Got an absolutely fantastic response from the VintageRadio forum on this topic.
Not so much on here though, which is a bit disappointing.
Not particularly having a rant but it seems we get pages and pages of posts on here on the differences between Honda candy gold paint finishes, so I am beginning to wonder why I bother posting this tech stuffas there seems to be very little discussion.
Anyway, moving on, I bought a high voltage insulation tester to 'weed' out condensers , before I build a pukka JIS spec. tester.
Interestingly, I took a dual capacitor from a Honda twin and a NOS similar one.
Testing the capacitor at elevated temperature, akin to what you get in service, the capacitance was bang on the correct value given by Honda on both the old and the NOS parts.
Also testing the the insulation firstly at 250v, then 500v, then 1000v, on a single 'static' test, the insulation was fine on both the used one (Genuine part .. taken from a 1973 bike) and on the NOS one. I also tested a NOS electronic component 'wire ended' one of similar capacitance and voltage spec.
However, there is a 'continuous' test mode on the new tester I bought and when tested on this function (it repeatedly applies the voltage), the used one started displaying a value of a couple of megOhms leakage, whereas the NOS part and the 'reference' capacitor didn't.
So I guess the used condenser is leaky, dynamically and explains why the owner of the bike was reporting excessive arcing at the points, even though on a static test the capacitor tested out fine.
My next 'experiment' is to dig out a brand new Daiichi pattern one and put it through it's paces. This brand has a very bad press but is the pattern part sold by lot's of reputable parts suppliers including David Silvers.
[...] I think the tacho function is probably a crocodile clip onto the points type signal though.That's the way I'd prefer to have it. Congrats again!