Reading multiple strain gauges (10 to 20 strain gauges)

[cyotani]

I am working on my ME senior project and I am attempting to read the strain on suspension A-arms and push rods for our school's FSAE vehicle. I want to read and data log 5 strain readings per corner so 20 strain readings in total. The system needs to be relatively compact to fit in the car. 

From my research, each strain gauge will need 4 analog inputs (2 analog channels, one for strain voltage reading and one for excitation voltage reading). This would mean I would need 80 analog inputs to accomplish this and I can not find a suitable DAQ.

I'm trying to find a cheap and easy solution to accomplish this and am looking into things like multiplexers to switch beetween channels to require less DAQ inputs but this is getting very compliated.

Can anyone recomend a cheap and easy solution to accomplish this task?

Reference for strain bridge channel example: (http://zone.ni.com/devzone/cda/epd/p/id/6417)

Thank you in adance for any input!

[RavensFan]

Why do you think you need 4 analog input channels for each strain gage?  I think you'd only need one for each gage to measure the voltage.  And possibly one more to measure the excitation voltage (which could be shared among all the bridges.)

What do you consider cheap?  Generally the signal conditioning required for strain gages causes the price per channel to be higher than normal analog input channels.  But maybe you can find some relatively cheap strain gage signal conditioners that could convert to a voltage analog input that would be cheaper per channel.

[cyotani]

I was under the impession the excitation was to correct for the wire resistance. Which could varry per channel but I guess if I use the same and equal length wire I could get away with one excitation voltage channel?

Cheap would be all materials under $1000 at most. I hope that is possible. Educational strain gauges are about $2 a piece. The precision resistors to complete the Wheatstone brige are expensive. And I still need to amplify each signal prior to reading the analog channels into my DAQ correct. I was looing at the LabJack U3 which reads 16 analog inputs as a possible DAQ if it can get the job done.

If all 20 strain readings is not possible, the fall back would be to read 10 at a time (half the car) or 5 (one corner) at a time.

[johnsold]

I agree with RavensFan.

Get an EE student who needs a senior project and ask for a small circuit with one differential amplifier and suitable filtering per strain gauge (or maybe 5 per corner). This sounds like a project which needs collaboration between disciplines. The schools where I have worked with senior projects have encouraged such cross-discipline projects. If yours has not, maybe you can start something good.

Lynn

[johnsold]

We were posting at the same time.

$1000 dollars should be easy. Two LabJack U6 (14 channels at 16+ bits) at $300 each. Four copies of a simple printed circuit board - about $100 if you have to buy them. Many EE departments can make their own for the cost of the materials. The amplifiers, resistors, capacitors, etc. will be another $100.  Add connectors, cables, power supplies, enclosures and you will be close to $1000. If you can scrounge, reuse, recycle, and adapt the way many unversities do, you will probably have money left over.

If you decide to stay with the LabJack U3, I still recommend that you get two.  LabJack does not specify the accuracy of the devices. With your own strain guages you need to do calibration anyway so standard 1% metal film resistors are plenty good enough ($0.15).  You will be more interested in stability and reproducibility than absolute accuracy.  The system accuracy will depend on your calibration.

Lynn

[cyotani]

That is a good suggestion. I'll look into getting some help from an EE major at my school who is willing to help. 

But assumming worst case scenerio and I am on my own for this project...

If I only need one excitation voltage and 1 channels per strain gauge, an 8 channel (16 analog input) DAQ device with built in amp, would be capable of reading 7 strain gauges at a time, correct (8 channels - 1 channel for excitation reading)? And this could give me the capability of reading the 5 strain gauges per corner of the vehicle that I am looking for.

Am I on the right track? 

Also, my professor mentioned a relay bank to solve this, but that sounds like its also takes me down a complicated road where an EE expertise would be most likely required.

[cyotani]

Thank you for the input Lynn!

From my understanding the LabJack U6 has 14 analog inputs, meaning 7 analog channels(+ and - per channel). 

[johnsold]

You asked:

---

I was under the impession the excitation was to correct for the wire resistance. Which could varry per channel but I guess if I use the same and equal length wire I could get away with one excitation voltage channel?

---

The excitation is the voltage applied to the bridge. It causes current flows in the strain gauge elements and fixed resistors, if any. The voltage drops in the two branches are compared in a differential amplifier. If the bridge completion resistors are mounted close to the strain guage elements and have nominal values close to that of the strain guage, the excitation voltage is largely irrelevant because you will be measuring the difference in the ratios of two pairs of resistances.  With typical 350 ohm strain gauge resistance you could use 100 feet of 20 gauge (AWG) wire and have less than 1% voltage drop in the connecting wires. And when you do the calibration mentioned in my other post, any wire effects will be eliminated.

Lynn

[cyotani]

Lynn,

Thankyou for clarification of the excitation voltage. My confusion came from seeing the picture I attached for a myDAQ strian gauge configuration showing the excitation being read into channel 1. This led me assume each individual strain circuit needed its own excitation voltage reading.

My initial plan is to build the wheatstone bridge circuits for the 5 channels per corner and mounting it near the center of the car and reusing that for each corner. This would require about 3-4 feet length of lead wire between strain gauge and bridge. But based on your explanation this dow not seem be an issue if calibrated correctly which is good.

[LabJackSupport · Accepted solution]

First, you need complete bridge circuits.  If you are planning to use a single active strain gage at each location, then you need to connect 3 other elements (resistors or dummy gages) to form complete bridges.

If you had a handful of bridge circuits, the U3 would be the least expensive option.  Along with a U3-LV, you need 1 LJTick-InAmp for each 2 bridge circuits, so that would be about $170 to handle 2 bridges.

For 20 bridge circuits, though, the least expensive option will be a U6/U6-Pro with the Mux80.  For a little more you could go with the T7/T7-Pro to get Ethernet and WiFi connectivity in addition to USB.  These solutions are going to be about $450-$650 and give you 84 analog input channels (you need 40 analog inputs for your 20 differential signals, plus at least 1 analog input to read the excitation voltage).

What is the max scan rate you need?

There are a few more details we can get into, so you might want to email us to talk more.