Re: Timepod (was - Morion OCXO)

Andy G4JNT

"Cross correlation" suggests  mixing of the two sources and averaging (filtering) the result followed by digitisation of the baseband output.   Or at least its equivalent using DSP    Quadrature mixer to resolve ambighities.

I've been intending to make something do that for ages- but its not high up the priority tree.

I've measured such reference sources by multiplying both the test signal and the reference  up to 2.5GHz, but 900Hz apart, then mixing the two signals and using a sound card to monitor the resulting 900Hz.   An FFT plot shows short term stabilities and frequency offsets.


On 3 August 2016 at 18:29, Grant Hodgson grant@... [ukmicrowaves] <ukmicrowaves@...> wrote:

The Timepod is a very high performance instrument for measuring the
stability of signa lsources. It uses cross-correlation to reduce the
system noise floor to very low levels. It was designed by John Miles
KE5FX as a commercial product which is sold by Symmetricom.

Timelab is the analysis program that does the back-end processing,
also written by John and performs time, phase amd frequency analysis.

The Timepod uses LTC2216 ADCs and great care was taken with the signal
layout, use of low-noise regulators, signal conditioning etc. Digital
data is sent to the PC for processing in Timelab, which does the FFTs,
user displays etc.

Timepod retails for around $5k, out of the reach of most amateurs but
representing and exceptional price/performace for commercial
customers. The schematic is provided in the manual if you want to
roll your own!

Grant G8UBN

Quoting "Andy Talbot andy.g4jnt@... [ukmicrowaves]"

> Interesting link - in a way. That 'TimePod', described as a 'Programmable
> Cross Spectrum Analyser' [whatever that is ] needed to provide input to
> the measurement software is an enigma. They state it uses "... high
> performance host-based DSP techniques ... " which really means everything
> that matters is done on the PC and the Pod could be something quite simple.
> My guess is, its a digitiser as in any direct sampling SDR (and judging by
> the upper frequency limit of 30MHz, an average-range one).
> 'jnt
> On 3 August 2016 at 08:54, Stephen Tompsett stephen@...
> [ukmicrowaves] <ukmicrowaves@...> wrote:
>> For comparing oscillators you might find Timelab useful
>> On 02/08/2016 19:29, Edward Harland g3vpf@... [ukmicrowaves]
>> wrote:
>> My MV89 came still attached to the PCB. Someone had gone round the edges
>> with tinsnips. No attempt had been made to de-solder the unit. It has full
>> output and works at all orientations.
>> I tried comparing the frequency of the MV89 with the Leo Bodnar GPS and
>> found that for about an hour after both units were switched on the relative
>> frequency was changing pretty much continuously but then slowed and
>> stopped. After adjusting the frequency trimmer on the MV89 it held the
>> frequency for the next 30 minutes. Had to stop the test at that point. I
>> have no means of finding out which, or both, of the units were changing,
>> but the result looks very promising.
>> Will now build a unit with built-in batteries so the MV89 is always on,
>> even when travelling to a /P location.
>> Thanks for all the comments received - very helpful.
>> Ed G3VPF

Join to automatically receive all group messages.