Not sure that is really a problem. If you locate the rotor contact so it's in the middle of the curve, you are not talking about much leeway
Yes. But you can also use filtering. Back in the day, with 2 way radios, it was common in some rigs to see coaxial/ feed-through capacitors. I cannot even find a photo. These were generally a physically large cap, all metal, about 2x the diameter of a typical coil radio filter cap, with a stud out each end. But they did not need to be that way. Just putting a cap to ground from a terminal can be just as effective. Capacity is not too critical, but sometimes a lower value is more effective than higher, depending on the type of noise. A high narrow pulse is an example. The old school noise caps in the alternators (generally all makes) and across the + side of the coil, and on Mopars with gauge voltage regulators, across those. The trick is to have the cap fairly close to the noise generator and or at the noise "victim."
Don't dismiss just plain twisting the wires. That is why CAT3, 5 and whatever is "now" wiring is twisted. Also you can use a Faraday shield, which means practically, a shielded cable with ONE end of the wire grounded. You can do the same with cat5 and the phone company used to. Let's say you have a cat 5 with 2 live pairs. You take the remaining, unused terminals, and ground them all together. This makes a Faraday that helps shield the in-use pairs You do NOT want to create a ground loop, causing current to flow ON the shield.