In article , George george@least wrote:

"Robert Bonomi" wrote in message
...
In article , George george@least wrote:

"Morris Dovey" wrote in message
...
Dave O'Heare wrote:

So, what kind of watch was it? You make a brand name
suggestion, was that what you bought?

Yes - it's a plain Rolex from the days when the company built
watches accurate enough that they could be used for navigation.
The case, stem and band are plain old stainless steel. And even
though it doesn't understand Daylight Savings Time or time zones,
it hasn't missed a tick (5.00000/sec) for over forty years.

That's better than any other machine I've ever owned.


Yet the buck three eighty quartz jobs keep as accurate time, if not
better,

No, they do _not_.

A "good" digital will have a claimed error of less than 10 seconds per
day.
If you're lucky, for a specific watch, it will be under 3 sec/day.
Budget brands claim 30 seconds/day. The _really_ cheap ones don't make
*any*
accuracy claims whatsoever. And keep significantly "different time"
depending
on the temperature.

Furthermore, I have _yet_ to see a sub-thousand-dollar quartz watch that
can be 'adjusted' for minimum error.

A reasonable quality mainspring-based watch -- e.g. Seiko, Benrus -- is
capable of accuracy that is orders of magnitude better. Mostly because
it *can* be adjusted (the technal term is "regulated"). The trick lies
*entirely* in finding a good watchmaker; one who has the proper tooling.
I have mainspring wris****ches that keep time accurate to a fraction of a
second *per*week*. If the error rate gets to around 30 seconds
_per_month_,
it goes back to the watchmaker for tweaking. Typically, every 5 years or
so.



I'm sure you believe what you're saying, but, the "wrist chronometers" were
regulated for the average of three axes, and too long in any one reduced the
accuracy. They even included directions for regulation by placement in the
owner's manual. A true chronometer was regulated for one position only,
which made a great accuracy possible, but also made it a pampered, shock
isolated charge for a junior officer, who kept a log of errors to estimate
inaccuracy when a time check was not available.

I'm speaking from direct personal experience, over a span of more than 30
years, with 'quality' manufacturer, moderately priced (i.e. retail-priced
in the $50-125 range) "wris****ches" -- *NOT* the high-priced (and mis-named)
'chronometer' types.

My 19.95 Timex is still doing 1 second/day as it nears the end of its
battery life at four years old.

Congrats! You're one of the lucky ones who got a watch where the crystal
is close to the middle of the 'acceptable' range.

The Rolex (O-P) did the same for about two
years at a time before it had to go in for regulation and cleaning. As I
never had anyone who did the job - authorized only - guarantee accuracy
after the fact, I wouldn't know about your technician.

The jewelers (plural, in several states) I've used didn't "guarantee"
accuracy, either. They merely *delivered* it. grin It did help
_considerably_ that I could tell them fairly precisely how much it gained/lost
per period. e.g. I'd go in saying "it's running about 8-1/2 seconds/week
fast", and I'd get it back running 8+ seconds/week slower.

With a 'new' watch, it's typically taken me 4-6 adjustments, over a period
of several months, to get it fined down to the 'seconds/month' accuracy.

My 2nd watch, a mail-order $50 Seiko (in the early 1970s) came out of the
factory running 87(!!) seconds/day fast. 3 trips through the local
quality watchmaker/jeweler shop later, and it was approx. 1/2 sec/week.


I was/am involved in SWL as a hobby, so I would check against the broadcast
NBS/NIST time reference on WWV. This gave me 'interval' accuracy in the
sub-millisecond range, although "absolute" time had an offset in the tens
of milliseconds.