Strain gauge data acquisition

[jfkunstmann]

Can anyone help? I'm trying to take the dynamic signal from a single
strain gauge and input it into my computer so I can record and analyze
the data (maybe dump it into Excel). I can't affort to spend a lot on
data acquisition hardward, so I'm wondering if there is a way I could
do this for less than $200. I've read some info about using the
computer sound card for data input. Any ideas how this might work?

 

[Al]

In article <9b7e0210.0405010729.17044185@posting.google.com>,
jfkunstmann@yahoo.com (jfkunstmann) wrote:

Can anyone help? I'm trying to take the dynamic signal from a single
strain gauge and input it into my computer so I can record and analyze
the data (maybe dump it into Excel). I can't affort to spend a lot on
data acquisition hardward, so I'm wondering if there is a way I could
do this for less than $200. I've read some info about using the
computer sound card for data input. Any ideas how this might work?

Check this out:

http://data-acquisition.globalspec.com/Industrial-Directory/how_to_build_
h_bridge_amplifier

Al

--
There's never enough time to do it right the first time.......

 

[Rene Tschaggelar]

jfkunstmann wrote:

Can anyone help? I'm trying to take the dynamic signal from a single
strain gauge and input it into my computer so I can record and analyze
the data (maybe dump it into Excel). I can't affort to spend a lot on
data acquisition hardward, so I'm wondering if there is a way I could
do this for less than $200. I've read some info about using the
computer sound card for data input. Any ideas how this might work?

Since the signal is in the microvolts, you need a chopper amp
to amplify it to useable levels.

Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net

 

[Joerg]

You could also build a V/F converter around chips such as the LM331 which
costs a few Dollars. That generates a frequency proportional to the
voltage input which should be easily readable by a sound card. Amplitude
errors or uneven frequency response in the sound card won't matter much
now because all you register is a frequency. Of course you'd have to run a
good FFT on the computer but those program should be very easy to obtain.

This V/F scheme also alleviates many noise problems on the link between
measurement location and PC.

Regards, Joerg
http://www.analogconsultants.com

 

[Rene Tschaggelar]

Joerg wrote:

You could also build a V/F converter around chips such as the LM331 which
costs a few Dollars. That generates a frequency proportional to the
voltage input which should be easily readable by a sound card. Amplitude
errors or uneven frequency response in the sound card won't matter much
now because all you register is a frequency. Of course you'd have to run a
good FFT on the computer but those program should be very easy to obtain.

This V/F scheme also alleviates many noise problems on the link between
measurement location and PC.

An FFT to get the frequency of a squarewave ?
What a waist.

Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net

 

[Joerg]

Hello Rene,

It sounds like a waist but since this has to happen inside the PC it may be the
most straightforward option. FFT programs are pretty commonplace and with the
fast processors of today the PC won't even flinch.

How else would you do it using a soundcard and a PC?

Regards, Joerg
http://www.analogconsultants.com

 

[James Meyer]

On Sat, 01 May 2004 20:47:09 GMT, Joerg <notthisjoergsch@removethispacbell.net>
posted this:

Hello Rene,

It sounds like a waist but since this has to happen inside the PC it may be the
most straightforward option. FFT programs are pretty commonplace and with the
fast processors of today the PC won't even flinch.

How else would you do it using a soundcard and a PC?

Regards, Joerg
http://www.analogconsultants.com

The original question stated "dynamic" signals which I assume means AC
voltages. Although he didn't say so, I suspect some sort of vibration analysis
is being attempted. In that case, simple amplification and direct processing
through the normal soundcard inputs would be practical.

Jim

 

[maxfoo]

On Sat, 01 May 2004 22:20:57 +0200, Rene Tschaggelar <none@none.net> wrote:

Joerg wrote:
You could also build a V/F converter around chips such as the LM331 which
costs a few Dollars. That generates a frequency proportional to the
voltage input which should be easily readable by a sound card. Amplitude
errors or uneven frequency response in the sound card won't matter much
now because all you register is a frequency. Of course you'd have to run a
good FFT on the computer but those program should be very easy to obtain.

This V/F scheme also alleviates many noise problems on the link between
measurement location and PC.

An FFT to get the frequency of a squarewave ?
What a waist.

Rene

Come on now, it's not nice to make fun of someone because of the weight problem.













Remove "HeadFromButt", before replying by email.

 

[Joerg]

Hi Jim,

Yes, if dynamic means frequencies above a few Hertz the FFT approach many not be the
ticket.

