Those are indeed awesome and I have several of them for frequency standards in my home lab. They go for about $100 US in a known-working used state on eBay. Unfortunately, I don't know whether the OP would be willing to put a piece of used equipment in these units he is building (presumably commercially). Also, Rb standards have a definite running life - my understanding is about 10 years of continuous operation. This is due to an interaction of the Rubidium vapor with the glass bulb it is contained in. The vapor slowly diffuses into the glass, if I recall correctly, which eventually no longer leaves enough Rubidium to produce the glow discharge from the lamp. Most of the used oscillators on the market have only about half their life remaining, sometimes less. This could mean only 2 or 3 years of service before it quits. It does provide a signal which indicates the strength of the emission from the lamp so you can get a warning pretty far in advance that the unit will need replacement. All that said, a clock based on one of these wouldn't even lose 1 second over the entire 10 year life of the unit! Sean On Tue, Jan 1, 2013 at 6:25 AM, Nily wrote: > May be Rubidium Atomic Frequency Standard will help, not so expensive. > Cheers > > Rodolfo > > -----Mensaje original----- > De: piclist-bounces@mit.edu [mailto:piclist-bounces@mit.edu]En nombre de > Bob Axtell > Enviado el: Martes, 01 de Enero de 2013 02:44 a.m. > Para: Microcontroller discussion list - Public. > Asunto: Re: [EE]Re: GPS Modules > > > On 12/31/2012 9:12 PM, RussellMc wrote: >>> The field units are mains powered and draw hundreds of Watts each, so >>> the added power consumption is not an issue. >>> >>> As long as the GPS locks onto the time within 5 minutes, I think the >>> client will be happy. They are also fixed position. Would this help? >>> >>> Maybe using the GPS to correct an on-board RTC would be a better option= ? >> Off the cuff low practical experience comments: >> >> GPS sounds like a good fit. >> A typical GPS (all modern?) maintain ephemeris (short term) and >> almanac (long term) satellite constellation descriptions to allow >> location of the satellites. If the system is always on the time signal >> is essentially continous. If you turn the system on regularly and >> frequently (say hourly to maybe a few hours) the ephemeris information >> is usually relevant enough to allow reacquisition in under about a >> minute and sometimes much less. If you turn the system on daily or >> weekly it can take minutes to acquire enough satellites to start >> producing information. (1 =3D somewhere apparently :-), 2 =3D a torus, = 3 >> =3D where you are or out closer to lunar orbit, 4 =3D where you are, 5+ = =3D >> better.). >> >> Away from cities you are unlikely to seldom experience catastrophic >> loss of enough satellite signals to lose the time signal. >> >> So, if you have a RTC that has enough stability for at least 10 >> minutes (not exactly hard even on the far side of the black stump on a >> hot Australian summer day) then a GPS fix "when required' will keep it >> stable enough. >> An RTC less stable than that or no RTC would benefit from permanent >> GPS operation. >> >> >> http://en.wikipedia.org/wiki/GPS_signals >> >> >> R > > h > I am involved with solar tracker design. In order to determine sun > position, I must determine the exact UTC time. We keep a supercap-backed > RTC which is updated daily at 4AM. > > After a lot of research, I determined that WWVB ("Atomic" time service > at 60khz) was too sensitive to cellphone and lightning interference. A > pity, since WWVB receivers use only a milliamp or less, while GPS use as > much as 100ma. We are forced to use GPS receivers. > > We keep the GPS unpowered until 4AM local time, which is usually a quiet > period. We keep the unit powered until we receive at least 3 good nmea > time passes. I then average those, update the RTC, then turn off the GPS. > > The time signal is accurate enough for solar purposes from just one > satellite. As a consequence, we use older less-sensitive $16 GPS modules > crammed into a tiny plastic, sealed waterproof box. We get time signals > in 3 minutes or so, but we check the source to make sure the pass is > from a satellite. > > -- Bob A > > -- > http://www.piclist.com PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > > -- > http://www.piclist.com PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .