1922 Masterpiece 920 Pocket Watch

The other day I posted about a 910 dress pocket watch from 1923 that I had waiting for attention in my workshop for several years.  While researching it I was reminded that I had a 920 in my collection, or was it the 922?

Looking on my blog, I've covered the 922 but not the 920.  They are very similar as the 922 replaced the 920 as Hamilton's premier 12 size caliber.  It was regarded, at least by Hamilton, as the finest 12 size watch made in America.

The 920 is a marvel to behold in person.  The movement is nickel, there are 23 fine ruby jewels set in solid gold settings.  The train wheels are solid gold.  It was adjusted at the factory to temperature, isochronism, and five positions.  It really was the finest workmanship that Hamilton could put out.

My recent focus on pocket watches made me dig out the various examples I have to see what sorts of things I've kept over the years.  Sure enough, I do have a 920, as well as the 922 that I posted about back in 2014.   Since I haven't covered the 920 yet, I thought it would be a project to take on.

From a collector standpoint, I've failed miserably at record keeping.  I have no idea when I bought this or even how much I paid.  I do know which examples were family pieces, so at least that part is documented.  If I could go back in time and do it all over again, I would keep better records.

Anyway, my project watch caught my attention for a couple of reasons.  First, it's a 920 and I've not covered it.  Second, it's cased in a rolled gold plate case - which is unheard of for such I fine watch.  It would have originally been in a solid gold case or at least a case with heavy enough gold fill that it would be "permanently warranted".  

So how did it get into this case?

I have a couple of theories - one of them is the Great Depression.   Times were very difficult in the 1930s and it's very possible that the original watch case was sold for it's gold content and recased in whatever the least expensive option was.  Another theory is an extension of the first... the original case was scrapped sometime between 1922 and the time I got ahold of it.

I find scrapping watch cases to be very sad - at least for a working and desirable watch.  It's happening a lot today thanks to gold prices being ridiculously high.   Demand for a watch is less than the demand for gold, I guess, so a lot of cases are being scrapped and movements sold for parts.

It's an interesting dynamic.  Models that recently might have been considered pretty common are now becoming scarce.   I'm sure more than a few $900 watches have been scrapped for $400 in gold melt, maybe making those $900 watches worth more like $1,200 in the future.  Who knows?

When I picked up my project watch out of it's drawer I though it might have been a 917 or other 10-size, based on the case.  RGP would certainly have been an option in the 1950s.  The case back is unremarkable, other than a couple of minor dents.

The movement inside is in great shape.  There are several different bridges and quite a few more screws than in a 910 movement but they are very similar from a geometry standpoint and things like dials and cases are interchangeable.  It's really the materials of construction and finish work that made the 920 a masterpiece.  Two of the jewels support the "motor barrel" - a unique design where the mainspring and barrel are supported by jeweled setting.

The inside of the case back makes me think this is a generic 12 size case.  I do not see any indication that I have ever serviced this watch before.  There are quite a few other marks inside though.

Two case screws hold the movement in the case and then the assembly can be pushed out the front of the case.  I'll pull the hands and dial next.

The dial side of the main plate looks identical to a 910 movement.  I'm sure all of these parts are interchangeable between the two calibers (as well as the 900, and 914).

The first thing I'll remove from the back is the balance and then the pallet fork.

Once the pallet fork is lifted out of the way, any remaining energy in the mainspring will be released.  At this point I can start disassembling the rest of the movement, starting with the barrel.

Just the train wheels are remaining.  There are three bridges, one for the center wheel and third wheel, one of the fourth wheel (the second hand), and lastly, one for the escape wheel.

Check out the solid gold train wheels - I bet these are pricey today if you had to. make them.

Everything is cleaned and ready to be put pack together.

All four train wheels have to go be in place in order to put the bridges back on.

Next the barrel bridge goes on and then I'll add the jewel that supports the arbor.

Lastly, the keyless works goes in and I can replace the final winding bridge.

Once the pallet fork and it's bridge are installed the movement can be wound up.  I'll install it temporarily in the case and use the crown to wind it up.

I carefully drop the balance in place and it starts ticking away with a good motion.  Let's see what the timer has to say.

Oof - the beat error is a bit high at 6.2ms.  That's an indicator of how well centered the balance is and measures the difference in how much it swings to one side or the other.  Ideally it would be zero but I usually draw the line at 3.0ms. as it's a tricky adjustment.

My first attempt gets the beat error to 2.9ms.  I should stop here but I'll roll the dice again to see if I can do better.

I got it down to 1.1ms.  It's a little wavy though so I may need to tweak the hairspring a little and see if I can get it a little better.

One of the reasons I decided to take on this project is I accidentally dropped the 910 I covered the other day.  It slipped out of my hand and landed on the floor - perhaps you heard my cry of anguish as it echoed across the Blue Ridge Mountains of Virginia?

I decided to case the 920 in the green gold filled case that housed the now-damaged 910 movement.  It's a much more fitting case than the RGP version.  Alas, my previous 910 movement will now be used for parts... sigh.  It happens.

