The Watch Site banner

1 - 8 of 8 Posts

·
Registered
Joined
·
15 Posts
Discussion Starter #1
Up on the bench today is a scrappy Seiko Actus I inadvertently bought at an online Japanese auction having neglected to cancel my lowball snipe bid. With no other takers, I walked away with it at the starting price, although shipping costs to New York, via Tokyo, totaled considerably more than the hammer price. Even so, the appeal of that Japanese day wheel made it all worthwhile.


(Courtesy of unknown auction seller)
I’ve included the auction photo here (on the shaky legal premise that in purchasing the item I also gained copyright to its image) so you too can see it as I did on that blurry Sunday morning.



The Google-translated description of ‘junk’, ‘scratched windshield’, and ‘unoriginal belt’, was a little harsh (while simultaneously endearing) as, once delivered, unwrapped, and de-cased, it was all aglow…


Nobody does blue starburst dials like Seiko – and nobody can truly capture them in photos either.


The bright blue minute track adds a vivid highlight, matching the kanji day wheel – but only on Saturdays!

As an added bonus, the signature of a Japanese watchmaker’s service was evident inside the caseback, but alas, so too was the telltale scarring of a loose rotor at the bullseye and circumference:


Quite a bit of “human history” got under the rim of the caseback, but not much made it past the water resistant gasket.
A blast in the ultrasonic cleaner will clean this up nicely.


As best I can make it out, the inscription reads 511011, which I’d hazard indicates the 11th of October, 1976 (the 51st year of Shōwa era in the traditional Japanese calendar system), and that ties in neatly as a 5 year service after the manufacture date of June or Sept 1971 – decoded from the serial numbers on the dial and case backs:




Seiko had launched the world’s first quartz wristwatch in 1969, but they were still priced at the luxury end of the market as this watch rolled off the production line in 1971. By the time of that first service in 1976 though, the tide had turned and quartz-powered watches, with their futuristic digital displays, were more modern, more accurate, and now cheaper than their automatic watch brethren of yesteryear. Someone dutifully paid for that 5-year service, but that was to be its last, no doubt relegated to some dusty drawer by the eighties, where it lay in a 40-year slumber – from which it shall duly be reawoken this very day.

I’ll spare you all the details of an uneventful strip down, and get straight to the glory of the Part Art:



Can you spot the critical missing part that got left behind in the cleaning basket as I laid out this photo? Answers, for no prize whatsoever, in the comments below, ( excluding gaskets, which are still in their packets at this point ).


The Build

Tradition dictates that the keyless works go in first, although this is not really necessary as the 7019 movement does not hack (tsk, tsk!), but why not, as I’m refitting the jeweling on this side anyway.



The minute wheel is not really necessary at this point.

The lower end-piece (crescent, just below center), containing two jewels for escape and third wheels is a real timesaver over the diafixs favored by other Seiko hi-jeweled trains – just two straightforward screws rather than troublesome springs, but, on the downside, totally inflexible. Then the diashocked balance wheel jewel is clipped in, prior to oiling them all from the other side of the main plate.

A little lubrication is applied to the bearing surfaces ( not the pinion teeth ) of the center wheel…




… prior to its placement in pole position in the main plate, secured by its bridge.




At this point it’s wise to flip the movement over briefly and attach the cannon pinion, lest the incoming fourth wheel staff pivot should get pranged if the movement slips from its holder.
Up next – the mainspring, having been cleansed of decades of congealed black grease, is wound up…





… and pressed back into its barrel. Greasing protocols have evolved since the 70s, so it receives just a thin application of braking grease on the inner rim and a sparse coating of modern synthetic oil to the inner sides and arbor.





The barrel is placed in its main plate port along with the silver-colored escape wheel, then the brass third wheel top center, and finally the fourth wheel in the very center.




With a little application of (expensive) oil to the tip of the fourth wheel pivot, and to the barrel arbor (different oil, also expensive), we can gingerly refit the bridge that holds everything in place. Although a little inelegant visually, the integration of the automatic winding mechanism (the pincers) into the same train bridge elevates it to an object of design simplicity and efficiency.
Also of note are the two brass diashock springs that hold the jewels in place, chosen perhaps to add a little shock absorption to the third and escape wheel pivots, because, as noted earlier, the other ends are mounted in inflexible jewel caps.



Note the bridge has been visibly eroded – most certainly by the underside of that same loose rotor that had scarred the inner caseback.

So with the train bridge replaced (carefully – all four train wheels must be aligned before fastening down the screws), its time to fit the palette fork, here shown in place, with its bridge awaiting politely in the wings.





