COSC Spec. Please explain. :)

[Dicko]

I have been looking at what COSC spec is.
When I google it, the definition seems to vary.
Some say -1 to +5 secs a day, some say -4 to +6 secs a day.
Some different again. How much does the age of the watch have to do with it all?
So where better to ask than here.
Being a new Rolex owner I am just wanting to know more about these beautiful timepieces.
So if some one could enlighten me I would be very appreciative.

[GradyPhilpott]

To gain chronometer status from Contrôle Officiel Suisse des Chronomètres (COSC), the watch must average -4/+6 over a range of positions and temperatures over a specified period of time.

Some watch companies, like Rolex, have even higher standards.

This is a good write-up on the subject.

http://www.watchonista.com/2914/watc...onometres-cosc

Like all mechanical devices wear will to some degree affect operation, in which case a regulation or even a full service might be necessary.

[Dicko]

Thank you.
I just wanted to be sure the watch I just bought was within limits.

[Dicko]

Well after originally checking I thought my watch was losing about 1 to 2 seconds a day.
I have since discovered it was my computer clock going a little fast.
After setting things up again and checking after 2 weeks and re-syncing my computer clock I am losing about 1 to 2 seconds a week. 3 seconds over 2 weeks.
Must say I am VERY happy with that.
Also checked against a website world watch and with my phone.

[Cerachrom12]

Unfortunately in real life, not every new Rolex runs at COSC specs. Plenty of forum stories about new watches either gaining more than 6 a day or losing more than 4 a day. And supposedly Rolex preaches to their watchmakers to never send a watch out the door running slow. But many do including one of mine, albeit it loses about 2 seconds a day. No biggie

[GradyPhilpott]

You need to stop using your computer to time your watch.

You need a serious reference like this one:

https://time.is/

[Dicko]

Thank you. I will sync using that link. :)

[DJRikki]

I just use my radio controlled clock and now set my watch about 3 seconds slow so by the end of the week it will be 3 seconds fast max

[padi56]

Originally Posted by Dicko

I have been looking at what COSC spec is.
When I google it, the definition seems to vary.
Some say -1 to +5 secs a day, some say -4 to +6 secs a day.
Some different again. How much does the age of the watch have to do with it all?
So where better to ask than here.
Being a new Rolex owner I am just wanting to know more about these beautiful timepieces.
So if some one could enlighten me I would be very appreciative.

Now many members have seemed to have caught this latest -2+2 syndrome plus the many others APS {Alignment point syndrome} CCTS {constant checking time syndrome}.Now regarding the latest -2+2 syndrome what in the real world does this mean.Well first the movement is tested to the Swiss chronometer standard at the COSC which is a AVERAGE of between -4+6 seconds over any 24 hour period. And when testing in the first 10 days of said test the movement could vary by 10 seconds in a single 24 hours and still pass the test.After movement gets certified and passes its shipped back to Rolex in there hundreds then stored till needed to match to a case.Now if its true Rolex must re-regulate and test on a machine to this new specification a AVERAGE-2+2. What does this mean well in the real world it was tested on a machine and passed said test at the time of testing much like the certification at the COSC. Now this test does not mean the movement will be -2+2 every single day for life.As on the wrist its a bit different to what the test on a machine,already some on forum wanting to rush out and regulate because there watch was only showing a consistent +4 seconds.

The COSC test refers to the movement only because uncased movements are all that the testers get,and every COSC movement is tested. These blank movements without dial, hands and winding crown are received and given standard hands and dials. Depending on whether manual or automatic winding, they received a standard wind and are tested many many at one time. The test procedure uses photography time lapse at set times daily, synchronised by a master computer type clock. Any deviation of the seconds hand from master clock, seen on the photo, is the daily variation recorded.The movement is then returned to manufacturer passed or failed,the failed ones are re-calibrated and re-oiled and sent back for testing.Now the passed and certified ones Rolex can take as long as they want to finally encase the movement in your "chronometer watch".And then machine re timed by Rolex before shipment. Some reasons why the COSC procedure IMO can be meaningless in real terms:in reality its just a bit of paper,afraid words like superlative are meaningless just Rolex talk very much like the COSC just a big marketing ploy IMHO.Test costs around £120 -£140 which is passed on into the price profit on watch

As I already indicated, only the bare movements are submitted for testing. Not even hands winding rotor and dials are present, these are added at the institute movements are wound up by a machine hundreds at a time Rolex have there own machine at the COSC. It is common knowledge, however, that additional shipping and handling, and above all, encasing the movements, can and could drastically affect its accuracy from original test.And today nearly all modern day movements Swiss,Japanese,European and even some Chinese,can with a bit of patience and fine tuning reach COSC standards.But like most machine made movements a Rolex is built to be a chronometer,whether tested or not,[now they are all tested].But many others are not,and to keep there timing to say the COSC standard will need more in the care of regulation service in its life.