Since professional AD cards or USB converters are above budget here, maybe another
idea could work: How about modulating the strain gauge signal onto a 1KHz "carrier"
and then do a peak detect on the sound card data? That would be an AM protocol which
is subject to noise but it won't cost much and now it is more dynamic than an FFT.

Anyway, got to know how dynamic it has to be to figure out what could work. If it's
vibration analysis into the tens of KHz I believe a soundcard won't work here.

Regards, Joerg
http://www.analogconsultants.com

 

[Larry]

jfkunstmann@yahoo.com (jfkunstmann) wrote in message news:<9b7e0210.0405010729.17044185@posting.google.com>...

Can anyone help? I'm trying to take the dynamic signal from a single
strain gauge and input it into my computer so I can record and analyze
the data (maybe dump it into Excel). I can't affort to spend a lot on
data acquisition hardward, so I'm wondering if there is a way I could
do this for less than $200.

Greetings,

There are several low-cost A2D products, most using USB Try a google
search on USB data acquisition. One such is
http://www.labjack.com/downloads.html
(which I have not used) I see on their FAQ that have a
utility/demonstration app that dumps data into an excell file. I
don't know if this'd work for you, but it (a) sounds reasonably close
and (b) since this prod has interface software avail for multiple
languages (assuming the SW works) you'd have more flexibility. These
would give you multiple channels of analog input + misc other handy
I/O abilities. Another mfgr. is http://www.measurementcomputing.com/
they've got both USB and serial modules in your pricerange. I've used
boards from them, but not these modules. YMMV.

Since sound cards are not well spec'd re low-frequency response,
acquiring data via something that goes down to DC may save you time
and hair vs. the soundcard approach.

This doesn't yet address getting from a single gauge to a usable
signal. Some suggestions:
If you possibly can, mount a second gage -- ideally such that it
strains in the opposite sense for the loads of interest. Use these in
a half bridge config. (Even if you can only mount a second gage to
respond only via poisson coupling, or no strain at all, you'll at
least gain on thermal stability, which can be a big factor, even for
foil gages.) Whether you end up with 1 or 2 gages bonded, complete
the bridge with precision (1% of better), low tempco resistors,
located as close to the gage(s) as you can (to help with common-mode
noise). You can use higher value resistors (in parallel) to trim out
the passive bridge elements to better than their nominal resistor
tolerance.

Avoid connectors if you can (gold plated contacts if not, and
mechanically secure them!); and *shield* your wires. Make sure your
excitation is low-noise and can drive the bridge without sag. (Remote
sensing/control of the excitation would help, but may be more
complexity than you want to take on.)
Choose the excitation voltage so that the mounted gages do NOT feel
notably warm to the touch, even after say 10 minute warm-up. A linear
supply with remote sensing is a quick hack for consistent excitation,
if you have one handy at a voltage that doesn't heat your gages too
much.

I suggest using an instrumentation amplifier, preferably one with
single-resistor gain setting. FYI, digikey stocks several. I used
the Burr-brown (now part of TI) INA-2126 (a dual) in my last
strain-gage project. Several companies (incl Melexis, I think) are
offering single-chip solutions that include excitation and (I think)
A2D, These might be even easier, but I haven't yet tried them.

HTH,

Larry

PS I post via google, and I cannot figure out how to get google to
post my posts with an obfuscated address (along the lines of
realname_remove_this@somehost.com) Google seems to require a valid
address, which it makes visible.) How are you folks doing your
posting?

 

[The Real Andy]

On 3 May 2004 09:22:23 -0700, ursine@gmail.com (Larry) wrote:

%<

I suggest using an instrumentation amplifier, preferably one with
single-resistor gain setting. FYI, digikey stocks several. I used
the Burr-brown (now part of TI) INA-2126 (a dual) in my last
strain-gage project. Several companies (incl Melexis, I think) are
offering single-chip solutions that include excitation and (I think)
A2D, These might be even easier, but I haven't yet tried them.

Analog Devices, Crystal and others are making 24bit sigma delta AD
converters now that will interface directly to a load cell/strain
gauge.

HTH,

Larry

PS I post via google, and I cannot figure out how to get google to
post my posts with an obfuscated address (along the lines of
realname_remove_this@somehost.com) Google seems to require a valid
address, which it makes visible.) How are you folks doing your
posting?

Don use google or if you do, make up a bullshit hotmail/yahoo address
and then post you real, yet disguised email in the signature.