Hi > On May 21, 2016, at 7:18 PM, Magnus Danielson <[email protected]> > wrote: > > Hi, > > On 05/22/2016 12:11 AM, Richard (Rick) Karlquist wrote: >> >> >> On 5/21/2016 11:25 AM, Attila Kinali wrote: >> >>> I'd give a lot to read the design documents of the 5071. >>> There must be a lot of knowhow and techniques in them. >> >> Read papers by me and my colleagues at the 1992 Frequency >> Control Symposium. There is nothing else in the public >> domain. >>> >>> >>> What reference did you use to measure its absolute accuracy? >>> Did you have access to the NIST standards? >>> >>> Attila Kinali >> >> At that time, HP had an ensemble of 5061B's that they >> compared to LORAN and it was somehow traceable to NIST. >> It was claimed to be the most accurate clock on the >> west coast at the time. >> >> Also, one of the first pilot production units was >> actually taken to Boulder and compared directly with >> NIST's frequency reference. I remember that it was >> off by a few parts in 10^13, after correcting for >> the gravitation effect of Boulder at 5000 ft vs >> Santa Clara at sea level. >> >> Over time, customers like NIST and the Naval Observatory >> accumulated data on 5071A's to establish their accuracy. >> At one time, 85% of the weight of the TAI (International >> Atomic Timescale) consisted of 5071A's. Jack Kusters said >> that, as a group, the 5071A's were accurate to around >> 10^-14, because there were no significant systematic errors. >> Hard to argue with that. > > Never the less NIST and USNO keeps them in temperature and humidity > stabilized environment with well stabilized power. Just to keep such > systematics out of major harm.
But then again, not everybody can afford a cave to mount their gear in …. hmmm … maybe … Bob > > Cheers, > Magnus > _______________________________________________ > time-nuts mailing list -- [email protected] > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
