Post by thread:[PIC]: Detection of peaks in a wave using a PIC

Hi all. Basically I am sampling in a waveform using an 8bit ADC. I have some memory to play around with etc so that is not important. What I need to do is detect the peaks in the waveform which I suppose is roughly sinusoidal, probably looks more like speech. I can probably figure out a way to do this but I can`t help thinking it`s probably been done before. Can anybody point me in the right direction. Cheers Andrew Seddon Struggling Student -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Andrew, > What I need to do is > detect the peaks in the waveform which I suppose is roughly sinusoidal, > probably looks more like speech. Approximately what frequency or range of frequencies are you expecting in your waveform? Do you want to detect the time/frequency of recurring peaks or do you want to detect the amplitude? Assuming amplitude, do you want to do peak detection in software (hard unless your waveform is slow) or hardware? Do you want to get data at every peak (probably noisy) or an average of several peaks? Do you plan on resetting your peak detector after every signal cycle, ADC sample, or do you plan on just trickling a little current off of your hold capacitor? Regards, Dave -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

At 02:42 PM 10/6/00 -0700, David Minkler wrote: >Andrew, > > > What I need to do is > > detect the peaks in the waveform which I suppose is roughly sinusoidal, > > probably looks more like speech. Speech is more sawtooth, and hardly a pure waveform. You can take the first derivative and get there, but beware of smaller peaks riding on bigger peaks. You need to define "peak" so as to get the ones you want. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.