[time-nuts] Another hydrogen maser success story

[Magnus Danielson]

Hi Skip,

On 2020-01-17 01:41, Skip Withrow wrote:
> Congratulations to Magnus on his new toy.
Thanks!
>
> I would like to relate my recent hydrogen maser experience.  Hope you
> enjoy the story.
Oh yes.
>
> Probably about a year ago I was introduced to Dave Howe at NIST.  Last
> fall at a hamfest we were talking and I mentioned that I was recently
> retired and wondered if it might be possible to 'play' up at NIST. He
> mentioned that there was a hydrogen maser that had experienced a power
> outage and did not come back up.  It had been taken out of the time
> scale and was sitting in what was essentially a back store room.  He
> asked if I would like to try and bring it back up?  Duh, of course I
> said yes.

Yes. Who wouln't? I would for sure. I'm just not local enough.

Dave is also a ham, but I think he has not been very active lately, at
least as I recall when I met him a year ago.

> So began what has been quite a journey over the last six months.  I
> have gone up to Boulder for a day every week or two over that time.
> First, was getting the maser from its storage location to a lab where
> it could be left and I could work on it.  The unit is a Sigma Tau
> model NIST-2001 s/n-1 manufactured about 1990!  It resembles Corby
> Dawson's maser kit, and is the precursor to the current Symmetricom
> (which bought Sigma Tau) MHM2010.
Cool.
>
> The unit had been sitting unpowered for quite some time.  Turns out
> both of the 24V supplies had failed at some point, so they were
> missing along with the power cords and fuse holders.  It had a cart
> with it for powering the unit from the External DC and a kludge to
> charge the batteries.  Next all the covers were removed, and the
> vac-ion pumps and hydrogen valved off.  At this point power was
> applied and the DC fuse blew.
24 V is really the standard voltage in this business.
>
> Turns out one of the DC/DC converters in the power section was toast.
> That assembly was dismantled and a new DC/DC converter was fitted for
> the unobtanium blown part.  No more Ext DC fuse blowing.  Currents on
> the vac-ion pumps are high (as expected) but after a week they came
> down to reasonable levels.  Next the valves were opened up to the
> maser vacuum system and the vac-ion currents take off again (as
> expected), but after another week looked good so no vacuum leaks.
You should maybe have considered using a turbo pump at first as it has
been sitting there for a long while, as it will eat on the ion-pumps.
>
> At this point all the heater fuses were plugged in to get the cavity
> up to temperature (another week).  Now all is ready to try and light
> the source bulb.  The Pirani guage and source heaters were powered up
> and the hydrogen turned on.  But all attempts to light the source bulb
> with the RF generator fail.  The RF generator was dropped from the
> unit and after several weeks of trying lots of different things was
> finally rebuilt.  After installation the source bulb still did not
> light.
>
> We resorted to a cattle prod and a piece of aluminum tape up to the
> source bulb to get anything to glow.
ESD guns have been known to be used too for the same purpose. It's a
common problem it seems.
> The palladium leak heater was
> operated open-loop and the hydrogen pressure was monitored with the
> vac-ion currents.  Things seemed to get better with higher H2
> pressures.

There should be a recommended H2 pressure. Mine runs at 1.5 bar prior to
palladium valve.

Did you close the loop on the paladium valve from the Pirani-gauge?

> So, the RF generator was adjusted some and reassemble and
> YES, there was light.  In fact, the hydrogen pressure could be lowered
> all the way down to its normal level and the source bulb was still
> glowing.
Good.
> But alas, there was still no IF signal meaning that it was still not
> oscillating.
>
> So today the top of the maser was disassembled to get to the 1.240MHz
> circuits and the vacuum chamber RF connection.  There are about a
> dozen different layers of electronics, magnetic shielding, and
> insulation to get through.  A -100dBm signal was stuffed into the
> receiver front end and there was IF signal indication - good news, the
> RF chain is good / bad news, the cavity is not oscillating.  So all of
> the top was re-assembled (I meant to get a picture of all the goodies
> on top of the vacuum chamber, but spaced it).
That is indeed good news. Then you can focus on getting it to mase.
> Where to you go from here?  I figured more excited atoms into the
> cavity could not be a bad thing and figured I would try cranking up
> the source RF generator power.  I had installed a trimmer cap that
> would help me control the operating point of the oscillator during the
> rebuild.  When tweaking the cap all of a sudden, BAM, lots of light.
> When the maser started masing if was obvious!  IF signal was
> immediately at the right level.  After a few minutes the green LOCK
> light was even solid!
As it should. One needs to make sure the cavity is tweaked to the right
frequency for it to mase, and primarily this is done with temperature,
then fine-tuning using varactor. The amplitude depends on how well tuned
it is. Some cavities also depends on pressure, just to help making the
issue more difficult. On mine the cavity is within the vacuum system, so
pressure sensitivity is essentially completely gone.
>
> Along the way both of the 24V power supplies were replaced, along with
> the fuses and power cords.  Batteries are now back on line on the
> internal supplies.  Next steps are to get the covers back on and let
> it settle for a while.  We will check the ADEV against the (local)
> NIST time scale (don't have the issue of a good source as Magnus
> does).
When you have more of these, it's much easier as you have more of these
available. Mine was measured against other good masers when serviced,
but I don't have that luxuary with the current hardware, so therefore I
need to improvise. NIST should not have to improvise as I have to do in
my private lab, if they have to, then they are severly underfunded and
not just underfunded.
>   The goal, I believe, is to get it moved to Colorado
> University.
That will be a nice place for it.
> I would like to thank Dave Howe for the opportunity and all the
> insight, help, and suggestions along the way.  It has been an
> invaluable educational experience. 
He is one very nice and gentle gentleman!
> I have attached one picture of the
> bottom of the unit.  There is a mirror laying in the bottom reflecting
> the bottom of the RF generator where you can see the nice purple light
> of the source bulb through the hole.  The other picture is the front
> of the unit (with the two new Lambda power supplies) showing the IF
> level reading and the green lock light.

I have the pink-purple light visible out front, as the EFOS-B maser was
built with a nice little window for it, so I get to see it whenever I
walk by. :)

The EFOS-B does not have a lock-light, you need to see that on the
serial interface that pull out other state.

>
> Sorry for the long post.

No problem, fun story. :)

Thanks Skip!

Cheers,
Magnus