The electronic balance-driven movement: horological cul-de-sac or brilliant piece of lateral thinking? I am inclined to think a bit of both. In the late 1950’s when electric watches were first developed, the concept of a battery-powered balance-driven movement will have presented a number of compelling advantages over a mechanical movement. From a marketing point of view, the most obvious of these is that it will just keep running, regardless of the extent to which the owner interacts with it, for as long as the battery lasts. An automatic mechanical watch may not need hand-winding but it will stop within a day or two if it is not worn.
One of the benefits of an automatic watch compared to hand-wind is that it maintains a constant state of wind, once fully wound, for as long as it is on the wrist of a reasonably active wearer. However, once removed and set to one side, the mainspring winds down, the power to the gear train diminishes, the amplitude of the balance reduces and the accuracy of the movement is affected. Of course you can regulate a watch to take account of faster running overnight, and fluctuations in timekeeping with the state of wind may not be particularly apparent from day to day. An electronic balance-driven watch however can maintain a constant amplitude day and night because the power being transferred to the gear train does not fluctuate with the state of wind of a mainspring. It is a function only of the capacity of the battery to provide an unwavering impulse to the magnets on the balance wheel.
The earliest electric balance movement developed by Hamilton used delicate contact switching that proved unreliable and very difficult to work on but diode-switching (LIP) and subsequently transistor-switching (Citizen and ESA) solved that problem and the electronic balance-driven movement reached a sufficient degree of refined maturity to allow confidence in it as a reliable option for those attracted by the benefits of an electronic watch.
In a mechanical movement, the mainspring provides the source of power to the watch and it is the job of the escapement to regulate and measure out that power to the gear train. An electric or electronic balance-driven movement turns that concept on its head by assigning responsibility for powering and driving the gear train to the balance. That drive comes from the interactions between the magnetic fields produced by the current passing through the coil and the permanent magnets fixed to the balance wheel. The transistor’s role is to facilitate the periodic switching of the current to the impulse or induction coil as the balance swings back and forth. The instruction to switch the current on and off to the impulse coil comes from a current induced in a trigger or primary coil as the magnets on the balance pass over the coil.
Seiko’s entry to this market was a little laggardly compared to their main domestic competitor but nevertheless, by 1968 they had introduced their own electronic balance-driven movement in the 3100 series, superceded quite quickly by the 3700 series. Seiko marketed the EL-370 electronic watch from about 1969 through to about 1972 when it was replaced initially by the 3200 and 3300 series and then the runout Elnix series of electronic watches fitted with the 0703A caliber. I’ve written in some detail before about electronic watches in general and the Elnix in particular and I don’t propose to revisit too much of that back-story again here now. Instead, let’s get to the nub of this post and introduce two examples of Seiko’s first foray into electronic watches.
To the left, we have a Seiko 3702-7000 from December 1970 and to the right a 37-7000 from January 1971. The two watches were produced within a month of each other and differ most conspicuously in the colour of their dials and in that the older watch sports a date complication and the younger, none. The black watch came on its original mesh bracelet and the white-dialed watch as head only. Both watches are in decent cosmetic condition, with the only real external flaw some marks on the surface of the dial of the 37-7000. Both watches exhibited signs of life with a fresh battery inserted but neither ran properly. I propose to catalogue the revival of both watches in a single post but we’ll separate the two to avoid confusion. Let’s begin with the 3702-7000.
You can see from the case back that there is no separate battery hatch but there is a battery hump to accommodate the protrusion of the battery above the level of the movement. The case back itself is secured in place by a separate screw-down ring, much like the arrangement we’ve met before with the Seiko Cronos Sea Horse and Silverwave.
If we open the caseback, we are greeted by the still slightly incongruous sight of, on the one hand, what appears in part to be a conventional mechanical watch movement given the presence of a balance, but also clearly some electrical componentry and space for a button battery.
The movement is held in place by a substantial case ring and with that removed, and the crown and stem released, the complete movement drops out onto a case cushion and is then transferred to a dedicated Seiko S-620 movement holder.
The dial and hands are largely free from signs of wear or abuse, although the latter just a little grubby (but that is easily sorted with some Rodico putty prior to refitting).
