Just for grins I ground all the paint off both sides of the box and ground off the paint on the fender. Also installed a new ballast resistor. AND of course it's raining out so I have to wait for it to stop to do a check ride. Before doing all that I did check continuity of the box to ground and it was a direct short.
just an FYI: you can measure 0 ohms resistance but the connections ability to handle Current is another thing all together.
Grounds can also sometimes be intermittent and a challenge to identify. Increasing the engines RPM may raise the alternators V and current (I) output enough to mask a bad connection or ground. A small wire braid ground strap between the motor and control box is ideal.
Also I have seen more than one ballast that is on it's way out cause issues like your experiencing. The resistor is breaking down internally resulting in the issue occurring. Again raising the RPM may provide enough current to "bridge" the internal resistance. Measuring them when cold may show good Ohms wise, but hot may be another reading
keep us posted
