When I introduced the thread about near empty tubes of super glue, and although it was a significant irritation at being ripped off, I was actually in the middle of some final experiments to make a usable balance spring for my skeleton clock.

In previous threads about balance springs, I searched and tested several ideas to try to find (with my limited equipment), a reliable method. As my own notes show, I had determined that a wire diameter of 0.2mm (0.008″) would replace the flat section as used in the original clock design. This has since proved to be fairly accurate since the clock is now beating with an accuracy of less than 1 sec/hour, and responds well to fine adjustments.

That said, I can briefly describe my approach to making my balance springs. The most significant problem I have had when winding the spring onto a 4.0mm diameter mandrel has been how to uncoil it successfully after winding. Once released, the usual result has been for the spring to fly open and to wrap itself into an unuseable mess. There have been some very useful comments posted, some of which pointed to additional twisting which takes place (unseen) during the winding process.

Michael Williams wrote :- The traditional method of making very small springs with limited equipment is to pass the feed wire through a hole in a piece of flat soft wood and arrange for the flat part of the wood to bear down on the coils that you have already made . Until such time as you deliberately remove the wood the coil remains tight and under total control . You cut off surplus feed wire with the wood still in place and then slowly lift it away . As an added bonus there is an element of screw-cutting going on in the wood and after a few turns; a neat set of grooves form which help to keep the winding pitch constant . Please be very careful when winding springs under power – it is a terribly dangerous process.

Perhaps it was the hidden twist and how I anchored both ends of the wire which were responsible.

To avoid weakening the mandrel I had drilled the 1mm anchor hole through the larger 1/4″ diameter section. What seemed to be happening was that the wire was skidding down the side of the step, thus introducing extra twist. This was rectified in my final attempts.

The free end was weighted and the wire was passed over a simple pulley. I took the trouble to stop the weight from spinning, expecting that any back and forth twisting would be transferred to the spring. Taking note of all the warnings, winding the wire onto the mandrel had always been done by hand with the main power switch turned off. This was very tedious and tiring for my old hands and fingers, since reaching the three-jaw was awkward. I had also chosen to wind the spring ACW so that I could see a little more of what was happening to the winding process.

It was at that point that I decided to wind the wire under power with the lowest lathe speed (267 rpm), and in reverse. The chuck was bolted so there was never any danger from it unscrewing. I also calculated that I had at least 15 seconds before I had wound all of the 1 metre length of guitar string onto the mandrel. I stood ready to switch off, and hey presto, a nicely wound spring. That was until I removed the wood which Michael had suggested.

Another tangled mess!

That was when I went off at a tangent, and began to explore the possibility of gluing the wire to the mandrel, and using a solvent to let the spring slowly uncoil. The traditional `glue’ in clock making is shellac which dissolves in methylated sprits. This would be ideal, or so I thought. So I tried again giving the tightly wound wire a good coating of shellac, waiting overnight for it to set.

It didn’t work so I figured that the shellac had not found its way under the wire and onto the mandrel. With that fixed, I tried again. The wire instantly flew into another mess, and another guitar string went into the bin. I interpreted this result to mean that the shellac was too brittle, and/or hadn’t bonded sufficiently. It was also possible that the meths had not fully evaporated completely from under the coils.

That’s where the super glue came into the picture. I should (again) point out that super glue (cyanoacrylate) sets by combining with moisture from the atmosphere or from your breath. It is also supposed to soften in acetone, which would be ideal, or so I thought. With a very thin coating of super glue applied to the surface of the mandrel, I tried once more, applying two coats to the tightly wound coils. Several hours later, I cut the wire.

When I fished the spring from the acetone, it became clear that although the glue had turned white(ish) it was still present. Very carefully, I teased the coils free, and found that I had to strip off whiskery lengths of glue still attached to the wire.

Remember, the diameter of the wire is only 0.008″.

This posting may provide food for thought, but it has become more of a saga, so I must end here.

One question. Why didn’t the glue dissolve in the nail-polish remover?

Edited By Sam Stones on 01/07/2011 09:00:17

Hello again, especially Dick, Ian, and Norman,

Thank you for all your ideas, I really appreciate getting your type of assistance.

