Saturday, December 7, 2019

In Which Your Host Captures A Unicorn....

I have written about the 607 and 618 "Bumper" automatics, which came out in about 1950.  These are generally considered to be the FIRST American-made automatic wrist watches.  But before them, there was the 605.

The 605 appears in the serial number list in the 1950 "Elgin Genuine Material" catalog.  Three runs appear in the list:

       1.  J835001 - J838000         3,000 units

       2.  V751001 - V770000    19,000 units

       3.  H600001 - H612000    12,000 units

A total of 34,000 watches, starting in about 1947.  But the parts list does not contain any entries for the 605, and nothing appears in the databases about the missing 34,000 watches or the 605 grade.  Schlitt's Elgin database lists the 605 in the Grade Numbers list, but the entry only says

grade total runs   first yr  last yr class  size  code   jewels Adj/name
----- ----- -----  --------  ------- -----  ----  ------ ------ ----------
605   34000   3        1947     1949   ?      ?s  ?????      ?j ?

Pocketwatchdatabase is similar.

Years ago, when I first started collecting Elgins, I THOUGHT I saw a watch with a 605 on Ebay.  It was a Bumper automatic movement.  My recollection is that it looked pretty much the same as the 607.  I was just getting started and I was focused on watches of the late 50s and early 60s, so I let it go.  Only later did I realize that it wasn't just a 607.  But I wasn't sure I really saw it, because I didn't snag any pics from the listing.  So, it has bothered me for 9 years.

Last year, my friend Mick Grocott acquired a big lot of unbuilt Elgin watch plates and bridges, which included a number of Train Bridges marked "Elgin 605 USA; (serial number);18 Jewels 4 Adjs"

These bridges were the same shape as the train bridge in the 607 and 618, but those grades carry their grade and serial numbers on the Autowind Module.  This confirmed that the 605 was a Bumper Automatic in the 607 family, as I had thought!

Last week, 9 years of searching FINALLY paid off!  A watch with a 605 came up on Ebay!  I set a snipe.  Then I raised it.  Then I raised it twice more - after all, how much is a watch you see once a decade worth?  Luckily, I won the auction for far less than I was ultimately willing to pay!

So, here it is, my Unicorn, my Holy Grail - the Elgin 605 Bumper Automatic!

A couple observations off the bat:

First, this case, a 10k Gold Filled Wadsworth, does not match any of the cases in which the 607 or 618 were sold.

Second, there appears to be a jewel and pivot on the Barrel Bridge, which is NOT on the 607/618.

Third, the oscillating weight is different.

Okay, but what's it like on the inside?  So I carefully uncased the movement and set about disassembling it - CAREFULLY! because there are no spare parts for a Unicorn.

Once the Autowind Module is removed, I noticed a number of detail differences.

The Ratchet Wheel screw is countersunk, while the 607/618 uses a flat wheel with no countersink.  The Crown Wheel is thinner, and the washer sits flush with the wheel, versus the thicker 607/618 Crown Wheel with its sunken washer.  And that is, without question, a jewel and pivot on the Barrel Bridge!

Okay, so what's it look like underneath?  What's that other wheel?

It's an off-center Center Wheel!

These are pretty common in automatic watches.  It saves space and lets you use a nice big balance and barrel.  Usually, the offset Center Wheel has a shorter extended arbor, and a low Cannon Pinion frictions onto it, which drives the Minute Wheel, which in turn drives a slip-on Cannon Pinion on a hollow center post which carries the minute hand.  But there's no evidence of that on the dial side of the movement!

No Drive Wheel here!

The answer becomes clear when you remove the concentric Fourth Wheel, a feature shared by the 607/618.  The extended Fourth Wheel arbor runs through the hollow shaft of a pinion that meshes with the Barrel teeth.  Call it the Center Pinion.  The Center Pinion extends through the Pillar Plate and the Cannon Pinion frictions onto it.

This is an arrangement I've never seen, offset Center Wheel with a second pinion driven by the Barrel to carry the hands, but concentric Fourth Wheel to provide directly driven center seconds (hence no jumps and skips).  

The 607/618, by contrast, put the Center Wheel in the center, with the extended  Fourth Wheel arbor running though the hollow Center Wheel arbor:

You can see one of the Center Wheel spokes peeking out underneath the Fourth Wheel.

