A right of passage for those that dabble with vintage watches is to succumb to the charms of something that turns out to be not quite what one would have hoped. The defining such experience for me came in 2008, when still very green around the gills, I bought my first vintage Seiko, a 6105-8110 dating from about 1976. The price I paid was more or less par at the time for a nice example of this iconic divers watch.   The artfully executed photographs in the auction were good enough to fool me and even with the watch delivered and in my hand, it was a day or two before I realized I’d been sold a pup. This is what it looked like when I received it:

Superficially, it looks clean enough but the educated eye will see through this one very quickly: The most conspicuously dodgy component is the horrid aftermarket dial, complete with shiny plastic markers filled with flat, smooth lume facsimile. The dial is the single most objectionable component for me, not just because of its very poor quality, but because its Seiko branding tips it from aftermarket replacement to fakery. Next up is the bezel insert, which is as wrong as it is possible to be whilst still being close enough to fool the unwary: it’s wrong not just because it is rubbish quality but also because the font choice for the numerals is incorrect – the zeros are too narrow and upright and the numeral 1 in the 10 marker has a serif tick where an original is just a plain vertical line. On the upside, the hands are original, although in poor condition, the bezel turning ring, case and crown all look pukka. The question buyers of these blinged up watches should be asking themselves, above anything else, is the following: ‘Why does this watch have a fake aftermarket dial and insert?’ The obvious answer is: ‘To obscure the horrors that lie beneath’. When I got around to opening her up though, to my inexperienced eye, what I saw was nothing much more objectionable than a slightly grubby looking but most importantly ‘working’ movement, ticking happily away and keeping decent time into the bargain.

But it was the exterior that grated and at the time I was dismayed, sick at having rushed into a purchase only to regret it at my leisure. However, my error turned out to have been the catalyst which turned an idle, ill-judged toe-dip into the vintage watch world into a full blown adventure in watch smithery in which I find myself immersed presently. My dissatisfaction with this particular watch was amplified by the not insignificant cost of the thing and so rather than fold and walk away, I resolved to see if I could address at least some the cosmetic issues. This process started with a change of a clapped out date wheel, replacement of the incorrect winding weight (above), a series of increasingly better quality bezel inserts and, most drastic, a decision to replace the dial and handset with items sourced from the MkII website. Over the years, I’ve nipped and tucked this one, a tweak here, new insert there, a swap of the mid-case with one fitted with a triplock conversion, leading to the settled configuration we see here:

In spite of the shaky start to our time together, this watch has become my default go-to, a staple beater and a firm favourite. But somewhat ironically, with all of the restorations that I’ve worked on over the years, I had never serviced this one, never even taken a proper look at the movement to see what that dastardly Ebay seller might have been hiding with all of that superficial surface gloss. Well, the prompt to do so has come finally with my recent purchase of a slightly scruffy, but wholly original 6105 dial. My objective is to take it back to something approaching what I was hoping for from that original 2008 transaction. That process needs more than just a dial though, the hands being an essential component. With some uncharacteristic foresight, a while back, I’d ferreted away a 7005 dress watch fitted with the same hour/minute hands as those used in the 6105 divers watches and I had also kept hold of a traffic light seconds hand taken from a 6306 sports watch:

The dial is largely free from hour marker lume rot but is pretty dirty and the SEIKO emblem is corroded:

There is plenty of promise here but we’ll wait to see what I can do to smarten the dial up until after I’ve given the movement the once over.  Let’s start as usual with a quick initial gander at the movement in situ, fitted with a (now) correct but rather grotty winding rotor:

That autowinding mechanism has remained in place since I bought the watch and I am interested to get an unobscured view of the movement proper.

Something is not quite right here. That balance wheel and cock look familiar, but only because I’ve seen loads of them fitted to 6309’s, 6306’s and 6319’s – but most definitely not to 6105s or indeed any of the other 61xx family of movements on which the 6105 is based. The 63xx and 61xx families both run at 21600 beats per hour, so a balance from a 6309 will allow the watch to run at the correct rate – but it’s still a bodge and has to go. The train wheel bridge looks like it’s seen better days, in particular in the region close to the scruffy looking barrel ratchet wheel. It should be clear at this point that we have a bit of a cobble job here but just how cobbled becomes apparent at the next step. The train wheel bridge is secured to the main plate by three screws, shown in the photo above. The two top-most screws unscrewed without drama but the third, at the bottom, broke off as I unscrewed it. It broke, not because I had sheared it off, but because someone else had and to disguise the fact, they had glued the screw head into position in the recess in the bridge. In fact the bridge was secured at that point using glue alone, which I think explains the tarnishing to the finish on the bridge close to the screw head.

