Average ultrasonic waveforms

[downunder]

I am developing an application to process waveforms from a digitizer connected to an ultrasonic transducer. Presently I am playing with the ultrasonic starter kit example from the NI site. We will want to average say twenty waveforms and to prevent smearing the waveforms a consistent zero point for each waveform to be averaged will obviously be required.
Is it a valid approach to detect the frontwall for each scan using the threshold trigerred peak detection in the labview vi, then realign the zero point of each scan to a position a set few sample points before there, and then average the modified scan samples.

tia

Chris

[hbin]

You don't have a trigger signal for digitizing the waveform?
e.g. you can use the pulser (or PRF signal ) as a trigger, after each trigger, collect a waveform?  If so, you have a consistent zero point.

Also, if you don't have real-time requirement, you can use "averaging function" of the oscilloscope to digitize.

[denyart]

My guess is that an electrically timed zero point is not the problem. It is the fact that due to scanning and other factors, the reflection signal isn't always going to be at exactly the same point on the trace. It will be exactly the same in reference to something physical such as the front surface signal, but that same signal has some problems with using it for timing. If you can set the scope/digitizer to trigger on this reference and still average data "pre-trigger" it may work for you. I beleive that is what you are suggesting. I have seen this also done with 2 gates. One gate is set to capture the front surface signal, with a positive threshold trigger. The second gate triggers off of the first gate (with a negative delay time). This second gate's data can then be aquired and averaged, etc. With GaGe digitizers this method seems to work (although in other software, haven't tried it in Labview).

Message Edited by denyart on 04-03-2006 02:30 PM