Intrigued, I also took apart a spare 607 I had (the original one from the earlier blog posts on the Bumper Automatics), and began comparing parts.  These are the Fourth Wheels, with the 605 on the left and the 607 on the right:

Not only is the arbor of the 607 a lot longer, but the wheel itself is larger!  Just to be certain I photographed them from above.

I thought about this, and I realized that for the Fourth Wheels to be different diameters, there must be geometry differences between the two movements that I hadn't seen.  So I took pictures of the Pillar Plates and carefully lined up the centers from the Balance to the Escape Wheel, and then from the Escape Wheel to the Third Wheel on both.  Then I drew an arrow from the Escape Wheel pivot to the center of the movement of the 605, and copied and pasted that arrow on the 607.  Sure enough, the 605's Escape Wheel is closer to the center on the 605!

L-R: Elgin 605 Pillar Plate; 607 Pillar Plate.

Notice how the angle on the 607 is more acute, and that the arrow from the 605 is to short to reach the center of the 607.  

So, how did Elgin manage to stack the Center Wheel and the Fourth Wheel in the 607, when they hadn't managed in the 605?  They made it THICKER! About 1/2 mm.

All in all, there are surprisingly few interchangeable parts between the two grades.  Maybe the Balance and Balance Cock, maybe the Pallet, but the Pallet Cock has to be just a hair different to put the Pallet pivot in the right spot.

Here's the cleaned, reassembled, and lubed movement, running again with good amplitude!  I had to ever-so-carefully tweak the hairspring, which was being pushed down onto the balance arms. Once lifted just a bit, the coils lie flat, and the balance spins beautifully!  

The original case screw was missing, replaced with a larger screw that didn't really fit between the two buffer screws.  Over time, it had stripped the case screw threads inside the buffer block.  Luckily, the buffer block, screws, springs, and the case screw from the 607 is the same, so I simply scavenged the whole assembly from a spare 607.

Add a new crystal, and a strap, and my unicorn is now a wearable watch!!

BTW, the case is unique to the 605.  It is not one of the cases the 607/618 were sold in.  While those are all two piece cases, this has separate bezel, center, and back.  It's not waterproof AT ALL.  It "wears" a bit smaller than the other Bumpers, at least partly because it's thinner by about 0.5mm.  That doesn't sound like much, but on the wrist, it definitely is!

As far as I can tell, it's the original case, since it is stamped "Cased And Timed by Elgin National Watch Co."

Here's the 605 with all but one of the other Bumper Automatics - the remaining one is solid 14kt gold, too rich for my blood!

L-R:  605; 6801 with 607 mvt.; 6802 with 607 mvt.; 6803 with 618 mvt.; 6804 with 618 mvt.
My best guess is that Elgin completed at least half of the first run of 3000 watches (mine would be 1537th of the run), and at least SOME of the plates for the second, larger run.  The three serial blocks are 800-900,000 apart, about a year at that point of Elgin's production, with another 800,000 before the first run of 607s, starting at N433000.   The first ad I've seen for the Bumper Automatics was in 1950.  My conclusion is that after building 3000 605s and doing some wear-testing by Elgin employees, they decided that the 605 was not ready for sale and probably scrapped them.  This example was likely kept by one of the employees testing them.  Years later, like many vintage watches, children or grandchildren sell off 'Dad's Stuff', and that's how this watch came to be on Ebay.  It did, after all, come from a seller in Illinois, where Elgin was located.

But we'll probably never know the whole story of the 605.  At least now we know that it was built, and that it does indeed work as intended!  Because it IS so rare, I don't think I'll wear it much.  Even my 1870s keywinds are common by comparison!  If a part breaks, I'd have to find a highly skilled watchmaker to replace it! Similarly, I won't be trying to straighten the bent 4th wheel post, or reduce the enormous 7ms beat error, or fix the 50 second/day positional error.  It's enough that it runs, and runs pretty well.  

What's my NEXT Unicorn?  I don't know yet.  I'm still basking in the glow of watching my 605 ticking away, scabby dial and all!

EDIT:  One thing that I think supports this as a factory prototype is the jewel count.  I've gone through the pictures several times, and I keep getting 19 jewels, despite being marked '18 Jewels'.  The extra one appears to be in the Autowind Module, on the upper pivot of the Winding Sector, which transmits the motion of the oscillating weight to the Crown Wheel.  In the 607/618, that pivot gets a brass bushing.