The calendar side of the movement initially presents as conventionally arranged, the date dial held in place by a guard and the date driving wheel visible beneath. However, with the date dial and its guard removed, we see that this movement features a separate calendar plate, whose role appears to be to host the date corrector finger, date jumper and its spring and to secure the remaining calendar and setting parts. Removing the calendar plate reveals the setting parts, date corrector, hour wheel and so on.
With the movement stripped down as far as the setting lever spring, we can turn it over and tackle the business side.
It is not immediately obvious from this bird’s eye view but the coil sits between two balance wheels and so we cannot remove the balance without first extracting the circuit block and coil. The white cover is removed first followed by three screws securing the circuit block to the main plate. As the last screw loosens its grip, the block gently moves outboard, aiding its removal.
With that out of the way, we can make out the second setting lever (indicated below) as well as the retaining pawl just below the balance wheel.
The job of the retaining pawl is to prevent the escape wheel from rotating in reverse due to reverse torque created by actions such as setting the hands.
We get a decent view of the escape wheel and pinion in the photo below. A very different looking device than what we are used to seeing in purely mechanical movements.
The friction spring serving the sweep second wheel, the bridge and train wheels are all removed next, exposing the centre wheel bridge and centre wheel.
We are pretty much done now with the disassembly. The only remaining tasks are to separate the Diashock springs and cap jewels from their settings, remove the cannon pinion and centre wheel and deconstruct the remaining setting parts.
With that little lot dispatched to the cleaning machine, I can turn my attention to the case.
No great dramas await here. Light quantities of aged slime in the bezel recess but very little in the way of corrosion.
A round or two of pegwood, toothpaste and elbow grease results in a spick-and-span set of case components.
A new Seiko OEM crystal clicks into place without protest, capped with the cleaned bezel.
Back to the now clean and dry movement parts. First order of business is to reassemble the setting parts.
The centre wheel and its bridge come next followed by a start at arranging the train wheels. In the photo below, I’ve placed the escape wheel and the third wheel, with the seconds sweep wheel waiting in the wings.
I’ve paused at this point because the fitting of the sweep seconds wheel and pinion requires the third wheel to be tilted to one side to allow the toothed shaft of the seconds wheel to clear the teeth of the third wheel. The relationship between the two with the train bridge now fitted is shown in the photo below.
The pallet fork and its bridge come next followed by the retaining pawl. The latter requires a little care in fitting to make sure that the pawl screws are properly centred in the channel before tightening down fully.
We are now close to the point of being able to test whether this thing runs. Fit the calendar side Diashock, the balance and its Diashock, then the circuit block and all we need is some power.
Simply pressing a battery into place in the insulating cap will not close the circuit. You need the special power cell holding spring and screw that comes with the S-620 movement holder.
The movement sprang into action as I was tightening down the holding spring screw and a little regulation later was showing clean, steady output with negligible beat error and a flat timing curve. The only minor issue was that my timing machine struggled to come up with an amplitude value, requiring a good minute or so before it would produce a figure. The technical manual suggests greater than 200 degrees is a sign that all’s well but initially I was getting readings around 165 degrees. Flipping the movement over however saw this increase to about 220 degrees (again, after a long wait). I didn’t know at this point whether to trust entirely what the timegrapher was telling me and so devised a means of measuring the amplitude directly. I’ll describe that process when we get to the other watch shortly. For the moment though, I was quite happy with the performance of the movement and pressed on with the remaining calendar parts, followed by dial and hands.
This pretty much marks the conclusion of the work required for this 3702-7000 and all I need to do now is fit the movement into the cleaned case and survey the result prior to fitting a fresh case back gasket and then case back.
Having cleaned the original mesh bracelet in the ultrasonic bath, we are ready to refit and take a look at the complete watch.
From the perspective of 2019, it manages to pull off the combined effect of sci-fi futurism and old-fashioned at the same time. I’m pretty sure that the old-fashioned vibe derives as much from the mesh bracelet as anything else.
On to part 2, and our classic three hander 37-7000.
The dial looks a little grubby but overall the watch makes a good initial impression. The movement looks fine but I note that an additional contact tab has been attached to one of the screws on the train wheel bridge. Perhaps this has been added by a watchmaker to test the movement with the caseback removed and had been left behind in error.
The main cause for concern with this watch was the condition of the balance hairspring. In the photo below you should be able to see that it is twisted quite badly out of plane at the point that it enters the terminal curve towards the index pins and stud. I suspect that this is going to be beyond my abilities to correct and so I set about locating a replacement balance wheel.
