Yes, "Ira" I do acknowledge that there is also long term frequency drift, too (I think I mentioned that, in my earlier piece). But my own experience overall is that long-term drift is a smaller error and over a longer time frame than absolute frequency or temperature effects.
Maybe anyone who "needs" their nixie clock locked to a cesium referrence needs to get out more? :) LW On Jun 2, 6:05 pm, Instrument Resources of America <[email protected]> wrote: > MrNixie, > Also keep in mind that regardless of temperature compensation, etc. > etc. a crystal and therefore crystal oscillators exhibit a frequency > change with AGE. This is why all crystal oscillators, no matter how > good, how well compensated for temperature and other factors, will > always have to be readjusted and or locked via one method or another to > a Cesium Beam reference. Ira. > > On 6/2/2011 2:01 AM, MrNixie (UK) wrote: > > > > > > > > > See how active this thread is, and how many options you have! My two > > cents (sorry - pence).... > > > Longer term (days, weeks), a mains referenced clock will be your best > > bet - your utility suppliers see to that in the longer term, for the > > benefit of the many MILLIONS of legacy mechancial devices that use the > > AC mains as their only reference. However, such a clock will gain and > > loose a few seconds throughout different times of the day, when the > > grid slows down a little under heavey load (think car going up hill), > > then speeds up again to make good the "lost time" during the night and > > late mornings. > > > If you want good absoute time, all the time, even with the actual of > > day time itself, then GPS or WWVB/MSF time code receivers are the way > > to go. But beware - radio signals can be unreliable, you will still > > need a way of maintaining time locally in the event that your external > > source fails for some reason. Also, you are looking at more cost, more > > wires, more software. > > > Simple Crystal oscillators that are designed just to clock your CPU > > can be PANTS (quaint English expression meaning "not particulay > > good"). They will let you down in three ways; absolute frequency, long > > term drift and Temperature. Sure, some clever software can correct for > > absolute frequency errors once you know what the error is, but no > > simple solutions that I have seen will accomodate temperature drift, > > which can be significant. > > > Which brings us to Temperature Controlled Crystal Oscillators (TXCO). > > These can be bought now as IC's that run cold (no heating > > requirements). The on-board precision crystal is already laser tuned > > to be pretty damn near spot on (and is also still controllable in > > absolute terms using internal software settings), AND an onboard > > temperature sensor adjusts the capacitance across the crystal to > > compensate for temp changes. Check out the DS3231 from Maxim. It is > > only available as an SMD part, I beleive, but I have hand soldered > > these without issue. > > > This little device is quite capable of offering better than 15 seconds > > a year accuracy - more if you have the time to fine tune it in the > > first place. Maxim call it "Extremely Accurate", which for a general > > clock, it is. It leaves similar simpler crystal controlled RTCs in the > > dust for accuracy. And you get battery backup, of course. > > > I have built clocks around ALL of the above options. They all work, > > within the limits mentioned. My favourite is now the DS3231. For all > > practical purposes (of mine), it makes the GPS or WWVB route somewhat > > redundant. > > > That was probably THREE cents' worth :) > > > On Jun 2, 3:39 am, threeneurons<[email protected]> wrote: > >>> in the US, line frequencies are its adjusted > >>> whenever it hits 20s/10s of error. > >> :http://en.wikipedia.org/wiki/Utility_frequency. > > >>> - Robert > >> I've used line frequency for decades, and I only touch those clocks > >> twice a year. And usually that's only to adjust the hour (Daylight > >> Saving). Rarely touch the minutes button. The correction scheme > >> described on Wiki should be satisfactory for long term accuracy. Short > >> term, who cares. > > > > IRACOSALES.vcf > < 1KViewDownload -- You received this message because you are subscribed to the Google Groups "neonixie-l" group. To post to this group, send an email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/neonixie-l?hl=en-GB.
