You've omitted the following pertinent data:

1. The idea is to correlate photon arrival times at pairs of stations with lateral separations of up to 1 km or so.
2. The Signal to noise ratio of the resultant measurement is proportional to   
       A) The geometric mean of the areas of the pair of telescopes          This cannot be too large as it smears out the details in the Fourier plane.  
       B) Square root of the electronic system bandwidth (or equivalently the reciprocal                of the photon relative arrival time stamp accuracy) 

    C) Square root of the observation time.
    The differential light path delays between the telescope pair can be up to 3.3us or so.
The effect of differential noise between the timestamps at a telescope pair is to in effect reduce the effective electronic bandwidth or equivalently increase the observation time for a given SNR. For example with a time stamp resolution of 1ns (~500 MHz bandwidth) the differential timestamp noise should be much less than this (ie 100-200ps or so). With a 10ns (~50MHz bandwidth) timestamp resolution 1-2ns of differential timmestmp noise is tolerable, however the required observation time increases by a factor of 10 for a given SNR.
The SNR is independent of the optical filter bandwidth so that with a sufficiently narrow optical filter bandwidth its possible to ensure that photon counting is feasible.

The original Hanbury-Brown Twiss interferometer used 10m light buckets running on a circular railway track with interconnecting coax cables. The position of the collector pair was adjusted to maintain the baseline at right angles to the light path to the observed object thus ensuring zero differential light path delay between the collectors. The two signals were multiplied together using an analog circuit with a bandwidth of around 40MHz. Observation times of several hours were required even for relatively bright stars.
 For further details see:Long baseline optical intensity Interferometry

|   |
|   |  |   |   |   |   |   |
| [1506.05804] Long-baseline optical intensity interferome...Submission historyFrom: Dainis Dravins [view email] [v1] Thu, 18 Jun 2015 20:00:34 GMT (4473kb,D)  |
|  |
| View on arxiv.org | Preview by Yahoo |
|  |
|   |

Bruce

On Tuesday, 26 April 2016 11:03 AM, Bruce Griffiths <bruce.griffiths@xtra.co.nz> wrote:

There's no spec for th 10MHz option you originally specified.What's the jitter bandwidth of interest?

Still need a spec for the data logging system timebase error.This is for your intensity interferometer?

Bruce

    On Tuesday, 26 April 2016 10:00 AM, Ilia Platone info@iliaplatone.com wrote:

I haven't bought anything yet, so feel free to recommend me any
components/setup.

here is the datasheet I am referring to:
http://www.foxonline.com/pdfs/FVXO_HC53.pdf

It contains some informations about phase jitter on page 5.

The

I selected this because of the CMOS output, since I didn't find any
Connor Winfield osc with these levels yet.

Actually  want to achieve less than around 78ps jitter at 125MHz. (this
should be achievable using this also:
http://www.digikey.it/product-detail/it/fox-electronics/FVXO-HC53B-125/FVXO-HC53B-125-ND/2153894)

Let me know,

Ilia.

Il 25/04/2016 22:51, Bruce Griffiths ha scritto:

time-nuts mailing list -- time-nuts@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@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Thanks Bruce,

I know this, the delays due to the light path difference will be
calculated later by software, and i can have filters with a bandwidth of
3-5 nm, two 0.4m diameter telescopes. the distance of the two telescopes
will be around 2500m and the maximum delay between these two signals may
be 8us. The expected photon flux (the actual center frequency of the
data logged) will be less than 10M/s, but the clocks that record these
data must be synced with a maximum error of some ps, otherwise I'll not
find any correlation.

I would propose a software capable of aligning these two (or more,
sequentially) logs by finding correlation by a slighty variable
time-base, excluding pseudo-correlations by a discriminating algorithm,
but this is for an interferometry forum/list maybe.

Regards,
Ilia.

Il 26/04/2016 02:50, Bruce Griffiths ha scritto:

--
Ilia Platone
via Ferrara 54
47841
Cattolica (RN), Italy
Cell +39 349 1075999

On Mon, 25 Apr 2016 20:45:53 -0400
Bob Camp kb8tq@n1k.org wrote:

Not really. In order to prevent oscillation you have to ensure
that the lowpass filter formed by the input capacitor and the
bridging resistor has a low corner frequency. Putting the corner
frequency in the range of 10kHz-1MHz is usually good enough, the lower
the better. Though there are of course limits to this:

• from a certain point on, the surface resistance of the PCB
  dominates the resistor. A normal FR4+solder stop board has a
  surface resistivity of 1M-10G depending on the actual solder stop,
  humidity and dirt/grease on the PCB and the geometry of the wires.
  (I usually go with 10M for close wires when i'm too lazy to calculate)

• The larger the capacitor the lower its self-resonance frequency
  becomes. Ie from that poin on the capcitor behaves more like an
  inductor than a capacitor. Rule of thumb: a 4.7uF 0603 X5R has a
  self resonance frequency around 1-3MHz

For my squaring gates, I usually use 100nF+1M (~100kHz) on a
single gate inverter (usually LVC, but not always) and have
not seen any oscillation yet.

