[time-nuts] AN/URQ13 reference AT cut crystal? ==> Crystal Robot

[Bob kb8tq]

Hi

One thing with quartz is that the acoustic Q gets better as 
the frequency gets lower.  If a 5 MHz 5th overtone has a Q of 
3M (say), A 1 MHz 5th *could* have a Q of 15M. That assumes
you fully scale the package to keep the blank at the same 
relative size. It also assumes some other effect doesn’t come
along and mess with a simple approximation ….. (simple 
approximations often fall victim to that sort of thing ….).
Often the package grows as you discover this or that. 

The obvious issue here is that your ~ 22 mm diameter cold weld 
package is now up around 100 to 150 mm. Nobody seems to want 
to buy a part that big ….. Which isn’t quite true, sapphire resonators 
(with a higher Q) do get sold. 

In order to do an experiment, one would need to grow high quality 
quartz bar 5X larger than what we have today. It grows at a 
“some number of mm a month”. You go from a (maybe) 6 month 
grow process to (maybe) 3 years. The whole autoclave is tied up 
for that period. You would pay the same total $$$ for one bar or 
a whole lot of bars. Tooling to accommodate the larger bars likely 
would also get into the act. 

Moving to 3 years is only a guess. The process has been “optimized”
to deliver good Q at 5 MHz and above. There is no guarantee that 
this or that “speed” would be adequate for 15M Q. There is a lot of
evidence that slower is better. ( = if you use fast quartz in your 5 MHz
5th parts, their performance is poor ). 

You would be chopping a lot of quartz to do all the work to come 
up with a “production” ( = useful) design. Having more than one bar 
probably *is* a good idea. 

To the extent that Q and ADEV are related, one could expect other
useful specs to improve along with the Q. You also might find out 
that for this size / precision  blank, 3 years is “to fast” to get good 
ADEV …. how would you know that … hmmm .… run another autoclave
 .... (all that after you had tooled a giant cold weld package /process 
machine /welder to do your testing in … yikes ….)

Fun !!

Bob

> On Feb 18, 2021, at 10:40 AM, Poul-Henning Kamp <phk at phk.freebsd.dk> wrote:
> 
> --------
> Bob kb8tq writes:
> 
>> Even if you went with a different material, moving any of the specs listed
>> above by 10X would be pretty amazing.
> 
> That was the answer I expected :-)
> 
> What about geometry ?
> 
> Would things like whispering gallery improve things if we found a reasonable way to produce them ?
> 
> -- 
> Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
> phk at FreeBSD.ORG         | TCP/IP since RFC 956
> FreeBSD committer       | BSD since 4.3-tahoe    
> Never attribute to malice what can adequately be explained by incompetence.