Expectations of accuracy in a mechanical
Like Tree2Likes
  • 2 Post By Rdenney

Thread: Expectations of accuracy in a mechanical

Results 1 to 8 of 8
  1. #1
    Member Apollo83's Avatar
    Join Date
    Mar 2012
    Location
    Ireland
    Posts
    653

    Question Expectations of accuracy in a mechanical

    Hi everyone.

    So as mentioned on a previous thread I recently received my first mechanical; A Speedbird 3 automatic with ETA 2824-2 Top grade movement.

    I've been monitoring its accuracy for the past couple of weeks and I thought I might share the data as well as ask a few questions for the experts out there.

    To measure the accuracy I measured the deviation of the time on my mechanical from the time on my Digital Casio PRW3000 which receives a radio controlled time update at ~2am every night.
    I did this at ~6pm every day so that I had the deviation over the last 24 hours and I did this for 16 days.
    On receiving the watch I manually wound it twice and then I wore the watch almost continually for the duration presumably winding it automatically as I did so.
    Here's the graph of deviation versus time:
    Name:  Speedbird_Deviation.JPG
Views: 461
Size:  27.5 KB

    As you can see the mechanical watch was always running fast compared to the atomic clock updated watch.
    The first 24 hours it was +7secs and eventually it seems to have settled into ~ +1secs per 24 hour period.

    I've read a number of other threads about 'settling' and I'm not sure that I've seen a consensus, so I've decided to start this thread hoping someone might clarify things for a mechanical newbie.
    Here are my questions:
    1) Is the above behaviour normal/typical?
    2) What is the reason for the initial inaccuracy and the asymptotic behaviour as it 'settles'? Is it something to do with the increasing tension in the main spring as it is wound? I also read something about a new movement 'settling'?
    3) Assuming the watch settles to a regular deviation (say +1 sec) then how confident can you be that it will remain reasonably constant? Should you check its accuracy at some interval to detect any trends?

    My hope:
    If I know the watch has settled into ~ +1 sec/day behaviour then I think I'd be happy.
    When changing the date I can set the time at ~ -30secs and then leave it for 2 months until I need to change the date again knowing the watch is unlikely to be more than about 30 secs inaccurate at any given time.
    This level of accuracy would seem perfectly acceptable to me.

    Thanks in advance for any help.

    Finally, part of me would have liked to have taken the back off the watch and included a picture of the movement in this thread... but most of me knows that would be a bad idea!
    So instead a quick hello from the Speedbird behind the data, sporting it's casual RAF strap for everyday wear
    Name:  Speedbird_RAF_Bond.jpg
Views: 452
Size:  143.6 KB

  2. #2
    Member
    Join Date
    Dec 2012
    Posts
    7,144

    Re: Expectations of accuracy in a mechanical

    The above performance is superb. Enjoy the watch and fret no longer.

    Running a few seconds/day fast is optimal for a movement that hacks. When you need to reset it, just pull the stem (stops the seconds hand), wait for accurate time to catch up, and then push the stem back in.

    Rick "a little fast is better than a little slow" Denney
    Apollo83 and Rigel like this.
    Zenith: Captain Chronograph 03.2110.400*; Cartier: Santos 100 XL Concord: Mariner, C1 Big-Date, C1 v.2 Chronograph; Ebel: Chronosport 1134901, Tekton 9137L83*, Type E 9137C41* (*=COSC)
    Ebel: 1911 BTR 9137L73* and 9139L71*, 1911 1120L41*, 1911 Senior 9080241, Brasilias 9120M41 (2), Aquatica 500 9120K61, Classic Hexagon GMT 9301F61, Classic 100 LE 9120R41; Baume & Mercier: Capeland World-Timer
    Heuer: Carrera 1964 Re-Edition CS3110; Maurice Lacroix: Masterpiece MP6439; "Seagull": 1963 Reissue cal. ST19; Seiko: Black Monster SRP307; Poljot: Sturmanskie cal. 3133; Tissot: T-Touch Lew and Huey: Acciona
    Vintage: JLC: ref. 2953, ca. 1946; Longines: Flagship cal 285; Zodiac: SST cal. 86, Aerospace GMT cal. 72; Favre-Leuba: cal. 253; Tianjin: Dong Feng cal. ST5; Elgin: Gr. 152 (1898), Gr. 384 (1919); Ebel: ca. 1962 ref. 9214955
    WUS: ST5 Project Watches (Black and Blue), F72.2014.DG3804 (Gray and Cream); Swatch: Sistem 51 Blue; TNT: Rattrapante cal. Rochat 7750+RAT-1

  3. #3
    Member sticky's Avatar
    Join Date
    Apr 2013
    Location
    South Yorkshire
    Posts
    57,322

    Re: Expectations of accuracy in a mechanical

    It's pretty good to say the least.

