Finally, we're in the home stretch. This Bosch ZEV magneto has passed all my electrical and mechanical tests so far, giving me good reason to be confident it will be function without problem for some thousands of miles. But, there are still a few more things to check before I am ready to ship it back to the engine builder for installation on the bike.

Low Speed Test

It's great that this magneto works so well at high rpm, but first the bike has to start. So, I next moved it to the lathe, whose continuously variable speed would allow me to run it as low as 40 rpm (80 rpm engine). The point of doing this is to determine the lowest speed the magneto will reliably produce 6 kV sparks with it fully retarded (where it will be for starting).

[Linked Image]

The photograph shows the magneto mounted on a bracket that I substitute for the lathe's compound and that sits 45 mm below the center line of the lathe. I also have a 10 mm spacer for mounting magnetos having a 35 mm spindle height, as well as a vertical plate with appropriate holes for holding flange-mounted magnetos. Although I don't expect anything in the magneto to seize, the lathe's 1-1/2 h.p. motor could cause quite a bit of damage if it did, which is why if you look closely you will see a short length of plastic tubing is part of the drive train for the magneto. I ran this test for several minutes and the magneto continued to spark reliably across a 5 mm gap down to 135-145 rpm (270-290 rpm engine) with the cam at all positions between fully advanced and fully retarded. A Lucas Manual says 300 rpm is a the low end of kick starting speeds, with 500 rpm normal, so the magneto passed this test.

Distributor Tester

My final test used a modified distributor tester to see if during operation the sparks are precisely 157.5/202.5 degrees apart as they need to be for this Harley-Davidson engine, and as my static measurement of the cam profile indicated they should be. However, there are several reasons why under dynamic conditions the firing could be off, or even fluctuate around the correct values, and the only way to know for sure is to measure it. To some extent the Strobotac already addressed part of this issue, since I would have seen fluctuations in the positions of the points when they opened if larger than it a degree or so.

Ten years ago I made several modifications to an Allen distributor tester, including adding an adjustable platform that accommodates both platform- and flange-mounted magnetos. However, when the Bosch arrived the tester had been partially disassembled for a few months to make upgrades to it (actually, most of that time it had been just sitting there waiting for me to find the time to finish), so the temporary configuration I used for this test had the sparks strike just outside the markings on the large protractor. This tester spins the magneto either CW or CCW, whichever is appropriate, at up to 2500 rpm (5000 rpm engine). In addition to testing magnetos with Manual advances, like this Bosch ZEV, a digital tachometer allows me to determine the advance curve of a magneto's auto-advance unit, to make sure it is operating properly. I also have adapters that let me check the advance curves of auto-advance units from newer motorcycles that don't have magnetos.

Ideally, the sparks always would happen at the same angles. Since their positions can be easily read to a fraction of a degree on this tester, any variation in spark timing larger than this is immediately apparent. The quantitative results from this tester allow me to decide what further work on a given magneto might be required (e.g. stoning the cam to alter the timing by a specific amount).

The first photograph shows what this tester looks like, although with a ZE1 magneto sitting on the base. The second photograph shows the sparks from cylinder #2 of the ZEV, with the protractor adjusted so cylinder #1 was at 0 deg.

[Linked Image]

[Linked Image]

The magneto was running at 1250 rpm (2500 rpm engine) for the above photograph and I used a 1/4-sec. exposure to capture 6 sparks. The protractor is slightly blurred because of vibration of the tester during the long exposure. Although it might appear that the timing was wandering by a degree, the apparent variation is largely an illusion due to the spark finding a different route to earth each time. Note that all the sparks radiate from the same point (to within ~0.2-deg.), which is the tip of the spark wire passing by much too quickly to be photographed with this long exposure (some of the ~0.2-deg. variation could be due to vibration of the tip of this wire in the temporary configuration I used, since the last ~3/8" was unsupported). This test shows that the timing of the magneto wanders by no more than ~0.2-deg. from one cycle to the next. Although this photograph only captured 6 sparks, I watched it closely for several minutes, seeing no sign of problems.

As I wrote in an earlier post, the points for cylinder #2 should open at 157.5 deg. on the magneto (315 deg. engine) for a 45-deg. V-twin engine, but this test shows that the spark is 1.3-deg. early (2.6-deg. engine), at 156.2. Again, this is with the protractor adjusted so #1 is at 0 deg. In terms of engine timing, what this means is #1 would spark that cylinder 2.6-deg. late if #2 were set to fire at the perfect spot. Since the spacing between #1 and #2 on the cam is 1.3-deg. too close (and between #2 and #1 that much too far), it might seem I should improve the timing by stoning a slight amount from the #2 ramp (if they were too far apart, it would require a new cam). However, the static measurement I made on just the cam, described in an earlier post, found it to be good to better than 0.1-degree. This indicates the source of the 1.3-degree problem is a buildup of tolerances of the several components, so grinding one of the ramps would be attacking a symptom rather than the actual problem. And, once ground away, metal can't be put back on the cam.

