How to regulate the DG 2813 movement

[fcrvincent]

Until last year my watch experience was limited to quartz watches including Casio, Citizen Eco-drive and Seiko Kinetic. I always wanted to try mechanical watches due to their heritage and this year finally found the time to build up some knowledge and experience with them. My first experience was not so fruitful with a £34 Vostok Kommandirskie manual winding model that I finally got refunded after receiving three defective watches and replacements - one would not hold its power reserve more than 24 hours - it probably had a poorly oiled mainspring - one with a date change that would stuck in between two dates - probably a loose date wheel spring - and finally one with a cracked crystal. My second attempt was with a £9 automatic Soki submariner-type watch from ebay with a low quality bezel, bracelet and crystal and a stiff rotor that would barely keep the watch powered for a couple of days on the wrist before it stopped unless manually wound. And finally my last attempt seems to be much more positive with a sterile dial automatic submariner-type from dajiwatch for about £30. Overall quality is good except for the rubber gaskets and the Nato strap that I had to replace; the watch features a DG 2813 movement which reading from several forums including RWG seemed to be an excellent movement.

However the regulation of the DG 2813 was much more difficult than I expected and at some stage I even thought that the movement was defective because its accuracy would be unstable and nearly threw away the watch in the dead watches drawer.. Finally after much experimenting I managed to regulate it to about +2s per day. The procedure below describes how I regulated the DG 2813. With hindsight the regulation procedure is simple. But given that I have not found any such detailled procedure on any forum yet I thought that it could be useful to anyone who thinks - like I did at some point - that the DG 2813 is impossible to regulate.

Here is the regulation procedure:

1. Fully wind the watch with about 40 turns of the crown
   
   
2. Let the watch lay dial up for a few minutes
   
   
3. Record the watch beats with the (free) Audacity software and a basic external microphone against the back cover of the watch - the watch should lay dial up again -. Audacity should be set at a 48000 Hz sampling rate and recordings transformed into 'spetrogram' in order to clearly reveal the watch beats.
   
   
4. If necessary adjust the isochronism lever (the one with one dot on it) so that the beats are equidistant - this is important otherwise the watch accuracy will vary significantly depending on the state of winding of the mainspring - with Audacity isochronism is easy to adjust and the lever is not too sensitive so it can be done without a magnifier.
   
   
5. Adjust the rate lever (the one with two dots on it) to get the rate at about 1/100s fast per minute - you need to use a magnifier and move the lever very very gently with a toothpick for example so that it is not perceptible with bare eyes but so that you can see it move with the magnifier. Pushing the lever toward the center of the watch make it go faster, pulling it makes it go slower.
   
   
6. Wear the watch and enjoy its new accuracy - to my surprise after this procedure the watch would even only gain/loose about one second after a one-hour jogging session!
   

This procedure also worked with my £9 Soki submariner and possibly could apply to other movements with a two-lever regulation system.

Edited July 25, 2012 by fcrvincent

[LMPAZ]

Fcrvincent, thanks for taking the time to provide that explanation. This give me a little better understanding about the internal workings of the movements. I might even get brave and try it for myself!

[fcrvincent]

You are welcome Lmpaz; one thing I forgot to mention is that ideally one should apply this procedure after the so-called break-in / settle-in period for the movement which from my experience was around 15 days of wearing the watch continuously after purchasing it (during this period the daily accuracy tends to slow down by may be around 20s/day after 15 days).

Also the isochronism adjustment described above should be made quite precisely ideally to millisecond (i.e. 1/1000s). This can be done with Audacity by zooming in the recorded graphs with [Ctrl-1] ([Ctrl-3] to zoom out). Again isochronism should be adjusted only if necessary - i.e. only if you see that the beats are not equidistant.

Another note that may be helpful: in my case the sweet spot for an accurate daily rate is around 1/100s fast per minute but depending on individual wear pattern it could be around 2/100s per minute or otherwise. If you are not completly satisfied with the daily accuracy you can always repeat the procedure after measuring the accuracy a couple of days later; once you are familiar with Audacity and the procedure it takes about 15 min to execute it.

Good luck and let us know if you had any success.

Fcrvincent, thanks for taking the time to provide that explanation. This give me a little better understanding about the internal workings of the movements. I might even get brave and try it for myself!

[2camsam]

Also, the watch back can be placed against the internal microphone included on most laptops these days (it's usually right next to the camera lens in the lid).  My HP did a much better job using the internal mic than an external I tried first.

Edited May 7, 2013 by 2camsam