[time-nuts] Determining Time-Nut infection severity.

Magnus Danielson magnus at rubidium.dyndns.org
Sun Oct 24 22:40:12 UTC 2010


On 10/24/2010 11:57 PM, jimlux wrote:
> Mark J. Blair wrote:
>> On Oct 24, 2010, at 1:58 PM, Poul-Henning Kamp wrote:
>>
>>> In message <53601812-FC62-4098-9266-55EDF50AFB2C at nf6x.net>, "Mark J.
>>> Blair" wri
>>> tes:
>>>
>>>> Does anybody know how many bits of precision are used
>>>> on tombstones these days?
>>> Most tombstones are engraved with CNC machines and while I have yet
>>> to see a company advertise it as a competitive parameter, your vital
>>> lack of stats will be engraved with better than 1/50mm precision.
>>
>>
>> Oh, I was mostly concerned with the bits of precision used for the
>> birth and death timestamps.
>>
>>
>>
> I should think that if you have left instructions with your executor,
> they can have any arbitrary number of digits engraved, including an
> estimate of uncertainty, if needed. You could even have them estimate
> the Allan deviation of the various clocks involved, and have the
> requisitite plots engraved as well.

Considering that this is a time thing, you want the TDEV rather than the 
ADEV plots, alongside an indication of time since last calibration which 
is the index into the TDEV plot. This then adds to the TDEV of the chain 
giving the calibration along side a separate traceability of systematic 
errors in the calibration chain and ...

Proper calibration of time is a beast.

Cheers,
Magnus



More information about the time-nuts mailing list