So, a couple weeks ago I’ve ordered some mechanical watches on eBay,why?


-Good looking

-Kinda reliable

-Really nice mechanical interior,fun to watch and show

Ok so now,everybody know that often,cheap good looking stuff can’t be reliable.

So upon receiving 3 of em, I wondered just how accurate they were at keeping time….

First obvious point : why don’t you set them and just check the next day?  It’s pretty simple,they are mechanical watches,if you don’t wear them,they won’t crank back up and stop pretty quickly. If you wear them 12h/day,they will last forever.

So, a device of some kind was needed to check the Accuracy…

Unfortunately, I’m not at home,where I would have used an arduino to do data acquisition and count up the “ticks” on a definite time length.

Then I got an idea : my pentax WG-1 GPS has a pretty cool intervalometer built-in that can take pictures every minutes for a long period…why not use it to gather “stroboscopic” images of the watches running, if they are accurate,the seconds counting hand should stay still while the minutes and the hours go on.

Second Obvious point : What tells me that the camera is accurate?

Solution was to make a first “video” at 1min intervals of a Real time clock,the only one I could find near me was an online website : http://atomicclock.org/

Here’s the video of the snapshot taken of the clock :

YouTube Preview Image

As you can see,it’s pretty darn accurate,somewhere about 0,2 s.

From that point,I was ready to go on and test those :

Here is the resulting video:

YouTube Preview Image

Analysis base on a timescale of 2hours and neglecting the camera’s error:

Watch #1 :

Lost 2 seconds

-2/7200 = -277.778E-6 seconds

Error of -0.02777 %

Error for one day : -24 sec

Conclusion: Pretty accurate.

Watch #2 :

Gained 14 seconds

14/7200 = 1.944E-3

Error of 0.19444%

Error for one day :  2 min and 48 sec

Conclusion : Not accurate enough

Watch #3:

-2/7200 = -277.778E-6 seconds

Error of -0.02777 %

Error for one day : -24 sec

Conclusion: Pretty accurate.

General conclusion:

It is clear that W1 and W3 bears the same clockworks thus have the same error minus the variations in springs.

So this is how you find out about watches guys. Don’t get scammed,I’m returning W2 to the supplier. I have to time to waste. ;)