1951 Kirby - this was a tricky one

The early 1950s was like a change of season for the Hamilton Watch Company.  The designs and traditions of the 1930s and 40s was giving way to a new era.  Very soon new, Swiss-made movements would be introduced, including automatics.  Less expensive materials would be adopted too, like rolled gold plated cases and stainless steel backs. 

I wouldn't go as far as to say that Hamilton's standards were reduced.  The fine traditions of Hamilton watch making did not go out like a lamb with new standards coming in like a lion.  Instead, a greater array of products were offered, including watches at lower prices points to compete with the evolving watch industry, in general.

One of the hold outs for "traditional styling" was the 1951 Kirby.  It was produced for three years.  It has a strong family resemblance to several 1940s models.  I think the dial is very Midas-like with two types of solid gold markers in addition to solid gold numerals inside the chapter ring at 12, 3 and 9.

The Kirby is on the smaller side, like a 1940s watch.  It was cased in 14K gold filled.  By 1953 the overall design was the same however the movement inside switched from the 14/0 sized 19 jewel 982 to the 12/0 sized 19 jewel 753 movement.  Technically, when that change occurred sometime in 1952, the name changed to the Kirby B to accommodate the differently shaped movement.  So the movement inside helps to date the year the watch was made.  The price was the same, regardless of the model year... almost $900 when adjusted for inflation.

My project watch arrived in good aesthetic condition but not running.  I have to open it to see what movement is inside.

The case back is contoured from side to side and it's very evenly worn.  The gold filled layer is worn through on the lower left of the photo below.  I won't do too much polishing to the case.

The shape of the case opening and the name are clues that this is the 14/0 version.  If it was the 12/0 version it would say Kirby-B and the opening would be slightly wider and slightly shorter.  The dial feet locations are different on the two movements so a Kirby dial won't fit a Kirby B - even though they look identical from the front.

The 982 was highly damascened at this point in time.  Even the 17 jewel 980 was too.  It's a nice looking treatment that only watchmakers would have admired.  Based on the serial number on the barrel bridge, this is a 1951 model.

You never know why a movement won't run until you check it out closely.  Using the tip of my tweezers I can see the balance swings freely and without a wobble.  So the balance staff isn't broken.  What I don't see though, is the pallet fork moving when the balance swings passed it.  So there's either a pallet fork issue or an issue with the balance.

The balance is a very complicated part.  It's the "heart" of the movement and it controls how the energy of the mainspring is released, tick by tick.  It has several integrated parts including the hairspring on one side, and the roller table on the other side.  Installed within the roller table is a tiny precision lapped jewel called in impulse pin.  This jewel is what engages the pallet fork, or vice versa.  The energy from the mainspring goes through the train wheels and pushes on the pallet fork.  When the balance moves the pallet fork, the energy releases and the pallet fork then pushes the balance.  When it swings back the process repeats, over and over.  

It's sort of like pushing a child on a swing set, you give them a shove, they swing out, then swing back past you and you shove them again, over and over, until all your energy is exhausted.

You may be able to see there is just an empty hole in the roller table of this balance.  The impulse jewel is missing.  Normally it's held in place with shellac.  If you click on the image it will get bigger.

There's already a white alloy "lifetime" mainspring installed in the barrel so the only thing I see that needs to be addressed is the balance.

I remove the balance from the balance cock.  Now what?

There are LOT of ways to screw up from here.  Some of my options include; replace the balance entirely, replace the roller table with one with a jewel, or try to find the impulse jewel and then reinsert it.

Option 1 is replace the movement but in order to do that another movement must be sacrificed.  So that's really my last option.  Option 3 is extremely challenging because the jewel is so tiny and getting the oddly shaped jewel to fit in the D-shaped hole, then glue it, is very time consuming.  So I'll go with option 2 and attempt to replace the roller table by using a donor balance that's unusable.

Replacing the roller table is a very challenging task.  First you have to remove the hairspring without ruining it's shape or knocking it out of level.  I'll have to put the hairspring back on again later, so I mark the wheel with a sharpie to say where the hair spring stud should go.  I also mark the wheel on the opposite side to indicate where the impulse jewel should point.

I have a donor balance with a messed up hairspring but a good roller table.  I can remove this roller table and install it on my project balance.  It's the exact same process, I just have to do it twice.

To get the hairspring off I use my green Bergeon oiler to slowly and carefully lift the hairspring up the balance staff until it comes free.

Now I can use my staking set to hold the balance by the roller table - this is very precise work, I don't want to break the pivots on the balance staff.  You can see why I needed to remove the hairspring.  It wouldn't fit in the tool.

With a special punch of the exact size needed, I use a hammer to push the balance staff down and out of the roller table.

Voila.  Now I do the same with the other balance to remove the damaged roller table.

Here's the good roller to install on the good balance.

Now I have to change my staking setup to support the balance staff.  I install the roller table with the impulse pin pointing toward the mark I added to the wheel so it goes back with the same orientation.  I added a tiny amount of red locktite to keep the roller in place once I seat it.  A punch on top with a couple of light blows from a small hammer will seat the table.

