What I knew, but hadn't accounted for in the winding, was that the cores of the two coils had to fit absolutely flush with the base plate and likewise, so too must the blocks which contact the magneto's magnet. For this, the ends of the cores were machined as smoothly as possible because the slightest air gap created by machining marks could reduce the efficiency of the magnetising process. The coils I'd wound wouldn't allow this fit because as the windings increased and pressure on the end plates went up, so the windings pushed the plates out beyond the ends of the core.

What I should have done at the start of the exercise, was inset the end plates and secure them in place so that they wouldn't slip. There was no question I'd have to begin again. But first I had to unwind everything to get the cores out.

If you can imagine a 350 metre 'Slinky' going AWOL, you get the picture. Chaos! Not knowing what to expect when I took the tape off the first coil, the wire sprang out in all directions and contrived to muddle itself into a bird's nest. It took me 7 hours to unravel and wind it all back onto the original spool.

Resigned to my fate, I made up a winding fandango...

... and stuck at it...

... until the first reel was complete.

The core was removed, the end plates glued to the metal and the whole process began again.

Knowing then what to expect, I was a bit more careful with the second unwind, though it still took a couple of hours to complete. Despite my best efforts, the wire would still behave badly given the slightest inattention.

All's well that ends well. The coil on the right is obviously neater, it being the second to be rewound, but uniformity is more to do with aesthetics; they'll work just the same.

The upcoming Lacy Scott Classic Auction is attracting some interesting vehicles and Counsel and I are out and about getting those that have been idle over the last few years, ready for sale. This extremely original 1906 Wolseley-Siddeley, once we'd worked out what was what, chuntered into life with relative ease. It has been in the same family for 67 years.

The advance and retard lever, instead of adjusting the magneto timing, altered the timing of the valve opening via a cam mechanism built into the rocker block on each cylinder.

Also unusual was the throttle pedal arrangement. It was directly connected both to a throttle plate in the inlet tract and a governor driven from a train of bronze gears starting at the crankshaft and which in turn powered the water pump, the fan, and magneto. The trick to getting the engine to idle was to balance the valve opening with the governor. It was almost like running a rotary - controlled chaos.