WatchUSeek Watch Forums banner

1 - 18 of 18 Posts

·
Registered
Joined
·
16 Posts
Discussion Starter #1
Hi guys,

Im kinda confused about the accuracy of mechanicals, I keep getting varying answers.

How many secs does the avg mechanical (not COSC) lose or gain a day?
Is +/- 10s normal? Ive been told by my jeweler that +/- 2 or 3 secs a day can be expected.

Also, what is the range for a COSC certified watch?

And after wearing the watch for a while and getting it regulated how much more accurate does it become?

Sorry for all the questions but I would be grateful for some solid ans.

Thanks
 

·
Registered
Joined
·
174 Posts
It's rare to find a mechanical watch accurate to less than 10 seconds a day. Someone correct me if I'm wrong, but ten seconds might even be considered very good on the accuracy range for mechanicals.

But even if was very accurate, it could easily lose its accuracy with a minor shock such as bumping into a wall, which would knock everything out of whack.
 

·
Registered
Joined
·
8,826 Posts
The accuracy of mechanicals varies widely from crazy erratic for cheap watches that have not been regulated and have poor tolerances to precision chronometers. If you want the highest precision then by a $5 quartz watch (the folks that talk about $K quartz watches are talking about such small increments of improvement its amazing).

The Eta 2824 is the workhorse auto movement among the Swiss movements today and considered a pretty good movement. The standards for its various grades (yes all 2824s are not created equal) vary according to the components in the movement.

This is taken from a review I wrote on the Affordables forum at
https://www.watchuseek.com/showthread.php?t=46610

"The performance differences are the big differences between the various grades: ("The limit values are subject to interpretation: 95% of the pieces delivered in a lot must be within the specified limits.")

Standard:
2 positions (CH, 6H)
daily rate: +/-12 sec/day
Maximum positional variation: 30 sec
Isochronism (between 0 and 24 hours): +/- 20 sec

Elabore:
3 positions (CH, 6H, 9H)
daily rate: +/-7 sec/day
Maximum positional variation: 20 sec
Isochronism (between 0 and 24 hours): +/- 15 sec

Top:
5 positions (CH, FH, 6H, 9H, 3H)
daily rate: +/-4 sec/day
Maximum positional variation: 15 sec
Isochronism (between 0 and 24 hours): +/- 10 sec

Chronometre:
As per COSC specifications, which as far as most owners will notice, isn't much different from Top grade."
 

·
Registered
Joined
·
16 Posts
Discussion Starter #4
Thanks a lot for the information much appreciated.

I wasnt planning on buying a bad mechanical so that reference is irrelevant.

I was looking at the Hamilton jazzmaster or the mont blanc timewalker auto.

The mont blanc styling really appeals to me but I wasnt sure what accuracy I should expect in the first few mnths before regulation and then after having it regulated. I have always relied on my trusty tissot so I was just wondering.

BTW a watch is considered COSC if it can do +-4 secs a day?? Am I right?
 

·
Registered
Joined
·
725 Posts
BTW a watch is considered COSC if it can do +-4 secs a day?? Am I right?
Yes and no. A COSC watch will be expected to be within +6-4 sec/day, but to be considered a COSC movement, the movement has to pass COSC testing in the COSC labs.

And as to your earlier question, the wildest of my mechanicals runs within about 6 sec/day. Most of them are within about 2 sec/day.

Enjoy!
 

·
Registered
Joined
·
95 Posts
There is a survey in the Seiko forum of the accuracy of a cheap Seiko movement called the 7s26 or something like that.

Basically accuracy seems to depend on price (surprise surprise). Cheap watches can vary by 15" or more per day, but often can be regulated to a few seconds. More expensive watches with a decent ETA movement often seem to achieve a few seconds per day. Accuracy will depend on the quality of the components and the amount of effort put into regulating the watch. There seems to be a lot of discussion in the affordables forum of DIY regulation.
 

·
Registered
Joined
·
8,842 Posts
The 7S26 and 7S36 are the Seiko workhorse autos and are comparable to ETA 2824s, although, I don't think Seiko bothers to fancy them up beyond what ETA would consider standard grade. It's performance is typical of that grade movement which also includes the FHF 96 serie handwinds, and the Miyota 8215 automatics.

The gain/loss on the wrist depends greatly your habits, and a watch regulated to +2 seconds on one person might lose 10 seconds on someone else's wrist.
 

·
Registered
Joined
·
351 Posts
I've had a Tag and Rolex in the past, both of which have been COSC certified Chronometers. Accuracy has always been no worse than +2-3 seconds per day, as would be expected when spending that amount of money. I have also purchased a couple of Seikos with the 7S26 movement, and sure enough the accuracy is debatable. I get variations ranging from +5 to +25 seconds per day. This is no good in my opinion, although I have heard some other forum members claiming much higher accuracy from the same movements. (there's a separate thread on this in the Seiko forum). i managed to pick up a vintage hand-winding Omega Seamaster recently and it has an accuracy of +5 seconds per day. Not bad for a 50 year old watch.

