[time-nuts] BBB DDMTD - analysis

[Chris Caudle]

> From: Simon Marsh <subscriptions at burble.com>
> Subject: Re: [time-nuts] BBB DDMTD - analysis
> Message-ID: <54524F04.7060906 at burble.com>

> The first few pics (DFlop-unsync-floating-*) show
> the Q output, which was unconnected to anything
> other than the oscilloscope. They show a few glitches
> at the edge then a load of funky stuff going on later.
> Connecting Q to an input pin (DFlop-unsync-connected)
> cleared the problem suggesting that in this period, the DFlop
> output wasn't being driven and Q had been left floating.


I find that description really confusing.  Can you draw up a schematic?

If you connected the Q output to an input later, what had been driving
that input previously?  Was the input driven by an oscillator, but you
disconnected the oscillator and connected the output to input instead?  Or
the input was floating prior to connecting the output to the input?

> the DFlop output wasn't being driven and
> Q had been left floating

If you have a standard 74AC74 the output never floats, it is always driven
as far as I know.
Even if the output had been floating, if it was truly not driven you would
have seen it as low due to the input impedance of the scope pulling down
the output.
On the other hand, if you leave a CMOS input floating you get all kinds of
weird behavior on the output, every little disturbance can change the
input charge enough to toggle the output.  Are you sure you weren't just
seeing oscillation because the input was floating?

Are CLR_ and PRE_ pulled up always?  Are the input and clk pins always
driven by the two oscillators?
Have you looked at the two oscillator outputs at the beginning of each run
to see how close the edges are to each other?  The timing relationship
will of course drift (unless the two oscillators get injection locked to
each other), but that will at least give you an idea of how likely you are
to hit setup and hold problems at the beginning.

You mentioned 17us periods in the glitching, have you checked the
frequency offset between your two oscillators?  You aren't just seeing the
beat frequency? 17us is just a hair under 59kHz, that seems really far off
for a decent OCXO, just wondered if you had ruled it out.

-- 
Chris Caudle