[time-nuts] GRAIL USO
Jim Lux
jimlux at earthlink.net
Sat Jun 1 20:03:10 EDT 2013
On 6/1/13 4:50 PM, Hal Murray wrote:
>
> jimlux at earthlink.net said:
>> It's also the knowledge of the process yield at each step which means you
>> can stay in business. APL knows how many to start at the beginning to
>> insure they'll have 4 at the end, 2 years later.
>
> I assume there is a distribution. Anything published on that? How good are
> the next 4?
>
What's in Weaver's paper(s) is what's publically available.
As I recall, they had quite a few candidate oscillators to choose from
that were going through long duration burnin. There's a sort of complex
process to pick the best of the litter.. do you want the one that shows
the lowest drift rate, the one that's closest to the desired frequency,
the one that has the lowest DC power, the highest RF output power, etc.
Turnover temperature is a critical thing, because it drives the DC
power required to run the oven, and whether it can meet the mission
requirements.
Low temp is good from a power standpoint (minimum heater power
required), but bad from a "margins" standpoint: If it has to work at
+70C, and the turnover temp is only +75, that's a challenge. If the
turnover temp is +100C, then you've got 30 degrees to work with, but
you're going to draw a lot of heater power, esp when you're at the -55C
test temp.
So the ones chosen for GRAIL were "best" by some criteria (e.g. best for
the measurements GRAIL was making), but perhaps not "best" by some other
criteria.
You can bet the "spares" at every step of the way aren't just thrown in
the dumpster. They wind up being used for future applications,
sometimes because they're actually better for that than for GRAIL, or
perhaps the application is willing to take what they can get.
I don't remember all the details (it was several years ago), but the
burn in and test is of part of the system, and then they choose some to
take the rest of the way and put into a package, etc.
More information about the time-nuts
mailing list