In classical physics, a foundational principle is that it is possible to assign precise values simultaneously to all physical properties associated with any object.  We can define, for example, the speed and position of a motor vehicle with seemingly perfect precision, sufficient certainly to result in a penalty notice for exceeding the speed limit on a tempting stretch of the York ring road.  In quantum physics, however, the Heisenberg Uncertainty Principle states that quantum objects have certain pairs of complementary properties, such as momentum and position, that cannot be measured simultaneously: the more precisely we measure one property the greater the uncertainty with which we can know the other.  By that principle, the radar gun may well be able to record the speed of a particular racy white hatchback but it would not then be able to place the car at the scene of the crime (and vice verse).

An analogous principle applies to the recording of the processes documented here:  the greater the effort I make to capture the repair and restoration of my latest horological basket case, the greater the likelihood of mishap or misadventure.  To record even static photographs requires that the process is interrupted at every step, often involving the transport of parts from workbench to lightbox and back again, and that carries with it significant jeopardy.  Worse still is the perturbing effect of recording video because this requires that the execution of every operation be carried out under controlled lighting and typically beneath a suspended camera where access and mobility is impeded.

There may well be less intrusive ways to record the action, but my method of video capture makes the process massively more difficult than it would otherwise be and, in truth, sucks a great deal of the enjoyment out of the activity.  You might ask why bother then?  Well, the loss of pleasure in undertaking the restoration work is compensated, at least in part, by the rewards that come subequently from the compilation and editing of the result, regardless of whether anyone has any interest in viewing the outcome! All of that brings us to the subject of this post:  the filmed documentation of the restoration of a Seiko 6139-8040 from November 1974. 

I do not propose to produce a step-by-step written account of the contents of the resulting video here, but it is probably worth noting some of the bumps in the road that I encountered in working on this watch, one of which self-inflicted and a direct result of the perturbations caused by the documentary process, and others occupational hazards of working with Seiko 6139 automatic chronographs:

Mainspring Options

The original mainspring fitted to this example was out-of-flat and so I was forced to consider what other options might be available.

Although there is plentiful supply of used examples of the lesser 401615 mainspring native to the non-chrono, low-beat versions of the 61 series movements (as well as new, if you are prepared to pay inflated prices), these mainsprings do not have the requisite torque to drive the more power-hungry chronograph and so are only worth employing as a desperate last measure.  Supplies of the correct, higher torque 401616 mainsprings has dried up almost entirely and used options require the sacrifice of another 6139.  What about aftermarket?  I’ve had mixed successes with Generale Ressorts mainsprings but the G25341X has shown some promise in this application.  Its specified dimensions of 1.05 x 0.115 x 420 x 10.5 mm appear to be a good match to the 401616.  The struggles I have had in the past with this and other GR mainsprings have been related to disappointing amplitude.  I suspect that this is related to how well the GR slipping attachments grip the walls of Seiko barrels and so to give the mainspring a fighting chance, I opted to use a strong braking grease (8217) on the barrel wall and that seems to have helped to achieve decent full-wind steady state amplitudes.

The consequence of carelessness

One of the improvements to the design of the original 6139A calibre made in the later 6139B revision was the addition of a mobile hammer transmission lever on top of the flyback lever.

This lever was added as a safety-feature but its addition created the potential for mishap when reattaching the pillar wheel to the barrel and train wheel bridge after cleaning.  Distracted by the video-recording process, I failed to notice that the hammer transmission lever had swivelled by 180 degrees from its correct position, and I proceeded to reassemble the movement with the lever incorrectly oriented.  I did not notice my mistake until I’d fitted the chronograph bridge and realised that the hammer click spring was not properly engaged.  Correcting my mistake required me to disassemble the movement all the way back to the first coupling lever.  Live and learn.

Centre Chronograph wheel vulnerabilities

In spite of the fact that a very large number of these iconic Japanese automatic chronographs was produced between 1969 and 1975, there is one well known vulnerability in particular that conspires, together with the finite resources to deal with it, to making their repair potentially difficult or uneconomic.  That vulnerability is the vertical clutch spring in the centre chronograph wheel. 

