From: "Bill Sloman"
Subject: Re: Pulse Height Measurement
Date: Wed, 22 Jan 2003 01:31:16 +0100
Organization: Planet Internet
NNTP-Posting-Date: 22 Jan 2003 00:31:33 GMT
X-Newsreader: Microsoft Outlook Express 6.00.2800.1106
"John Devereux" wrote in message
> Greetings to all you fine SED folk.
> I am looking or ideas in the area of pulse height measurement. We
> have an instrument which produces a noisy background signal (1/f
> noise), with pulses superimposed on it whose "height" we wish to
> ____________________ ___
> ' ' ^
> ' ' ' ' |
> ' ' ' ` Pulse Height
> -- ' ' `` ` - |
> ` ' ' ` v
> `-- _________________________ __
> The pulses can be from a few mV to a few V, and arrive randomly.
> The duration is ~10-100 microseconds, (but faster would be nice).
Try a search on "constant fraction discriminators", which should clue you
into the physics literature.
Sergio Cova of the Milan Polytechnic has published some nice stuff over the
years, in Review of Scientific Instruments and Applied Optics (in English).
Most of the work is on pulses coming out of fast focussed photo-multipliers,
with a full-width at half height of about a nanoseond or so, but the ideas
should translate down to your relatively slow pulses.
Are your pulse widths variable? Which is to say, is the full width at half
maximum height largely independent of amplitude?
If you can set up a constant fraction discriminator, you should be able
sample the peak of the pulse with a fast A/D converter - you may need to
delay the pulse in an analogue delay line to accomodate propagation delays
(which is a trick Sergio Cova has used to advantage).
You can , of course, use a multi-tapped analog delay line and a fast summing
amplifier or two to implement an analog finite impulse response filter (the
classic digital filter) which is a fun trick for blowing minds at
For extra credit, make the filter output insensitive to the DC level of the
IIRR, John Larkin knows about this sort of stuff.
Bill Sloman, Nijmegen