I'll soon be getting an internet connection to my observatory, so I'm looking for advice on a better way to set my computer clock, possibly via the internet.
The observatory camera/mount control computer runs XP. CCDSoft and SKY6 get the time from the computer clock. Currently, at the start of the evening, I set the computer time manually to an "atomic clock" which connects to WWV by radio.
I have satellite based internet at the house. The connection to the observatory will be from the house via "ethernet over power" for about 1000 feet. What internet based time services are observers using? Would my proposed new connection still provide accurate time via the internet?
I'd also like to know if it's possible to use GPS directly into the computer. I already have a Garmin GPS unit which connects to a Losmandy mount (Gemini I) to set position and time. Perhaps this could be adapted to work directly with XP. I wouldn't be averse to buying another GPS unit if there is already one made specfically for this purpose.
Phil
Especially on an XP machine, use Dimension 4. Free. Works well and many observers use it. A few of us have found an issue on Win7 machines but fine on my XP machine. I have used it for years. My accuracy is reportedly 0.02s
Ken.
Meinberg is another option. It is a good option if you plan to use your own GPS.
Barbara
I agree with Barbara that Meinberg NTP (a port of Unix NTP daemon) is a real NTP client/server and in principle would be a Good Thing. I'm using it at one of our observatory's telescope.
Dimension 4 and other similar programs behave a bit differently - they just force time change. NTP client determines computer's clock behaviour and corrects clock rate according to that. In principle, a NTP client should give better results.
However, several image acquisition programs and/or camera drivers are written in such way that during the image read-out from camera, computer's clock is more or less effectively stopped. Taking thousands of short frames in many hours, one may find that without synchronization, time drifts even few tens of seconds (my most awful experience with that was a SpectraSource HPC-1 camera + Win95 + 486 computer loosing 1 second during ~20 second readout time). Probably this is not a big issue when observing Miras or Be stars.. but how about delta Scuti and other short period stars? Unfortunately, NTP daemons usually are not able to correct such behaviour well enough and IMHO D4-like programs are the only way how to deal with the problem.
I have Andor Ikon-L camera at one telescope, Andor offers a software called Andor Solis to control it. With this program, I have no problems with computer clock and NTP works just fine, taking into accound clock changes due to temperature changes etc (that computer runs 24/7). In the same computer, using MaximDL to control the same camera, clock slows down... Should I blame MaximDL? Camera driver? I just don't know.
Best wishes,
Tõnis
I manage time with the NIST (National Institute of Standards and Technology) Internet Time Service You can downoload their free Windows application - nistime-32bit.exe. I use it on my 32 bit laptop running XP. I don't know if they have a 64 bit version. At the begnning of an imaging session I launch it, query one of their servers, and it sets my computer clock. Then I set it to do the same automatically every hour and log the clock deltas to a log file.
Scott
For those using Dimension 4, is there a rule of thumb for how often you should synchronize?
Blake
I use 15 minutes for the refresh rate in Dimension. My reported clock corrections are about 0.02s. At some point you annoy the time server folks if you're always pinging them!!?? ;-) The corrections should be probably be less than 1/100 of the period of your target. So not really a problem unless your doing something really fast and sophisticated!?
Ken
I am using NetTime 3.14. It works on any Windows system from 95 to date. 32 or 64 bit. You can set the intervals between re-synchs. I normally set for once per hour. I am using an SBIG camera and Maxim DL 5 or 6 and I don't have the extreme loss of time problems that were mentioned in one of the posts above.
It is free and open source. I am no expert on time synchronization but this seems to work just fine.
Ken, I think I got the recommendation for it from you at the Spring AAVSO meeting, but memory is increasingly unreliable as the years accumulate. What was the question again?
Brad Walter.
See, I forgot the attachment with info on the program
HI all
If you want to see just how accurate your system is, Fountain Computer Products has a nice program to test most systems. You can download it from http://www.fountainware.com/Products/TimeCalibrator/index.htm.
Robert Jenkins
That you all for your responses. That's all good information. I'll probably start with Dimension 4.
Phil