Just a couple of points about the use of guitar strings, and also in making extra equipment. At the risk of repeating my earlier comments, my target is to finish the clock while I still have some dexterity left. I also consider my workshop to be a very temporary affair, and am reminded on occasions that the Hobbymat lathe is, after all, still on loan.

As some of my pictures show, the mechanism of a skeleton clock is intended to be fully visible. That’s the primary purpose of piercing the frames and wheels. In the case of John Stevens clock, the most obvious animation is the balance wheel and spring sitting inside the top `cage’. The nice thing about steel guitar strings, is that they are already polished. So if I can keep them in that condition, there will be less effort needed to bring them up to scratch. I would agree however, for those who like that sort of thing, that the spring could be a nice blue colour after tempering, but once again . . .

I like Ian’s idea of passing a low voltage current through the wire to soften it. I did notice that being rather fine, the wire burns immediately (like wire wool) if placed in a naked flame. I could also imagine that with care, the use of electrical current in this way, could be employed in both the rehardening and tempering after the wire has been wound to size. Holding it to shape would be quite a challenge however.

In my next comment, if not before, there is likely to be an element of `egg-sucking’ for grandmas. However, I’m inclined to believe that winding the wire around a 4mm diameter mandrel to get a finished diameter of about 8mm, does actually subject a percentage of the steel section to stresses above the elastic limit. How much of this will be in tension (or compression), I could not imagine. Clearly, there would also be a certain % of the section which does not get stressed beyond the yield point, or the spring would come off a 4mm mandrel at 4mm without any `snap-back’.

Machinery’s Handbook offers a table for Arbor Diameters for Springs Made from Music Wire and although it doesn’t actually mention the OD I needed, by extrapolation and my own tests, I seem to be on the right track at 4mm.

In terms of `annealing’ (I have difficulty with the definition of annealing), early tests of heating previously wound coils to about 200 deg C showed very little relaxation, and all the wire that I have wound since remains very close to what I wanted without this treatment.

Finally, and switching to Norman’s comments about the case (thank you Norman), and where to buy them, I have elected to follow John Parslow’s approach. Refer ME p202 15 August 2008. It concludes John’s article called “15-day skeleton timepiece”.

With the help of a friend with some good woodworking machinery, I now have a pine base suitably stepped, together with some pieces of Tasmanian oak beading to trim the surround. The ends of the beading have been mitred which is a task I have never enjoyed via the hand-sawing method.

I’m currently waiting for some shellac to dissolve and before I glue I will begin to apply this to the oak in multiple coats alternating with a rub-down with fine sand paper. I trust that PVA will hold everything together. The centre area of the base (inside the glass) will be covered in a dark blue material, the sort of cloth that is turned into tracksuits and the like. There’s a glass supplier down the road who has made fish tanks and display cases, so I’ll be talking to him next week. Alternatively, I’ll get him to cut the five pieces of glass, and glue them in place with clear silicone. Depending upon how the silicone runs between the glass, I imagine knifing off the excess to leave a very clean joint.

I reckon the French polish will be ready by now.

Regards to all,

Sam

Edited By Sam Stones on 02/07/2011 02:06:00

I’m also left wondering about the surface stress, and the possibility of fatigue failure. I have only come across shot peening (academically) as a process applied to helicopter blades. As for this tiny spring of mine at 0.2mm diameter, I can understand that if the surface is in compression, it will contribute less to the spring’s performance, but is also less likely to facture in use. Since peening would have to be down to a micro scale, the only option (should I proceed) would be heat treatment. All I would be comfortable with would be to stick the spring in the kitchen mini-oven for an hour or two, but at what temperature?

Instead, I’ll take a chance and trust that the spring in the clock will outlive me. There’ll be less embarrassment that way round.

I also take your point about preparing the end connections. Not that I’ll be able to do much more then carefully curl the ends into line with the top (balance-wheel stud), and the bottom (balance-wheel collet) anchor points. This has led me to make a cylindrical template, such that one end was turned down a fraction to lineup with the stud hole, and the other end to match the hole in the collet.