Having levered the bridge off, we see the train wheels, a very worn mainspring barrel and the hacking lever, present but not necessarily quite correct. A closer look at the rear of the movement shows the corroded remains of the sheared screw stubbornly refusing all attempts at extraction:

This little screw looks like stopping me in my tracks. 6105 main plates are simply not available anywhere, 6105B main plates even less so! I do have a spare, not quite complete 6105A movement, but to use that would mean losing the hacking function on the watch and the whole point of this exercise is to return this watch to something close to its original spec (yes, I know the crown is wrong, but the watch is properly water proof as a result). The potential solution to the problem is to attempt to dissolve the screw in a solution of alum, a tactic I’ve had some success with in the past on an old 1945 Tudor (see here). The batch I used then, I think may have been somewhat augmented with other ingredients and so I try a different supplier in the UK, receiving what looks like very pure, crystalline sample of aluminium potassium sulphate.

I’ve described the principle of how this works before, but the general idea is that the acidic solution of alum will react with the iron in the steel screw, hopefully dissolving it and liberating hydrogen gas in the process. The much lower reactivity of the copper in the brass protects it from the action of the acid, leaving the bridge unaffected. I made up a saturated solution in boiled water and left the main plate partly submerged over night. The last time I tried this it took a good week for the screw to dissolve but I could see when I checked on the bridge the following morning that I’d made a rather more potent brew this time than last. The nickel plating on the submerged part of the bridge had pretty much disappeared but importantly the screw thread looked thoroughly blackened. Some digging around with the head of a pin and the remnants crumbled and with a little more work with a toothpick and it looks like we are back in business.

The nickel plating had disappeared too because the concentration of the solution was sufficiently high to exploit the fact that nickel is only a couple of places below iron in the metal oxidation scale. My feeling at this point, having saved the main plate, was to remove as much trace of the abuse this movement had been subject too, starting with that tarnished train wheel bridge. The 6105 is closely related to the 6119 and 6106 movements but I wanted to stick with 17 jewels and so I went with a bridge from a 17 jewel 6106C. The additional bonus for me was that this bridge features a beveled finish around its edge, which in my opinion lends the movement an aesthetic edge over the more workmanlike cliff-edge finishing on the 6105B and some versions of the 6119.

The balance from the same donor movement ought to serve as a perfect replacement for the 6309 balance. If we look at the incoming (on the left) and outgoing (on the right) below, we can see that the balance wheel on the 6106 balance is significantly larger than that of the 6309:

Intuitively, you might expect the smaller balance wheel to run at a higher beat rate than the larger, in much the same way that an ice skater will spin more rapidly with their arms pulled in than thrust out, but both balances run at 21600 bph. However, the beat rate depends on more than just the size of the wheel. The beat rate or natural resonant frequency, f, of the balance given by

where k is the force constant of the spring (a measure of its stiffness) and I is the moment of inertia of the rotating balance wheel. The moment of inertia is a measure of the resistance of the wheel to rotation and is proportional to the mass being swung about during the rotational motion but also to the square of the radius of the wheel. So a smaller wheel yields a smaller moment of inertia and since the frequency is inversely proportional to the moment of inertia, this translates to a higher frequency of oscillation. The balance is redressed of course (if you’ll excuse the pun) by the force constant: a softer spring means lower force constant and a lower frequency. So the smaller wheel used in the 6309 balance runs at the same frequency as the larger wheel in the 6105 balance by virtue of a softer hairspring. With a fresher barrel with cleaned mainspring installed, I part assembled the movement to check all was well

but at this point it occurred to me that the hacking lever may not operate correctly with the larger balance wheel, having seemingly worked properly with the 6309 balance. Removing the hacking lever and comparing it with one taken from the 6106C, we can see that my concerns were justified:

The stop lever fitted to my watch had been bent outwards to enable the lever to make contact with the outer edge of the smaller 6309 balance wheel with the crown pulled out to the time setting position. With the correct part now fitted, the hacking function now works as intended with the correct balance wheel installed.

From here on in, it’s relatively plain sailing. I had given the dial a jolly good clean and, rather than risk attempting to replace the tarnished Seiko emblem, I elected instead to give it a light sanding, exposing the brass finish and lending it a slightly steampunk look, in sympathy with the ordeal the main plate has been through. With the hands fitted

we can pop it into the case,

and see how the timekeeping looks, following regulation:

I must say, I am very pleased with this, particularly given the parlous state the main plate was in and particularly given that I’ve not used any brand new parts in servicing the movement. I checked the timekeeping again, after a couple of days of wear, dial side up and got almost identical figures of 0 s/d, 210 degrees amplitude and 0.0ms beat error. Autowinding mechanism refitted and she’s done:

Three to finish, one face up

and two on the wrist

 

Caveat emptor.

Post script:  I could not live with the green hue if the lume on the hands from the 7005 donor and so this weekend relumed the hands in a slightly grey toned white lume, which while not a perfect match to that of the hour markers, I think looks a great deal better.