Bull Dog gets its spark back
Webster magneto from a Bull Dog instruction leaflet. This drawing is not correct as the push finger needs reversing for the Bull Dog.
Editor’s note: This two-part article is a supplement to Peter’s four-part series on restoring a 1-1/2 HP Bates & Edmonds Bull Dog, which began in the October/November 2008 issue of Gas Engine Magazine.
Having made the main body of the igniter bracket, attention then turned to the magneto and the other parts needed to complete the assembly.
Webster tri-polar low tension magnetos are known as oscillating magnetos, and are regarded as the simplest and most efficient low-tension ignition device made. This is demonstrated by the fact that these magnetos were fitted by numerous manufacturers.
The Webster does not turn continuously but operates by the armature or inductor being turned 30 to 40 degrees by the trip lever before it snaps back through the tension applied by the two springs on the front.
Different models were used on Bull Dog engines: the “M” with one magnet or the “MM” with two magnets for the smaller 2 HP engines, increasing to “J” for the 16 HP engines.
One reason for the success of the Webster was the fact that the only moving part was the armature, with no brushes or collector rings. The Webster was sold as a unit, comprised of the igniter bracket designed for a particular engine complete with a magneto. In the vast majority of cases these magnetos were set up for clockwise rotation when tripped.
The problems with a Webster more often stem from the springs wearing so there is not enough snap or the bearings getting sloppy so that the inductor moves sideways and touches the poles rather than rotating.
An important feature of the Webster is the starting lever. The engine is primed with fuel, moved to compression, and then the lever is used to cock the magneto. When the lever is released the engine should fire. This was particularly useful for the larger engine, whereas the smaller engines could still be started using the starting handle.
With a Webster, the igniter points are normally closed, opening when the magneto is tripped, which ensures that carbon does not build up on the contacts and they are kept clean. Should the contacts need cleaning it is a simple matter to push the movable electrode into the combustion chamber then pull it back, thus cleaning off any carbon or dirt.
While I had been able to acquire a Webster AMM magneto I needed one that turned counter clockwise when tripped, as the Bull Dog magneto must fire on the return, pull stroke. There is normally a chisel mark directly across one end of the shaft. If this mark is on the inside, by the trip finger, it is set up for right-hand operation. If the mark is on the outside, it’s set up for left-hand operation. You can check this by looking at the inductor. When moved from rest in the desired direction, the wide faces of the inductor should align with the pole pieces of the coils. The shaft was checked on my magneto and it was indeed a clockwise model.
I understood from my research that all that was needed to convert the Webster to work on my Bull Dog was to reverse the inductor and obtain a new trip finger. This is not something to be undertaken lightly as it is so easy to lose the magnetic charge when stripping magnetos, but if I hit a problem with this I had access to a friend’s re-charger to boost the magnets.
As a first step I removed the springs, spring arm and push finger, taking care not to lose the small keys that were fitted in both ends of the tapered section of the shaft.
One of the end plates was carefully removed by undoing the four screws and gently prying it off, taking care not to damage the edges of the thin pot metal. After checking that the inductor was free, I mentally rehearsed what I had to do several times then very quickly removed it, turned it around and replaced it, all in a matter of a second or so.
When the end plate was removed, I noticed that the insulation on the wire from the coiIs to the terminal block was frayed. Rather than remove the coils to replace it I cut off the old insulation and used some heat shrink sleeve to replace it, using two layers for added protection. A soldering iron was used to heat the shrink sleeve rather than a heat gun to prevent any heat affecting the coil and other old wires. The end plate was then replaced and the next step was to test the magneto.
With the inductor springs and spring arm still removed, an AC voltmeter was connected and a fixed speed drill was used to drive it. The slowest speed of this drill was only 400 RPM, but it still managed to achieve a reading of just over 9 volts, in proportion not far short of the 13 volts that 500 RPM should have generated, so all was well, and the magnets would not have to be re-charged.
While the original spring arm could be used, a new push finger was needed. I decided to make this from ordinary steel and then case harden it. Thinking out the sequence in which to make the finger, the taper was cut first before shaping the arm, so that there was a regular shape to hold in the chuck. Finally, to make it easier to copy the shape of the original, this tapered hole was used as a reference point.
However, the first problem was to measure the short taper at the end of the shaft. This was achieved by taking some rough measurements with the calipers before off-setting the top slide using a dial indicator to duplicate these measurements. A piece of scrap steel was then turned to this estimated taper. The inside of the old push finger was given a thin coat of engineers blue and the test piece was fitted in the push finger with the blue highlighting any error.
This was corrected by adjusting the top slide and the test repeated until a good fit was achieved. Leaving the top slide on this setting, a block of steel thicker than needed was held in the 4-jaw chuck. This was drilled and bored to the required taper, the test piece being used to ensure the correct depth of cut. If a mistake was made and it was cut too deeply it could be corrected by removing some of the surplus metal from the face; alternatively, if the taper was correct, the surplus could be milled off the back.
