Accelerometers for >1000g measurements

Ceramic resonator maybe?

No. But there's a new kid on the accelerometer block.

The proximity fuse is a famous example of something ostensibly fragile that could be re-engineered to survive being fired out a gun.

I actually met - and worked for - William S. Butement, who had the original idea for that, probably because he didn't have enough sense to realise how impractical it was.

Attach an adhesive label printed with a black and white wedge to the rod. Illuminate with uniform light and view with a photodiode lens slit combination. Capture the velocity waveform on 'scope. Calibrate at low speed. I expect the acceleration peak might only last a few microseconds but could be many thousands of g.

John

No, just measure the voltage or the charge that it generates under acceleration. It's a polarized, piezoelectric ceramic thing for 30 cents.

I wonder how one would test it. At 2000 g's.

Even ceramic caps are notoriously microphonic. Maybe just look at your own power supply rail.

The guys who make expensive crystal oscillators go to great lengths to make them G-insensitive. I suspect that the folks who make cheap oscillators don't.

You could lay out a strain gauge on the PCB too.

During WWII, we fired proximity fuze shells out of rifled cannons at

It's not that hard to estimate. The hammer part is about 130 mm tall, and the speed of a compressive wave in steel is 6 km/s or so. Thus the impulse will last at least 22 us. If the hammer and the electronics assembly have about the same mass, and everything is perfectly elastic, in the collision the hammer will go from speed v to zero, and the electronics will go from zero to v (the same speed as the hammer).

If the hammer speed is 10 m/s at impact (probably an overestimate), the acceleration will be about

In real life the hammer won't be that fast, but the electronics assembly is heavier than the hammer and won't react as a unit. It'll be that order of magnitude in the limit of very stiff material. A compliant layer would reduce that, and so will the right choice of potting material, but the effects on the cables have to be considered.

Anecdotally it can take as much as 150 hammer blows to get the 60 mm tines fully inserted in the soil, so we're going to need this gizmo to survive at least 1E6 impulses over a 10-year life.

And then there's the ferrite-loaded shotgun balun to go from coax to the balanced line. ;)

Cheers

Phil Hobbs

They were very special small tubes and the whole circuit was encapsulated in wax.

John

Compare to a single expensive calibrated accelerometer? Not familiar with Measurement Specialties, other possibilities are:

I've measured big fast forces using strain gauges. If you can be sure the place where you place that on the metal rod stays elastic during the acceleration the bandwidth should not be a problem. Now, computing a calibration and amplifying that signal is the next problem...

Arie

Thanks, Glen, good leads.

The PCB Piezotronics folks (formerly Endevco) have the single-axis,

Cheers

Phil

Thanks. The issue there would be calibration, as you say, as well as the fact that it would measure the stretch at the surface and not the actual acceleration.

The objects in question don't have simple shapes, so going from the one to the other is a bit of a science project in itself.

If I can buy one with a cal sheet, that'll save me a lot of time.

Glen's other lead, Kunstler, advertises the model 8742A-50, also +-50k gees. I have a quote request out to them too.

Next question is how to make SMT PCB survive a million cycles of that. Those fuze guys had it easy. ;)

Cheers

Phil Hobbs

Do you only need one. or do you want to put one into every probe?

If it's one, buy something expensive and calibrated.