Accutron 214 - coating coils after rewind?

The potting/coating on Accutron 214 coils is soluble in acetone, allowing me to strip the coil form without damaging it. I have successfully rewound my first coil, and now could use advice about what to recoat the wire with to protect it. I think after hardening it must be soluble in acetone or otherwise be removable, like the original, so some future person can rewind it again if necessary. A lot of the potting products I am finding say they must be removed "mechanically" which is not acceptable here. I would rather leave them uncoated than to make it permanent. "First, do no harm" applies in watch repair too! Any thoughts about what modern product I could use, or what Bulova used originally?

New feedback coil:


New completed coil which could really use a protective coating:

 

Hmm I wonder if the UV-curing glass crystal cement is soluble... That would be really convenient to use.

 

Dave do you mean Esslinger part 12.202? It says "only cures in the repair area" which seems weird.

 

Thanks everyone, I've ordered some of the 12.202. I'll rewind a second coil today and if I get the resistances closer to what I want, I'll coat it. Then we will know in about 2050 whether it's as good (or bad) as the original coating.

Do any of you know the design resistances? I have a huge variation in the coils I've measured out of my pile. For the 3-wire cell coils (most of my dead ones are 3 wire so I started on those) I have 5.61, 5.61, 5.62, 6.02, 6.29, 6.73 kohm on the drive coil, and 1.81, 2.21, 2.26, 2.38k for the feedback coil.

Everything I find online has different numbers for the coil resistances. I suppose they all come from people measuring coils like I am (or worse, just one example). Or maybe the design spec was just number of turns and that's why they vary so much in resistance? I could try 2000 and 6000 turns and see what I get.

Do any of you have a bigger pile you could measure?

 

Skutt50 you are right except you have it backwards: obsolete mercury cells were 1.35v and today's silver oxide cells are 1.55v. This does cause problems in some 214s and I think you are right and I might be able to compensate by altering the number of turns.

 

My second rewind with 2020/4890 turns, 2.03k/6.47k looks much better on the current draw. I think this is about what to expect with an unloaded fork (it's not turning a gear train):


When my UV glue gets here, I'll try coating this one.

 

Thanks Samantha, I've seen that on the web in several places, but not from a primary/contemporary source. I'm suspicious that their wire is a little thicker than that, though, from my empirical results. But to be fair, I don't have a way to measure the original wire or my new wire directly. I've only confirmed my wire is 0.015 by measuring a length of it and checking its resistance. (This works better than you might think, because there's a square law in effect: I am confident mine is between 0.014 and 0.016.)

Do you remember where you saw those numbers? In a quick google image search I see an ad that says "Over 500 feet" which would be 76m/coil, not 90 -- a lot closer to what my results suggest might be the case. If it were 90m, might the ad have said "About 600 feet"? It's hard to say...

 

How cool! Does it say anything else related? Can you share a picture with us?

 

Dave I got that glue and the UV lamp Esslinger sells. (I've used UV glue for crystals and eyeglasses before, but always inconveniently used sunlight.)

The glue is a perfect viscosity and goes on really nice and looks so smooth and shiny on there. But I put it in their light and the darn stuff doesn't set up, even after an hour. I'll let it sit under the light overnight and hope for the best, I guess. I wonder what the deal is! Maybe there was something to the weird warning.

 

That could be! After sitting in the light overnight it's pretty solid, but still dents when I poke it. I'll leave it longer I guess...

Maybe to make this thinner they've cut it with something that isn't really UV-curing. I'm going to keep looking, I don't think this is my solution yet.

 

Well that's no good. Which one do you have? I have 13.523. There is also 13.526 which looks like it has a completely different kind of bulb.

 

I did finally get it to set up. I like the result.

I got advice from a gentleman off the forum that adding 800 turns to the drive coil should compensate for the higher voltage of new batteries, and I'm going to try that on the next one. I think there will be room if I'm careful to get the layers nice and flat, but if not, maybe I can remove some turns from the sense coil to make room.

