Measuring Drift

Wrong track altogether. PC clocks are junk at the resolutions you're talking about. They are explicitly banned for use as a time base in every sport for which there are timing hardware rules, because they are crap.

NTP is also completely useless at the resolutions you're talking about, due to the vagueness of a TCP/IP round trip.

Any drift analysis which uses NTP or a PC's clock is immediately compromised, before you even start.

A pretty good tool for measuring drift is a FIS-homologated start clock, which you can program to send a GPI at every TDC.

If you want to manually adjust the oscillator on the start clock to an even higher rez than the max for which it is certified (so that it is a "perfect" reference), you're going to have to buy or borrow some really good equipment, or else send it to one of the FIS-certified homologation labs (such as University of Berne) for some tinkering.

Yep, got it.

I totally agree about using PC (even with NTP or GPS sync) for general sports timing, but I'm not sure it so clear cut for measuring drift.

A well-disciplined clock (define well?) will be constantly corrected to maintain zero drift. Over small regions there is significant jitter and noise, but over a long period, the drift will be elimininated. For example: Once synced, your pc might be less than 0.1s off the reference time. One week later, one month later and one year later... it is still about 0.1s off.

An undisciplined clock will not be corrected. With enough samples over a long enough period, it's drift will begin to 'appear' through all the noise and jitter.

I'm just trying to find a nice 'backyard' way if identifying problems with ageing equipment, that does not rely on comparisons to other ageing equipment and which does not require purchase of new. It's only a dream...

Thanks for the infos!

It all depends on your budget and on your time.

You may want to put some $9.400 and have an ExactTime GPS device. A 'must have' for professionals.
http://www.symmetricom.com/pro...frequency-receivers/

If, as I may imagine, you want a simpler and cost-effective device, excellent choices are the Digitech 'Syncro GPS' and the Microgate MicroSync.
http://www.digitechtiming.com/...i.php?id_category=13
http://www.microgate.it/ita/ti..._sport/microsync.asp

Both provide excellent locking to the GPS constellation and you may then check on the long term the drift of your equipment. Both sell for approx $800-$900.

Wanna go "el cheapo"? Boy, you are lucky. There is also a $25 solution.
You need any 'PPS-enabled' receiver such as those from FasTrax, Flavus or Navman.
That makes 99.9% of your cost. Through a dealer, you may get it starting at $22-$24. If you go into any online shop, you may need up to $40.

You need then a NPN transistor (2N1711 or similar), 2 caps, some resistors and an iron.

You have now got the simplest but still quite good 1 Hz source.
Reliable and cheap. You will see how much even the most sophisticated sports clock will drift from the GPS time base.

You should however be aware of this facts:
a) GPS is considered not very good in the short term but performs quite well in the long term. For sport clocks.... far more than you will ever need;
b) Remember that you now own a $25 receiver, nothing less, nothing more.