Anyone who enjoys restoring things might be able to relate to this situation: seeing an unloved and worn out object and having a feeling that it could be something special if restored to its original condition. That’s been my feeling on a project I’ve been working on over recent months, and now that it’s finished it has completely surpassed my expectations. A rather simple and unassuming dress-style watch has been a real surprise, charming me with its design and giving me a new perspective on what smaller-sized watches have to offer.
Assessment
The watch arrived from the USA back in September 2020, one of a few cheap items I bought on an online auction around that time. Specifically it was a Seiko reference 66-9990, a small stainless steel model offered by the Japanese brand in the 1960s.
The most obvious issue upon seeing the watch physically for the first time was the overall condition of its exterior. It was fitted with a non-original expanding bracelet, which although of the right period, wasn’t really suited to the watch. The case and acrylic crystal had both sustained quite a few scratches, and looking closely at the hands there was evidence of some slight damage. I also discovered the true size of the case as I’d only been able to judge previously by its photos – at 33 mm it was quite modest by today’s standards (whereby 38 to 42 mm is more the norm).
But in spite of all this there seemed to be real potential for a good outcome with this project. The case was stainless steel meaning it would respond well to a light polish, as would the crystal as it wasn’t in too bad condition. The design of the watch was also interesting as there wasn’t much of a bezel – the dial face and domed acrylic crystal extended to very edge of the case, giving it an ‘open’ appearance and helping to make it look larger than it really was. The dial had an attractive sunburst effect that caught the light in nice ways, and was embellished with applied indices which were long, narrow and elegant. Overall it had a pleasing, dressy vintage vibe.
Popping open the case revealed Seiko’s popular 66A movement, this one dating to 1969 near the end of the 66A’s production run. The 66A (and its 66B relative) were used in many Seiko watches and can therefore still be found pretty easily today. It’s quite a simple movement: hand-winding only, no seconds stop hacking feature, and no day or date complications. But Seiko sought to maintain a level of quality and reliability by fitting their Diashock protection system for the balance and Diafix jewels for the escape wheel, contributing to a total of 17 jewels.
I tried winding it up to see if it was working, and was happy to see it start ticking away. Putting it on the timegrapher, however, all was not well – the amplitude was quite low between 150-160° and the timing rate high at +40 secs/day. Hopefully a clean of the parts followed by proper oiling would be all that was needed.
Getting into it
The first challenge would be to remove the movement from the case. The movement was secured in place with two case screws, but removing the screws followed by the crown and stem didn’t release the movement as expected. Turning the watch over, it became obvious what was causing the hold up – the large dial! It was too big to fit through the case. There was also no way of releasing the dial from the movement while it was all still installed.
I found out that this case design requires the crystal and bezel to be removed first, allowing the dial and movement to be extracted through the front of the case. There was a small indent between the bezel and the case, and with a firm shove of the case opener I was able to get them to come away. The movement was now free to come out.
The movement could now be dismantled. First to be removed were the motion works containing the wheels that transmit power to the hands. Next were the main parts on the reverse side – the balance and pallet fork, combined barrel and train bridge, and the gear train. Finally the keyless works were removed. While these were being done, some of the issues in the movement had become clear.
One was that the movement hadn’t been serviced in a long time. There was lots of evidence of dried grease around the barrel arbours, and opening up the barrel showed plenty more black gunky stuff to be cleaned out.
Another issue was more unusual – many parts were magnetised, especially the ratchet wheel and its screw. Magnetism can speed up the timing rate so it would need to be removed – this would be a job for my demagnetiser.
All the parts were cleaned, demagnetised and then inspected for wear and damage, which is when the most serious issue was observed. The arbour hole in the combined barrel and train bridge was worn, displaying the dreaded “gold ring”. This would allow the barrel to have excessive free play, affecting the amplitude and possibly the regulation. One option to remedy this was to ream the hole and insert a bush of the correct diameter, but it seemed easier to try and find a replacement bridge instead.
Slightly at odds with the wear to the arbour hole, there was also wear to the balance lower cap jewel (on the mainplate) but none on the upper cap jewel (on the balance cock). If one jewel was going to be worn I would have expected it to be the upper jewel as that could be explained by a long period running in the dial up position after maintenance was due, and would be consistent with the wear on the arbour hole. It was a bit of a mystery.
EDIT by “future me”: It’s not unusual to see the barrel arbour hole wear differently to the cap jewels as it’s subjected to different forces. The difference in wear between the upper and lower jewels could be a result of the lower cap jewel not having the same amount of oil as the upper jewel, and once the oil had dried up it began wearing out sooner. Another possibility is that as the lower jewel is located on the dial side of the mainplate and needed the movement removed to access it, it might not have had as much attention as the upper jewel which is exposed on the balance cock (although it’s poor practice to oil only some jewels and not all, and especially without cleaning them). A less likely explanation is that the watch was continually allowed to run face down for really extended periods…
Taking a look around eBay, I happened to find a rusty 66A movement that was cheap to buy so I ordered it as a source of spare parts. I then came across a “new old stock” barrel and train bridge which was less of a risk than the one on the rusty movement, so I ordered that as well.
While waiting for these to arrive, I spent a bit of time cleaning the case and polishing the acrylic crystal with both finishing up very nicely.
