Computer program that shows p-p voltage?

--- Determine what voltage corresponds to 3dB by inputting a signal (within the bandwidth limitations of the card) which drives the indicator to 3dB. Measure that peak-to-peak voltage with an oscilloscope and write it down.

Next, increase that voltage by a factor of two and read the sound card display. If it indicates +6dB, then the display is calibrated to read

E2 dB = 10 log10 ---- E1

where E2 is the peak-to-peak voltage causing the +6dB reading and E1 is the voltage causing the 3dB reading.

However, if it indicates +9dB, then it's calibrated to read

E2 dB = 20 log10 ---- E1

where E2 is the peak-to-peak voltage causing the +9dB reading and E1 is the voltage causing the 3dB reading.

Knowing that, you ought to be able to figure out the peak-to-peak voltage causing the value of the display in dB.

-- John Fields

Wouldn't peak to peak be the greatest change in acceleration and not the greatest acceleration?

Regardless, are you *only* interested in the peak values or also in the instantaneous acceleration? If the former, then an appropriate op amp "peak hold" circuit might work, with the value then read using any inexpensive data interface; a multimeter with RS-232, or a small Dataq or Measurement Computing device.

If you need a higher sample rate then you're looking at more specialized equipment or "rolling your own" with a buffer amp and a microcontroller.

There's a nominal voltage of ~1.2 V, and the voltage varies about 23 g/V (.043 V/g). So, the peak to peak will give the difference between the two extreme accelerations (it's vibrated vertically). If the p-p is 1 volt, then it's experiencing ~11.5 g's each way, right? This is in the 20 - 2000 Hz range. I think I will try the calibration exercise suggested in the above post (putting in a known p-p, then doubling it and seeing what that does to the dB's). Will having a wave that isn't sinusoidal adversely affect this?

How would I design an op amp "peak hold" circuit? Am I correct in assuming this would simply hold the peak voltage, so I could measure it with a multimeter? I would then have to assume that the variance from the nominal voltage was symmetric both above and below that nominal voltage (I don't think it would be - it would be 1g higher on the positive side, right?). Is there a way to design a "trough hold" circuit that would hold the _lowest_ voltage? Thanks for the help!

Only if it's symmetrical. If it changed direction more slowly on one end of the travel than on the other it will still be moving back & forth but the acceleration wouldn't be the inverse.

Google for { "op amp" "peak hold" } for some ideas. You'll probably want some amplification in any case, so peak hold circuit followed by a modest amplification stage might work.

A multimeter would work to measure the result.

Basically, just flip the diode.