more than most having been very lucky enough to witness the facilities and meet some of the team who perform it. There is also no doubt that it was an additional influence in my decision to purchase a JLC as the test exceeded mere rate measurement including some shock testing, power reserve test, and most importantly testing the whole assembled watch, conducted over a 41.5 day period rather than the 15 of the COSC. It is a significant investment on the part of JLC that, frankly, most customers will rarely understand or appreciate (remember we are the minority) and it is a laudable effort to improve the delivered quality.
I am concerned however, that everyone seems to prejudge the COSC as "easy" and "obsolete" and take no account of some other less well known elements of the COSC test and their (actually quite stringent) timing standards (although we have debated their upgrade here
Everyone has assumed MC 1000 hours is better. Based on what exactly? Well, probably, the usual situation on the Internet; where ubiquity equals "correct", where self-proclaimed hyperbole and website-cum-wikipedia meets to create "the-truth"? On the one hand the COSC is performed to a documented, published and publicly accessible procedure, on the other the MC 1000 hours is a self administered quality procedure for which there are no easily understood limits to the tests performed that I can find (someone once mentioned "-1s/d +6s/d") . (An explanation of part of the 1000 hours test report I made is here
I cannot comment on the tests outside of rate measurement and I can only make a comparison of the timing data from 1000 hrs test sheets, but I have made an interesting observation. Let's look at COSC first. It's less well known, but there are a lot more criteria than the simple "-4 to +6s /day" that are needed to be satisfied to qualify as a chronometer
. ISO 3159 1976
sets it out to be precise.
Now, I have entered the data into a simple spreadsheet to simulate the watches likely performance, based on the timing sheets, over the 15 day COSC test (as close as possible considering JLC use slighty different temps 4, 22 and 40 deg C etc).
To plug these numbers into the ISO 3159 I've made the following assumption - The daily rate will be an average of the timings in any measured position for full wind and after 24 H. In the above example:
6H (VG) = +1 s/d. (average of 0 and 2)
3H (VH) = 0 s/d
9H (VB) = +3.5 s/d
DD (HB) = +2 s/d
DU (HH) = +1 s/d
We cannot get any variation in the same position (as we are using spot timing data), so we cannot get any sense out of Vbar or Vmax in fact! It should always be zero. However, we can get pretty solid numbers for Mbar, and particularly the positionally important D and P parameters. So for the above watch:
Mbar = 1.5 (-4/+6 s/d) Pass
D = 0 (-6/+8 s/d) Pass
P = 2 (10 s/d) Pass
Now take another MC 1000H watch. The timing results as follows:
6H VG +1 s/d
3H VH -4 s/d
9H VB +5.5 s/d
DD HB + 2.5 s/d
DU HH +0.5 s/d
The same process gives us:
Mbar = 1.1 (-4+6 s/d) Pass
D = 0.5 (-6+8 s/d) Pass
P = 5.1 (10 s/d ) Pass
So clear passes for both cases.
The MC1000H temperature test is from 4 deg C to 40 deg C. This compares to the COSC range of 8 deg C to 38 deg C. The variation in daily rate obtained on my RDM was +0s/d at 40 deg C and +4s/d at 4 deg C. That's -0.111 s/d per deg C compared to a COSC limit of 0.6 s/d per deg C. The COSC test is performed dial up, but I'm not entirely sure of the procedure for the MC1000H. Again, essentially I see this as a COSC pass.
It would be very interesting to assess some watches with the new Autotractor style balance that should exhibit better positional performance. I would imagine they will perform much better for 'P' than the older movement designs I have data for.
Here are the results for a calibre 822:
6H VG -6 s/d average between full wind and 24 hours running
3H VH -12.5 s/d
9H VB -0.5 s/d
DD HB 3.5 s/d
DU HH 2 s/d
12H VD -8.5
Gives us the COSC ratings as follows:
Mbar -2.7 limit of -4+6 s/d Pass
D -8 limit of -6+8 s/d FAIL
P 9.8 limit of 10 s/d Pass