APPENDIX VIII: Rebuilding a Lucas Magdyno

Since the internals of the magneto portion of a Magdyno are very similar to the Lucas K2F of Appendix II and Lucas KNC1 of Appendix VII, as well as the Bosch ZEV, much of the information for rebuilding it is the same. Because of this I won't repeat the description of disassembly and replacement of the condenser. I'll take a different approach and instead of the usual instructions to "reassemble in reverse order of disassembly," I'll suggest you disassemble in reverse order of assembly.

Although the resistances of the primary (0.72 Ohms) and secondary (5.5 kOhms) were what they should be, there still could have been an internal short between windings. If there were a current would flow through that path likely causing the damage to increase with time until the coil ceased to function. To check for this I used a growler, which most people think is only good for checking dynamos. I forgot to photograph this test, but the following shows an armature from a Dixie magneto I recently checked as a favor for someone who will be riding a 1916 Indian in the upcoming Cannonball Rally.
[Linked Image]
A growler generates a 60 Hz AC magnetic field in the 'V'-shaped cradle. If a closed loop of wire were in the cradle that field would generate an AC magnetic field in it. Although an armature has 10,000 loops of wire, unless there is a short there is no electrical connection between any of them so no magnetic field will generated in a good armature. However, if there is a generated field (i.e. a short) it will result in an AC force on any nearby steel. A hacksaw blade serves as a convenient detector since any such force can be easily felt when it is held next to a bad armature. The armature in this Magdyno passed the growler test so there are no internal shorts in it.

After replacing the condenser the end cap of the armature was screwed on, followed by the slip ring, oil slinger, appropriate shims (more on this in a moment), race, and a retaining ring that I forgot to have in place for the following photograph of another Magdyno armature:
[Linked Image]
The shims are necessary to give the proper end float of 0.001"-0.003" for the ring cam. This sandwich is assembled with a press and tool in the shape of a deep-drive socket, although a socket would be fine. I don't recommend beating the race on using a hammer instead of a press except in an emergency.

The next photograph shows how the outer race goes into an insulating paper cup in the end cap. The same is the case in the main housing. A new paper cup is at the left.
[Linked Image]

The next photograph lays out some important items. At the front is the assembled magneto with an indicator showing how the end float of the armature is measured. At the top right is a pile of armature shims. The ones in this pile vary in thickness from 0.007" to 0.049". At the top left are shims for the end cap of 0.004", 0.005" and 0.006" thickness. These two types of shims can be used in combination to give the correct end float.
[Linked Image]

Note that the fitting for attaching the advance/retard cable, in the endcap at the left of the unit in the photograph, is on the same side of the housing as the HT pickup. In this location tightening the cable advances the spark for this magneto (which rotates counter-clockwise viewed from the end with the drive gear). Endcaps with the fitting on the other side, where tightening the cable retards the spark, also exist so you need to pay attention to both the direction of rotation as well as action of the advance/retard cable in order to set the timing correctly. In the case of this magneto the lever on the handlebar will have to tighten the wire to have it in the fully advanced position when setting the static timing.

After the bearings were lubricated with Sta-Lube 'Hi-Temp Disc Brake Bearing Grease' the armature was placed in the housing, the end cap attached, and the end float measured. If there had been negative end float, i.e. the end cap didn't sit flush with the housing, the gap would have been measured with feeler gauges and the correct 0.001"-0.003" float obtained by use of one or more shims under the end cap. Or, the race could have been removed from the armature and thinner shims installed there to remove the excess float. If there had been positive end float larger than 0.003" the only choice would have been thicker shims on the armature.

If the race is removed to change the thickness of the shims the end float has to be measured again because even if the same shims are used it is difficult to reproducibly position the race to 0.001". Because of this it is typically faster to try to achieve a negative end float with the armature shims and then make the final adjustment using end cap shims.

Don't forget to replace the earth brush:
[Linked Image]

After the magneto had been reassembled it needed to be remagnetized to recover its full strength. As the next diagram shows an iron plate with four holes in it needs to be added to the magnet pole pieces to clear the pegs in the bottom of this type of magneto.
[Linked Image]

However, while Lucas says a flat plate is sufficient when using their recommended 65,000-70,000 A-turn electromagnet, it's possible to do even better. In my case, I designed my electromagnet to produce a ~15-20% greater field of 80,000 A-turns. Further, the next photograph shows the iron plate I machined placed against one of the magnet's pole pieces. As can be seen the side of the plate that fits against the base of the magneto isn't flat.
[Linked Image]

The left side of the next composite shows the bottom of the Magdyno housing and the right side shows it with the plate in place. The shape milled into the plate significantly reduces the "air gap" between the pole face and the internal magnetic structure of the magneto which increases the field that magnetizes the Alnico. The positioning of this plate isn't arbitrary; I machined it so the pole pieces would align with the location of the Alnico magnet within the magneto body.
[Linked Image]
This difference in "air gap" might not seem significant, but it is. Think about the force needed to pull a magnet from a refrigerator versus the much smaller force needed when there are only a few pieces of paper between it and the refrigerator.

