From: John Devereux
Subject: Re: Pulse Height Measurement
Date: Tue, 21 Jan 2003 18:14:54 +0000
References: <firstname.lastname@example.org> <email@example.com> <firstname.lastname@example.org> <email@example.com>
X-Newsreader: Forte Agent 1.8/32.548
NNTP-Posting-Date: 21 Jan 2003 18:15:00 GMT
Spehro Pefhany wrote:
>On Tue, 21 Jan 2003 17:06:25 +0000, the renowned Don Pearce
>>Are you really trying to measure the peak - even if it occupies just
>>one sample? If not, perhaps you would do better with a sort of rolling
>>average that smooths off some of the noise. This would actually be a
>>lowpass rather than highpass function. But it would give you smoother
>>and more reliable data to measure.
Yes, I think I do want to measure the peak (the signal has
already been low pass filtered to some extent, to clean it up a
bit). The rolling average sounds good, but I only know how to do
that in software, so would have to make to with a bit of low pass
>Seems to me that he wants to measure the difference between the
>signal and the signal passed through a LPF, assert an interrupt on
>differences that exceed a +/- threshold from that. Pretty
>straightforward, a difference amp or in-amp and 2 comparators, plus a
>DAC and gain-of-minus-one amplifier to get -DAC.
>He wants to detect the peak(ish) bits of +/- signal, so that's the
>only tricky bit. Perhaps a pair of the classic precision rectifier
>circuits with caps on the outputs. Then have the micro check
>that the pulse isn't there anymore (if necessary) and reset the
>caps (TGs across the diodes) after grabbing the reading, then
>re-enable interrupts and go back to whatever it was up to.
>Seems easy enough. Have to keep the sample cap reasonably
>small for it to be able to follow fast enough, maybe <= 0.01uF.
Yes, something like that sounds right.