[time-nuts] hm H Maser

Bob Camp kb8tq at n1k.org
Mon Jan 9 07:56:28 EST 2017


> On Jan 9, 2017, at 1:05 AM, Bruce Griffiths <bruce.griffiths at xtra.co.nz> wrote:
> For a rubidium vpour standard a cavity is essential, one could always use a 
> microwave horn to illuminate the cell in an anechoic chamber.

The cavity in an Rb is not the ultra high Q monster that you have in a Maser. There is 
no magic “minimum Q” requirement to get the beast to work. That’s by no means saying
that the cavity is unimportant or trivial. The point is only that it’s about 2 orders of magnitude
easier to make up the required cavity for the Rb.

It should be noted that cavity != shielding and that cavity != temperature control. It is simply 
the microwave resonant structure that gets the electromagnetic doing the right thing. In both
cases you still need (very) good magnetic shielding, pressure shielding, and temperature 
control. In the Rb case, you need to set up specific temperatures to get things to work in
each cell. In the Maser case you simply need the “right” temperature for your setup. 

Traditionally one of the big deals about both devices was the synthesizer required to convert
the physics based frequency to something useful. With the Maser the frequency is pretty
much always the same number. That gives a simpler synthesizer in terms of tuning. A whole
raft of this and that give you a range of answers for the Rb. That used to make the synthesizer
a bit of  a pain to design. These days, the synthesizer tuning the Rb requires is easily done with
a cheap DDS chip. Take a look at the 5065 manual if you want to see how much fun that
used to be …

Indeed the whole electronics side of both standards is easier than it once was. The temperature
probes in the compact Maser still are $1K each, but most of the parts you need on the electronics
side are pretty common items. Again, common parts != trivial design. You still need to get the
details right. Signal to noise *does* matter. You need to use the right design with the right parts. 

So how do you do this? The normal approach is to get a dozen or so people together and work 
on it 40 hours a week for about 5 years. You build up a series of batches of prototypes and 
get to the point you believe you have a design (1 in 10 sort of works). You then spend roughly 
another three to five years knocking the rough edges off of that design and making the first 
batch of real units. Beyond time and people there is the cost of parts, software licenses, normal
test gear, really weird test gear, and all the other stuff.  This assumes it is run as a business with
somebody managing the whole thing. Try to run it as a committee of the whole, both the cost
and the time will go up. Try to do it without the right tools, at lest the time will go up. I’d bet
the cost will go up as well…


> Using an integrating sphere can enhance the contrast of the optical signal 
> significantly.
> http://www.princeton.edu/physics/graduate-program/theses/theses-from-2011-1/bmcguyer_dissertation.pdf
> https://infoscience.epfl.ch/record/178228/files/IFCS_Invited_Talk_Finalpdf.pdf
> https://doc.rero.ch/record/32317/files/00002318.pdf
> http://tf.nist.gov/general/pdf/1154.pdf
> http://tf.boulder.nist.gov/general/pdf/1663.pdf
> Bruce
> On Sunday, January 08, 2017 10:20:33 PM Bruce Griffiths wrote:
>> Possible sources of Rubium vapour
> cells:https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=1470
>> http://www.precisionglassblowing.com/custom-solutions/optical-glassware/vapo
>> r-wavelength-reference-cells/
>> https://www.sacher-laser.com/home/lab-equipment/spectroscopy/reference_gas_a
>> nd_vapor_cells/reference_gas_and_vapor_cells.html
>> Bruce
>>    On Monday, 9 January 2017 11:14 AM, Bruce Griffiths
>> <bruce.griffiths at xtra.co.nz> wrote:
>> Bob
>> As long as one stays away from CPT and merely uses the laser as a
>> replacement for the traditional rubidium lamp plus filters it should be
>> easy enough as one doesnt need to modulate the laser at 3.4 GHz.I was
>> thinking something along the lines of the recent PhD thesis that gave all
>> the detail required to duplicate their low noise rubidium standard that 
> was
>> quieter than am HP5065.One could easily substitute ones own ECDL 
> (These can
>> easily be constructed from commercially available parts) and improve
>> somewhat on the performance (The oven design of most commercial 
> ECDLs seems
>> suboptimal). Bruce
>>    On Monday, 9 January 2017 10:23 AM, Bob Camp <kb8tq at n1k.org> 
> wrote:
>> Hi
>> The large diameter Rb cells are a bit harder to come by than the more
>> generic telecom sized cells. I suspect you are correct and they are out
>> there from somebody.. The real advantage you would have with an Rb is 
> that
>> the design you do is gigantic compared to what everybody is doing 
> today.
>> Their constraints are not your constraints.
>> Based on the laser driven Rb on my bench …. don’t bother with that part 
> of
>> it. It is indeed doable. Doing it in a fashion that gives you a better
>> standard …. not really easy at all.
>> Bob
>>> On Jan 8, 2017, at 3:55 PM, Bruce Griffiths 
> <bruce.griffiths at xtra.co.nz>
>>> wrote:
>>> The rubidium standard appears much more manageable given that 
> the cavity
>>> dimensions are somewhat more compact and rubidium vapour cells 
> are
>>> readily available. Substituting a laser for the lamp should also help in
>>> improving the reliability. However an ECDL laser locked to a rubidium
>>> line is required for a double resonance setup. Building ones own ECDL
>>> doesn't appear to be particularly daunting, however low noise drive
>>> electronics will be required. All the necessary optics are off the shelf
>>> items. One still has the issue of  the frequency pulling due to the
>>> presence of the vapour cell. Bruce
>>> On Sunday, January 08, 2017 10:22:54 AM you wrote:
>>>> Hi
>>>> I guess the question then would be:
>>>> Is a H Maser that runs 6.6 x 10^-12 at 1 second worth the trouble?
>>>> With 100 KHz / C temperature coefficients running around, getting
>>>> good stability in a real world setting at 1 day will be “interesting”.
>>>> Just for reference:  The MH-2010 data sheet shows 1.5x10^-13 at
>>>> 1 second for the “cheap” version and 8x10^-14 at one second for
>>>> the low noise version.  Data showing the 5065 Rb at 1x10^-12 at
>>>> 1 second is running around on various web sites.
>>>> The NIST paper suggests that they made several prototypes before
>>>> they got one good one working. That’s a lot of “fun and games” with
>>>> ceramic machine lathes and Rb magnetometers…..
>>>> The punch line being - would the same effort / cost / many years of 
> time
>>>> be
>>>> more fruitful (ADEV wise) doing a large package Rb (like a 5065) ? 
>>>> Based
>>>> on the number of people making them in volume over the years, 
> Rb’s
>>>> appear
>>>> to be the easier item to debug, design, and build.
>>>> Bob
>>>>> On Jan 8, 2017, at 6:01 AM, Bruce Griffiths
>>>>> <bruce.griffiths at xtra.co.nz>
>>>>> wrote:
>>>>> You could try a cavity like the one
>>>>> in;http://tf.nist.gov/general/pdf/156.pdf
>>>>> This avoids the requirement for a fused quartz storage bulb.
>>>>> Bruce
>>>>>   On Sunday, 8 January 2017 11:33 PM, timeok 
> <timeok at timeok.it> wrote:
>>>>> Hi,
>>>>> the thought of being able to work on building a H Maser has 
> always
>>>>> accompanied me in recent years. I fully understand the many
>>>>> difficulties
>>>>> of this project and also the necessity of a work team. Maybe a 
> Passive
>>>>> Maser would be easiest to implement, but I do not know in detail 
> the
>>>>> processes of construction of the physical part of the 
> interrogation.
> _______________________________________________
> 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 mailing list