From the meandering posts, I was led to understand that the op's bike now had three relays kludged in. One bypassed the start button switch and key switch headlight current, and two more at the blue and white wires near the headlight respectively.
The ones at the headlight reduced the current flow through all of the headlight current consumption circuitry (including the starter switch/key switch current bypass relay). As explained earlier, the less current flow through faulty switches and connectors, the less voltage drop will be incurred in those faulty circuit path components which are now used as logic control circuits rather than power routing. So, with simple measurements, the marginal switches and connectors will now "measure" better, hiding their true functional capability, as they only have to pass signals now rather than the current loads they were designed to carry.
I learned the OP would rather ride the bike, than locate faults in an electrical system he does not wish to understand. So, the kludges meet his needs perfectly, for now.
I don't think I want to know if the wire splices were cut, solder, and shrink wrap, or twisting wires together with electrical tape, for even more lurking gremlins to leap out in the future. If you want an unreliable SOHC4, do scrimp and cheat regarding the electrical system.
---------------
Someone, mentioned grease becoming an insulator inside switch mechanisms.
All the sliding switches use a mechanical wiping action to displace grease during the "make" cycle of the switch. This also self cleans the actual electrical contact patch. If the spring pressures are weak, or the contacts are worn from repeated cycles so much as to reduce or eliminate contact patch area, the resistance goes up. The point here is that if the contact patch has been reduced so as not capable of carrying full rated loads, then mechanical wear will eventually, prevent ALL contact so that low current signals can't pass either (or they become intermittent). Now when the headlight doesn't work, the technician has to find out if it is the headlight, the relays, or something in the undocumented headlight control system (including wires twisted and taped together) that is at fault. More parts can actually reduce reliability ratings, due to more opportunities for something to fail.
What's the old add campaign slogan? "Pay me now or pay me later", (usually at the side of the road, in the dark of night in the rain).
When you make something more complex, you also reduce the number of people able to find faults in that something, even in a magnificent opinion forum such as ours.
I should mention that the later model SOHC4 "headlight interrupter" start button switch is a bit of a different animal than described above, as a spring forces contacts together with very little "wiping" action. Carbon, oxidation, or very congealed grease can cause some issues, (which is why I use pure silicone grease on these contacts during rebuild, which displaces easily and doesn't change its physical properties over time, yet still prevents oxidation of the contacts, and has no destructive effect on plastic).
These switches are "born" with electrical contacts in the 10-15 amp make/break capability and hundreds of cycles before significant degradation. Failures of these switches are far more likely to be caused by switch housing deterioration, which is the part that holds the internal contacts is proper orientation.