After the fitting of the ratchet wheel, lubrication of the exit palette stone and transfer of said lubricant to the escape wheel teeth – the details of which I’ll leave to a later post (once this blog has a comments mechanism robust enough to withstand the deluge of opinions that any mention of escapement lubrication is sure to incite) – all that remains is to replace the balance wheel and we’ll hopefully get a heartbeat from our long-dormant patient.

Not so fast! An issue this watch has, which likely contributed to its decades-long purgatory at the bottom of a drawer rather than atop a wrist, is that its hairspring – which provides the main pulse of the timekeeping – is far from its spiral ideal and has taken on distinctly eccentric profile:




Towards the bottom of the photo the coils are very widely spaced, while towards the top, they are all bunched up to the point of touching – this will result in very poor timekeeping.

The usual reasons for this are one or more of:


Magnetism – these hairsprings are made from ferrous alloys that can become magnetized making the coils stick to one another. The rest of the watch, screws aside, is either brass or stainless steel and is thus immune, but until recently the spiral hairspring (along with the mainspring) could only be made from steel. This magnetism can strike all too easily in the modern world, but fortunately, the remedy is simple – the whole watch can spend 20 seconds in a demagnetizer and it’ll be right back to normal. Any jewelry store can do this, or they are available cheaply on ebay.


Oil – stray oil from elsewhere in the watch gets flung onto the hairspring causing it to stick to itself. This is the inevitable final curtain when someone unskilled in the art tries to oil their watch at home – “I change my car’s oil – how hard can it be really?”


Dirt or rust – if the seals are no longer water or dustproof, grime and fluids can get in, which will cause havoc everywhere, but if the watch is running really fast then it likely got to the hairspring.



After demagnetizing and cleaning this hairspring, it stubbornly retains its lopsided posture – there’s really no option but to go in and operate.

Open Heart Surgery

First step is to remove the balance wheel from its cock, then carefully prise the hairspring from the balance wheel staff (being sure to take note of its orientation). Freed from these constraints the hairspring opens out into more or less concentric circles, indicating that there is nothing seriously amiss.



Blurry hairspring, with a ballpoint pen photobomber indicating scale

For this hairspring to work effectively though, it needs to be re-shaped such that it is correctly positioned at all 3 of its contact points. Let’s attach the bare hairspring back to the stud of its balance cock and see what needs to be done:



It’s already correct at point 1, the stud, and due to the pointed set screw and v-channel design will be reset at this same orientation everytime. So far so good.
The hairspring will need to be bent inwards so that the inner face of the outermost coil is ‘just not touching’ the regulating pin (white dot at 2), which it isn’t by a long way right now. An added complication is that the regulator arm, which 2 is attached to, has to be able to rotate (around the axis 3) about 90 degrees, all while maintaining the ‘just not touching rule’. This is how the timekeeping of the watch is adjusted, and we need to ensure our hairspring maintains its concentric spiral shape whatever rate adjustments we’ll need to make.

Finally, the brass ring at the center has to sit concentrically over point 3.

Let’s switch to the microscope. Using the shaft of a needle (literally – a fine sewing needle), we push the hairspring at the point of the red arrow – it’s actually more of a stroking action along the spring length while exerting pressure in that general direction. The spring is, as its name suggests, springy, so we have to push it quite a bit further than needed in order to have it bounce back most of the way, while leaving a little permanent deformation so that it ‘just doesn’t touch’ the white dot. With experience – mostly bitter – one gets to sense the faint pushback the spring exerts on the needle. Little by little is the mantra here.


I should point out, for this exercise the hairspring is loaded in the stud at a higher elevation than when it’s mounted for real,
hence it won’t connect with the regulator pin, merely hovering above it.



Once that is dialed in, we can move the regulator index to the 45 degree mid-point, and holding the spring steady at the previously set position (in blue) with the finest tweezers they make, again push down along the red arrow with the needle.


Needled at the red arrow, until the outer coil just doesn’t touch the regulator pin (white dot). Looks good here.

And once again with the regulator at the 90 degree extreme:



All that remains now is to coerce the quarreling centers back into peaceful co-axial relations, achieved by an artful combination of tweaking the spring’s angle alternately at the two elbows below:


At last, everything is how it should be.

Scroll down to the following post on this thread.........
 

·
Registered
Joined
·
15 Posts
Discussion Starter #2
A quick check from a side-on angle that the hairspring is still planar and parallel and we’re good to go.


Note the extended elevation of the stud for this hairspring manipulation exercise.