So far so good.  One fascinating aspect of balance wheels is the mass of the balance is matched to the length of the hairspring.  Swapping a hairspring from one balance to another is doable, but if the mass of the two wheels is different, the beat rate of the resulting assembly could be way off.  You might get lucky but the odds aren't in your favor.

At this point I can flip the balance over and reinstall the original hairspring using the same setup in my staking set.

The last stem is to place the collet on the balance staff with the stud pointing toward the black mark I added to the wheel, so it has the original orientation.

A few gentle taps with the hammer seats the collet on the staff.  As you can imagine, there are myriad ways to screw up this very precise task.  I think I'm in good shape so far.  Now I can clean the rest of the movement and the final test will be when I reinstall the balance.

Okay - everything is cleaned and dried.  This can be a challenging movement to reassemble but time and patience will see me through.

Well looky here... I found the impulse jewel for option 3.  Imagine trying to pick this red speck up with tweezers and install it in an equally tiny hole without losing it a dozen times.

It's the moment of truth... the movement is assembled and wound up.  Time to add the balance assembly.

Success!!! The movement is ticking away with a good motion.  Now let's see if it's close to accurate.

Hurrah!  The beat error is within my personal spec of 3.0ms.  The amplitude didn't register yet but it was a healthy 250+ degrees.  I could try to lower the beat error by removing the balance but that's risky business and I've made it this far without goofing up.  I know when to call it a day - and it's been a day.

The glass crystal on this watch wasn't too bad so I just reinstalled it after cleaning the case.  Paired with a fresh black leather strap, this little Kirby is ready for some more wrist time.

I'm glad I was able to save the balance - it took a bit of time but it's better than sacrificing another movement that could be used elsewhere.

1920 956 Pocket Watch

I've been on a pocket watch kick lately and have posted several that have been waiting in the wings for several years.

My latest one is a 1920 956 model.  The 956 is an interesting caliber.  It was introduced in 1914 and produced through 1924.  By 1925 the variety of 16 size pocket watches decreased considerably.  The popularity of pocket watches was giving way to wrist watches, unless you needed a railroad watch... you could get those all the way through 1969.

The 956 was the entry-level 16 size model for Hamilton, or should I say, the "moderate priced" model.  It was cased by Hamilton in various cases and different dial options were available too.  You can see the catalog snip shows dial pattern no. 8.

The 1923 catalog broke down the differences between the various 16 size models, starting with the 23 jewel 950, which is a beautiful watch.

The 992 was the entry-level railroad-approved model.  You purchased the movement only and your jeweler would case it for you.  It's not unusual to find 992 movements in very basic cases as a "working man's watch" or cased in very fine cases.  The 952 was typically cased in solid gold or very heavy gold filled cases.

The 956 (above) was also sold cased by Hamilton.  It was a the least adjusted model - plenty accurate but not adjusted to 6 position, temperature, isochronism, et. 

The next step up from the 956 would be the 974, then the 978 and finally, the 972.  The 972 was approved for some local rail lines.

The catalog describes the functions of lever set, or pendant set.  Lever set involves a small lever on the side of the dial that you slide out to move the watch into the time-setting position.  Pendant set uses a stem sleeve in the case that allows you to pull the crown out to the time setting position and then you push it back in to the winding position.  The 956 was pendant set only.

I've had my project watch for many years.  In fact, I saw that my mark inside said I serviced it in 2012 - hard to believe that's 14 years ago.  You can see it has a No. 3 dial.

I don't know if this is the original case but it's an interesting design.  You unscrew the bezel to reveal the dial, then pull the crown out to the setting position and the movement hinges outward to reveal the inside of the case and back of the movement.

The movement is ticking away and looks pretty clean for being 14 years since the last overhaul.  The enamel in the lettering is missing, I'll re-enamel it after it's been cleaned.

Three screws on the side of the movement hold the dial on.

The microgliss lubricant I put on the clutch, etc. is a bit yellow and gelled.  Time to replace it.  Pendant set movements are also called negative set.  Notice the clutch is in the time setting position.  It requires the case to go into the winding position.

Everything is cleaned, dried, and ready to be put back together.

I added black enamel to the lettering so now you can read the details better.  At this point I can put the movement into the case and use the crown to wind up the mainspring.  Then when I reinstall the balance it should start ticking.

With the dial out of the way and crown pulled out, notice the clutch is in the setting position.

When I push the crown in the clutch is moved toward the winding pinion.  Now when I turn the crown the watch will wind.

The movement comes back out of the case and I can drop the balance into place.... voila - it starts ticking.

It's ticking away a little fast but a quick adjustment will slow it down.

Now I can put the crown and the rest of the keyless works in place.  Last on will be the dial and hands.

I cleaned up the case too and now this 106 year old gem is ready for pocket-time.  Wrist watches are fun, they are relatively loud tickers, especially when you have several of them sitting together.