The centre chronograph wheel comprises the fourth wheel and its pinion, a clutch ring, clutch spring, seconds heart (to align the seconds hand to zero when resetting) and the centre chronograph wheel axle and finger.  The main difference between the 6139A and 6139B is that in the 6139A, the clutch ring contacts the fourth wheel directly whereas in the 6139B, the clutch ring contacts a separate clutch plate mounted above the fourth wheel.  The figures below all pertain to the 6139A but the principle of operation is essentially the same between the two variants.

When the chronograph is in its running state, the clutch is in contact with the fourth wheel (6139A) or clutch plate (6139B), and the fourth wheel and centre chronograph wheel axle rotate together.

When the chronograph is stopped, the two coupling levers come together to lift the clutch ring away from the fourth wheel (or clutch plate), against the force exerted by the clutch spring.  In this condition, the clutch ring is no longer in contact with the fourth wheel and the centre chronograph wheel axle comes to a halt while the fourth wheel continues to rotate.

The correct operation of the chronograph when running requires the clutch ring and fourth wheel to rotate in perfect synchronicity and this is only achieved if there is no slippage between the two.  However, with time and in particular, if the clutch spring is left for extended periods in its compressed state (i.e. the chronograph in its stopped state), the spring can weaken with the result that the clutch ring can no longer maintain a perfect grip on the fourth wheel when engaged.  Consequently, the torque provided by the forward momentum of the seconds hand can overcome the coefficient of friction between the clutch ring and fourth wheel, the hand flings itself further forward than it should and the chronograph runs fast.  Such slippage can be pronounced or subtle.

When I first tested the watch before beginning its service, there did not appear to be any clutch slip but having completed the reassembly all the way back to fitting the dial and hands, I re-timed the movement and discovered the following:

At 45 seconds, the chronograph appeared to have gained 0.25 seconds, but that falls within what you might reasonably attribute to experimental error.

However, at 7 minutes and 30 seconds, there was no doubt.  The chronograph had gained 2.4 seconds which means that the clutch was slipping and that the chronograph centre wheel was not fit for purpose.

As Captain Haddock would say “Billions of bilious blue blistering barnicles in a thundering typhoon!”  My own profanity was rather shorter. Under the microscope, the relative weakness of the clutch spring was apparent, with only the lightest force required to lift the clutch ring against the restoring force of the spring.

Ordinarily, at this point, a reasonable conclusion would have been that this restoration was dead in the water, with no realistic prospect that a replacement centre chronograph wheel would present itself at a reasonable cost.  I did find a one option from an eBay seller in Portugal at an asking price of more than £200 delivered including taxes, but while it claimed to be new and unused, it was not in its original packaging and there was no way I was going to take the risk even if I thought the price worth paying (which I didn’t).  The project was therefore stalled.

A few days later though, I noticed an auction on Yahoo Japan for an 888612 chronograph wheel for a 6139B, sealed in its original packaging and with a much more reasonable buy-it-now price.  Unfortunately, that auction was unavailable for bidders using the Buyee proxy-buying service that I used habitually and I was forced to conclude that it was out of reach.  It niggled away at me though and so I started to wonder if there might be a back door route to securing its purchase.  A cunning plan hatched, and within a week I had managed to bid and secure the part and about 10 days later, it landed in North Yorkshire.  The project was saved.

The wheel was tarnished and the finger needed adjustment but importantly, once fitted, the chronograph ran true with no slip and I was able to complete the reassembly and re-case the movement.

This particular variant of the 6139 is not one that has held particular appeal to me in the past, perhaps because it seemed a little ordinary besides its more flamboyant siblings.

However, fitted with its new crystal and brimming with life once more, I am rather taken by it.  In fact, I think it looks rather special and its more compact size makes it more wearable as an everyday watch than some of the larger examples that I’ve worked on in the past.

Those of you who are interested seeing more of the detail from this particular project are very welcome to take a look at the accompanying YouTube video linked below: