[time-nuts] Advise on building a DIY GPSDO?

EWKehren at aol.com EWKehren at aol.com
Wed Apr 6 09:37:29 EDT 2016


Having been a time nut for almost 50 years knot knowing it I  like to add 
my thoughts. Starting in the late 60's I won second place at IEEE  student 
contest for my counter design using 7 segment lamp and Motorola RTL in  the 
70's I buried a Sulzer using bank coin bags filled with sand to close the  
hole and application notes at TI using TI LED and IC's, in the 80's bought an  
Efraton FRK and did my first DO using a Tracor Omega converted to 60 KHz  
and a Philbrick DAC. I was the u processor  using every 6 month using the  
Tracor to recalculate the time for pulses to the DAC counter. I had then and  
still do now have problems with computers and processors. When Brooks came 
out  with his GPSDO i bought 10 A&A boards and started dialog with Brooks. 
Thanks  to Corby I joined time nuts and continued work with Richard McCorkle. 
One of the  results is a GPSDO for FE 5680, 5650 and 405 that is readied for 
release by Cash  Olsen. It took extensive testing fine tuning and more 
testing. During that time  I also had the opportunity to communicate with pro's 
on that subject and came  away even for the pro's it is not a walk in the 
woods. Tom's 2008 comparison  mentioned in this thread 
http://leapsecond.com/pages/gpsdo is very informative it clearly shows the relationship between OCXO 
 and digital loop. Few HP units are as good as his and there is a wide  
distribution when it comes to 10811's. The tbolt shows clearly the impact of 
the  fact that it changes the frequency to correct the 1  pps.
Miller is very impressive and we have done one using u-lox-6 and a MV89  
with similar results mainly intended for Ham's. Working on a 100 MHz version 
for  Ham uwave work
Summary true GPSDO is complex and long time work, Miller is simple and  low 
cost and maximum bang for your buck. 
One more comment on u-blox. In my opinion unless you plan on doing saw  
tooth correction the T version is a waste of money. Also if you plan on using  
the Miller circuit buy a $ 10 version and use a PIC to program in the higher 
 output frequency a version that stores it in EEROM costs three times as 
much  relying on the battery unless it is always powered up may be a problem 
once you  are out in the field and have no PC to change from the 1pps  default
Bert Kehren
 
 
In a message dated 4/6/2016 4:20:14 A.M. Eastern Daylight Time,  
hmurray at megapathdsl.net writes:


albertson.chris at gmail.com said:
>  I may never get  around tooth's as the Rb is very good all by itself to 
the
> limit of my  ability to measure it.

Wait longer.  :)

If your Rb doesn't  have a PPS, make one.  One of tvb's divider chips is 
probably the  simplest approach.  You can probably do it with an Arduino if 
you  have one handy and like writing that sort of code.


Compare the PPS  from the Rb with the one from your GPS.  It's probably 
simplest if  your Rb PPS is offset a bit from the GPS PPS.  The idea is to 
make  the difference always have the same sign.

Feed both PPS pulses to some  equipment that can measure the time 
difference.  
Collect  data.

If you have a typical low cost (non GPSDO) GPS device, you should  be able 
to 
make graphs like the ones that show a hanging bridge.   Look at a sawtooth 
region, not a bridge.  Pick the top or the bottom  or the middle or 
whatever 
you can remember.  Call that the  offset.  Wait an hour collect more data.  
Has your offset  shifted?  If not wait longer...

You are measuring the average  frequency of the Rb.  It might be 
temperature 
sensitive.  Can  you see the offset change over a day?  (That assumes you 
turn 
the  heat down at night...)  Or maybe you are seeing the GPS night/day  
shift.


-- 
These are my opinions.  I hate  spam.



_______________________________________________
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.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: Time nuts links.doc
Type: application/octet-stream
Size: 24576 bytes
Desc: not available
URL: <http://www.febo.com/pipermail/time-nuts/attachments/20160406/ccb44393/attachment.obj>


More information about the time-nuts mailing list