[hpsdr] High Accuracy Timing for HPSDR

[Robert McGwier]

Alex wrote:
> GIBRALTAR is the perfect name for the board. I agree that we are on the verge of a design.
>
> Here's a potential issue
>
> Steve, your design has a header for the SiRF daughterboard?  Wouldn't we have clearance problems with the card spacing on ATLAS if GIBRALTAR is installed in anything other than slot 6 (opposite end from power connector)? The answer might simply be only install GIBRALTAR in slot 6.
>
> Bob, What GPS engine is AMSAT using? Are you using the SiRF units?\
>   

This is really an aggravating question.  I will bore you with some 
history.  Tom Clark (K3IO) and Rick Hambly (W2GPS) helped Sirf produce 
the worlds best timing receiver.  They did it openly and the end result 
was that someone bought Sirf in my OPINION to squash that effort.  What 
these two friends wanted was to be able to buy the receiver they helped 
produce.  There are numerous papers where you can see the implications 
of this at www.gpstime.com.  Tom is the 2nd president of AMSAT and its 
current president emeritus and "chief scientist".  Rick is the 
president.    I HOPE Sirf does come out with the receiver and that it 
incorporates good stuff.   Tom came up with a neat idea and Rick 
implemented it in his CNS clock II.  ALL sawtooth error,  with its "mean 
zero but large local deviations" problems is removed in an approach done 
in this clock down to truly small variance "nearly gaussian" noise.  CNS 
produces the CNS-II and in my opinion, is the single best clock of its 
type in existence and it is under $3,000 if you buy every option the 
table.   It has a 10 MHz output that with a FIVE DOLLAR TCXO produces 
one part in 10^12 accuracy!  With the one hundred dollar TCXO,  we have 
not found the bottom yet.   I helped design those tracking algorithms 
because I need the clock professionally.  Now let's see,  the president 
emeritus, the president and the VP engineering all are involved in using 
a particular approach to taking GPS and making superaccurate clocks and 
references at "bargain prices" (compared to the alternatives).  ;-). I 
suspect that we will in the end, roll our own and use Motorola or Sirf 
or whatever receiver allows the best final results which requires: 
publication of the sawtooth error, publication of oscillator frequency 
and phase numbers to allow DGPS,  and modification of the issuance of 
the 1 pps to be emitted a fixed and known amount of time early.  The 
Motorola receiver does that.  I like the cyclone II but on the timing 
board, based on Steve's current implementation of Luis work,  the Max II 
is cheaper and is not nearly full.  We don't need the hardware 
multipliers that are on the Cyclone II.  Frankly,  my opinion is that 
the basic design of Reflock,  Reflock II, and Gibraltar as a variation 
such not break what works unless someone can think of a real reason to 
"fix what ain't broken".

> Would it make sense to use the Altera Cyclone II instead of the Altera MAX II for commonality? Would be easier to purchase larger quantities of parts...
>
> My offer still stands to be persuaded into ponying the money up for a couple alpha boards.
>
> Alex, N3NP
>
>   


-- 
AMSAT VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats,
NJQRP/AMQRP, QRP ARCI, QCWA, FRC. ARRL SDR Wrk Grp Chairman
Laziness is the number one inspiration for ingenuity.  Guilty as charged!


 1147441404.0