SEQ: Naked eye asteroidal occultation 10 Oct

[citivolus (Ric)]

Hi Paul.

Basically the way that windows sets its clock when you use the Internet time sync option is that it periodically polls a time server and steps the system clock to match the time that it was told that it is. This results in discontinuities in the time, or the same time occurring twice, etc.

The better way to control the clock is to adjust the actual clock rate, so that the clock is running at the proper speed in the first place. This is the method used by NTPD, however it can take a fair amount of time to do clock adjustments using this method. In some systems, such as my laptop, the system clock is bad enough that NTP can never establish a steady state and NTPD can be a lost cause. This will happen whenever the system clock rate is off by over 0.5%, or whenever it does not run at a consistant rate.

GPS receivers are capable of providing an accurate time stamp as they are constantly receiving a time signal from each satellite that they can see. Most of them provide the current time in NMEA format over a serial connection, however, this time can be off by almost a second due to various factors, and is not reliable. To adjust for this, some GPS receivers have an output available on them that will proved a signal, once per second, which can be used as a time reference. This is known as the PPS signal (pulse per second), but it isn't available on most consumer grade receivers. The Garmin 18x LVC is one low cost receiver that does provide this signal, though. It comes as a bare receiver which you can connect to a 9 pin serial connection to your PC, or you can buy it from some places (such as the one linked) with a serial and USB connection on it, serial for data/PPS and USB for power.

Your average USB GPS unfortunately won't be able to provide the signal that we need.

Your easiest option if you simply want a fairly accurate system clock and don't have a PPS GPS receiver would be to install NTPD for Windows, and configure it to use the Australian NTP pool servers. This would be acheived by using the following ntp.conf file (in C:\Program Files\NTP\etc\ntp.conf):

server 0.au.pool.ntp.org
server 1.au.pool.ntp.org
server 2.au.pool.ntp.org
server 3.au.pool.ntp.org

driftfile "C:\Program Files\NTP\etc\ntp.drift"

However, your system clock could still potentially be off by a significant amount if it has not been running for several hours with a reliable network connection. The Meinberg NTP monitor allows you to see how much your "offset" (clock error) is in milliseconds from the servers, so that you can see if you have a good synchronisation.

I'd be happy to go over details if you want, via PM

Regards,
Eric

[citivolus (Ric)]

I should probably mention that none of the above is necessary to simply observe this event. All of that is intended for super-accurate timing, but binoculars and a stopwatch in this case will get you 90% of the way there

[[1ponders] (Paul)]

Cheers, thanks for that Eric. My connection is not stable enough for NTP I don't think. I can get more than a dozen dropouts a day on my satellite connection so I think if I want to get into this I'll need to look at the GPS option.

I don't think I'd be able to get the Garmin one in time for this event, but it is certainly an option I'll look into. I will be intouch. Thanks.

[[1ponders] (Paul)]

Also I guess if my capture frame rate isn't too high and the event is short enough frame drops could be kept to a minimum.

[jjjnettie (Jeanette)]

Originally Posted by [1ponders]

I don't think I'd be able to get the Garmin one in time for this event, but it is certainly an option I'll look into. I will be intouch. Thanks.

I've one you could borrow, but getting it to you in time, that's the problem.

[[1ponders] (Paul)]

Do you have the Garmin one? If you do then express post would get to me by mid week. I've got a GPS unit at work I could use (don't know if I can USB connect it though ) but I don't know if its one that is suitable and accurate enough. I wonder if one of the GPS units that plug into mounts would be suitable and if it could be used to plug into a laptop.

[jjjnettie (Jeanette)]

Sorry to get your hopes up, it's not the right unit.

[[1ponders] (Paul)]

Ah well, thanks for the offer.

[AstroJunk (Jonathan)]

You guys have been busy!

I think that beeperSync to set the pc clock and Gstar3 to record will get a +/- 20ms accuracy, which is outstanding. I'll give instructions later of how to achieve this (and I need to do some calibration tests too).

This star is brighter than a Galilean satellite - so there will be no issues of capturing it live on a gstar/dmk/toucam/malincam/watec live even on a 80mm scope i would have thought.

More later - I have chores to do now

[[1ponders] (Paul)]

Jonathan, I've downloaded beeperSync but I'm having trouble getting a Sync. Maybe you can tell me what I'm doing wrong.

1. set priority high
2. set connection broadband (I'm non satellite. I've tried the dial-up option as well)
3. selected ntp0.cs.mu.OZ.AU & ntp1.cs.muOZ.AU & tick.usno.navy.mil (which is what I normally sync my laptop clock to)
4. Click Get Tim Sync
5. Graph at the top (Syncronising and calibrating) goes and goes and goes (a few minutes)
6 Message "Syncronization was not successful. Please try again."

Just out of interest I tried to sync my laptop normally through tick.usno.navy.mil and had np.

[AstroJunk (Jonathan)]

Interesting...

I have just tried with mine and no issues with getting the time - However, on my Win7 machine i had to run as administrator to actually set the time. I can't test on vista, but my xp machine was fine too.