We’ll get back to that later. Let’s get the movement out of the case and note some differences compared to the 3702.
I am not too concerned with scuff marks on the dial. I think it will clean up well enough to make for a presentable watch. The hands are decent too. With dial and hands removed, we see that there is a great deal less to take in on the dial side of the movement with none of the calendar features present. In particular, the lack of a calendar plate means that the centre wheel, cannon pinion, sweep seconds pinion and hour wheel are all shorter than in the 3702A, not having to protrude through a calendar plate.
The balance side of the movement looks identical to that of the 3702A, other than the different calibre numbering inscribed on the balance cock.
With the movement parts cleaned, the first order of business is to deal with the bent hairspring. While I was working on the 3702-7000, I’d located an old stock balance wheel from an eBay seller and had also bought 3702 spares watch from Yahoo Japan, the latter not just to provide a balance wheel back-up but to provide a source of any other parts I might need now or in the future. In the end, I opted to use the balance wheel from the spares watch rather than the ‘new’ part. Here they both are, having just emerged from the cleaner.
The reconstruction is executed in the same way as with the other watch, minus the calendar componentry.
The circuit block is fitted in the same way as before and with a fresh battery in place, the watch springs into action.
As with the 3702A, when it came to performing the initial regulation using the timegrapher, I was able to get a nice clean timing curve with negligible beat error quite easily but the timer would take a good minute or two to produce an amplitude and when it did, it ranged from as low as 145 degrees to a high of about 170 degrees (see below).
The technical manual is very clear in suggesting that anything below 200 degrees is below par but the behaviour of the timing machine suggested that the indicated amplitude may not be the same as the actual amplitude. How then to measure the amplitude for myself? Well, I have a decent camera with a 120 frames per second x5 slow motion mode. So I set up my lights and camera and filmed the running movement, first at regular speed (@4k 25p) and then at 5x 120p.
The amplitude of a watch movement is defined as the maximum displacement from the mean position of the balance wheel. There are two ways to measure this, assuming one can capture the balance at one or other or both of its turning points. The first is to measure the angle of each turning point relative to the position of the balance at rest. The photo below shows the balance at rest, with no current flowing to the circuit block.
The second approach is more direct in that it simply requires us to measure the total displacement from one turning point to the other and then to divide that angle by two. I opted for the second approach. In the photo below, we see the two turning points captured from the slow motion video and the measured angle between the two.
The one crucial point to appreciate here is that the balance travels through more than a full 360 degree rotation between each turning point and so in order to determine the total rotational angle, we must add 183.5 degrees to 360 degrees which yields 543.5 degrees. Dividing this value by two gives us our amplitude of 272 degrees, give or take a half of a degree. That’s more like it! I am reassured and pleased and can press on to complete the watch.
The dial cleaned up quite well and the hands look sharp after their clean.
All that remains is to deposit the movement into the freshly fettled case, fit fresh gaskets to the crown and case back channel and we are ready to close her up.
Turning the watch over and I am surprised at just how clean and crisp it looks and indeed quite how different it feels from the other watch.
The final step is to identify a suitable strap. In the absence of a great deal of thought other than to locate something that would fit the 18mm lugs, I selected a shiny black lizard strap that I think works rather well.
I have enjoyed working on these two watches, in part for the modest challenges they have presented, but also because in spite of the fact that they represent a developmental blind alley in the horological timeline, they are nonetheless important watches and in use, something of a delight. I’ve been wearing one or other of these two pretty much continuously for the past three weeks. I’ve enjoyed them as a pair too because in spite of their obvious close similarities, the contrast of black to white dial, date/no date, mesh bracelet to lizard strap permits two quite distinct experiences on the wrist.
Thanks for your support in 2019. I hope you all have a relaxing holiday season and a happy and prosperous 2020.
Very interesting post about these electronic watches, I did not know, in my memory were only the first quartz whose needle jumped from second to second. Happy holidays and an excellent 2020 I wish you.
Thanks again for your inspiration Martin – it is largely your fault that I have just managed to clean and reassemble my first movement, and entirely your fault that I have a birth year Seiko weekdater on the way!