I have been testing each rewind with the same circuit side and tuning fork, and I can't be sure I've caused any amplitude differences yet in my trials. It seems like to be sure I'd need some kind of optical measuring system, ideally to directly measure how far the drive jewel is traveling. Does anyone have ideas about something I could get that's not too expensive? If I turn my microscope up all the way it's easy to see that the jewel amplitude is maybe 30% of a jewel length, and the regulator clips oscillate maybe 80% of a regulator width, but I have no confidence I can tell (say) 80% from 70% this way, especially if I take the eyeballometric measurements on different days.

 

For the latest test rewind of a 3 wire cell coil, I used 1800 turns on the feedback coil and 6895 on the drive coil, giving resistances of 1.80k and 9.22k (a factory coil is around 5.6 to 6.5k). The form is very full but it does fit. There is still room for the UV-cured coating. Here it is running in my test rig (same tuning fork and component coil I've used for all testing):



This gives clearly reduced amplitude (by eyeballometric measurement) and current draw (3.0uA vs 4.5uA). This will give a more reliably-running watch that doesn't triple index (gain time) when bumped, and uses the extra voltage of today's batteries to give extended battery life instead of giving trouble. I would expect it to add months to the battery's life.

This is a huge win! I'm so excited!

 

Mostly for protection during servicing.

The watch would work fine with the uncoated coils, because as you say they don't move at all, but they would be very vulnerable to damage at the next servicing. Even correct servicing procedure, like flipping the tuning fork base up to remove the coil lead retainer plate, could damage the wire, especially where the ends leave the coil and attach to the soldered contacts. Also the coil screw is very close to where the leads connect, and they are vulnerable to a screwdriver slip or other mistake.

I think it's also possible the coils would vibrate a very little bit, and the windings might rub against each other and eventually wear through the polyurethane. I do not know how real that risk is.

My goal is for my rewinds to last longer than the originals, as they clearly had some kind of defect. I think coating them is an important part of meeting that goal.

I'm currently servicing a 214 to see what amplitude my low-current coil gives when running a clean gear train. I ran into some rust on this one so it's taking more time. I will let you all know!

 

Awesome news! The 3-wire cell coil I rewound with 1800 and 6895 turns that runs 3uA unloaded takes 4.3uA running a freshly-serviced movement.

As best I can tell, the amplitude is exactly 2 index wheel teeth, right in the middle of the amplitude range that gives correct indexing (just over one tooth to just under three teeth). This is with a new silver oxide cell.

The Energizer 387s data sheet says the cell has a capacity of 63mAh. A typical stock Accutron 214 will draw 6uA, giving about 14.3 months of life from a cell. This is a pretty familiar cell life.

But in addition to indexing properly, an Accutron 214 with this low-current coil I've wound should last 20 months!

 

After running 24 hours, it's settled down to 4.0uA.

 

I'm now satisfied with my results on 3-wire cell coils, and am working on finding the right number of turns on a 2-wire. The first attempt gave amplitude too low, and it wouldn't self-start (1365 and 5485 turns). Second attempt self-starts and amplitude looks better (1467 and 4773 turns). For the third attempt I'm going to shoot for 1.5k and 5.5k, probably somewhere around 1500/4330 turns.

For the two-wire, I want to do the same thing -- reduce the amplitude and current a little bit, but not too much, giving amplitude right near two index wheel teeth with a silver oxide cell.

I took a little video of winding the first 2-wire yesterday:

 

No you are correct - the 214 coils are 15µ and the ESA9162 coils I'm also rewinding now are 12µ. Alas the wire is expensive. Considering that and the hours I put into each one (it is a lot of hand work) I'm not able to sell them very cheaply. But the alternative is "ebay untested" coils and we all know how likely those are to be good... And even if they are good today they have the same 50 year old wire coating technology that is failing.

Also (this thread is old!) I've settled on a silicone-based conformal product and I'm really happy with it.