Part 2;
Just after the 1st Covid lockdown I found someone selling a brand new Tony Griffiths riser block for the Centec; arranged to meet, brought it home and stared to fit it.
This means first removing the vertical head for obvious reasons.
What I hadn’t realised, was just how much strength I’d lost during my bout of Covid, so when I slid the head off forwards, I very nearly dropped it; previously I’d been able to remove it and swap for the horizontal arm without much issue.
Blocks of wood etc allowed me to install the riser block and re-fit the vertical head.
Recently I’ve been having issue with work not quite being square, and put it down to carelessness in not allowing for vice jaw lift.
More investigation followed, and I found that I could get a thin feeler gauge under one side of the  head, and under the other side of the riser block.
By coincidence, the two gaps matched each other, so when I check left right tram after fitting the riser, everything still matched.
It did however leave a forward nod.

There’s another thread on here posted after I’d completed the rectification, which also points out the cause of the issue.
Effectively, the sharp edges of the milled dovetails in the new riser block were fouling in the inner corners of the vertical head, and also the main body of the mill.
Filing these back a bit on the riser block allowed the mating surfaces to fit flush.

To remove the vertical head without injury I jury rigged as below.
Sorry these photos are poor, I was set up for the job, rather than photography

Wooden block clamped down to the table for support, with table fully backwards

As you can see from the above, I used an old bearing outer to provide support for the nose, as well as a second block to support the sleeper end.

Head now screwed down with some perforated galv strip, so I don’t lose it, and then the table raised and wound out.

The Riser can now be removed backwards and modified slightly; about 50 thou taken off the sharp edges, on top and one bottom, opposite the clamps.

Riser block removed, and everything still quite stable and safe(ish).

Left Right could now be trammed fine, which is much as before, but I still had a slight nod forwards.
Far too much to be acceptable at about 5 divisions on the metric clock.

Setup for initial rough tramming with 2 thou shims, above left hand clock

Note that Now, I can recess the riser block slightly and spin the spindle gauge through 90°

2 thou shims were just too thick, but after much experimentation, 3 layers of green cigarette papers proved to be about spot on.
In theory, that comes out as 3 thou, but the paper distorted into the machining marks on the riser block and works nicely.

It now trams quite well as regards front to back nod, with a more precision granite U block.

Left right can now be adjusted quite accurately, just by the conventional swivel locknuts

To mix my units a bit, over the 7¼” long bar between the clocks, I’ve gone from about 10 thou nod, to a bit less than 0.005mm, which will do OK for me
Similarly left right tram is also close to spot on as well.

I’d also made sure the table top was square to the column, just using an engineers 12″ square, to check as well as I was able.
I later fitted an MT2 test bar to the spindle which showed negligible runout and no real visible light between it and the square.

The original inspiration for the spindle gauge was from John “Bogstandard” on Model Engine Maker
I think the article is reprinted on this website somewhere, as part of a list of John’s projects, but I can’t currently find the link

Bill