[time-nuts] Next upgrade

Thu Nov 23 18:18:35 UTC 2017

The Next upgrade has touched several subjects we deal with on a dayle  
basis. Allow mw to add my thoughts. First we are time nuts trying within our  
limits to advance time and frequency generation and measurement with 
affordable  resources. We are at least a factor four orders of magnitude behind  the 
big boys in performance and  combined available resources.
As a market we are to small potato's so products tailored to our needs are  
limited and are getting less. An example SRS FS740 a nice box but no help 
to get  past E-12. New GPSDO's for Telecom do not have 5 or 10 MHz output.
PRS-10 a nice Rb, originally intended for FRS replacement has ADC and DCA  
control limeting it to 2 E-12. I have one, have not said anything because I  
want to get rid of it.
The question of why Rb. For our work there are only two reasons. Not living 
 next to NIST the only precise frequency reference is GPS. To overcome its 
limits  you need long time averaging. Aging compared to a OCXO is orders of 
magnitude  less. Second because of aging and tuning range a 16 bit DAC will 
do the job.  There was recently discussion of the use of the LTC1655 which 
is still my # 1  choice. What was not mentioned on the 1650 is that in needs 
a external reference  discontinued and double ling the cost. 18 bits would 
be nicer but we have not  found one affordable. Using a modified Shera with 1 
E-14 steps yields a  range up to 6.4 E-10. We are presently running three 
Tbolts, standard, OSA 8600  and the Efratom M100 with 40 000 seconds and 6 
E-17 per uV direct C field drive  through a 10 K resistor. Temperature, 
pressure and some start up issues make it  a less desirable option.For temperature 
and pressure we have an interface board,  but automatic recovery after 
power or GPS loss is not there after two weeks.  That is why we keep going back 
to 1655/Shera. Going to 6 E-17 keeps the jumps we  all see below ! E-13. 
Attila recently caught up with me on my Europe trip at  Juerg's home in 
Switzerland and can tell you about all our projects. Using the  standard Tbolt we 
also showed him the consistent jumps with a Tracor 527E and a  Cesium 
reference.
Yes doing a Cesium GPSDO again with 1655/Shera is doable but the question  
has to be: is it necessary. I worked with Richard McCorkle on it 4 years 
ago.  The question is does it make sense. My HP5061B with the new smaller tube 
shows  no aging but my GPS measuring capabilities are limited to 1 E-12. We 
have  decided to monitor 1 pps from a Tbolt against a 1pps derived from our 
respective  Cs and if necessary manually adjust.. To do it with le least 
amount of equipment  we decided to take a PICTIC II a divider chain and a 
custom V drive  using USB stick and blue tooth monitoring low cost, no fan and 
low power. I  never built a PICTIC but did combine Richard's and my boards for 
cost reason.  Bill Riley did an extensive evaluation and recommended to 
consider 4 layers.  This was the first time I looked at the board. Richard used 
schematic  capture resulting in no ground plane and long ground runs not 
what you want when  you chase nsec. I took a look and decided 4 layers is not 
necessary but was able  to add a nice ground plane. I have boards of PICTIC, 
divider and V drive. Right  now we still have work to do on M100/FRK GPSDO 
and the high resolution Austron  counter. If some one wants to build a unit, 
please contact me off list, if I am  convinced you will do it, I will send 
you the boards free of charge.
Richard did a PICTIC III and was working on a 4 and I was working on a 5.  
Sadly contact with him has stopped, the last correspondence was two years 
ago,  he had fallen and broken some ribs and had prostate cancer. His last 
words  where, he would get back with me after health was back under control. I 
have  tried what I did with Brooks, send a letter with USPS which resulted  
in Brooks Wife responding, you know the details, no response from Fairbanks. 
Has  any one heard from Richard?  He did a lot of work for us, brilliant, I 
 would ask a question, because of the time zone the typical next day 
response was  working PIC code.
Bert Kehren

In a message dated 11/22/2017 7:26:57 P.M. Eastern Standard Time,  
k8yumdoober at gmail.com writes:

For the  most part the SRS-10 is a nice choice, although I'd always be wary
of  buying a
used one.

My only real beefs are that the tuning  granularity is rather coarse, about
2E-12, and the
disciplining loop  seems to be a bit aggressive so that the poor oscillator
gets  jerked
around quite a bit by the GPS.  This makes for rather  ugly-looking plots of
time error
over time.

The above comments  are derived from about 3 years of operating one as a hot
emergency spare at  the Arecibo Observatory against the day when the H-maser
crashed  abruptly.  In this case the SRS-10 was embedded in an FS725 which  
we
bought new.

Dana

On Wed, Nov 22, 2017 at 4:16 PM,  Jerry Hancock <jerry at hanler.com> wrote:

> Three  questions:
>
> 1) Now that I’ve split my Lucent RFTG-U into a REF0  and REF1 unit with
> both supplying 10Mhz and 1PPS, is there a way to  combine the outputs or
> some other technique to improve the short  and/or long term performance?
>
> 2) I’ve become interested in  Rubidium Disciplined Oscillators recently 
and
> was now thinking of  purchasing one of the PRS-10 that I see on Ebay. If I
> did that and  replaced one of the DOCXOs from one of the Lucent boxes, 
what
> impact  would this have on the overall performance both with and without
> (when  in hold-over)?  Basically, is it worth the money to upgrade one  of
> the boxes to a Rubidium disciplined oscillator assuming the GPS  signal is
> rarely lost?
>
> 3) Figuring the PRS-10 will  cost around $250 when all is said and done, 
is
> there a better option  to improve my GPSDO system?
>
> I basically use the GPSDO as a  reference for any equipment that takes an
> input. I have no monetary  need for a reference, just an interest.
>
>  Thanks.
>
> Jerry
>
>  _______________________________________________
> time-nuts mailing list  -- time-nuts at febo.com
> To unsubscribe, go to  https://www.febo.com/cgi-bin/
> mailman/listinfo/time-nuts
> and  follow the instructions  there.
>
_______________________________________________
time-nuts  mailing list -- time-nuts at febo.com
To unsubscribe, go to  
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the  instructions there.