A couple of questions if I may? Any photography tips? I try to take lots of photos as I am working (to aid the reverse process), but often plagued by flash reflections, or poor depth of field – usually rendering invisible the part that has become critical!
Also, can you recommend any Seiko reference books? There appears to be one authoritative one, but I have only seen it in Japanese.
Best wishes for the New Year
Hi Brian, off-camera flash works well to avoid unwanted reflections, or use soft, strong continuous light source. Small apertures will help with depth of field but of course that cuts down the amount of light. If you are using a small sensor camera, then that can help too.
Wow! This article just blew my mind. Such an amazing read.Thank you for the value offered here:-)
No problem. Glad you enjoyed it.
JONATHAN BENTHAM said:
Great write up
It was what tempted me buy the non date Jeanette case as in your article
I have a question?
Will all EL 370 movements fit in all the EL 370 watch cases?
They should do. The only caveat would be whether the case is crown at 3 or 4 in which case you may need to think about alignment of printing on calendar disks. But in terms of fit, should be ok.
JONATHAN BENTHAM said:
Thank you so much for your rapid reply
I have a day date battered case but good movement arriving soon
May buy a stainless case NOS and slip the movement in
All for around $100 us
Thank you again
Thanks for this teardown, and for the Elnix one.
My interest is in pocket watches with electronic movements, which is pretty much as niche as you can get. So far, I have found out about two (or three, depending how you count) EL-370 models, & two Elnix.
The pocket Elnixes (Elnices?), 0703-3000 & 0703-3010, are the same except for the case material & finish, which is stainless steel on the -3000 and silver with a kind of golf-ball texture on the -3010, & the dial finish, which is heavily textured on the -3010. The odd part is, both dials carry the same number, 0703-3000R. Have you ever encountered that? They have an awkwardly cramped date window at 4:30, which I’ve also seen on other Seiko pocket watches of the mid-70s. Somebody must have thought it was a nifty stylistic choice. Both models appear in the Seiko JDM catalogues I’ve seen, from 1974 to 1977.
The EL-370 pocket models, on the other hand, could hardly be more different. The 3702-0010 uses a kind of “flying saucer” steel case also fitted to some of the pocket Skyliners. It exists with both a black & a white dial, but I have not yet obtained a white-dial specimen, so I can’t say for certain whether they have the same dial code. The date window is at 6:00, which to me is the best position on an open-face pocket watch. The 37-3000, on the other hand, is a bizarre stylistic monster. In a silver case, with a reeded edge like a coin, it has a dial with a displayed double-headed heraldic eagle embossed on it. Overall it looks like a thaler of the Holy Roman Empire. Also no seconds hand, which I find annoying. Most specimens have a caseback decoration (not engraved, but embossed or something of the kind) commemorating the 70th anniversary of the Dentsu advertising agency, but I have seen one offered for sale on Yahoo Japan without this. Neither of these seems to appear in any of the catalogues I have seen, either the regular JDM ones or the “Special Luxury” editions.
I have found no evidence of an EL-310 or EL-330 series pocket watch, but that does nothing to tell me whether one exists. The 33EL movements are unique so far as I know, in that the electromagnet acts on the pallet fork, rather than the balance. And as for the EL-320, a couple of ladies’ models show up in the catalogues, including a pendant in a solid 18K gold case, but I’ve never seen one of any kind for sale.
The only Seiko pocket watch I have experience with is a railway pocket watch powered by the 7550 quartz movement (see the post on the 7549 Tuna divers watch). It is always so interesting to hear about fascinations of others with niche corners of the watch world. It all makes for a more interesting tapestry.
As to your dial question, 0703-300R will be the dial code but I don’t know if those would have changed according to the surface finish. Do they otherwise have the same appearance?
Yes, the two dials are the same in all other aspects that I can discern. The location, shape, & appearance of the printed numerals & Daini logo, as well as of the applied Seiko name & Elnix logo, & the location of the date window are all the same. The date wheel also appears identical, what I can see of it through the window.
The hands, however, are different. The -3010 uses black hour & minute hands, with “spade” tips, whereas the -3000 uses bright metallic straight hands. As a result, the -3010 is a great deal easier to read.
I actually bought a friend one of those railroad watches, by the way. It made a great birthday present for a railfan with fond memories of a visit to Japan.
That all makes sense. The dial code must just reference the layout, font, numerals and so on.