Fwd: Re: Frequency Stability of Trimble Mini-T
----- Forwarded message from hasweb@has.org.nz -----
Date: Fri, 17 Oct 2008 12:02:48 +1300
From: hasweb@has.org.nz
Reply-To: hasweb@has.org.nz
Subject: Re: [time-nuts] Frequency Stability of Trimble Mini-T
To: Mike Monett XDE-L2G3@myamail.com
Quoting Mike Monett XDE-L2G3@myamail.com:
"Lux, James P" james.p.lux@jpl.nasa.gov wrote:
[...]
Even without TMR or other similar schemes, the probability of
upset IS pretty low. However, as Black or Scholes said (I can't
remember which), "One should not confuse very low probability with
impossible". If it absolutely, positively can't take any hit, then
some more work is involved.
James Lux, P.E.
How do you do that? Any web links to study?
As far as I know, it is impossible to absolutely guarantee against
metastability. Do you wait a week for the metastability to settle?
If zero probability of failure is so important, you would also have
to include the probability of a solder joint opening, or a chip
failing due to metal migration or latent ESD damage. That is never
zero.
Of course, after the system is perfect, someone will take it and put
it on a destroyer running Windows:)
Best Regards,
Mike Monett
Mike
You could always look at: United States Patent US4820939.
But dont bet on it.
Bruce
----- End forwarded message -----
hasweb@has.org.nz wrote:
Mike
You could always look at: United States Patent US4820939.
But dont bet on it.
Bruce
There's lots of them. I don't believe them.
Philips even has one on their 74F50729: US5,789,945. I have not
tried it but some reports say it doesn't work.
OTOH, I used the digital mixer described previously to lock as close
to the metastability condition of a flip-flop as possible for my
second patent.
The circuit tried to lock on logic level 0.5, but of course could
never accomplish this task. But this tied the edge of a delay to a
reference so I could measure the offset and window width in the data
separator with a TIA.
The circuit used Motorola MECL 10KH ECL. I could see the attached
waveforms on my Tek 7104 fairly often. It was barely possible to
measure the jitter between the clock and d inputs with a HP 5370
TIA. I don't remember the value but it was quite small.
Later I upgraded to Motorola EclInPs. Using the same circuit, and
others like it, I could no longer get any metastability condition on
a MC100EP52 D flop no matter how hard I tried.
So the faster chips really are difficult to get into a metastability
condition. But I continue to use cascaded shift registers to keep
them at bay.
Best Regards,
Mike Monett
You could always look at: United States Patent US4820939.
But dont bet on it.
I'd go so far as to suggest betting against it.
Back in the 80s when word about metastability was reaching trade rags and
real designers, I went to a trade show that had a panel on it. One guy got
it totally wrong. 1 or 2 mumbled and didn't contribute much. The only guy
who got it right was John Wakerly from Stanford. (Of course, maybe I drunk
his cool-aid and not theirs.)
Back in those days, there was a lot of activity in designing kludgey circuits
to "fix" metastability. I could usually find the flaw. It got boring after
a while. The classic was a circuit to detect metastability and reset the FF.
That reset signal would sometimes have runt pulses.
There are several problems with such kludges. First is that they don't work.
Second is that they usually make things worse by adding logic in the
critical path thus reducing the settling time. On top of that, they are
usually a pain to analyze. (If the designer analyzed them correctly, he
would have thrown it out.)
It's much better to keep the circuit simple. Then you have a chance of
correctly estimating the MTTF.
The key is settling time.
The best analog of metastability that I know of is rolling a ball over a
speed bump. If the ball is slow, it bounces back. If the ball is fast, it
goes up and over. If the speed is just right, the ball gets to the top of
the bump and stalls. If you have a frictionless bump and such, the ball will
eventually fall off one side or the other. (You can't adjust the speed to
the exact value.)
Johnson and Graham's book on Black Magic has a good section on metastability.
I think there are some good scope pictures, but my copy isn't handy.
I used to use metastability as a calibration on logic design books. If I
couldn't find metastability or synchronizer in the index I didn't expect
much. If I did find something, I would scan that section. A lot of them
weren't very good.
--
These are my opinions, not necessarily my employer's. I hate spam.