Emily was away this weekend, so I cracked on with the car. First up, I put the boot floor back, then put some protective rubber trim around the immobiliser as requested by IVA. I thought I'd better check I hadn't pulled a wire loose when meddling with the immobiliser, so gave it a quick test start and... nothing.
A few checks later and it was clear that there was no immobiliser fault code, which was good, but also no fuel, which was a problem. Pretty quickly I figured that I'd accidentally pulled out the plug to the fuel pump when I was putting the tank strap in, so I drilled out all the rivets again and this indeed proved to be the case. Kind of wish I'd used more rivnuts instead of rivets on the panels now, for ease of removal, but it's too late to change without getting all new panels.
Unfortunately, the package of bits from Peter (mainly containing a modified headlight bar, indicator mounts, steering column bracket) hadn't arrived, so I couldn't do any more things needed for IVA. Since I still had a quick rack waiting to be fitted, I decided to fit that and play around with steering setup and see if I could locate any stiffness which might be hindering self-centering.
I trial fitted the rack without springs, but the cycle wings were (just) hitting the bodywork, so I dismantled both racks, and transferred the springs over to the new rack. These limit steering travel too much, I think, so after IVA I'll replace them with a suitable number of washers - but for now they solve that problem (if a bit too much) and may assist self-centering, so I'll keep them on for IVA. Note to people doing the same: if your rack has the large round bit that is hard to grip when dismantling, I recommend doing as I did and grinding a couple of flat bits on it so you can get a spanner on it.
The hard to unscrew round bit.
The new rack has one immediate benefit: it permits far more toe-in than the old rack. The old one maxed out at about 2º of toe in each side, by my crude measurements; the new one goes far beyond that, which may help me get self-centering.
Toe measuring setup
To measure toe, I taped a couple of broom / mop handles to the wheels, then measured the distance between them at different points. A bit of simple trig then gives toe = sin-1(((w1 - w2) / 2) / d) where w1 and w2 are the two width measurements and d is the distance between them. Modulo the inherent inaccuracies in measuring angles of a few degrees using broom handles (although it did seem reasonably repeatable), I currently have about 3º toe in and will see how that goes.
After this, I took a crack at measuring camber (the amount the wheels lean inward), which was harder. Eventually I used the iphone to measure wheel angle and ground angle, and established that one wheel was at 1º (which is probably about right), and one is at 4º (which is too much). Fixing this involves removing a ball joint, which did not want to come out. After battering the end of the ball joint with a hammer, it was still stuck, but with a damaged thread preventing the nut from going on. Progress of a sort. Things being a bit desperate, I got the dremel out and managed, eventually, to cut the end off which left just enough thread remaining to get the nut on properly, and left it at that.