[kb_bkk]

In real life, when a solid material (such as a piece of wood or plastic - not perfectly isotropic and homogeneous, but close) is impacted by another solid object at the surface (in a short time interval), how perfectly radial does the generated surface sound wave propagate outwards from the impact point (consider only the first 30 microseconds of propagation)? Can you please give me an idea how imperfect do you expect it to be? Does angle of impact influence how perfectly radial the wave propagates?

Thanks!

[ebaines]

Not knowing any specifics of how you're trying to model these two collidoing bodies -

A material like wood has a grain that structure that is very pronounced, so I would expect that the speed of sound along the grain is quite different from the speed of sound across the grain. This will tend to spread out the sound waves in a non-spherical way, and of course imperfections in the material will further jumble the waves. In 30 miliseconds the sound will have propogated a few inches at most (assuming the speed of sound is about 500 ft/sec), so as long as there are no boundaries or major imperfection within a few inches of the point of impact a spherical boundary wave is probably not too nbad an assumption.

Second - the point of impact for two objects colliding in the real world is not a point, but typically involves some bit of deformation in the bodies that tends to deform the bodies at the point of impact (otherwise the material would have to be able to handle an infinite amount of stress). This deformation would have a big effect on how the energy of the collision is coupled into the bodies, and would be very difficult to model dynamically. Think of the slow motion movies you see of car crash tests - pieces go flying every which way and even th solid pieces like the chassis frame rails of the car get deformed in ways that are very difficult to predict with any precision. So if the collision is at an angle, or is violent enough to cause significany deformation you will probably find that the sound waves are much more chaotic. The speed of the sound waves across the surface in different directions may not vary, but their magnitiude would.

[kb_bkk]

Not knowing any specifics of how you're trying to model these two collidoing bodies -

A material like wood has a grain that structure that is very pronounced, so I would expect that the speed of sound along the grain is quite different from the speed of sound across the grain. This will tend to spread out the sound waves in a non-spherical way, and of course imperfections in the material will further jumble the waves. In 30 miliseconds the sound will have propogated a few inches at most (assuming the speed of sound is about 500 ft/sec), so as long as there are no boundaries or major imperfection within a few inches of the point of impact a spherical boundary wave is probaly not too nbad an assumption.

Second - the point of impact for two objects colliding in the real world is not a point, but typically involves some bit of deformation in the bodies that tends to deform the bodies at the point of impact (otherwise the material would have to be able to handle an infinite amount of stress). This deformation would have a big effect on how the energy of the collision is coupled into the bodies, and would be very difficult to model dynamically. Think of the slow motion movies you see of car crash tests - pieces go flying every which way and even th solid pieces like the chassis frame rails of the car get deformed in ways that are very difficult to predict with any precision. So if the collision is at an angle, or is violent enough to cause significany deformation you will probably find that the sound waves are much more chaotic. The speed of the sound waves across the surface in different directions may not vary, but their magnitiude would.

Thank you for your reply. But as much as I know, maybe sound travels much faster in solid materials than 500ft/s? Sorry that it seems my question isn't sufficiently clear. Let's say, we have a fine quality, smooth wooden table - (ofcourse, all tables have a large flat surface - table as in dinner table). Please note that everything here is talking about surface of the table. There is an impact (very small area, almost a point) and suppose it travels, like you said, 500ft/s. I put a sensor (eg. An accelerometer) 0.25 feet in distance away from the point of impact. In theory, it should arrive in 0.0005 seconds or 0.5 milliseconds. However, in real life, would you expect an error as much as +/- 0.1 milliseconds (or less or more) due to grain irregularities and grain boundaries? Of course it is impossible to be exact, but any estimation (from accumulated experience and knowledge) is greatly welcomed.

*I thought the waves travel in wood at around 3000meters/second?

[ebaines]

Oops that's a typo on my part! I actually used an estimate of 5000 ft/sec in my calculation of how far the sound woud travel in 30 micro-seconds, but I mistyped the number in my response. I just guessed at this value - thinking 4 times the speeed of sound in air seemed about right - but if you have a source that says 3000 m/s then by all means use that.

As for your 0.1 microsecond error - not knowing your experiment set up I can't say why this is, but that's "only" a 20% error, and so could indeed be due to irregularities in the material, boundary conditions, or perhaps systemic error in your setup. I suggest placing your sound pickup device at various distances and see if you get a linear relationship in the time differences, as you woud expect.