The magneto with the extra plate was placed in the electromagnet and the field ramped up to its maximum value several times to leave the largest possible remnant magnetization in the Alnico.
[Linked Image]

As an aside, while the Magdyno is vertical in the electromagnet whereas Lucas shows it horizontal, magnetism doesn't care about gravity. I designed my electromagnet to use various spacers and pole pieces to be able to magnetize every automotive magneto whose dimensions I could find. As a result some magnetos have to be vertical (e.g. Lucas K2F and Magdyno), while others are horizontal (Bosch ZEV and Lucas KNC1).

After magnetizing the magneto, but before attaching the dynamo, I wanted to make sure the magneto was working well so I clamped it to my long-term tester and installed a pulley that spins it at 2000 rpm (4000 rpm engine). Assuming that corresponds to about 60 mph my 6-hour test was equivalent to ~350 miles.
[Linked Image]
The reason for this extended test is that if there were any loose parts or other issues there's a good chance they would have revealed themselves in this 6-hour, 2000-rpm test. Also, since all the electrons and magnetic monopoles were now comfortably in their final state another test would reveal how hard my friend was going to have to jump on the kick starter in order for the magneto to start the engine.

After running the magneto on the long-term tester I put it on my modified distributor tester to make sure the spark occurred at the same point each revolution (i.e. didn't wander over a range of firing angles). Unfortunately, I was rushing because I still had to get ready to go out for dinner so I didn't take the time to set up to take photos and when I came back to it the next day already had removed it from the tester before I remembered this. Anyway, the firing wandered by less than a degree.

After the distributor tester I put it on the lathe to check the lowest speed at which it would provide a reliable spark. I should have been able to make this measurement in the distributor tester but I'm in the process of installing a different tachometer on it and don't have it wired in place yet. Lucas gives 500 rpm (engine) as the normal kick starting speed and 300 rpm as the lower limit. This magneto reliably sparked down to 276 rpm (engine; 138 rpm magneto). I didn't photograph this since the setup is identical to the one I used earlier in this thread for the low speed test of the Bosch ZEV.

I already had refurbished the dynamo so it was time to install it on the magneto. As the next diagram shows attaching the coupling gears is straightforward.
[Linked Image]

First the fiber gear and clutch are bolted to the armature. To tighten the nut requires a special tool, which is simply a 1/4"-dia. rod of the right length bent into a 'U' shape.
[Linked Image]

After the nut is fully tightened the lock washer holds it in place. To make sure the clutch slips within the right torque range Lucas, having not yet discovered the existence of torque wrenches, suggests the following setup.
[Linked Image]

However, rather than raid a fishmonger's stall to get the necessary scale, I used an actual torque wrench. This test still requires a way to lock the magneto armature in place, for which Lucas suggests another jig shown in the above photograph that places all the force on a single tooth of the fabric gear.

In another document Lucas suggests a diamond-shape tool for this task, as shown at the left of the next composite photograph jammed between the fiber gear and a dynamo gear placed in position for the photograph. Clearly, this also puts a lot of force on a small area of the fiber gear. Because of this, also shown next to the magneto is the jig I made for this which engages 10 teeth of the fiber gear, reducing the force on any one of them by a factor of 10. This jig is shown in place at the right of the composite photograph.
[Linked Image]

With the fiber gear locked in place and a deep drive socket on the torque wrench to clear the end of the armature I found the clutch slipped at 7 ft-lb. Lucas calls for it to be anywhere in the range 4-10 ft-lb. so all is well with the clutch and the magneto is now ready for the dynamo.
[Linked Image]

The dynamo is pulled into position by a nut at the front but the actual holding force is provided by a strap. I first pulled it into position with the nut and then, as can be seen from the next composite photograph, tightened the strap bolts on both sides in a way that left approximately equal gaps. This allows the greatest flexibility if the strap needs to be further tightened in the future.
[img]http://i1151.photobucket.com/a...ectrics/Magdyno140_zpsplozw6lw.jpg[/img]

After the dynamo was firmly in place I removed the top nut and replaced it after installing the cover. At this point the Magdyno was now finished and awaiting installation on the bike.

p.s. I don't know why the last image won't appear, and shows up as "does not exist" in Photobucket if the code is clicked on. I've tried uploading it again to receive a new URL, and slightly changing the image in Photoshop and then uploading that one. Nothing I've tried has worked. Anyway, the image you're not seeing is a composite that shows the Magdyno without the cover beside it with the cover.