[time-nuts] HP 5065A performance vs. others
lists at rtty.us
Mon Mar 22 02:25:46 UTC 2010
Extra mu-metal shielding on the physics package might do more than just help with external fields. You have all sorts of variable currents running around in the box. They will generate fields that could indeed disrupt things at some level. Greater room to space things out would help with this kind of thing as well.
Has anybody characterized the noise performance of the photo detector in the 5065? As I recall it's one of the things that creeps into the signal to noise equation and thus into stability.
On Mar 21, 2010, at 8:21 PM, Magnus Danielson wrote:
> John Miles wrote:
>>> -----Original Message-----
>>> From: time-nuts-bounces at febo.com [mailto:time-nuts-bounces at febo.com]On
>>> Behalf Of Bob Voelker
>>> Sent: Sunday, March 21, 2010 1:59 PM
>>> To: time-nuts at febo.com
>>> Subject: [time-nuts] HP 5065A performance vs. others
>>> As several postings have indicated, the performance of the HP
>>> 5065A is better
>>> than many of the other rubidium standards. What enables the HP 5065A to
>>> achieve better performance?--Is it the physics package or the particular
>>> control system implemented in electronics? Would it be possible to
>>> achieve the HP 5065A's performance by modifying a more commonly
>>> available rubidium such as the LPRO? Would an ensemble of LPROs
>>> match a single HP 5065A in performance?
>> I think those are all open questions, because it's not immediately clear
>> what limits the performance of the smaller telecom-grade physics packages,
>> or what low-hanging fruit might be left on the tree.
>> For instance, how important is the length of the path the light takes
>> through the filter cell and/or resonance cell? It's obviously a lot longer
>> in a 5065A.
> I think you would find this article of interest:
> Especially page 446-447.
> Another mentioning of optical path length is here:
> This Litton contribution goes into more detail about ligth-shift and path length:
> Read the Ligth shift chapter starting at page 679, then check figure 4.
> These is more if you care to search for Rubidium and light shift.
> There is considerable amount of efforts spent on reducing light shift. It has many sources, including lamp temperature, filter cell temperature, filter cell length etc. etc. Filter cell temperature-stability is a key issue from what I understand.
>> the difference? Is there something special about HP's lamp? Is their microwave synthesizer that much better?
> Could be part of it.
>> HP's temperature stabilization is better than the LPro's -- so maybe it
>> would help if you just moved the LPro's Rb assembly into an outer oven,
>> separate from the rest of the electronics.
> Temperature stability is a key issue as it connects via filter cell light shift, lamp intensity (which spectrally needs to match that of the filter cell... they do not match very well actually and the filter is a bit steep), then buffert gas frequency pulling alongside of the wall-shift. Then toss in the cavity detuning and the frequency pulling effect that it causes.
>> How important is all that
>> mu-metal shielding on the 5065A, given that most people these days would
>> care more about stability than absolute accuracy (thanks to GPS)?
> The uniformity of the C-field is an issue, and if it is allowed to change (due to externally shifting fields) well...
>> Someone with more free time should tackle these questions. :) F. G. Major's
>> book would be a good starting point, and this paper on laser-pumped Rb
>> clocks also has a lot of hints about what limits the performance of ordinary
>> sources: http://tf.nist.gov/general/pdf/1219.pdf . They used the cell from
>> a commercial Rb standard in their experiment, although they didn't say which
>> one. If nothing else, you can infer from this paper that the path length
>> through the resonance cell isn't a huge deal.
> Depends on how it balance with other things. See above links.
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts