vibration sensor

Microphone?

Good Luck! Rich

In what manner is a $2 accellerometer expensive?

$2 isnt expensive at all, what did you see and where at that price? the ones in farnell are more like £30

Ive probably got a microphone I can try somewhere, but then the piezo acts like a microphone allbeit a very resonant one, I think the moving element needs a reasonable amount of mass, or it might be more sensitive to the noise as sound than the vibration I need.

thanks Colin =^.^=

Surely they can be had even cheaper than that. My Radio Shack "universal" remote lights up at the slightest movement... so slight I can't trick it ;-)

...Jim Thompson

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

lol :)

btw this was the one that was discontinued

Colin =^.^=

The extra 49,999 pieces you have to buy to get that price? ;-)

Best regards, Spehro Pefhany

of

de

no

What you need is an inertia activated floating capacitor. It can be built with silicon on insulator on 6 inch wafer using 1 micron process. Projected cost is approximately $2000 for 1 or for 1000. Want to spit a wafer?

well ive just come accros ADXL311 at £9 is not too bad, also it will allow me to precisly level the thing too, resolves to 0.1deg. two birds with one stone.

hmm i dont like those legless chips however, oh well.

Colin =^.^=

Check out the 3-axis accelerometers as well, some of them are not too expensive.

Best regards, Spehro Pefhany

Do a search on DigiKey for MSP1007-ND. Maybe that will get you started.

Good luck.

John

The possible meanings of 'vibration' include all three axes as well as multiple torsional modes. Probably you have some particular load-bearing forces in mind, and a single axis force sensor (like a load cell) under your rotating machinery might be enough.

So that's my input: use a load cell. It can be as easy as cementing a strain gage to one of your struts. Or you can get calibrated load cells with preamps. Make a variable-frequency (digital pulses) output, and feed to a digital processor for detection.

The important frequencies are going to be multiples of your rotation, and lots of averaging schemes can reject out-of-band information while building up a solid number from noisy input. Look into phase- locking detectors (not a phase-lock amp, which is a tricky analog gizmo, just a phase-lock-loop with multiple counters).

So, in phase 0-90 degrees of frequency A you increment counter-1; in 90-180 degrees you increment counter-2; in 180-270 degrees you decrement counter-1; in 270-360 degrees you decrement counter-2. With the force sensor running into a voltage-controlled-oscillator, you can average over hours of operation and get a good measure of even a small signal. Counter-1 and counter-2 add in quadrature to give a vibration amplitude, and the relative values and signs will give a phase...

It sounds harder than it is. There's a nifty seismology gizmo called a hammer seismometer; you pound with a hammer, and detect each blow with a microswitch on the hammer handle. A digital accumulator builds up a synchronized composite signal from a seismometer placed nearby, and after a few dozen strokes the summed signal isn't noisy any more.

I have a vibration meter specifically designed for motor maintenance. It is practically brand new with less than 30 minutes total use. Still in box, w/pkg, etc.. Was several hundred dollars (US), & I would part with it for a lot less. Model: Entek IRD, model # 808 Trobleshooter. Email me offline if interested and we'll talk. -mpm

thanks, ive seen a reasonable signal from a piezo tranbsducer, but only when i held onto a metal slug glued to the center, the slug alone was not enuogh.

The signal wil be fed to the dsPIC wich drives the BLDC motor, so the signal can be corelated to the motor phase. probably average the signal and display it on the pc, or do an fft.

I think i just need the fundamental, particularly the phase relationship, so i can balance the shaft.

I think out of balance forces should be detected by 1 axis on the bearing housing, of course il have to balance the shaft in several places so il move it from one houising to the next.

Colin =^.^=

a phonograph pickup ? (used is much cheaper than new)

Bye. Jasen

Circuir Cellar, last issue, last page, see advert for Parallax IO devices, note one of them is a vibration sensor.

SioL

Piezo transducers need physical contacts for detection.

Floating capacitor does not. It can be totally isolated. Most accelerometers are based on floating capacitors.

You can mount several bare sensors on the motor, and plot the inertial vector. Depends on how many you need. $2 to $3 each is doable.

well ive put one of those piezo transducers with mounting lugs and mounted it onto the bottom of the rubber mounted chasis then a screw from the non isolated chasis section is adjusted to just make contact with the brass disc inside the transducer via the top hole. works quite well with a 2stage 200hz rc filter. get realy strong 83hz signal at 5krpm.

mechanical resonance is at about 5300 rpm, and the phase of the vibration varies about 180' either side of the peak.

all ive got to do now is work out how that relates to where ive got to put some weights on the shaft, any ideas ?

I have a strobe wich will flash at the peak, gona try fixing some small weights with pvc tape on the shaft and see how it affects the position.

its 1.2M long and I have a sensor at each end. interestingly it seems to differ depending on wich direction the shaft is rotatiing. i need to be able to reverse direction and still be in balance. I have a motor in the middle and an optical disc at each end all of wich probaly contribute to inbalance. The shaft itself is quite straight as much as I can tell with a 0.01mm dial guage, although at 12mm dia its quite heavy.

thanks Colin =^.^=