[sdiy] Moogey jitter

[Magnus Danielson]

From: harrybissell <harrybissell at prodigy.net>
Subject: Re: [sdiy] Moogey jitter
Date: Sun, 16 Apr 2006 20:32:30 -0400
Message-ID: <4442E21E.F170A05D at prodigy.net>

> I'd suggest that ~we~ trigger the scope with the actual reset pulse
> of the VCO. That should eliminate a lot of uncertainty.

That would be great for analyzing a single VCO, it would not help very much
when looking at it as it has passed through other circuits (such as VCF, VCA
and other "sources" of "warmth").

> Best thing to do from there would be to run a really really really
> fast counter, and store the count for every cycle.  

Which is what I have. For longer runs will not every cycle be stored, but my
counter counts how many cycles has passed as well as how much time. It stores
them as back to back measures so that the end of one cycle is the beginning of
the next. It can do this trick up to 10 MHz (13.33 MHz if you skip some of the
longterm data), so I am confident that it can handle my Mini and other synths.
My single-shot resolution is 200 ps (so each measured edge is measured with
that time resolution). Nice to have all that in one box with FFT, statistics
etc. The reference oscillator is a nice oven oscillator, which is heated all
the time. If that doesn't suffice I have more stable sources available, but I
doubt that any synth would outmatch the oven oscillator to start with. I have
one of these babies at work and one at home. A counter on steroids.

> Then there would be very little doube if the system was
> jittering.

Indeed.

> Another way would be a fast comparator set at 1/2 the ramp peak
> amplitude... but if the voltage was not stable it would be
> a bad measurement

Indeed. My game-plan is to measure square voltage or sawtooth voltage and use
the fast slope. Then I have come up with an arrangement of linear amplifiers,
a LM311 and then digital gates to shape up the slope of my trigger-point.
Basically a new trigger circuit adapted to low frequency (<100 kHz) which
creates high-slewrate output WHEN it flips. Very interesting little side-
problem. I never really had to face this problem before since I've been working
on so high frequencies that the main issue have been the trigger voltage.
There is always time to learn something new.

Cheers,
Magnus