[time-nuts] ensemble of standards
lists at rtty.us
Sun Dec 30 03:39:46 UTC 2012
Having written code for a 4K PDP-8, I suspect that more of the heavy lifting was hand optimized assembler rather than Fortran. More or less, compile it / it doesn't fit / poke into it / fix it / recompile it / poke it some more / give up & hand optimize the problem areas.
On Dec 29, 2012, at 10:32 PM, Magnus Danielson <magnus at rubidium.dyndns.org> wrote:
> On 30/12/12 02:46, cdelect at juno.com wrote:
>> Also about ensembles,
>> This article describes combining 4 HP 5071A standards.
>> Description on page 2 and block diagram in figure 2.
>> Unfortunately no details on the controller are given!
>> It's what I'd like to do to two or more HP 5065A
>> Anyone know what they did in the controller?
> I'd assume that this is similar to what is described in Chapter 9 of NBS Monograph 140, by Dr. David Allan and friends:
> When reading the FORTRAN code, one has to realize that the ER calculations is hard-wired to the way that the NBS setup at the time where done, so that the 8 clocks was compared by a number of pair-wise measures and that the time-differences is combined as needed where some of the clocks becomes transfer standards in the comparison back to the primary reference standard. Clock 3 is the main clock in the hookup.
> A modern view on things is found here:
> The original time-scale algorithm ran on a PDP-8 machine with 4k word memory, coded in Fortran by Dr. Allan. Today they run using a pair of redundant 386 machines and using a blend of Jim Greys buffer and mixers and that of commercial time-interval measurement rig.
> 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