1931 Cushion B

Hamilton's first mens wrist watches were very simple and named by their shape.  The first 6/0 sized model was simply the Men's Cushion or Cushion Round, noting that the case was cushion shaped but the bezel opening was round.

Other varieties quickly followed and cases were "plain" or "engraved".  New shapes were introduced like the Barrel, Tonneau, and even the Rectangular.  The choice of materials spiced things up even further with solid gold, or gold filled, in green, white or yellow.  Even sterling silver got into the lineup from time to time.

In 1927 the new 987 movement was introduced and the 19 jewel 979 came along with it.  This new design offered a sub second movement at the 6 position.  By that time the Cushion Round had run it's course but in 1931 a new design was introduced, the Cushion B.

The Cushion B was produced in 14K gold filled in white or yellow only.  So if you see a cushion shaped watch with a round opening and a subsecond at the 6 position, it's most likely a Cushion B.

The Cushion was produced through 1934 and the only thing that changed year to year was the retail price.  The aviation dial listed in the catalogs is pretty rare to find in the wild but they're out there.  It's a simple dial with large luminous figures at the 12, 3 and 9 positions.

Generally speaking, the 17 jewel movement was used in the gold filled models and the 19 jewel 979 was reserved for the solid gold models.  However, there are lots of exceptions to the rule and sometimes you'll see the 987 in solid gold models (like the Piping Rock) and sometimes the 979 is in gold filled models (like the Lee).  The Cushion B was sometimes outfitted with the 979 but mostly you'll find the 987 inside.

One other bit of trivia is the 987 and 979 includes both the initial 987 and 979 but also the 987F and 979F.   The difference is in how the jewel settings are retained in the movement.  The F is in reference to friction fit settings while the earlier movements had setting retained by screws.  There is no other difference beyond that and Hamilton didn't promote the difference, 987 refers to the 987, 987F and even the 987E.

I recently had the opportunity to work on a Cushion B again.  The last time I saw this model was in 2013.

This project arrived in ticking order but it's been through a few hands, I suspect.  First, the hands are missing any sign of lume... I would call these skeleton hands.  The dial is missing it's lume too - or at least most of it.  It looks like a 90 year old dial, that's for sure.  What's remarkable is there is no wear through to the corners of the bezel.  The crystal is pretty tall though so it may have saved the bezel from shirt sleeve wear.  The crown is sitting proud of the case and you can see the stem tube.  This crown is not deep enough to accommodate the stem tube, or the stem is too long.  My guess is the crown is incorrect.

The back of the case is nicely engraved but I can't tell if it's just one letter or a couple of letters together.

Just for fun, let's see how it's running.  I normally don't bother with a baseline but I'm curious to see what I'm up against. 

Based on the shotgun pattern, this movement is making a ticking noise but it's not keeping any semblance of time.  Oh well, off we go into the unknown.

The Cushion B comes in a three-piece case with a back, a front bezel, and the center that holds the movement.  Two case screws hold the movement in place and once the crown is removed the assembly will come out the front.

This example has the 19 jewel 979F inside.  It's an upgrade over the 987F and includes a solid gold center wheel.  This would be an expensive movement to replace - I need to be extra careful with it.  Although, nice 987s are getting very hard to come by too.

Once the movement is removed, I can put the crown and stem up to the stem tube to see how it fits.  It's clear the crown won't accommodate the stem tube.  I've seen other examples where the stem tube is shortened - altering the case to fit a crown is the wrong decision.  I'll see if I have a better crown to fit.

Digging through my stash of crowns produces a winner.  This version has a deep enough well to cover the stem tube.  I'll probably have to replace the stem too though, as the crown and the stem must fit the movement inside the case perfectly in order to move from the winding position to the setting position.

With the dial and hands removed, I can start disassembling the movement.  In the shot below you can see the extra cap jewel that covers the escape wheel.  The other end on the train bridge has a cap jewel too - and brings the count up to 19 versus the 17 in the 987 movement.

Looking inside the barrel reveals an old blue steel mainspring.  It's likely set into a tight coil and lost most of it's energy.

Yup - called it.

According to the parts catalog, the 979 has a different mainspring than the 987.  The barrel is the same though so it's really the strength that's different.  I'll go with the 100 strength Dynavar mainspring.

Everything is cleaned and ready to be put back together.  Wish me luck.

I got the movement back together easy enough.  It's not a terrible movement to work on.  The worst part is the pre-Elinvar hairsprings are very easy to damage / bend and then you're screwed.  

The reassembled movement is ticking away with a decent motion - hopefully it's better than what I started with.

The two parallel lines had me hopeful until the 9.2ms beat error was registered.  That means the lines are actually really far apart, the one on the top is really on the bottom and vice versa.

Reducing the beat error requires removing the balance from the balance cock and then rotating the hairspring on the balance staff so that the balance impulse jewel is better aligned with the pallet fork.  If that sounds hard to understand, it's 10X more difficult to do in person.  As I said above, one false move and you'll bend the hairspring and that's game over.

My first attempt lowered the beat error to 5.1ms.  I'll give it another go.

Two more attempts and the best I've gotten to is 3.9ms.  Normally I prefer to get below 3.0ms and as close to zero as possible.  However, I'm going to call it here... after three attempts I have really pressed my luck.  A man's got to know his limitations.  A higher beat error mainly implies the watch will stop sooner than a watch with a lower beat error.  So this is still way better than what I started with.

The finished watch looks pretty good.  Personally I don't mind skeleton hands.  I could change them to black spear hands or I could even fill them with lume of some color.  However, since the dial is also naked, what's the point?  I hope I look this good at 95 years old!