[time-nuts] Re: Repairing an HP 5065A Rubidium Vapor Frequency Standard. Sharing experience and advice welcome.

[Ed Marciniak]

As far as the question on whether using a narrower band filter at 780nm:

The Rb lamp light would have four components…
A 780nm line pair separated by 0.014nm
A 795nm line pair separated by 0.014nm
Continuum from Rb and buffer gas
Background emission from nitrogen, noble buffer gas and contaminants

Additionally, a silicon photodetector has a limited range of sensitivity. Anything it can't "see" need not be excluded.

Using a 10nm bandpass filter, we’re excluding the 795nm line pair and most of the remaining light.

A quick check of a couple web resources suggests the 780nm pair is twice the intensity of the 795nm pair and dwarfs any other Rb lines.

Xenon (if used) has probably weak lines at 778.7, 780.3nm.
Krypton (if used) has lines at 785.5 and 774.7nm.
Argon (if used) has some lines around 772.3 and 772.4 and 794.8nm
Nitrogen doesn't have any close.

I'd expect continuum would be pretty weak at the low operating pressures.

The xenon lines are so close that it might hit or miss on filter center frequency tolerance as to whether there'd he any meaningful rejection.

I suspect most of the gain is gotten with the 10nm filter. A 3nm filter would have a steeper slope on the pass and and might exclude the krypton lines, but would be of near zero advantage unless the buffer gas actually contained xenon emission lines in the 10nm passband.

Putting the filter at the lamp output is smart as it reduces the light shift as well as avoiding excitation of the D1 line pair with nearly the same energy split.

I don't happen to know what the Schott noise is compared to the operating photocurrent, but perhaps there are incremental improvements to be made with newer op amps or detectors...or cooling the detector.

I'm assuming that HP didn't botch anything with power supply noise on the detector and amplifiers as far as noise centered around 127hz and 254hz.

I don't pretend to be an authority, but perhaps something here provokes thoughts or intelligent discussion.




________________________________
From: Jean-Charles BILLEBAULT via time-nuts <time-nuts at lists.febo.com>
Sent: Thursday, July 31, 2025 5:17:19 AM
To: time-nuts at lists.febo.com <time-nuts at lists.febo.com>
Cc: Jean-Charles BILLEBAULT <Jean-Charles.Billebault at rakon.com>
Subject: [time-nuts] Repairing an HP 5065A Rubidium Vapor Frequency Standard. Sharing experience and advice welcome.

Hello time-nuts,

First, thanks to this amazing community and especially the extensive archives which have been invaluable for my HP5065A work so far.

I'm knee-deep in another 5065A repair and running into some issues I'd appreciate input on, especially from anyone who's dealt with heater wire sourcing but also anyone that ever dealt with this peace of art.

So one day we smell that odor we easily recognize. After verification , there’s C1 short in the RVFR - no big deal, swapped it out with a 2499-003-X5W0-502PLF and that part's sorted. But of course, the short had enough time to cook the A12 assembly pretty good. The oscillator circuit inside got toasted but the component values are still in spec, just looks like it went through a barbecue.

Here's where I'm scratching my head - I'm seeing 85MHz from the A12 oscillator. Some pages in the manual mention 100MHz, others say 90MHz on the module description. Anyone know if this matters much? The lamp fires up fine after cleaning everything up, but should I be tweaking this oscillator? However it seems to me that it was brighter some years ago. Maybe I’m wrong.

The C1 failure also took out R2 and L2 in A15 from the current surge. Replaced those and the board seems happy enough despite looking a bit crispy. I'm guessing the protection circuit tried to do its thing but wasn't quite enough.

Now here's the real headache - after putting it all back together, HR1 decided to join the party and went short circuit. This almost killed the 1.5Ohms resistor and the driving transsitor had less luck.  Reading 1.2 ohms instead of 52.3 on the heating resistor, not good. Used thermal imaging to track it down and it's buried somewhere inaccessible in the heater assembly. Had to tear the whole heater body apart, found the damage, but it's toast. Tried to salvage the heating resistance during disassembly but that just created more shorts along HR1 , I think the insulation is cooked.

At least RT1 survived, so there's that.

So I contacted Peilican the wire manufacturer directly about 2332ADVFEP.009BL. Their response nearly gave me a heart attack - $700 for 300 meters or $500 for 30 meters. Has anyone actually had to bite this bullet? Please tell me someone's found a more reasonable source or figured out a way to buy smaller quantities. If we have no alternative we’ll go this way .

This is the 7th time I've had this particular unit on the bench - mostly caps, transistors, the usual suspects. Each time it's come back to these levels so for me it’s fine.
2nd harmonic
error
control
5MHz
photo I
OSC oven
cell oven
lamp oven
supply
30
0
-2
6
40
0
28
26
40



Once I get this thing breathing again, planning to do Corby's Super conversion on it.

A couple other things while I'm asking - are there any mechanical drawings floating around for the 780nm lens mount hardware for the Super mod? And has anyone tried different optical filters and found one that works particularly well? I mean there a reference for an Edmund lense 780nm CWL, 25mm Dia., Hard Coated OD 4.0 10nm Bandpass Filter , but would a tigher bandwith be better ? Like this one FBH780-3 from thorlabs which is 3nm bandwidth.

I’m not an optical guy so I can’t tell much about this.

Also, since my A12 oscillator is pretty well cooked, has anyone ever reproduced the A12A1 PCB and managed to salvage just the A12A2 part (lamp and coil)?

Any thoughts or war stories appreciated. This one's turning into quite the project.

Thanks,

Jean-charles BILLEBAULT

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