Because the hairspring is in the correct shape at the 3 critical points it can now be pressed back onto its balance wheel (recalling it’s previous orientation), the stud re-attached to the cock, and the regulator pins clamped around the hairspring, without becoming distorted, eccentric, or touching. We are good to go!

Mounting the resurrected balance wheel back onto the movement….



…and the heart is beating for the first time in decades. The patient is allowed to enjoy it’s convalescence overnight, giving those expensive oils time to diffuse deep into its aching joints.

At doctors rounds the next morning, after various pokings and proddings, the patient is showing healthy vitals on the ECG – pulse averaging +3 seconds per day, blood pressure at a steady 270 over 235, and arrhythmia under 0.5ms, but what caught my eye was the variation of rates across the vertical positions. Unlike its quartz contemporaries, a mechanical watch is susceptible to changes in its timekeeping depending on the angle at which it is held, an ailment we will now endeavor to minimize.


[/B]Dynamic Poising[/B]

Clearly, the angle of the watch changes as the wearer’s wrist goes through its daily motions, but in a clinical setting these changes are usually sampled at 6 key positions:

Two horizontal positions:

DU – dial up, as when the watch is resting on a table.

DD – dial down, as when it is turned over, resting on its crystal.

Four vertical positions:

3Up – as when the wearer is pointing up into the air, aka Crown Up – CU

6Up – as when the wearer is grasping a mug of coffee, aka Crown Left – CL

9Up – as when the wearer’s arm is down by their side, aka Crown Cown – CD

12Up – as when the wearer is looking at the watch in front of their face, aka Crown Right – CR.

I prefer the use of the NumberUp nomenclature, rather than the position of the crown, as it allows for easy naming of in-between positions, and avoids the issue that crown left and crown right depends on which side of the movement we are viewing from.
Looking at the timegrapher results, it is the vertical positions which are troubling in this watch as, when fully wound, the gold line on the graph below shows changes in timekeeping, (known as “delta”), that are a full 25 seconds between the fastest and slowest rates. I’ve included some extra in-between data readings as well to make the curves smoother. The green line is ‘perfect time’ – that’s what we’re aiming for (without buying a quartz watch of course).

Also shown in dark green are the results with a very low mainspring winding – the watch is an hour or so away from stopping altogether at this point. Amplitudes in these 2 vertical positions are averaging 235 degrees, (Seikos are designed to work at lower amplitudes), and 130 (anything much below this and the readings get noisy).


Rate variation at high wind (gold) and low wind (dark green), across 12 vertical positions.


The dark green line is on average a lot slower, as we might expect, but also shows a lot more variation between fastest and slowest – 201 seconds. More surprisingly, it is now slowest around the 5 position, where previously it was fastest, while 180 degrees later, around the 10 position, it is now fastest where it was slowest. The timing has been turned on its head!


After some trial and error, I home in the amount of mainspring wind, somewhere between the two previous settings, that flips the curves from one to the other, and it is just here, balanced on a horological knife edge that the watch keeps reasonable time – less than 10 seconds variation along the orange line.


Orange line is closest to flat – vertical amplitude averaging 211.

The prospect of monitoring my arm movement so as to wind the automatic watch no more nor less than this perfect winding is no more practical than keeping my wrist at some optimal angle in the first place, so we are going to have to address the issue that is causing these wild swings in rate in order that the watch runs optimally at most usual levels of wind and most usual positions.

And my diagnosis is that it’s a poise error – that is, the hairspring and balance wheel is not, if you’ll pardon the phrase, in balance. Switching shamelessly to an automobile analogy for a moment, it’s the same issue that is addressed by adding weights to the rim of a car’s wheel.



In a watch’s case, the tire is the rim of the balance wheel, but adding weight is no longer an option on modern smooth balances, so we’ll have to remove weight on the opposite side instead – an action that is irreversible. The underside of the balance wheel already shows the factory poise work – the neat scar at 4 o’clock machined nearly half a century prior – but this was likely done by static poising, which can only go so far in correcting the errors when the balance wheel is in motion. By undertaking dynamic poising, we aim to correct for the combined weight imbalance of the balance wheel, the hairspring (even as it winds up and down in motion), and the collet that holds the two together.