On the other hand, I also have a few Quartz watches, and these are always consistently accurate. My personal recommendation would be to go for a higher end mechanical if you are interested in accuracy, and if you aren't then a Seiko mechanical is great value for money.
 

·
Registered
Joined
·
3 Posts
I'd say that accuracy in a mechanical/automatic watch can sometimes depend on luck. My Sinn 657 (with Eta 2824 I believe) was always very accurate for me. Average was +1/+2 seconds per day when worn and +4/+5 seconds on a winder.

Earlier this month I dropped the watch on my bathroom floor (tile) and in addition to a small ding on the case back I noticed a very interesting thing - the watch now loses about 1.5-2 seconds per day when worn and gains 1.5-2 seconds during the night (dial down) giving me perfect time on average. Gains about 0.25 seconds per day when on a winder.

Needless to say, I'm pleased with the "upgrade" :-!
 

·
Registered
Joined
·
8,826 Posts
There is a survey in the Seiko forum of the accuracy of a cheap Seiko movement called the 7s26 or something like that.

Basically accuracy seems to depend on price (surprise surprise). Cheap watches can vary by 15" or more per day, but often can be regulated to a few seconds. More expensive watches with a decent ETA movement often seem to achieve a few seconds per day. Accuracy will depend on the quality of the components and the amount of effort put into regulating the watch. There seems to be a lot of discussion in the affordables forum of DIY regulation.
The 7s26 is NOT a "cheap movement". It is the basic Auto Seiko movement. 21 jewels, magic lever winding system. It runs +-15s/day on average but it can reliably be regulated to +-6 sec/day by most reviews. Price of the watch has NOTHING to do with the ability of movements to keep time reliably. I own an Eta 251.262 quartz that keeps terrific time - in the Victorinox case it cost $39 and in the Breitling case it cost around $2k. 7s26 movement watches range in price from $50 to $300 or so. I have one 7s26 that keeps excellent time and cost $60 (+-4 sec/day) and one that cost more ($150 or so) and keeps average time for the 7s (=-10 sec/day). We've seen complaints about Breitlings and Rolexs that run considerably faster or slower than those figures and they are anything but affordable. The most accurate mechanical I own is a $90 Poljot alarm watch that has not been regulated by anyone after leaving the factory and runs on average +-1s /day over a 30 day period.

My earlier post pointed out that even ETA offers their movements in different grades and that each one has its own number of positions - that means the number of positions is has been regulated in. A "Top" grade is regulated in 5 positions while a basic grade movement is regulated in 2 positions. If the factory did their job right then you shouldn't need to regulate the watch IMO.
 
  • Like
Reactions: Time On My Hands

·
Registered
Joined
·
193 Posts
When everyone keeps talking about bringing watches to be regulated, around how much money are we talking and where might one find someone that can do this?
 

·
Registered
Joined
·
8,826 Posts
When everyone keeps talking about bringing watches to be regulated, around how much money are we talking and where might one find someone that can do this?
Watchmakers regulate watches. Today they use electronic devices to do so quickly. The cost varies based on location and type of operation you are talking about - basically if they sell Rolex it will cost you a lot more for precisely the same service as it would if they had Hamiltons as their top-end watches.
 
S

·
Guest
Joined
·
0 Posts
Watchmakers regulate watches. Today they use electronic devices to do so quickly. The cost varies based on location and type of operation you are talking about - basically if they sell Rolex it will cost you a lot more for precisely the same service as it would if they had Hamiltons as their top-end watches.
Well, I´d add: a good watchmaker regulates your watch. The electronic devices they use look like this one (Witschi Zeitwaage Expert II):



Quote:

The Watch Expert II is based on a inhouse invented new measuring technology, called Chronoscope. The instrument offers the watchmaker all possibilities for testing mechanical watches. A diagram of the watch beats is presented completely silently on a LCD graphic display panel. Numeric values for the rate accuracy, amplitude and beat error are automatically calculated and displayed numerically, i.e. the diagram no longer has to be manually evaluated.
Operation for conventional watches is largely automated. The measurement parameters can be manually selected for special watches or for special measurements. With the new measuring mode, watches with a special escapement system can also be measured.

Unquote

The rest is done by the skilled watchmaker.:)
 

·
Registered
Joined
·
95 Posts
The 7s26 is NOT a "cheap movement". It is the basic Auto Seiko movement. 21 jewels, magic lever winding system. It runs +-15s/day on average but it can reliably be regulated to +-6 sec/day by most reviews. Price of the watch has NOTHING to do with the ability of movements to keep time reliably. I own an Eta 251.262 quartz that keeps terrific time - in the Victorinox case it cost $39 and in the Breitling case it cost around $2k. 7s26 movement watches range in price from $50 to $300 or so. I have one 7s26 that keeps excellent time and cost $60 (+-4 sec/day) and one that cost more ($150 or so) and keeps average time for the 7s (=-10 sec/day). We've seen complaints about Breitlings and Rolexs that run considerably faster or slower than those figures and they are anything but affordable. The most accurate mechanical I own is a $90 Poljot alarm watch that has not been regulated by anyone after leaving the factory and runs on average +-1s /day over a 30 day period.

