PJ:
I couldn't agree more with your point of these springs being poorly heat-treated when made. It's extremely common with automation machines for this to happen, too. I came across the "reason for" it being this way when we built a machine for a Swiss medical company in 1998 at MSA, and this machine included a smaller machine (about the size of a king-sized bed) within it, which made the springs (both less than 1 inch long, much like the ones in these bikes) for the rest of the assembly component that our machine made (2 different springs were part of the larger final product). This Swiss machine wound the springs from 'spring wire' and then quickly heat-treated them with electrical current by contacting both ends of the wound-up spring wire and passing a high current (15 amps) though it until it glowed to a certain color, which was monitored by a color vision system. This was said to be an upgrade from their previous method, which used a timer setting for the current and a monitor to make sure the current was sufficient: In those earlier machines they had used a (set time + an expected current) to ensure quality (or else it was rejected, usually for poor conductivity issues) while this new one altered the current to get the color they wanted for the spring's tensile strength.
All this is an (old engineer's) long-winded way of analyzing just how "quality control" is done in automatic equipment. Knowing what I do (and did) about Japan's economy and scale of manufacturing in the 1970s (and through the early 2000s) from having constant contact with them via Honda, SONY, Mitsubishi, Mitsuboshi, Toyota, Daicel and its worldwide affiliates, Panasonic/AIWA and its subsidiaries, and a dozen more Sino-Malaysian operations that were associates in one way or another, I am perfectly comfortable with the idea that Honda, in 1970, had Nippon-Denso, Hitachi and Tokyo Electric Corporation make their spark advancers with the least-expensive spring-wire they could get away with for an anticipated 10,000 miles engine life. And, it was not especially cleaned (for good electrical current contact) nor polished (for surface uniformity) like the spring wire that was found in USA-made automotive distributors that were expected (and specified) to last 100,000 miles.
So, yeah - in the end, these springs are cheap little toys that are currently 45+ years past their bedtime...
I have thought of (and may yet try) re-hardening these springs, although I don't presently have an adjustable-output, 40-amp-capable, low-voltage power supply available to build up such a toy with. (Most of my $$ are currently flowing toward my wife's serious health problems, caused largely by being a Type I diabetic for nearly 60 years, so my experiments out of curiosity are presently few and far between.) I could do it with a sinewave-output oscillator of variable voltage amplitude, driving the backside of a 20:1 transformer of 200 watts' momentary capability from an old model train DC throttle power supply for the controller, all of which I have: I just haven't the time nor $ on hand right now to try it out.
Then I could at least re-tension the springs closer to what Honda started with - although I will reiterate that in the case of the 500/550 engines, Honda screwed the pooch big time with the too-fast advance curve (the springs were always too soft) and too little full advance while trying to find a cheap way to meet NOx emissions by simply running the combustion process too rich in these engines: this was, and today still is, the Achilles' Heel of the 500/550 setup. Slowing its advance curve while extending the angle another 4-10 degrees above 5000 RPM really wakes them up if the local gendarmes will let you bypass the tailpipe sniffers in your State for a "vintage" plate. It's just not real simple to do?
