[time-nuts] TPLL secret reveled
sar10538 at gmail.com
Thu Jun 10 07:10:44 UTC 2010
I think I have already shown my position on this whole matter from my
previous posts and understand what Warren is alluding to but I'm just
trying to assist him make his case in a way that is acceptable to
other members on this list.
On 10 June 2010 17:35, WB6BNQ <wb6bnq at cox.net> wrote:
> I have been following this thread and I think I can answer your question about the
> 1^-15 thing. What Warren is stating is what he feels the base resolution is, that
> being 1^-15 OR whatever your reference oscillator is worth.
Well, taking the ref osc out of the equation, the 1^-15 seems to be an
arbitrary floor if you assume that other items in the implementation
are perfect and it's hard to see where this number comes from and it
would have no floor in that situation. Obviously, limits in the
measuring equipment and tolerances in the components used in the
implementation will limit the accuracy of the system but that would
have to be calculated to obtain a true measure of the limits of this.
> The key phrase is "OR. . . WORTH." Obviously if you had a high quality H-maser or
> perhaps a Cesium Fountain then you may be able to reach his claimed 1^-15.
> Otherwise you are stuck with what you got on hand.
Firstly, the design of the TPLL requires that the reference oscillator
is able to be controlled such that it tightly matches the frequency of
the unknown oscillator. This will exclude most forms of high quality
oscillators that you have mentioned. In fact, those high quality
oscillators only show higher stability for long Tau and for the
current range of Tau that this implementation has been tested, the
humble, or perhaps not so humble, xtal is the best choice. That being
said, a stability in the region of 1^-15 is not to be seen on an
undisciplined xtal. So why can we not use a disciplined xtal, say,
GPSDO? Well, the operation of the TPLL requires that the xtal is
directly controlled by the PLL mixer output filter directly and that
precludes a GPSDO being also able to control the xtal frequency, IE.
the two are mutually exclusive.
What I have tried to say in my posting about this base accuracy is
that let's not get too stuck on this number because more harm than
good can come from this being taken literally by some members on this
list and I felt that this claim would further add fuel to the fire.
> I agree with you on wanting to see an ACTUAL schematic detailed with real part
> numbers. Better detail on the software used would be nice as well.
> As far as his on-going argument with Bruce, well, at times it gets quite hard to
> follow. Particularly with Bruce refusing to insert BLANK lines above and below
> his responses. I think Warren is guilty of that too.
It seems to be a breakdown in communication between those parties and
that is why other have been trying to resolve these matters, both
on-line and via PM, with both parties.
> I can see where Bruce was coming from in the beginning because Warren was
> definitely being obtuse with his explanations. All of that just makes it harder
> for us less informed to follow along.
As are Bruce's cryptic comments which are frequently hard to
understand unless you have his mind-set. Both parties have become
entrenched in this matter and I feel that nothing can move forward
between them except through intermediaries. Both have taken an abusive
manner between each other and even others which have tried to assist
in the matter and this has really not helped things.
> Bruce, I know you are striving for the best in your efforts. BUT, if Warren's
> method provides a "GOOD ENOUGH" approximation for "us" less endowed, then perhaps
> it has a place with stated caveats ?
Bruce is, and always has been, a perfectionist, but that is not a bad
thing. This is a time-nuts group, after all, where the pursuit of
perfection is paramount and it is to that goal we should strive for.
That being said, there is a place for lesser mortals to enjoy and
expand their knowledge in this fascinating area without breaking the
bank or being looked down upon by some. We are ALL here to learn,
every single one of us, as if anyone thinks they know everything there
is to know about this subject, a) they are miss-informed, and b) whay
are they here.
As to this method producing "GOOD ENOUGH" approximations, this has a
very much a relative meaning to many people and these results in
themselves will also have degrees of relevant meaning to different
people. There are some limitations to the range that this "tester" can
work over but many will be surprised to find that the accuracy over
that range is al lot more than "GOOD ENOUGH" for their needs.
Examination of the graphs that John posted shows a close correlation
over quite a range between the two "testers" and outside this range,
say, at the top end, drift in the reference oscillator pays it's toll
in the results. If you have a reference osc xtal with very low noise
and very low drift, you will get very good results, even over an
extended range. It's a simple limitation that can be easily
> Hopefully, all this will get straighten out and we will all benefit from it.
This is exactly what I hope will happen soon as there is a lot to be
gained from this. A lot of us have been led to believe that this is a
very esoteric discipline and is impossible for us to take part in.
This current discussion only goes to show that this is not the case
even though some would seemingly cling to the idea that it is.
73 de Steve - ZL3TUV & G8KVD
Steve Rooke - ZL3TUV & G8KVD
The only reason for time is so that everything doesn't happen at once.
More information about the time-nuts