BTW: Does someone know about actual measurements of the small
signal transfer characteristics of a single CMOS gate?
It would be nice to put the above rule of thumb onto a more
solid foundation.

		Attila Kinali

--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neil Stephenson

Hi

Well, based on about a hundred breadboards, it always comes out this way
regardless of who does the layout or what caps are used for a buffer amp.
Consider that this is not an oscillator, so the only cap that is sure to be there is
a blocking cap.

Bob

Ciao Ilia,

On Tue, 26 Apr 2016 02:24:04 +0200
Ilia Platone info@iliaplatone.com wrote:

I cannot say. I don't know what your requirements are and thus
cannot say what a good design choice is.

Yes, the LEA-M8F does support the output of satellite phase data.
Yes you can use that for RTK. BTW: this is stated in the documentation
for the module. You just need to read it.

BTW2: is this a hobby project or is this a work project?

		Attila Kinali

--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neil Stephenson

Insufficient information to indicate if carrier phase data is externally available.Without this it appears that the position accuracy is around 10x worse than the requirement for 10ns timestamp accuracy.

The timing stability is also somewhat unknown.How well will 2 independent modules track in phase?In particular what is the effect of brief loss of GNSS signals?It would probably be necessary to obtain a pair of modules and measure their relative ADEV etc as a function of physical separation.

The phase noise floor is  somewhat higher than a Thunderbolt for example, which in itself is limited by its relatively noisy output amplifier.
Bruce

On Wednesday, 27 April 2016 12:08 AM, Ilia Platone <info@iliaplatone.com> wrote:

Thank you Attila,

Just for informations, as I must build the board from scratch, does this
design is good to implement into the datalogger?
I mean, for the second RTK purpose, does this u-blox module satisfy my
requirements (link below)?
http://www.opendigitalradio.org/lea-m8f-gpsdo

Regards,
Ilia.

Il 26/04/2016 01:36, Attila Kinali ha scritto:

--
Ilia Platone
via Ferrara 54
47841
Cattolica (RN), Italy
Cell +39 349 1075999

time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Hi Attila,

This is a hobby project, by now, but if these devices can be versatile
and small, they could be used for a work project. till now, hobby and
open source.
I am stuck at the timing issues. After finding a suitable stable clock,
on multiple and independent boards, for 8 hours, I'll begin designing
the software of these.

I must print each photon event captured by an APD at the focus of more
telescopes, these events must be timestamped with an accuracy of 2.5ns,
and the expected rate is lower than 10MHz, The overall capture runs will
take about 4-5 hours, it depends on the object observed, after capturing
I must compare the timestamps logs each other to find if are there
common events/timestamps in the various fluxes. I need a stable clock to
reduce the software needs, for example if the automated captures of each
device ends withinone second each other I think it would be too much,
but I am thinking I must be more flexible in the software side.

Regards,
Ilia.

Il 26/04/2016 14:40, Attila Kinali ha scritto:

--
Ilia Platone
via Ferrara 54
47841
Cattolica (RN), Italy
Cell +39 349 1075999

On Tuesday, April 26, 2016 02:40:14 PM Attila Kinali wrote:

Specs are:

1. Relative position of any pair of clocks located up to 2km apart has to be
   known to within 3cm or so. Post processing is OK, however differential Earth
   tides between the clock locations may need to be considered.

2. The difference in the time offset between any pair of clocks located up to
   2km apart shall not vary by more than 200ps (1ns time stamp quantisation) or
   2ns (10ns timestamp quantisation) over an 8 hour period (at night).
   Post processing of data to fit wander etc is not practical as the SNR is too
   low to support this.

Bruce

Errata..

Il 26/04/2016 22:24, Ilia Platone ha scritto:

...each telescope will have an avalanche photodiode ( plus amp )...

--
Ilia Platone
via Ferrara 54
47841
Cattolica (RN), Italy
Cell +39 349 1075999

BTW, the u-Blox NEO-M8N doesn't fit RTK specs (patched it returns RAW
data): it works on band L1 (20cm range), what I need is a GPS module
capable on wiorking on L1+L2 to get 1cm range precision (if it's correct).

Regards,

Ilia.

Il 26/04/2016 22:25, Bruce Griffiths ha scritto:

--
Ilia Platone
via Ferrara 54
47841
Cattolica (RN), Italy
Cell +39 349 1075999