[time-nuts] FE-5650A option 58 tuning word for 10 MHz output

Attila Kinali attila at kinali.ch
Wed Jan 4 05:08:12 EST 2017

Hoi Mathias!

On Tue, 03 Jan 2017 17:55:32 +0100
Mathias Weyland <mathias at weyland.ch> wrote:

> On 2017-01-02 12:18, Attila Kinali wrote:

> > May I ask what you want to achieve? Resp. what you need a 10MHz
> > reference for?
> I've always been wondering about those devices, I guess out of pure
> curiosity. Once I learned that those were available for cheap I knew
> I had to get one. But there is a secondary goal, which is to have
> a standard around to check my other gear against every now and then.
> Lately, I've been building and repairing 1.3 GHz radios and lost
> some time because I did not realize how much off both of my service
> monitors where. I could get away with a properly calibrated OCXO
> but the Rb was so much more sexy. So for the most part I need the
> 10 MHz for my own entertainment. :-)

So it's the usual curiosity and the need to have a reference for
your 23cm radios? For the latter I would recommend using a GPSDO
like the Trimble Thunderbolt or the Jacksonlabs Fury. There are also
a few Trimble OEM GPSDOs on ebay, but I have very little knowledge
about those. Advantage of an GPSDO over an Rb is that you know it's
on frequency, while with an Rb it can be off without you knowing it,
if the electronics or the physics package is defect. Of course, the
disadvantages are that most GPSDOs are made for fixed positions and
need a couple of hours of operation to achieve their nominal accuracy.
The short/mid-term stability (between 10s and ~1000s) is also slightly
worse than compared to an Rb.

To get the best of both worlds, there are a couple of GPSD-Rb around,
but commercial ones are usually quite expensive and modifying regular
GPSDOs to Rb is not always possible (at the very least you need to change
the control loop amplification).

BTW: there is a whole new way of looking at statistics of time series
in the field of time and frequency control, which you might find interesting.
The NIST Technote 1337[1] and NIST special publication 1065[2] give a nice
overview of the methods applied.

> > In general I would agree with Bill Houlne's comment that it's 
> > probably
> > easier to use a different reference that already has a 10MHz output.
> > Also keep in mind, that an output that has been designed as an PPS 
> > output
> > might not work as well as an output for 10MHz.
> Yes I agree with that, but I'd also say that based on the reading I
> have done, it would be foolish to assume to get something that is
> guaranteed to work anyway. With all the things that could go wrong in
> general (dead unit, worn-out discharge lamp, drifted sweeping 
> boundaries
> of the oscillator etc.), the 1 pps to 10 MHz conversion for this 
> particular
> model(!) seems fairly straight forward. 

Most of the units sold on ebay work fairly well. And even a worn out
discharge lamp can often be recovered using a hot air gun.

> On the output issue you are rising: This particular model has a DDS 
> that
> takes a 50ish MHz reference and synthesizes another frequency which is
> a power of 2. A coax cable is taking that to a divider board where it's
> brought down to 1 Hz. This is all happening outside of the physics
> package. The plan is to reconfigure the DDS to synthesize 10 MHz 
> instead and bring out that coax. 

Ah..ok. That should work then.

> I would have to open a hole on the face plate
> to do that, but I'm fairly confident that this outer enclosure does not
> contribute to the actual magnetic shielding of the chamber... All
> modifications are strictly outside of the physics package (i.e. no
> water jet cutting...)

Unfortunately, it's not that easy :-)
I don't know the exact construction of the FE-5650, but usually the physics
package is just a normal aluminium microwave cavity, without shielding.
The mu-metal shield is usually the housing of the whole device itself.
Why this kind of construction is usually employed I do not know, but it
makes any modification of the case cumbersome. Mu-metal itself is quite
sensitive to shock and vibration and will lose (part of) its permeability
when being machined. You can of course try to drill as slowly as possible
to keep the heat and vibration at a minimum, but I have no idea how well
this works and I do not have any hard numbers on the sensitivity of
mu-metal either.

			Attila Kinali

[1] "Characterization of Clocks and Oscillators", NIST Technote 1337,
by Sullivan, Allan, Howe, Walls, 1990

[2] "Handbook of Frequency Stability Analysis" NIST Special Pub 1065,
by W.J. Riley, 2008

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

More information about the time-nuts mailing list