[time-nuts] an interesting problem

jimlux jimlux at earthlink.net
Sat Feb 5 16:08:14 UTC 2011


On 2/5/11 7:30 AM, Poul-Henning Kamp wrote:
> In message<4D4D6BF6.8070202 at earthlink.net>, jimlux writes:
>
> First, this isn't that different from the analysis used in the NTP
> protocol, so you should read that with an open mind.

That is what got me started thinking it was doable at all.
(PTP 1588 as well)
>
>> Or, given that the interval between ticks is one
>> of 28 or 29 discrete values
>
> If you can reliably bin your measurements in these discrete values,
> do that, and calculate stats for all bins with sufficient data
> points to say something meaningful.

I guess a test is in order...
I've got my hardware
I've got my counter
I've got my logging computer with GPIB
I've got Matlab..

The way forward is distinct, if not entirely clear.

to the laboratory!
>




More information about the time-nuts mailing list