[time-nuts] GPSDO unclear situation

Attila Kinali attila at kinali.ch
Wed Jan 27 16:26:27 EST 2016

On Wed, 27 Jan 2016 12:17:32 +0000 (UTC)
Anton Moehammad via time-nuts <time-nuts at febo.com> wrote:

> Hi All, 
> My GPSDO based from VE2ZAZ design but I have a few question I do
> not clearly understand about GPSDO, like :

>  1.How important the quality of 1 PPS signal based from jitter etc what the
> difference I will get between timing purpose GPS module like Ublox 6T or
> trimble resolution T with a position GPS module like ublox 6M or other in
> term of stability, accuracy and phase noise.

I think the best improvement you can make is by not using a VE2ZAZ style
GPSDO :-) The design uses an FLL, which results in a slight frequency
error. If you change it to a PLL, you will only get a slight phase error.

Next, PPS resolution will be limted by the uC internal counter.
Ie your resolution is in the 10-50ns range. You will need to get this
down to <1ns and apply saw-tooth correction if you want to see any
difference between a standard GPS receiver or a timing GPS receiver.

As there is no LEA-6M I assume you mean the LEA-M8T. The Trimble Resolution,
the LEA-6T and the LEA-M8T are all timing receivers. The differences between
those are small. Smaller than all the other error contributions due to your
device design and your antenna position. Hence I wouldn't change the
receiver, once you have found one that fits your need.

>  2. I understand short time stability from OCXO is better than GPSDO ? 
> can someone give me a clue what the time constant need for OCXO control
> (in my application 30 second) is that enough ?

This highly depends on your OCXO, your GPSDO design and the performance
the GPS receiver delivers. There have been many papers written on this
topic and you need a fair bit of understanding on control theory to apply
them correctly. Usual time constants for the control loop are between
one second and a couple of minutes. 

> 3. I have few Rubidium oscillator and I read somewhere (I believe in KE5FX
> web) that GPSDO with rubidium osc is hard to beat, in what way ? short time,
> long time, phase noise.

The reason is because you can integrate over days using an Rb local
oscillator instead of just seconds or minutes. This filters out all
periodic perturbations induced by the GPS system.
The phase noise might or might not be better though. This highly depends
on how much care was given to the Rb's quartz oscillator output.
E.g. the ubiquitus FE-5680A has a very noisy output, worse than
a simple crystal oscillator.

> My goal is "simple" I have access to national time reference lab
> ( an active Maser H ) I want to compare my reference to them off course
> there is no way I have same stability but anything not so "embrace" is fine
> to me 

I am not sure I understand what you mean here. Do you want to compare
your GPSDO to their H-maser? Or do you want to compare some local
reference you have using the GPSDO as transfer standard?

			Attila Kinali

Reading can seriously damage your ignorance.
		-- unknown

More information about the time-nuts mailing list