My earlier post pointed out that even ETA offers their movements in different grades and that each one has its own number of positions - that means the number of positions is has been regulated in. A "Top" grade is regulated in 5 positions while a basic grade movement is regulated in 2 positions. If the factory did their job right then you shouldn't need to regulate the watch IMO.
Firstly, the original post is about mechanical watches, not quartz. I'm not sure why you mention quartz as everyone knows that a very cheap quartz watch can outdo most mechanical watches.

The 7s26 appears in watches that in my book are cheap i.e. inexpensive. As I said, look at the thread in the Seiko forum to see a survey of 7s26 accuracy. One sample is meaningless. The range of variation is quite large. You say that it can be regulated better. Well yes, and if the factory had done that, it would have cost more.

Many manufacturers, including Orient and Seiko, give the expected range of accuracy of a mechanical watch. In general, the more you pay, the higher the tolerances. That is also true of the ETA movements as you point out. As to where your watch falls in that range, that is down to luck. So you could have an expensive watch that gains 6" per day, and a cheap one that loses 1" per day. But on average the cheap watches would be less accurate than the expensive watches, and the maximum deviation of the expensive watches should be much less than that for the cheap watches. (This assumes that the manufacturers tolerances are credible.)

The reason that paying more can lead to greater accuracy is because of the use of better designs and better components, giving for example better temperature and positional stability. Thus the nature of the mainstring, and the balance play an important role. And of course the amount of effort put into regulating the watch plays a part, and that costs money, thus adding to the price.

Someone rightly pointed out that accuracy in use depends on how the watch is worn, especially for cheaper watches. Some people play weird games whereby they claim that their watch lost only 1" per 3 months. Read closer and you see that they took care to compare it to an atomic clock, and place it in an appropriate position overnight, to compensate for the days losses/gains. That is pointless IMO.

George Daniels makes a remarkable claim for the accuracy of his own wrist watch. I presume that is based on daily wear with no jiggery pockery. If so, then it shows that a wrist watch is capable of true chronometer performance (and not just the rather lax COSC pseudo-chronometer performance). Such a watch would cost a fortune.
 

·
Registered
Joined
·
16 Posts
Discussion Starter #16
Thanks a lot for your replies, especially Docrwm's post--really helped me put everything in perspective. I wasn't however talking about cheap autos. I was talking in the range of hamilton.

Anyhow, from my experience, and I have gotten my watches regulated....I have never had even +-4 secs a day performance. I simply cant believe 1sec a day performance...sure maybe incredibly lucky but I cant believe its a regular thing??? Am I right or have I just been unlucky.

Also, how do most measure the accuracy of their mechanicals?? I compare it with a cesium but is there a more sophisticated way of doing things?? Or what do most do to chk up on that?


Thanks a lot guys,
Appreciate it
 

·
Registered
Joined
·
2,792 Posts
Hi Faraday. I have downloaded a free atomic clock program on my computer and once a week, I adjust the time on my daily wear, Sinn U2 to what the atomic clock says.

My Sinn U2 is roughly accurate to +5 sec./day which is enough for me. My life is not that tightly regulated that a couple of seconds a day will throw me for a loop. Some days, depending on what I am doing or how long I am wearing my U2, the accuracy could shoot up to +10 sec/day (rarely and usually when whacked a few times or something similar) and other times, it's about 4 seconds a day.

I would suggest to those who are into automatic accuracies that you stretch your adjustment time frame to about a week and then divide by the amount of days since your last adjustment to get an overall accuracy as sometimes, like in the above paragraph, your daily activities could affect your accuracy.

BTW, my Sinn U2 has been through a helluva lot over the last 13 months and it's accuracy is pretty regular. For more info on what it's been through, visit the Sinn forum and the sticky U series ownership exp thread. Automatics may not be for everyone but they are capable and reliable. I base this on the ownership of, presently, 15ish automatics for which I have owned for several years.

All the best,
Tim
 

·
Registered
Joined
·
9,651 Posts
Hi -

Great post, dead on.

You can also think of it this way: lenses made by Canon or Nikon, or lenses made by Zeiss or Leitz.

All are excellent lenses per se: Canon and Nikon QC are excellent, but there is the risk of getting a lemon. Rarely happens, but it occasionally does.

Lenses from Leitz (Leica) and Carl Zeiss (for Hasselblad, high end Rollei, etc) are individually checked, i.e. each and every lens goes through the full QC. Canon and Nikon? Every nth, where n = minimum number for a statistically meaningful sample. Varies from product line to product line and over time.

No-name made in China for that no-name digicam? Every 1'000'000th.

The higher the level of QC, the more expensive almost anything is, as QC is invariably something that involves trained people.

Oh, and the difference, of course, is in the cost: a Leica or a Zeiss lens sells for 4 figures, a Canon or Nikon for 3, and the no-name for less than $10.

JohnF
 
1 - 18 of 18 Posts
Top