The milling mark on the right of the balance rim is the factory poising.
(Sorry, this photo was taken before the hairspring adjustments)

While a car wheel exhibits its poor wheel balancing with an unnerving vibration felt when driving at a certain speed, a watch’s balance wheel reveals it’s poor poise by changes in rate depending on vertical position – particularly noticeable at low amplitude. At low amplitudes the rate is slowest when the watch is orientated such that the heavy spot on the balance wheel is directly above the balance staff (axle). At high amplitudes meanwhile, poor poise will produce the fastest rate when the heavy spot is directly above the balance staff. And at middling amplitudes (theoretically at 220 degrees), there should be no rate bias in any position from poor poise.
This can all get a bit confusing, so I created a poising chart (only valid for Seiko 70** series movements) that allow a quick conversion from the low amplitude (dark green line above) graph’s slowest reading (around 5 in this case) to the point on the circumference of the balance wheel where weight needs to be removed:


The red dot shows the impulse jewel on the yellow balance wheel.


So a notch is cut on the balance wheel at the place corresponding to the poising chart (around 5).


And another round of measurements on the timegrapher:



The newly added brown line shows a distinct improvement as the trough is less deep and the peak less high. Too little is better than too much, so let go in for another round, noting that the low spot has now shifted to 4.



A second notch is cut at the 4 mark of the poising chart

Back on the timegrapher, reveals a further improvement with the pink line:



Just a little more filing at 4 and I think our work here is done, the new line (black) is still wavy, but it no longer looks like a poise issue, as it’s contours match (rather than mirror) those of the new full wind line (purple), which is showing a reduction in its delta of 10 seconds. Phew!



Poising Theory :


This is a chart from Hans Jendritzki’s Watch Adjustment showing the theoretical curves of the same poising rate changes vs position, at various amplitudes. The theoretical amplitude of inversion is 220 degrees, rather than 211, which was the closest I could get. My graphs show real-life data, noise and all, with lines showing rates at constant power not the constant amplitude of Jendritzki’s graph. Here are the corresponding amplitudes for the previously presented data, prior to poiseing:



The only way to get lines of constant amplitude would be to take multiple readings over the course of the mainspring naturally winding down, and then, for each position, select the datum at the correct amplitude.
Even I don’t have sufficient time on my hands for such an undertaking right now – but watch this space!


I should add that other hairspring & balance issues can cause this characteristic sine wave variation in timings, but only poor poise exhibits the characteristic flip in phase around the 220 degrees mark. Also, other prior adjustments needed to be made on this movement to produce timings clean enough to reliably detect its poor poise. Techniques for dealing with other issues will be covered in posts of future watch restorations – fear not, I’ve quite a backlog!


So with the movement ticking and healthy, we can turn over once more to rebuild the remaining calendar side:



First adding in the hour, intermediate and date driving wheels




Followed by the date wheel and its attendant finger (top left by the ’25th’), that advances the date every 24 hours,
and its jumper (bottom by the ‘3rd’), that snaps it into the next position, keeping it aligned in the dial window.


At this point in production, Seiko was still using all metal parts for the calendar wheels and fingers. They soon moved on to those new fangled plastics, which, while running smoothly without lubricants, (and lowering costs), broke with centuries of horological tradition, and, to my mind, just don’t provide that same sense of pride upon successful restoration.


Keeping everything in place with the date dial guard, the final part of the puzzle is to replace the day-disk with its finger and jumper.
(a snap clip, not pictured, then goes atop the day-disk)



For the same reason that Seiko even included an English option for days on this Japanese domestic market watch, I choose to use the kanji option despite my mastery of the language progressing no further than a passable ‘hello’.


Dial and hands are returned to the movement – it’s starting to look a lot like a watch now.

With the inner workings of the watch all but rebuilt, it’s time to take stock of the case. The crystal is made of Hardlex, Seiko’s branded strengthened glass – clearly not strengthened enough to survive the decades, as the original auction picture testifies, and with no feasible option of polishing, there is little choice but to seek a replacement. With the very few new-old-stock hold outs (Seiko part 310W13GN ) commanding ever-higher prices, I was pleased to find that a German aftermarket hardened glass crystal, designed as a drop-in replacement for a similar Seiko part, (310W17GN), was also a good fit for this model, albeit with a flat rather than slightly domed top.



The versatile Sternkreuz XMF 310.848 (right) to the rescue once more.
( do not assume it will also work in other 310W13GN cases though ).



The case and bezel cleaned up nicely, although no amount of ultrasonic cleaning, fiber brushing, or elbow grease was going to remove the pitting and staining under the caseback track.
My initial interest in this model was piqued due to my having a spare bracelet in good condition from a previous restoration, albeit in need of a little freshening up. This is easily done for brushed stainless steel with some heavy-duty scouring pad. In the way that blue scourer is used for delicate non-stick pans, and green for regular dishes, there is an extra heavy-duty purple variety, not often seen in stores but readily available online, that will cut through the surface of steel leaving a trim brushed finish.