  4. Remove Advertisements
    WatchUSeek.com
    Advertisements
     

  5. #4
    Member gtuck's Avatar
    Join Date
    Feb 2012
    Location
    Oregon
    Posts
    1,241

    Re: Expectations of accuracy in a mechanical

    Quote Originally Posted by Rdenney View Post
    The above performance is superb. Enjoy the watch and fret no longer.

    Running a few seconds/day fast is optimal for a movement that hacks. When you need to reset it, just pull the stem (stops the seconds hand), wait for accurate time to catch up, and then push the stem back in.

    Rick "a little fast is better than a little slow" Denney
    Totally agree. I have a Top grade hand wound and two COSC automatics which have been consistent throughout with no discernible "break in" period. I have three others--probably elabore grade which tended to settle in to consistent behavior after 4-6 weeks. I do not know the reason--distribution of lubricant perhaps?? Just don't know. All do their best when "on the wrist" rather than in a winder or sitting on the shelf.

  6. #5
    Member mpalmer's Avatar
    Join Date
    Dec 2011
    Location
    Middle of the Middle West
    Posts
    28,460

    Re: Expectations of accuracy in a mechanical

    That's pretty awesome accuracy. You clearly don't have anything to worry about. I don't even know how accurate my watches are. They just need to be accurate enough they don't gain/lose any appreciable time in the few days I might I wear them at a stretch before grabbing something else out of the box. I can safely say though, that +1 second a day give or take a couple seconds is a great place to be, and you cannot ask for much more.
    "The stuff that dreams are made of…" - RGM ‘801 EE Motor Barrel’

  7. #6
    Member Apollo83's Avatar
    Join Date
    Mar 2012
    Location
    Ireland
    Posts
    653

    Re: Expectations of accuracy in a mechanical

    Thanks all.

    So ~+1 is good - looks like I fluked it with my first ever mechanical!
    Somebody hand me a lottery card - I never do those either.

    So I think answer to my 1st question is the observed settling behaviour is typical.
    question 2 was 'why does it happen?'
    question 3 was 'Is the settled value stable?'

    I think gtuck's hinting that it may not be stable; time spent off wrist, etc.
    So maybe I'll ask it differently: 'With an automatic worn daily, is the settled value stable?'

  8. #7
    Moderator at Large stuffler,mike's Avatar
    Join Date
    Mar 2010
    Location
    The Hague-Amman-London-Berlin
    Posts
    51,083

    Re: Expectations of accuracy in a mechanical

    No, isochronism and positional variation both play a part as well as lubrication (aging). That being said there is some good reason that it will work within a certain spectrum when kept well.
    Kind regards
    Mike


    Instagram @mikestuffler

  9. #8
    Member Apollo83's Avatar
    Join Date
    Mar 2012
    Location
    Ireland
    Posts
    653

    Re: Expectations of accuracy in a mechanical

    Thanks Mike - Some good search keywords for me to look up there.

    So with my limited understanding:
    Isochronism is referring to the ability of the watch to have the same oscillation period across varying drive force as the mainspring moves from unwound to fully wound.
    Positional variation occurs when the oscillation period is affected by the orientation of the watch to gravity.
    Aging affects the oscillation period as friction at moving contact points changes over time (e.g. due to lessening lubrication, or wear on parts)

    So I'm guessing what I'm seeing in the first few days of putting the watch on is a lack of isochronism until the watch reaches a certain level of being wound.
    Across the week I may also see variation due to orientation in the watch (maybe it's horizontal more during weekdays typing at the computer) and more 3 o'clock down during weekends while walking?
    And finally over a number of years I may see the accuracy change due to aging (and hence will get a service every 5 years or so to counter that)

    Am I on the right track?

Thread Information

Users Browsing this Thread

There are currently 2 users browsing this thread. (0 members and 2 guests)

    Tags for this Thread

    Posting Permissions

    • You may not post new threads
    • You may not post replies
    • You may not post attachments
    • You may not edit your posts
    •