Path: utzoo!attcan!uunet!lll-winken!ncis.llnl.gov!helios.ee.lbl.gov!pasteur!ucbvax!hplabs!hp-pcd!hplsla!tomb
From: tomb@hplsla.HP.COM (Tom Bruhns)
Newsgroups: comp.sys.amiga
Subject: Re: Amiga Oscilloscope
Message-ID: <5160023@hplsla.HP.COM>
Date: 27 Jan 89 16:35:45 GMT
References: <8914@bloom-beacon.MIT.EDU>
Organization: HP Lake Stevens, WA
Lines: 32

>Just out of curiosity, does anyone out there know how to make a reasonably
>fast computer 'oscilloscope'?  Even with an immensely fast DAC, I wonder what
>a 7Mhz computer can do as for as processing nanosecond data, which would
>be necessary to debug something like, say, a 7MHz computer...  Can even the
>mighty blitter sling around data that fast?
>
>						-- Jim
>----------
Well, you DON'T use the computer memory to directly put the samples in!
Since you typically don't care about gobs and gobs of memory, you do it with
the ADC feeding a little (1k? 2k? 32k? pick your version of little) FAST
memory, then load that memory into the Amiga memory by your favorite
method (make it dual-port for fastest; hose it over serial for slowest;
or anything inbetween).  If you have some fancy trigger logic, you can
capture just what you want to look at in a relatively small memory.

Note also that most very fast digitizing scopes require a repetitive
waveform to operate at the fastest rate: they only get a "few" samples
on each repetitive cycle, but place the samples differently relative
to the start of each cycle, to build a good representation of the whole
thing.  Lots of tricks are played to smooth out digitizing noise, etc.

Bottom line is it isn't conceptually very hard to do what you want, but
will take some fairly expensive parts if you want to get real-time samples
at 8 or more bits at 20 MHz or more.

The computer and its graphics are a big help in the analysis and display
end of things; given the samples, you can do lots of signal processing
(FFT's, etc; you can even listen to the samples, played back slowly!)

Tom Bruhns
tomb%hplsla@hplabs.hp.com