Hi, On 2021-03-04 11:15, Hal Murray wrote: > [email protected] said: >> But since the units been off for at least 15 years and heavens knows how long >> before that. Any thoughts on how long it might take to stabilize days weeks >> months. Its pretty stable as is just wondering. > It's an exponential tail. How stable do you call stable? > > What do you have to measure it with? How stable is your environment? > These tails have been measured for years. As long as it is slower than the application can handle, it becomes only the issue of being "within range". This is the "aging" period. Environmental may cause more severe deviation eventually, but a power-down/power-up even cause retracing and that will restart the process.
So, it ends up being about being acceptably stable, as in not being annoyingly large anymore. There is fairly good evidence that crystal oscillators do not have much of wear mechanisms if done with a bit of care. There is oscillators out there that has been running 40-50 years in continuous operation. Getting any atomic clock live that long is more of a challenge, they need to be serviced because of explicit wear mechanisms. I think the PTB CS2 and CS3 clocks is the longest operations beam clocks, but they also have work done on them, and upgrades. We end up discussing some of that in the IEEE P1193 group on environmental effects on frequency sources. Cheers, Magnus _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com and follow the instructions there.
