Now, there is another inductor, L39 I think, that has a core that is fixed. If I insert a small screw driver into the coil, the inductance of course alters slightly and the width will increase. I am not sure of the purpose of this coil, but it goes from the wiper of the line shift control to the scan coils plus convergence circuitry. (L40 the width control is on the other side of the scan coils).
I had a closer look at this coil and the "apparent" fixed slug. The slug is a small smooth ferrite rod about an inch and a half long. The inside of the former is smooth too, without the usual thread.
It is not fixed, but was pushed or had moved all the way in as it was loose. ( I was pushing it, but it was at an end stop, so that's why I thought it was fixed.)
I temporarily removed the coil and ferrite slug. I put a small piece of rubber sleeve in the coil former and gently re-fitted the slug, but leaving enough sticking out so I could adjust it.
Then I re-fitted the coil and switched on.
I adjusted the slug for maximum width which is now more sensible.
Furthermore, the maximum width corresponds with a minimum AC voltage across both R229 and R230 to ground, which had previously overheated and I had to replace them both. They now run much cooler along with the line shift control which is mounted on the chassis and I had been surprised how hot this had been previously.
So, I think that the core of L39 was mal-adjusted and was causing the lack of width.
The proper width control can now be reduced to a more normal setting.