are used, they get very unhappy when the excitation voltage is above 10. So a regulated 10 VDC supply is pretty much a necessity.

A lower voltage will work, but the sensitivity of the bridge decreases (generally 3 mv/V).

Circuscurcuits - how does a chopper stabilized amp help with a differential signal? Strain gauges have to be used in either full or half bridge configurations. The 100 ms settling may be too slow to catch transient data from the suspension. See www.omega.com on strain gauges.

 

http://www.omega.com/literature/transactions/volume3/strainfig9.html

You mean like this schematic from Omega? I see no bridge.

Where did 100ms come from? Step response input settles in like a microsecond for the ICL7650.

A bridge requires many precision components and is a hangover from the days of lack of decent parts except resistors. A simple differential amp input as shown does not have balanced input impedance and is therefore subject to lots of common mode problems. Someone said use an instrumentation amp for this design and they are correct. If a bridge is used with 0.1% resistors to make it work the instrumentation amp is also required as that person stated.

It is grossly easier to place the strain gauge as the feedback resistor or gain set resistor in a simple op amp circuit and bias the plus input of the op amp from a well regulated supply such that the gauge has the appropriate bias current or voltage. If needed, the gain set resistor can be returned to some other supply. It is far easier and simpler to make good supplies with simple regulators than it is to make instrumentation amps and bridges that actually work.

Use a coax cable to reduce noise from sensor.

 

Well, I reread the "Practical Strain Gauge Measurements" by the people at Omega. Besides saying that there is no practical advantage to using constant-current rather than constant voltage excitation, the text around and about Figure 21 (the gauge under constant current excitation) dismisses the technique as having no practical value.

Yes, it is grossly easier to do some different placement of a strain gauge in circuit, but the result is simply not as good as what one gets from a bridge configuration.

The ICL7650 may indeed settle faster than I said. Good luck obtaining one, though. Maxim is chucking the line. They don't work well as difference amps, anyhow.

As far as -

It is far easier and simpler to make good supplies with simple regulators than it is to make instrumentation amps and bridges that actually work.

- goes, I can't say. I have done both with acceptable results each time.

By the way, I beat TI to a self-zeroing IA by 13 years. They did it on one IC and I had to do it on a PCB. That was an IA feeding a 16 bit ADC in a setup that got one gram sensitivity out of a total load of 25 Kg.

 

http://www.intersil.com/products/deviceinfo.asp?pn=ICL7650S

Buy from intersil. Maxim has planned to discontinue as you state and Harris may still be a second source but who cares? Maxims never worked as well as the original in spite of the fact it was supposed to be "improved" and more expensive than intersil. We digress.

The solution using multiple power supplies is pretty easy to accomplish bias and requires no special 0.1% parts for a bridge or instrumentation amplifier. Placing the gauge in feedback loop really simplifies building a working unit. Of course there are many ways to bias and read a strain gauge. Feedback loop is the simplest. Wheatstone bridge and instrumentation amp is the most difficult and complex. Why use 8 parts when 67 will do? I like really simple so the feedback loop works well for practical applications. Speed and stability are limited by the gauge and not the circuit as with the instrumentation amp. There are just a lot of advantages to simple. This is why it is not as popular as the "learned" method.

1 gram out of 25000 grams? A small amount under 15 bits or a 4.5 digit multimeter.

As for TI, around here any part or product that says TI goes directly to the trash can. "Never meets spec" and unpredictable and not reliable though many others have managed to make them work.

I hope some of this helps the guy with the spring. Does he know most auto springs are variable rate? He may be better off measuring suspension deflection angle rather than some strain on a intentionally non-linear (non-constant K) spring. Alternatively a load cell may be able to be placed under the spring stop and measure load directly. Moving the stain gauge to a different location on the spring will yield different results for exactly the same suspension deflection because of the non-linear spring rate. Maybe he has a torsion bar and it is a linear spring. He did not say.