This bracelet however, has alternating links of polished and brushed steel. I chose to polish all the center links in one go with a rotary tool and then laboriously mask off rows of links to brush with a pad one at a time….



Scroll down to next post in thread.......
 

·
Registered
Joined
·
15 Posts
Discussion Starter #3
With everything clean and polished its time to drop the movement back in the case, fit fresh, greased rubber gaskets to the caseback and crown, mount a replacement auto rotor in place of the loose original, and snap the back cover into place to seal it all up.



Note the arrow cut-out in the rotor aligns with the hole in the reduction wheel – thus ensuring that the winding system works at peak efficiency when the wearer walks with hands down at their sides.


And so the restoration is complete, and all that is left to do is ceremonially strap it to my wrist and enjoy its vintage charms.






Sporting the restored correct bracelet.



After all that polishing, I’ve rather taken to this simple single-pass seatbelt strap.


After a few days wear, the regulation of timing can be dialed in to suit my personal pattern of wear & activity, resulting in pleasingly accurate timings:



As a side note, these timings straddled the introduction of daylight savings time on March 8, but luckily this watch doesn’t hack so I was able to advance the hour without affecting the timing – there’s always a silver lining!


Daily wear shows timings between -0.4 to +1.6 seconds per day.


Overnight on the nightstand, resting dial up, shows timings between -0.9 and +1.7 sec/day – curiously a higher variance than when worn.


The overall average for this week of wear was +0.6 s/d, which I’m very pleased with for a timepiece of such humble caliber, and considerable vintage.



Some might say this is a lot of trouble to go to for quite a pedestrian watch – but such is the nature of obsession. It’s taken pride of place on my wrist this past month, dazzling in the spring sunshine, but I really must learn to exercise more diligence at those online auctions…..


Copyright © 2020 Horology Obsession. All rights reserved.


Suggested reading:


Hairspring manipulation: https://www.youtube.com/watch?v=idO5elKgFMA&t=9s
Dynamic poising: https://adjustingvintagewatches.com/category/dynamic-posiing/

New parts:
Crystal: https://www.cousinsuk.com/product/special-profile-by-brand-model-sternkreuz-xmfo?code=XMF310848
 

·
Registered
Joined
·
6 Posts
With everything clean and polished its time to drop the movement back in the case, fit fresh, greased rubber gaskets to the caseback and crown, mount a replacement auto rotor in place of the loose original, and snap the back cover into place to seal it all up.



Note the arrow cut-out in the rotor aligns with the hole in the reduction wheel – thus ensuring that the winding system works at peak efficiency when the wearer walks with hands down at their sides.


And so the restoration is complete, and all that is left to do is ceremonially strap it to my wrist and enjoy its vintage charms.






Sporting the restored correct bracelet.



After all that polishing, I’ve rather taken to this simple single-pass seatbelt strap.


After a few days wear, the regulation of timing can be dialed in to suit my personal pattern of wear & activity, resulting in pleasingly accurate timings:



As a side note, these timings straddled the introduction of daylight savings time on March 8, but luckily this watch doesn’t hack so I was able to advance the hour without affecting the timing – there’s always a silver lining!


Daily wear shows timings between -0.4 to +1.6 seconds per day.


Overnight on the nightstand, resting dial up, shows timings between -0.9 and +1.7 sec/day – curiously a higher variance than when worn.


The overall average for this week of wear was +0.6 s/d, which I’m very pleased with for a timepiece of such humble caliber, and considerable vintage.



Some might say this is a lot of trouble to go to for quite a pedestrian watch – but such is the nature of obsession. It’s taken pride of place on my wrist this past month, dazzling in the spring sunshine, but I really must learn to exercise more diligence at those online auctions…..


Copyright 2020 Horology Obsession. All rights reserved.


Suggested reading:


Hairspring manipulation: https://www.youtube.com/watch?v=idO5elKgFMA&t=9s
Dynamic poising: https://adjustingvintagewatches.com/category/dynamic-posiing/

New parts:
Crystal: https://www.cousinsuk.com/product/special-profile-by-brand-model-sternkreuz-xmfo?code=XMF310848
What a fantastic bit of magic and vision. Looking forward to the next reincarnation:)

Sent from my SM-G965F using Tapatalk
 

·
Premium Member
Joined
·
9,285 Posts
when's the movie coming out? (y)

excellent documentation and photos - thanks for the effort ...
 
1 - 8 of 8 Posts
Top