Revisiting an old thread of mine from ten years ago, **LINK** the result of the kindly interest of John Haine in this thread **LINK**, I thought I would share my Eureka moment.
Back then, this was all I could extract from my final attempt to produce an acceptable spring.

I wasn't too impressed, but had run out of patience and 0.008" (0.02mm) guitar strings.
Here’s an updated photograph of my method. I've added 'sharpening' via Photoshop to emphasize the bumps.

On the borrowed lathe I was using (a HobbyMat) I had no ready means of adjusting the pitch, so was obliged to close-coil the wire.
It can be seen that in passing the wire over the toothbrush handle (the red bit), I had allowed the wire do (some of) the controlling itself.
Once there were a sufficient number of turns to grip the mandrel, I fed the wire at a slight angle over the toothbrush handle. Feeding the wire slightly from the left, 'steered' the wire onto the mandrel to form a tight helix.
(The toothbrush handle just happened to be a handy and relatively soft piece of material to hold the coils (gently) against the mandrel.)
What I hadn’t considered, and which can be seen clearly in close up, was the uneven surface of the toothbrush handle, a result of having been injection moulded from an EDM’d mould surface.
As the wire passed over the uneven surface, its direction (angle) could (and probably did) change depending upon which set of bumps it ran over. In other words, the ‘steering’ was imprecise, and the most likely cause of pitch variation.
How it looks on the clock …

I’m sure with a bit more effort in shaping the spring ends and adjusting their position, it could be improved.
But that’s how it will stay.
Keep safe,
Sam
Edited By Sam Stones on 19/06/2021 22:21:58