Saturday, November 18, 2017


During Elgin's first decade, in addition to 18s keywound watches like the BW Raymond and others (see Luis Casilla's excellent presentation), Elgin also made 10s women's keywound watches, and for a while, 17s keywinders.  These seem to date to 1875-1878 (see slide 92 of Casilla's presentation)

 17s is an odd size.  These days, it's nearly impossible to find a 17s case, and the movements just don't fit either a 16 or 18s case.  I became interested in these after acquiring a handful of the 18s keywinders.  They're very hard to find as cased watches, but I did manage to grab a couple bare movements on Ebay.

The dials are one-piece, with no sunken sub-dial.

They're essentially full plate, with solid balances.  As far as I know, all of them were 7 jewel.

The TM Avery is named for Elgin's second President.  Note that the balance on the Avery is larger than the Leader, but they're otherwise largely identical.  The Leader's balance staff is broken, so I took the Avery apart for cleaning and hoping to get it running.  

No such luck.  This old boy will never run again.  The lower balance pivot is bent all to hell, and I can't find any replacement staff for these.  If I were a better watchmaker, and had a lathe, I might cut a new staff.  But I'm not!

Since the watch was dirty from neglect, I decided to clean it and then document the rebuild.

Like the 18s keywinders, the Avery has the serial number stamped on numerous parts, including the barrel floor and cap.

First things first - reinstalling the balance jewels.

Next, installing the balance in the cock.  If you've been paying attention, you may have noticed that every Elgin I've shown made before 1955 has a triangular stud, which goes into the stud hole in the balance cock and is held in place by the stud screw.  On these cheapies, Elgin went with a pressed-in stud (red arrow), requiring you to unpin and re-pin the hairspring to the stud each time it was serviced. To ensure you repinned the hairspring at the same point, there's a 'witness mark', by the green arrow.  Note also the gouges and crow-pecks in the foot of the balance cock to raise it for more endshake (blue arrow)

The balance is put in place in the jewel, with the hairspring in the regulator, then you carefully feed the hairspring through the hole in the stud.  Next, you feed it through till it reaches the witness mark, then insert the pin from the coil side.


Next, the ratchet wheel, click, and click spring are installed in the barrel bridge.


With those two tasks done, we're ready to put the whole thing back together!

Notice that there are only 3 pillars, versus the usual 4.  More economizing.

First we install the train.  Unlike most full plate watches, the barrel is not removable once the plates are together.  The balance bridge just holds the click.  So, we have to put the whole train in before installing the upper plate. 

Next the upper plate is installed.  Make sure all SIX pivots are in place before adding the first screw. And yeah, you can really only add one screw at this point...

As always,check all the pivots are in place before tightening down the screw.

Next, the barrel bridge is installed.  This means putting in the second pillar screw.  The smaller barrel bridge screw just threads into the upper plate.

The hole next to the barrel arbor exposes the click.  You let down power with a key on the barrel arbor, and insert something pointy in to push the click out of the teeth.  If there's any actual power on the mainspring, of course, you hold onto the key!!

I wound it up a few clicks, and sure enough, the pallet flicked to the opposite bank with a gentle nudge, in either direction.  Since the balance staff is irretrievably bent, AND there are no cases to fit it, this breaks my heart!

Next we installed the balance, making sure to rotate the jewel pin into the pallet fork.  This allows you to add the 3rd pillar screw, in the foot of the balance cock.

Now it's time for the dial side.  Key set watches are usually pretty simple - cannon pinion, minute wheel, hour wheel, and that's it.  No setting bridges, no clutch, no clutch lever, no setting lever.

And that's it!  Now for the dial. 

The dial is held in place not with dial foot screws, but rather with tapered brass pins that fit through holes in the dial feet which are drilled so that the pins pull the dial up snugly.

The hands go on like all hands, with the following caveat - you need to determine where the minute hand is, relative to the corners of the key square on top of the cannon pinion.  In this case, the corner lines up exactly.  Unfortunately, I had set the hour hand off just a little.  If this were to be a working watch, I'd correct it.

So, there you go!  An oddball, and an orphan, 140 years old!  Someday, I hope to find one cased, but for now, I'll keep this one.