question from Jon Giffen

A more informative title would be helpful, like "speed of a bullet?".

Your best bet would be to do some research -- go looking at reviews for representative guns, and find out what their muzzle velocities are.

Or, just design for anything that falls between the given 4000fps max and whatever is a sensible lower bound for what you're trying to do.

Only a VERY high speed bullet would be that fast. Pistol bullets may even be subsonic. Many are in the 1M to 1.5M range.

Bullet muzzle velocity varies greatly with the type of gun. Assume about 1200 FPS for a handgun, twice that for a long gun, but there are many exceptions, so plan accordingly.

There are no hard and fast constants in what you are doing, and it is very possible for the sound to take longer to travel to one microphone than to another one.

Now go work the math out again. The position in three dimensions will require FOUR microphones, because the shock wave is a sphere, and if all three microphones register the same time-of-arrival, there is no way to determine the position. It might be different if you could determine the local direction of the wave as well as its arrival time, but a simple microphone won't do that.

Seismology does this kind of thing all the time, and it requires careful timing of the multiple sensors' outputs; most USB links are terrible for timing, so your requirement will have to include synchronized clocks or timestamped ADC readings.

or 4 inputs to a single USB device