Hi

Welcome !!!

There are an enormous number of source types and signals you could be measuring. The
“right” gear for one probably is not the right gear for various others. There also will be a
lot of other gear that isn’t quite as exciting as a Cs standard. That gear may or may not
already be available.

A few examples:

An OCXO might be plenty good enough for “this” adev + tau combo and a hydrogen maser
might be needed for “that” ADEV + tau combo.

Many industries have “their own” measures of performance. Translating between these measures
often is near impossible. Gear that directly measures what’s needed is a good idea.

Inevitably temperature gets into the picture. Setting up to control it so it does not “mess up” the data
can be a major undertaking. It depends a lot on what you are trying to do.

Lab space and lab layout does matter. Packing this all in a corner of the machine shop :) :) is not a
great idea. Even things like “lots of foot traffic” may be a bad idea. The same is true of folks walking
in with their cell phones turned on.

Is this “lab” purely for R+D? If so, in what area? Is it for design evaluation? If so, how many designs?
(per week or whatever). Is it for production monitoring? If so, in what kind of volume?

There’s also the budget. Is a lab that comes in over $1M a realistic thing? Is $10M a possibility?
You can (and folks do) spend this sort of money doing this. In a lot of cases this is the “parameter”
that ultimately decides what the lab looks like.

This list could go on for pages and pages of guessing. It’s way easier if you can describe what
you are trying to do in a bit more detail. Based on seeing hundreds of outfits do this, no two of them
ever do it the same way. They all have unique drivers that impact the result.

Bob

Have you looked at the FSWP26? It's a pretty slick instrument.

Dave

On Tue, Mar 24, 2026 at 2:57 AM Sean McAllister via time-nuts <
time-nuts@lists.febo.com> wrote:

Hi all sorry for the very long delay in responding. I'm actually working on
coherent transponder applications in Ka band. I need a nice stable ground
reference I can use to drive a Ka band uplink bounced off a vehicle out by
the moon to estimate range and doppler.  To get performance like they get
on the deep space network you start needing Allen Deviation of O(1e-13)@60s

Sean

On Sat, Apr 11, 2026 at 4:38 PM David Bengtson via time-nuts <
time-nuts@lists.febo.com> wrote:

Hi

Ignoring the grubby details of doing a synthesizer at this point, let’s just say that you need
“at least” 1x10^-13 at 60 seconds out of the reference. How much the synth will degrade
things (if at all) is TBD.

Next there’s the basic issue that ADEV is easy to measure and that’s why we use it. Often
in real world systems is maps poorly to the actual performance needed. How much impact
that has in this case ….I certainly do not know. Again, ignore this for now.

There are exotic BVA based OCXO’s that will (barely) meet the spec. Finding one that does
is not going to be easy. Whatever you find will need to be tested to validate it’s performance.
Finding one that does better than 1x10^-13 at 60 seconds is unlikely.

Most OCXO designs will struggle to get to 3x10^-13 at 60 seconds.

GPSDO’s aren’t going to help in this case. GNSS signals “as received” simply are not stable
enough to help out.

A typical telecom Rb will be in the 1x10^-11 at 1 second range. It will get to 1x10^-12 by 100
seconds. That’s not going to help in this case. Even a lab grade Rb (that you can actually buy)
is not going to hit 1x10^-13 at 60 seconds.

The Cs standards that you can easily get your hands on do worse at close in ADEV than a good
Rb. Not much help there either.

This sorta kinda gets you to a hydrogen maser. The passive one in the basement here will hit
1x10^-13 at 60 seconds, but is not really spec’d to do so. (7x10^-14 at 100 seconds is the spec).
A modern active maser in good shape should  be able to do what you need to do.

The only thing on the list that gives you any real margin is that very expensive (and long lead time
and hard to find and no fun to set up) active hydrogen maser.

Fun !!!

Bob

Hi,

The "cheep" solution would be to try to find a good HP5065A (which seems
to be the hard part currently) and do the optical filter modification:
http://leapsecond.com/corby/Super-5065A-Project.pdf
https://www.timeok.it/wp-content/uploads/2016/06/HP5065A-Super-upgrade-v-1.4-1.pdf
That should do it until GNSS kicks in.

Cheers

Am 22.04.26 um 14:46 schrieb Bob kb8tq via time-nuts:

Yeah I've got an HSO-14 on order (8 month lead time!). Either a GPSDO or
Cesium would be used for longer integrations (up to 1000 seconds or so).

On Wed, Apr 22, 2026 at 6:46 AM Bob kb8tq kb8tq@n1k.org wrote:

Hi

Depending on which one you ordered, It probably should do the job. Note that the “full spec
after 28 days power on” the mention is indeed part of the process of using one.

I don’t know if it’s an available option or not: If they will supply an ADEV plot out to 1K seconds
on the part they send you, that would come in very handy. Even if it adds some time to the delivery,
I’d go for having them do that. The stock specs don’t get out to 60 seconds and you are obviously
interested in how it does at even longer tau.

I’m in no way suggesting I know how much that would cost :) :).

Bob

Hi

There is no guarantee what this or that 5065 will do after the filter mod. Some improve
quite a bit. Others do not seem to do quite as well.

You might well wind up buying several 5065’s simply to get one that does the job. Also
having run them for a while, you probably will need a couple simply to keep one up and
running while the other(s) are being repaired. That’s based on the stack of them that
“need attention” over in the corner of the basement :) :) :).

Bob

 
You know the DSN uses a Maser <grin>
And here at JPL, we have a Maser that gets distributed by optical fiber, and we have a cleanup loop in the lab where's used.
The maser is up in the Frequency and Timing Lab off by itself at the west end of JPL in Bldg 298, which I've been told has stainless steel rebar to reduce magnetic fields.  TDL (Telecom Development Lab) is in my building, 161, sort of in the middle of the lab. And the reference also goes to Bldg 298 (where a lot of experimental telecom/radio stuff is done) and Bldg 179 (Spacecraft Assembly Facility) and probably, although I'm not sure, up to the Mesa antenna range. (There is a map of JPL somewhere on the public web, but I can't find it now)

We have a ton of SRS FS725s Rbs around (there's one on my desk right now, keeping my office warm), but that's only good to around 2E-12. 
And a fair number of GPSDOs from all sorts of mfrs - Symmetricom/Microchip/Pendulum/Fluke. but those all tend to be in the same 1E-11, 1E-12 range at short tau (<100 sec)

What we do for measurements on transponders is compare the input and output, because what you're usually interested is the "added phase noise or deviation".
I'd have to go look at what TASI built for testing the KaTS for Juno - it might well have locked to an external source (e.g. the maser)

On Tue, 21 Apr 2026 09:54:06 -0600, Sean McAllister via time-nuts time-nuts@lists.febo.com wrote:

Hi all sorry for the very long delay in responding. I'm actually working on
coherent transponder applications in Ka band. I need a nice stable ground
reference I can use to drive a Ka band uplink bounced off a vehicle out by
the moon to estimate range and doppler. To get performance like they get
on the deep space network you start needing Allen Deviation of O(1e-13)@60s

Sean

On Sat, Apr 11, 2026 at 4:38 PM David Bengtson via time-nuts <
time-nuts@lists.febo.com> wrote:

time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe send an email to time-nuts-leave@lists.febo.com