Could it be to do with security?

[[1ponders] (Paul)]

Not sure. I'm the only one that uses this machine so it starts up as administrator anyway. I wouldn't know where to start to look for security issues, however I did have firewall, antivirus, spyware and adware and all other security type prog turned off.

I'll give it another whirl.

[citivolus (Ric)]

Originally Posted by [1ponders]

3. selected ntp0.cs.mu.OZ.AU & ntp1.cs.muOZ.AU & tick.usno.navy.mil (which is what I normally sync my laptop clock to)

tick.usno.navy.mil, being located across the Pacific, will get you a lot of "jitter" or instability within your time source, due to differences in routing between subsequent network packets. The Australian servers mentioned are technically stratum 1 (or rather, top level, highest authority) and use a GPS reference clock, as is my server at home. I've spoken to the person who originally was responsible for setting up and maintaining those servers, as we know each other from a separate pursuit, and he has suggested that they may be of some questionable stability at this time, due to being hand rolled and using aging hardware. It was suggested that one not become too reliant on the existence of those UM servers due to a number of reasons, these included.

The CSIRO operates a number of time servers around the country, but access to them is by private individual agreement as they have been abused in the past. There are also many servers in the NTP "pool" which provide public access, but most of them are just referencing the UM and CSIRO server time anyway. I would also be willing to open my server up for queries from the astronomical community, once I relocate my GPS antenna.

In the short term, I would recommend using one of the recently updated AU servers from this list:

http://support.ntp.org/bin/view/Serv...TwoTimeServers

Specifically, Brisbane area people would be well served by ntp.wfsltd.com.au and ntp.sjkwi.com.au

If you are on Internode, ntp.on.net is also available. It is generally sub-ms accuracy.

A number of the servers which I have tested against across Australia are of dubious time accuracy, being off by anywhere from 2 to 17 milliseconds. This includes many top level (Stratum 1) servers, which are supposed to be accurate to the microsecond level. We don't really have a good time infrastructure for public access, unfortunately.

Regards,
Eric

[AstroJunk (Jonathan)]

A quick guide to FREE occultation timing!

Ok, the very best that one can hope for is a time of +/- 10 ms. That is simply a limitation of the cameras. We use a standard called PAL in Oz, and that runs at a frame rate of 25 fps and each frame can be further split into 2 fields and measured individually. This gives in effect 50 20ms exposures every second.

If you happen to have the occultation hardware such as KIWI-OSD, then the whole thing is made simpler by a GPS derived time which is overlaid on the video stream and can be captured by whatever device for later analysis.

If you don't have such a thing - fear not as a couple of freely available programs are available that will do the trick very well indeed.

The first is BeeperSync: http://hristopavlov.net/BeeperSync/ This program is designed to set a great device called a beeper box wich is used for occultation timing. Luckily for us, the way it does it is by first setting the PC clock to within a very high tolerance using internet time servers. Run it and press the Sync Time button. The software will tell you how accurate it judges to have made the sync. Repeat until within single figures!

The second program is Gstar3 (very recently upgraded): http://www.myastroshop.com.au/guides/gstar/ Not content with just producing the most sensitive video camera, the guys have also produced a pretty good capture program too. This latest version has a check box to add UTC timestamp in the session settings. The time stamp is added to every frame and to millisecond precision. Nice. HOWEVER - be aware that GSTAR3 can't bust thought the old 2Gig file size limit yet. That equates to only about 2 mins of full speed video, so don't start running it too soon.

If you get lucky, Limovie will help you analyse your trophy: http://www005.upp.so-net.ne.jp/k_miyash/occ02/limovie_en.html
but more of that later

There are some inbuilt latencies in the system which need to be subtracted from any measured time - the Gstar camera retards the signal by 20ms and the GSTAR3 uses the time of the end of the frame I think so you will need to subtract about 30 or 40 ms to get the best approximation of the actual event time, but overall a time accurate to +/- 20 ms is easily achievable with any PC with an internet connection. This is outstandingly good

[[1ponders] (Paul)]

Any advice for us non Gstar users Jonathan? I'll probably use a DMK

[AstroJunk (Jonathan)]

Oops, forgot to make it clear that the Gstar3 program should work with any camera, not just our beloved Gstars!

Hopefully that includes your DMK

[[1ponders] (Paul)]

I'll give it a go Jonathan. Bit of training to do before the 10th then.

[coldspace]

Way, way, way,way over my head now all this timming stuff. Its a computer science that I have trouble following.

I will just be happy if I get to see it happen on my CRT and capture it onto a video feed .


Matt.

[AstroJunk (Jonathan)]

Originally Posted by coldspace

I will just be happy if I get to see it happen on my CRT and capture it onto a video feed .

So will I Matt; so will I

[jjjnettie (Jeanette)]

Originally Posted by [1ponders]

I'll give it a go Jonathan. Bit of training to do before the 10th then.

It's a very intuitive program Paul, if I can use it anyone can.