I'm going to make a contact probe similar to what was used with the HP117 Ultrasonic Translator. It was basically a stainless steel pointy rod spring loaded against a barium titanate piezo crystal.
I've seen these mentioned by some forum members and I 'd like to hear any comments as to how well they work.
The description doesn't say anything about frequency response and just kind of hints about amplification. They are cheap though.
A typical ultrasonic translator used for leak detection and mechanical noise detection operates over a range of approximately 36-44kHz so that would be the design goal.
Before I invest in some more expensive, better documented sensors, I'm just wondering if anyone has any suggestions.
Since they talk about hooking it up to A0 on an Arduino, I guessed it might be a receiver. 2-4 kHz would not be good regardless.
That was an alternate thought and probably more likely to work. I'll have to take one apart and see if it looks like one can safely press a probe against the crystal.
Piezo crystals are reversible transducers, so they can be used either as transmitters or receivers. High voltages are generated by rather small vibrations, but the impedance is very high so the generated current is tiny.
Yes, I'm aware of that. The circuit board seems to have a chip of some kind on it that I wondered might be some sort of amplifier. Doesn't matter really. I think I might have an old ranging board around somewhere.
The wide frequency response that the HP units had may be hard to match. There are some units I found on Mouser that are sensitive at the extremes but pretty low.
I saw that, and suspect that the component is a Zener or TVS diode, possibly with the idea of limiting the generated voltage to a safe level for a digital input.
That makes sense.
Now I'm looking to fill in my knowledge gap on how to flatten a frequency response. How does one make an amplifier that amplifies low and high frequencies but not the ones in the middle?
An "inverse bandpass response" sounds like advanced digital filter territory.
It is easy enough to implement a low pass or high pass filter, so make two sensors, with independent high and low frequency amplifiers?
Note that if you attach one face of the crystal to the end of a rod, that would strongly shift the resonant frequency, and/or dampen the peak, but I don't know how to predict what would happen.
That could be feasible but more work than I was hoping for. Actually, that might make for an improved tool though if a way to mix the two channels was used.
I hadn't thought about that. I've repaired several of the old probes, back when you could still buy parts. The most common problem was a shattered crystal. The probe was spring loaded against the crystal but I have no idea how HP compensated for the effect. There was a preamplifier inside the tube for the probe. I'll have to see if the manuals had the schematic for that circuit. It may give some clues.
An "equalizer" like for stereo amps sounds like the tool for flattening the curve. There is likely to be an explanation about how that is done somewhere.
Later:
Huh! So it looks like you just mix the weighted outputs of a low pass, band pass, and high pass filter together which is the way I imagined it but somehow seem not quite it. There might be some interesting wiggles in the frequency response curve that can be "tuned" out.
