I normally run long-term magneto tests at 2000 rpm (4000 rpm engine) because I want to generate a high internal voltage to reproduce what the coil will experience in later operation, as well as to put the largest number of "miles" on the magneto as fast as possible to reveal any problems. However, it's unlikely a 1920s Harley-Davidson V-twin will operate at 4000 rpm, so this magneto might not be designed to spin at 2000 rpm without the points bouncing. The reason this is a concern is shown in the next photograph.

[Linked Image]

On the left is the Bosch ZEV magneto and on the right is the Lucas KNC from my BSA Gold Star (note: I flopped the photograph of the KNC by 180-deg. to give it the same rotation sense as the ZEV, which is CCW when viewed from this end of the magneto). The KNC magneto spins to at least 3000 rpm (6000 rpm engine) and, at first glance, it might appear to have a more massive moving point assembly than the ZEV, so perhaps 2000 rpm will be fine. However, closer inspection shows that this initial impression is false. Most of the massive-looking KNC assembly is made of lightweight phenolic, while the entire ZEV assembly is solid steel. I didn't weigh them, but since steel has a density 7x higher than phenolic, a simple estimate is the moving point in the ZEV weighs 3-4x more than that in the KNC. This means that even 2000 rpm might be a problem for these points.

To see if this was the case, I installed the magneto on my long-term tester, removed the cover from the cam assembly (holding it in place with my hand, since the cover spring no longer was doing that), and used a General Radio Strobotac to watch the operation to see if everything was moving as it should.

[Linked Image]

Setting the strobe to precisely the rotation frequency of the magneto "freezes" the motion. Setting the strobe to a slightly different frequency makes the points plate appear as if it is rotating forward or backward in slow motion. This slow motion effect makes it easy to closely examine the operation everywhere throughout the entire 360-deg. of rotation, or to freeze it at any point, in order to spot any problems with the mechanical motion. This is why I always do this test. The next photograph shows that indeed there is a problem trying to make this magneto run at 2000 rpm.

[Linked Image]

For this photograph I had set the strobe to freeze the motion just after the points have been pushed open by the cam (rotation is CCW viewed from this end of the magneto). The reason it appears the points gap is quite a bit larger than it should be is because it is. The inset shows why this is the case. The rubbing block is "floating" above the cam because the inertia from its heavy mass has overwhelmed the ability of the spring to keep it in contact after encountering the ramp at this speed. This is the same phenomenon as valve float, except there is no piston nearby to wreak havoc. Unfortunately, this means I either will have to double the pressure by adding the spring from my ZE1 (as discussed in the previous post), or take the time fabricate a larger pulley to reduce the operation speed on my long-term tester. I decided to do the latter.

Another problem the strobe helped me spot, that I had not noticed earlier, is the spring comes within only a few thou. of touching the cam twice per revolution. Although I could slip a piece of paper between the spring and the cam at the distance of closest approach, which means they weren't actually touching, they were too close. I looked at the assembly in my ZE1 and saw the short "helper spring" that is on the inside of the main spring and attached at the 5:00 end in the above photograph instead is at the 11:00 end in my ZE1. Although I had no way of knowing if its location in the ZE1 is the correct one, after I moved it in the ZEV the clearance improved significantly. This seems to be yet another mistake to add to the long list of mistakes the restorer made when rebuilding this magneto. Send questions or comments to [email protected]