Reassembly
Eventually all the parts arrived. The new-old-stock bridge was given a clean as it had probably been in storage for decades, but it looked pretty good and had no wear as would be expected from a ‘new’ item. It didn’t come with a Diafix cap jewel and spring for the escape wheel so I had to try transferring them from the original bridge. This was no easy task as they have a tendency to break, or they fling away from your tweezers never to be seen again. But this time the parts complied with the prodding of my tools and were successfully fitted.
The rusty 66A was dismantled with great difficulty – the stem would only come out after a generous helping of oil and vigorous wiggling. Getting to the dial side of the movement I was able to take out the balance lower cap jewel and found that it was still in good condition with no dimple to be seen. I gave it a clean and oil, and installed it into the movement.
The rest of the movement could now be reassembled. I started with greasing the mainspring and getting it back in the barrel. As this is a hand-wind movement only, no special braking grease was needed on the inside wall of the barrel which made the process slightly easier.
There was a bit of difficulty in using my mainspring winder as it was designed for mainsprings that coil clockwise from the centre, but this Seiko mainspring coiled in the opposite direction. What’s meant to happen is a small hook on the shaft of the winder should catch onto a cutout in the mainspring, but in this case it wasn’t holding securely due to the hook facing the wrong way. It was lucky that the shaft was a tight fit on the mainspring as this helped hold it in place while I very carefully wound up the spring (going in the opposite direction). Amazingly it seemed to work, and after a short while the mainspring was back in the barrel.
With the barrel ready, I proceeded with installing the parts on the main plate, doing so without oil at first. The keyless works were first, followed by the train wheels and barrel. It was good to see them move freely when prodded with pegwood. Endshakes also seemed to be okay. The crown and ratchet wheels were installed, and finally the pallet fork and balance wheel. The movement had now come to life!
Checking it on the timegrapher, the timing rates were significantly different in all positions and the amplitude was low, but these would improve once the parts were oiled. What was important was that the display traces were relatively steady without significant roller-coastering (is that a word?), and in this respect the traces were looking good.
The parts were now removed again and each one properly oiled before being installed again. Another check on the timegrapher showed the amplitude had increased to 250-270°, and some small tweaks of the regulator arm on the balance resulted in a decent timing rate of +2 secs/day in the DU position.
DD wasn’t that great, though, losing around -20 secs/day, while the horizontal positions gained about +15 to +20 secs/day. The large loss in DD makes me suspect a worn balance pivot which would also explain the dimple in the cap jewel, or perhaps I hadn’t cleaned or oiled it properly. But at this point I was content with its overall performance and decided to continue reassembling the watch.
The remaining motion works were installed, completing the work on the movement. Now attention turned to the dial and hands. The dial was in pretty good condition, with the lacquer intact and only a few marks to detract from the delicate sunburst pattern. A piece of Rodico was used to wipe down the indices and restore some of their shine.
Turning to the hands, the minute and seconds hands were previously damaged and needed straightening out. These were difficult to work on as they were Dauphine-shaped and curved. As such they couldn’t easily be straightened on the flat surface of a staking block as I’ve done on other hands in the past. I ended up spending quite some time on each hand with the tweezers to make them look acceptable although not perfect.
With the hands done they were installed back on the dial, and then the movement went back into the case. The crystal and bezel were pushed back on, followed by the case back. Finally a new leather strap was fitted which finished the project.
Reflections
I think my main takeaway from this project relates to timing regulation. I’m now starting to see patterns occurring between amplitude and the timing rate when the movement is placed in different positions (DU, DD, 12U, etc.). Both factors sometimes inversely correspond to each other, meaning that a higher amplitude can lead to a slower timing rate, and vice versa. This would imply that by focusing on improving the amplitude in the movement positions where it is lower, more consistent regulation could be achieved across all positions.
But what causes the drop in amplitude in some positions? There could be a number of causes, like wear on some parts, such as the pivots and holes. Incorrect oiling techniques or quantities. Or perhaps incomplete cleaning of the pivots. Currently I do a visual check for wear with my eye loupe, and clean all pivots by sticking them into a piece of pith soaked in isopropyl alcohol (this is after they’ve already been through the watch cleaning machine). I may need to improve my methods, by increasing the magnification to check parts more closely and/or using a better technique for cleaning. There was an online suggestion to use a polishing compound on the pivots to get them really clean. I’d like to return to this watch at some point to try this out and see if it does lead to an improvement, and then perhaps try it on other recent watch projects.
Another issue was that hand-winding the movement was also a bit tough as the resistance seemed to get greater as the barrel reached maximum wind. This wasn’t made easier by the small crown not offering much grip. I might need to change the lubricant used on both the keyless works and barrel arbour from Moebius D5 to Molykote DX to see if that makes it easier to operate.
The greatest satisfaction has come from how the watch looks, now that the crystal and case have been tidied up. I’ve worn the watch to work a few times and it’s been giving me a buzz – I love the retro feel and the way it straddles being casual but also elegantly dressy. The 33 mm size has really grown on me, I think because the whole design has been thoughtfully proportioned and seems to suit my 6.25 inch wrist. This is supported by design elements like the large, simply-styled dial (relative to the size of the case), the domed crystal and the barely-present bezel – it all just works.
Recently I’ve also been seeing more people returning to smaller watches, which is a great turnaround from the trend for oversized models that has dominated watch designs since the early 2000s. For many people, a smaller, well-designed watch could offer a much more satisfying experience. I’m sure this surprising little charmer will now become one of my regulars when choosing which watch to put on for the day.
Sky Corner 