good way to couple contact acoustic measurement to microphone?

[CopperCone]

I made a probe before that was basically a tiny shielded microphone on a stick, but I am kinda thinking about field rejection and stuff. I used one that's buffered, and put filter caps and ferrite right against it (it is a SMD part), used double shielded cable, 4 wire twisted pair, copper tape and aluminum epoxy to seal it as much as possible, so only the opening of the microphone is not part of the shield (it is about 0.5mm or something).. but I am still concerned. The best solution seems to be to use a mechanical acoustic coupler. It's also not shielded against H-fields. I would need some kind of mu-metal tubes to deal with that I think.

Assuming I want to take actual measurements with instrumentation, whats a good way to measure mechanical waves on a circuit board?

By having the microphone near the part, like I did, I might be coupling fields into the coil so I am picking up electrical activity not real mechanical activity.

By using some kind of waveguide I am having some kind of bandwidth limiting and the microphone is further away from the part being analyzed, so I can possibly pickup other room noises better.I also expect to have some losses.

Whats a good method? What kind of tube should I use, dimensions, and how do I isolate it so its directional? I am completely lost in where to begin with this acoustic design.

[GeorgeOfTheJungle]

https://en.wikipedia.org/wiki/Electret_microphone ?

[CopperCone]

https://en.wikipedia.org/wiki/Electret_microphone ?

How do I bring a signal to the microphone, and impedance match the microphone to the coupler. I don't know what kind of microphones I have. I got a bunch of different ones I can use.

I wanted something I can probe individual parts with, including small ones like 0805.

So I can get a FFT that shows me what kind of vibrations are in the specific package on my board.

I just want to do it. There is no good reason to do it that I can think of yet.

To broaden the thread, I would also like to be able to measure ultrasonic vibrations.

[CopperCone]

Well I suppose I need to make a e-field capacitor parallel plate capacitor field thing (refereed to as e-field generator) to test my probes rejection and then see what a helmholtz coil does to it, but having some kind of understanding of how to use a tube could be useful especially if something is hard to probe or whatever, behind a heat sink, or I want to measure it with the chassis closed.

[Conrad Hoffman]

Well, a PI laser vibrometer is what you really need, but they're expensive. I'd use a small accelerometer, but only a few will get you up into the ultrasonics. I'd actually be surprised if you have ultrasonic vibrations going on in a PCB. Another completely overlooked way to measure small vibrations is with a phono cartridge. Gently rest the stylus on the PCB.

[CopperCone]

Do these options offer decent rejection from fields? Since the output is electronic and there might be high currents present on the PCB.

[CopperCone]

Can you put a little prism or mirror on a part so you can reflect a laser beam off of it to see vibrations? I assume that a laser vibrometer works kinda like a laser eavesdropper unit? Maybe something like that can be built for small scales with cheap parts? Maybe some kind of paint so it does not weigh it down? I mean you should not require too much right?

What if you lap the part flat and paint it with galium or something?

[Alex Eisenhut]

So get a contact microphone?

http://www.te.com/usa-en/product-CAT-PFS0013.html

[CopperCone]

what if i wanna use it in a place where the field strength exceeds its output and shielding? It would need to have a mixer inside of it.

I still wanna know about bringing sound to a transducer no matter what exists commercially.

[JS]

You might need to cancel external noise with some kind of humbacking, so you couple the vibration with two elements with opposed polarity and then add them with one reversed so the fields noise cancels out. HF cancellation gets trickier but you will be filtering anything over your band of interest.

Now, you could use piezoelectric, magnetic or capacitive mics for doing this. Even if the second element isn't picking up the vibration it will help as long as the predominant noise is from the fields and not from the element.

Maybe you could use a load cell, one end fixed to the peobe the other touching the DUT via some mechanical coupling,

The optical route has been mentioned, if you can couple a membrane you can shoot a laser at the membrane, use a photodiode to recieve and get away with it. I've put something like that some years ago, nothing fancy needed to do so. Also you could use the pads as a reflective area or even the body itself as you dont really need a lot of coupling, it's common to use a lock in amp with a laser in this applications.

JS

[LaserSteve]

take a look the Knowles BU family. Start with BU-21771-000

Steve

[CopperCone]

is that immune to getting nonsense from interference of the same frequency or similar frequency to the vibration though?

This thing might be located near MONSTEROUS fields.

[JS]

6mV/g sensitivity doesn't seem what you need, I don't know what you are getting but this seems pretty low sensitiivity. You could use two of them, in a small, maybe flex pcb, mount one on each side and do the humbucking trick, amplify both signas and make the difference, or even use a differential amplifier to star with and run from there. The 5k output impedance makes ir quite ideal to run with many opamps, 13μV noise could be a problem, but easy to archive what the sensor is meant for. PCB routing would be quite critical in your application.

JS

[Conrad Hoffman]

MONSTEROUS fields? OK, here's a sensor. You can buy it from Keyence or make one. Get a bundle of optical fibers. Polish one end. At the other end, randomly divide them into two groups. Send light down one group, LED or whatever. Put a photosensor on the other group. The thing will make a reflectivity sensor with a surprisingly long linear range. Put it near the board, probably 0.01" or so and with enough gain you should be able to pick up vibration signals to whatever bandwidth your receiver is capable of. Paint a white spot on the board for best sensitivity. The electronics can be a long distance away from whatever fields are present.

[CopperCone]

Would there be any benefits towards polishing the spot on the board or is the finish of a PCB good enough? Also any benefits towards using mirrored spray paint?

I love knowing stuff like this.

[Conrad Hoffman]

A diffuse white surface is probably best. A mirror gives alignment sensitivity. I've never seen "mirror" paint that actually works, though it might in this case.

[richnormand]

Assuming I want to take actual measurements with instrumentation, whats a good way to measure mechanical waves on a circuit board?


Whats a good method? What kind of tube should I use, dimensions, and how do I isolate it so its directional? I am completely lost in where to begin with this acoustic design.

In the past I was successful with an old piezo Astatic phono cartridge for playing records. I used a soft sponge with a slit to nessle the cartridge in it and leave the needle to contact the surface to be measured.
I was measuring valve latch noise on large engines so I just taped the whole assembly (with duct tape) on the engine block side. Routing the pickup coax near cylinder one also gave me an EM spike to get the proper timing on all the subsequent noises.
The audio recording was then played back while using my Tek 7D20 digitizer to get/transfer the waveform. This worked wonderfully to find which valve needed adjustment and also for rod bearing or piston slap identification..

Of course this approach assume you are in the audio range and dont care on response flatness too much.
That guizmo worked well for years to pick up surface vibrations. It was pretty immune to acoustic ambient noise (much better than a microphone) and pretty good for EMI if you were careful at routing or used a small 741 opamp right at the cartridge output to feed the coax in a low impedance setup

Probably can find it in the "archive" boxes in the basement if you need photos.

It could be easy to adapt to just contact the area or component of interest in your case.