(ultra)acoustic in circuit component level crack detector?

[rhb]

I just made a test of two xtals 5 mm apart immersed in cooking oil.  I learned two things from this:

the EM coupling is very strong

the SMPS noise from the Instek MSO2204EA is larger than any signal that might be present.

This and the strong 10 nS delay transmission mode signal  for two xtals glued to a 0.55 mm plastic disk suggest that the xtals are operating in shear mode.  That's not a viable mode for this application.

So I'm afraid that's the end of the experiments with the quartz crystals.

I whole heartedly agree with chris_leyson's comments.  I had no idea they made PZTs that extended up to even 20 MHz., much less 40 MHz.  But I don't want to ask the price.  I know from a friend's work building a one off lab instrument for measuring the propagation velocities in rock cores that devices made for the purpose are much too expensive to play with.

The mounting problem is a major obstacle.  I may pick up some cheap small PZTs and test them.  But the EM  coupling is a major problem.  I should have recognized that early on, but I was so amazed at getting a clean minimum phase under 10 nanosecond impulse response I wasn't thinking very carefully.

Edit:  Added a group photo of all the test specimens and the oil bath.

[coppercone2]

I am guessing the shield is going to be a mechanical problem to try to build? Also not sure how much I try the power system of a instek unit. are you running them with coax or twisted pair? I assume its mostly the capacitance of the plates though right?

and too weak to couple over something like a ceramic rod I am guessing..

[rhb]

The dominant signal is capacitive as there is no visible delay.  In oil a compressional wave should show up   1-4 microsecond after the pulse.   Water is ~1500 m/S and the plates are 5 mm apart.

The problem with the Instek is there is *no* shielding around the SMPS. 

What I really need is about 1" of 1/4" plastic rod so that the receiver arrival is well separated from the initial pulse.  I'm sure I've got something suitable, but can't think of where to look.  It as to be something I can glue to.  When I find something I'll uncan a couple more xtals just to measure the shear velocity.

The major downside to being in my line of work is that it is impossible to get access to data unless you are working for an oil company.  Which why my interest in being able to collect data myself.

Your question just prodded me to try some experiments.  Previously I did not have the test equipment.  The 33622A was expensive even used, but it is *really* nice.  It made setting up to test this very easy.

[coppercone2]

i only have a AWG2020 that works in this region, and some wavetek 50MHz AWG and 50MHz gigantic pulse generator with decent power specs. awg2040 would go higher and those look to be built decently but the PSU is not in a shield can either IIRC. 33622A looks real nice, but its also a car payment. I also think that SRS makes some kind of AWG that has a built in AA filter (the keysight 33622A spec sheet says most DDS don't have one). I actually got those function generators in the VHF range but I got bored with the VHF band due to lack of interesting applications, the only thing I wanted to do in that region is build the active circulator with fast op-amps.

the awg2020 is broken until I figure out which capacitors go where with a power supply schematic because I messed up during the repair documentation process

a nylon screw sounds like a candidate for a rod. you can also maybe widdle down one of those things that might be lying around for screwing a painting into a wall the expansion collet. or make a rod out of those auger things that also hold stuff on sheetrock, they are very stiff.

[rhb]

33622A looks real nice, but its also a car payment.

Actually more than the cash price of any of my cars except for my 1993 Toyota pickup which is the only new vehicle I bought for myself.

I'd had some doubts about getting that, but I was fed up with the F***Tech FY6600.  When I got that beautiful 5 nS pulse it made it all worth while.  I plan to get one of Leo Bodnar's pulsers modified to emit a 90-500 pS pulses at 10 million per second.  I've got to decide what pulse width I want.  I'd like 100 pS, but that would be rather attenuated by the limited BW of my gear.  So I may have to settle for 200 or 250 pS.

Nylon is going to be lossy, particularly shear mode propagation so I'm looking for polystyrene.  I know I have the plastic wand for adjusting window blinds somewhere.  I threw the blind away, but the plastic rod looked useful.  But a lifetime of that and it gets hard to find things.

Keep in mind that it will take an amplifier capable of generating 10V or more peak out up  to several GHz to make this crazy idea of yours work.  So you might want to start looking for a suitable MMIC and thinking about how to build the thing.  Board layout will be absolutely critical.

The concept may be simple, but actually making it work is wickedly difficult.  But you've given me an excuse to buy one of Leo's new toys.

[coppercone2]

i am not sure what material it is, but if you melt a cup down in acetone you can make a putty that will solidify and be pretty brittle. its some kind of styrene

with gloves you can kinda mold it too, i used to play with that when I was small

[rhb]

It's called polystyrene.

[coppercone2]

i don't think the grade you get from that is the same as a real polystyrene part, its pretty brittle. I thought about messing with it by reducing the pressure a bit but not too much or letting it evaporate in a dish slowly and dehydrating it a few times so it can maybe achieve a higher density but I never bothered doing serious experiments on it. and it might have a different composition and additives then a mechanical part because its from foam.

i don't want people getting a false impression that you can get great results doing that. its liquid string cheese

[coppercone2]

https://www.eevblog.com/forum/chat/sydney-tower-colllapse-fears/?topicseen

imagine a cheapie could check for structural problems or shifts, may be interesting even if you dont care about capacitors