I don't thinks so, since the failure occurs due to mechanical resonance of the component and at resonance the peak is probably 100 timer larger than t he exicitation. Also I do not think the resonance is at 50kHz, more likely at 1kHz so the 10um requirement should be ok.
Cheers
Klaus
What if it's a resonant system? That's the problem with PCBs, resonances all over the place.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
On Fri, 10 Oct 2014 10:37:49 -0700, John Larkin Gave us:
The Redstone and Apollo space programs proved that. The only thing changed is the substrate materials and binders, etc.
Circuit boards have come a long way since we used to be worried about via count, etc.
Notice how nobody worries about via count anymore. PCB Mfg technology has come so far, that we no longer worry about inter-layer interconnection issues, and their occurrence is rare.
Otherwise, we would not be able to have all these tiny micro-vias by the hundreds on our boards.
I really doubt that anything on a PCB has a mechanical Q of 100. If it does, just put an acoustic transducer someplace convenient, lay the board on something solid like a concrete floor, and hit it with a very small hammer, such as a 1 cm ball bearing. It'll ring like a bell.
The duration of the acoustic pulse is about D/c, where D is the diameter and c is the speed of sound, so for a 10 us pulse, you want a hammer of diameter less than about (3 km/s)*10 us = 3 cm.
Cheers
Phil Hobbs
Huh? units of P are kg m^2 / s^3; the 'mean' energy in a vibration is prop. to ks^2 i.e. M omega^2 s^2, so depending how you define power, it could be M omega^3 s^2. Am I misunderstanding something?
How much precision do you need (e.g. NASA grade)?
How much do you have to spend?
If you're thinking quick and dirty, I like whit3rd's mention of Chladni.
Consider mounting your DUT to a large speaker and sweeping the speaker through the audio range. Scatter talc over the board (less damage potential than sand) and take fast-as-possible video.
Watch the talc dance.
Mark L. Fergerson
Roight, force, thanks. Big number at high frequency though.
Cheers
Phil Hobbs
Could you please provide little more details of what exactly you are trying to achieve? Is the goal to measure the vibration frequencies of the PCB ? If so, how are the PCBs being excited ? Tears ago, while working at Lawrence Berkeley Lab, our team created almost from scratch, equipment to measure the vibration frequencies of very fine (human hair thickness) copper wires used in detectors for nuclear physics experiments. The scheme was simple. With the wire fixed to a rigid frame, shine a red laser at the wire. The reflected light would be detected by a sensitive photo transistor. To measure the vibration frequencies, an electronically controlled nozzle would blow a small puff of air at the wire, co make it vibrate. The light reflected into the photo transistor would be modulated by the vibration, and this information was used to measure the vibration frequencies. We built the who set up from scratch, and used a PC with an early version of Ubuntu to compute the Fourier transform and generate the spectrum. It was one of the greatest fun projects I have worked on.
onents flying all over the place during vibration tetsting
uess better than 10um resolution.
nting a MEMS sensor)
a XY table (looking down on the product), to measure the resonance over th e entire PCB
ture, so this can be challenging
hickness) copper wires
sure the vibration frequencies, an electronically controlled nozzle would b low a small puff of air at the wire, co
rate the spectrum. It
Sounds like an interesting project
The actual problem is a PCB that is mounted on stand-offs in the product.
The PCB is a standard product used for various different products and the P CB has a pot core inductor mounted in a corner.
However, in our product, the PCB is not supported by a stand-off in the cor ner of the location of the inductor and the inductor comes off during vibra tion tests (random 40G)
So we want to quantify the effect of adding a stand-off in the corner close to the inductor. Normally one would just do a test to see what happens and that's what we are doing shortly. I would just like to be able to deliver more data than just a statement of possible larger immunity to vibration, b ut instead I would like to quantify the vibration exitation and the vibrati on level on the inductor
That's the reason for wanting a displacement sensor, to measure the vibrati on level near/on the inductor with/without stand-off and other variations w e would like to try
Regards
Klaus
Check this out on page 3, which is vibration test of a PCB of less size than I am working on:
To me it looks like a Q around 100. It has Z axis displacement up to 250um..
Cheers
Klaus
PCB has a pot core inductor mounted in a corner.
orner of the location of the inductor and the inductor comes off during vib ration tests (random 40G)
se to the inductor. Normally one would just do a test to see what happens a nd that's what we are doing shortly. I would just like to be able to delive r more data than just a statement of possible larger immunity to vibration, but instead I would like to quantify the vibration exitation and the vibra tion level on the inductor
tion level near/on the inductor with/without stand-off and other variations we would like to try
How about this, Take the top off the pot core (to let the flux out). Place another coil above the pot core. Run a DC current through the pot core, Sense induced current in other inductor. Amplitude should be proportional to amplitude of motion.. at least until th e pot core flies off :^)
George H.
It makes perfect sense validating your FEM model with quantitative displacement measurements at key points. Two methods you might want to consider are the blurred fiducial method and equivalent time sampling, where in both cases the spot where displacement is to be measured is observed with a calibrated microscope, either using video or by eye with a microscope reticle.
For the blur method you need to have a very high contrast dot on a large enough uniform background that a streak can be seen by eye or with an exposure long compared to one cycle. Alternatively a line could be used as the fiducial for a single axis measurement, and some existing feature on the board might be usable as long as you can see the streak length. I have only seen this method used to measure the displacement of smoke particles due to sound in air, but see no reason why it would not work on a PCB.
For the equivalent time sampling method you undersample with a fast strobe light source (or gated laser or whatever) reconstructing the displacement waveform from your video frame samples in the usual way.
For a quick visual check of peak to peak amplitude only with constant excitation, trigger your strobe from the shaker excitation signal, divided down to suit the rep rate of your light source if necessary (all strobe systems designed for motion stopping I have seen simply skip all trigger pulses until ready to fire again, so dividing down is not necessary), with a manual phase adjustment so you can adjust the phase to see both ends of the fiducial displacement. Last time I did this I had to dial the phase through more than 180 degrees to be sure I had both extremes, but if doing it again today I would make every other strobe pulse 180 degrees of excitation signal from the previous so I could see both ends of displacement at the same time and verify the peaks with a small shift in strobe phase.
Regards,
Glen
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.