This is a direct quote from Watchbore they estimate that at least 15,000 Rolex movements failed in 2010-12. According to Rolex, the rejects are re-regulated or fixed, and sent back to COSC until they pass.Rolex Quote "We don’t use COSC to tell us how good our movements are," said a source deep inside the Wilsdorf Rolex foundation. "We test them ourselves All we want is the chronometer certification, it’s only for marketing."... I rest my case

The Rolex ladies Datejust with cal 2235 was the most consistently precise and accurate movement tested by COSC some years ago now.But don't forget that Rolex is the largest independent movement supplier for the COSC test.Movements like ETA 2892, ETA 2824, Rolex cal 3135,cal 3185 cal 4130 and the Valjoux 7750 gave excellent results.The most precise machine tools are only viable in high-volume production watches.While its possible to get chronometer standards to hand crafted watches ,its very time consuming and needs a lot of adjustment and failures are quite high making more expense.But your massed produced machine made movements are far more consistent in getting though the COSC test simply because they should be all the same.

But looking at the other sign of the coin today any modern day wristwatch chronometers are, by the almost 300 year old 18Th century navigational standards imposed on John Harrison H4 watch,quite laughably inaccurate even by today's standards.How about just 5 seconds slow after 63 days at sea,not bad for a almost 300 year old watch.When you think he had very primitive tools by today's standards all hand made no computer aided help now that was a real watchmaker.

[Fredrik]

Originally Posted by GradyPhilpott

You need to stop using your computer to time your watch.

You need a serious reference like this one:

https://time.is/

Well, all major operating systems can use NTP to sync the internal clock to an atomic clock via the internet. It is very simple, just go in to the date and time settings in windows or Mac OS and select "internet time" or "automatic" settings and your time will always be exact on your computer. If you are running Linux you are probably already doing it.

I have all my computers synced since many many years, Windows, OSX and Linux, and they are always exact.

What is a bit odd(to me at least) is that it is not the default setting.

[GradyPhilpott]

Originally Posted by Fredrik

Well, all major operating systems can use NTP to sync the internal clock to an atomic clock via the internet. It is very simple, just go in to the date and time settings in windows or Mac OS and select "internet time" or "automatic" settings and your time will always be exact on your computer. If you are running Linux you are probably already doing it.

I have all my computers synced since many many years, Windows, OSX and Linux, and they are always exact.

What is a bit odd(to me at least) is that it is not the default setting.

Yes, this is true, but the issue is how precise the internal clock is and how often the clock is synced.

I used to have an alarm clock that would sync every four hours. The problem with using that device as a reference was that it gained about 2 minutes in four hours.

My computer syncs about every seven days, so it can be off quite a bit in that time.

If however, the computer is synced manually every time the time is set and checked, it would be accurate enough.

The problem is that a lot of people who come aboard don't know these little quirks, so it's easier to just advise folks that just because a clock is called an "atomic" clock doesn't mean that it's accurate at any given time it's checked.

The same goes for cell phones and computers.

There are several good references for the time on the internet.

This is but one.

https://time.is/

[Fredrik]

Originally Posted by GradyPhilpott

Yes, this is true, but the issue is how precise the internal clock is and how often the clock is synced.

I used to have an alarm clock that would sync every four hours. The problem with using that device as a reference was that it gained about 2 minutes in four hours.

My computer syncs about every seven days, so it can be off quite a bit in that time.

On any unix like system it will be synced as often as it has to be to be precise. The NTP daemon monitors the drift of the internal clock and syncs when it is needed. My computers are always exact when I check with time.is.

I do not remember how Windows performs in this area, I might have tweaked some registry crap to obtain better precision. Nowadays I do not connect my windows machines to the internet at all.

Regarding cell phones it depends on who you use as operator. In Sweden all of them as far as I know sends accurate time and both Android and iOS are within a second always for me. I do know that in the US it never really works and clocks drift a lot.

Oh, I have a friend who used to own a radio controlled clock, when it drifted, it always just reset to twelve o'clock and waited until it was. So basically it always showed twelve o'clock...