It bothers me to accept this 1.3-degree error in magneto timing (2.6 deg. engine), and if I had more time I would develop a proper solution, but sometimes perfection is the enemy of perfectly acceptable. The magneto has a Manual advance, so if the rider hears pinging from one of the cylinders and retards the ignition until the pinging stops, the other cylinder will be 2.6 degrees further retarded from the optimum timing. However, given the low compression of the engine this magneto will be used on, this will not have a significant effect on performance, especially for its intended use in the cross-country Cannonball Run. Still, even though it won't matter much in practice for this particular engine, it bothers me not to have the time to resolve this.

Disassembly, Inspection, Remagnetization, and Reassembly

After having run the magneto for ~24 hours (~"1000 miles"), I disassembled it to inspect the bearings, measure the brushes for wear, and look for any signs of distress. Everything was fine, so I relubricated the bearings with Sta-Lube high temperature disk bearing grease and the rubbing block with Lubricam, reassembled it, checked that the gap was still 0.012", and then remagnetized it.

Although I previously wrote that running it on the modified distributer tester was the final test, the actual final-final test was to run it for another 15 min. on the long-term tester before packaging it up and sending it back to the engine builder.

Timing a Magneto Using an Inductance Meter

Timing this magneto to the engine wasn't part of the restoration, because that would be done 1500 miles from me after being delivered to the person rebuilding the bike. However, I wanted to at least briefly address how to do this using something quite a bit better than cigarette paper.

There are two aspects of timing a magneto to fire at the right moment. First, the engine has to be rotated to the correct position before top dead center (BTDC) where you want the magneto to fire. Most commonly, this is done at the fully advanced (high rpm) position, rather than retarded (low rpm). Finding the correct position -- most engines are between 30 and 40-deg. BTDC -- can be done to varying degrees of precision with a dial indicator and protractor, a ruler stuck down the spark plug hole, a factory mark on the crankshaft, etc. For the purposes of this post, assume the engine is now at the correct angle BTDC where you want the magneto to fire.

Next, the taper on the magneto's armature is loosely inserted in the gear or sprocket in the engine's timing chest. What you now need to do is to rotate the armature until the points have just opened by the slightest amount, and then tighten the armature to the gear to lock in that timing. Assuming for the purposes of this post that nothing slips when you do this and that there is no backlash in the gear train, the magneto is now correctly timed to have the points open when the engine is at the correct angle BTDC. But, how to find the position where the points have just opened? Forget using cigarette paper.

The resistance across the points of a magneto when they are closed is 0 Ohms, and when open is only ~0.5 Ohms, so a standard ohmmeter would barely register the difference. However, the inductance of a magneto's primary when the points are closed is some number of milliHenries, and when open is some number of Henries, i.e. ~1000x greater. So, an inductance meter across the points will register a huge change the instant the points have separated. Although measuring the inductance might sound difficult, it isn't.

If you search eBay for 'LCR meter' (without the quotes; also try 'LRC meter' for more choices) you will find precision ones go for over $1000. Luckily, though, you don't need precision, so Chinese-made ones that sell for ~$18 (delivered price) are just fine. Attach the leads across the points and set the scale to whatever mH value gives a reading when the points are closed. The actual value is irrelevant, and you don't even have to know the difference between a milliHenry and a megaOhm because all you care about is seeing the meter abruptly go over-range when you slowly rotate the armature. When that happens, tighten the nut to lock the armature to the gear, and the magneto is now properly timed to fire when the engine is at the correct angle BTDC.

Performance of the Magneto on the Road

A month after I shipped this restored Bosch ZEV back to the Harley-Davidson's rebuilder the bike was used in the 2012 cross-country Cannonball Motorcycle Run. Unfortunately, the builder was able to finish the bike only a few days before it had to be transported to the start in New York so there was no time for my friend to give it a proper shakedown. The bike suffered a variety of problems that would have been easy to fix under different circumstances, but that kept it from covering more than ~800 miles over the course of the Run. However, I'm happy to say that the magneto was trouble free. Added to the ~1000 simulated miles I subjected it to on the long term tester before shipping it back, this restored 90-year old magneto has "travelled" nearly 2000 miles without problem thus far. I have every reason to expect it to be good for many thousands more.

This post is the last on the actual restoration of this Bosch ZEV magneto, but there will be one final "Epilog" that I hope will help people identify someone who can properly restore their magneto. Send questions or comments to [email protected]