Once the taper had been cut it was necessary to form the slot for the woodruff key from a small half-moon piece of steel. A 0.062-inch-thick piece of high speed steel was held in a boring bar mounted on the top slide, still on the same taper cutting setting, and was used to broach this cutter into the wall of the taper to cut the slot. Using light cuts the slot was completed.
Once the block could be mounted on the shaft, it was possible to make a mirror copy of the original. A short length of steel, turned to the smallest diameter of the taper, was used as a locating pin. The side face of the new block was painted with engineers blue and the old trip finger was used as the template to scribe the outline. The milling machine was then employed to remove the bulk of the metal to get the rough outline before using files to complete the shaping.
Later, when the arm and magneto were assembled, the lip of the finger was checked against the trip finger before it was removed to be hardened. To toughen the trip finger case, hardening powder was used, and following instructions with the powder, the metal was heated, immersed in the powder for 15 minutes and finally quenched in oil. To increase the thickness of the hardened skin this process was repeated a second time. When finished, the finger was cleaned and given a good coat of protective oil. This oiling was repeated later as the salts in the hardening compound are corrosive.
Fixed and moving electrodes
Before making the electrodes, the sleeve for the moving electrode was drilled out to my preferred size – 0.375 inch – and reamed, although the original might have been 0.312 inch.
The shaft of the moving electrode was made from a length of 0.375-inch diameter drill rod onto which a cap was brazed and a slot 0.188-inch-wide cut for the contact arm. The bearing point of the cap, where it meets the inside of the igniter, was turned with a 45-degree taper. A countersink drill was used to cut a similar recess in the cast iron sleeve on the inside of the igniter. The fit of the electrode was then polished by lapping it using grinding paste; by not having fit the arm to the electrode, a screwdriver can be used in its slot to turn it back and forth. When satisfied with the fit, and a clean bright ring could be seen in both electrode and its seat, all traces of grinding paste were removed.
Before fitting the contact arm, the electrode was trial fitted to check its length, before turning it down in the lathe to cut the 0.250- inch thread for the electrode nut and the taper for the electrode arm. To finish this end of the electrode a 0.062-inch slot was needed for the key. Before cutting this, and while the top slide was offset to cut the taper in the shaft, a piece of round steel was turned to 0.687 inch, chucked and drilled, and then a matching taper was bored out to be used for the trip arm. The fit of this piece of steel was checked on the electrode before cutting the keyway. Again, the offset top slide was used with the same setting as the electrode, in a similar manner to slotting the trip finger.
To finish the movable electrode, a slot was needed for the woodruff key, and a woodruff cutter is normally used to get an accurate fit. Not possessing one of this size I had to find another way, as I was quoted $40 for a cutter which I might not use again. I had some small heavy duty cut-off wheels for my Dremel which were about 0.040 inch wide. One of these was held in a collet in the lathe, and the movable electrode was held in the vertical slide, which was centered under the disc. The slot was started, applying minimal adjustment to the depth of cut to allow the cutting wheel to do its work and prevent from shattering it, and gradually the slot was increased to the length of the key. The depth of the slot was not the same profile as the key, so the disc was worn down by using it against some scrap high speed steel, until it mirrored the diameter of the key. The disc was not profiled at the start of this task as it would have worn down in making the initial cut. The slot was then finished to the correct depth, before lowering the slide to its original setting to start taking 2-inch-by-0.011-inch cuts to the left and right side of the slot to get the required width of 0.062 inch.
A piece of 0.625-inch square bar was then drilled with a 0.274-inch hole to be tapped with the 0.312-inch-by-18-inch thread for the adjusting screw, before shaping the other end to the profile of a piece of 0.650-inch round and brazing them together. After rough shaping in the mill, files were used to finish it off. A stub of 0.250-inch steel was then brazed into a hole drilled in the arm, for the return spring. To secure the other end of this spring, a 0.125-inch hole was drilled in the support block of the retard lever and a split pin was fitted.
The fixed electrode was turned from a length of 0.500-inch steel, with a 0.312-inch body and 0.250-inch thread for the terminal nut, lengths being gauged by measuring the igniter block. The partially finished fixed and movable electrodes were then fitted to measure the length for the arm on the movable electrode and location of holes for the contact points. A piece of steel, 0.188 inch wide, was cut and shaped then brazed to the electrode for the arm.
The igniter contacts were made from soft iron with 0.125-inch stems which were brazed into similar sized holes in the electrodes. To finish the fixed electrode, some 0.438-inch outside diameter mica tube was cut to size and 0.125-inch stacks of mica washers were fitted at each end, together with a Fahnestock clip on the outside.
The last fitting needed was the priming cup. The body of this was turned from a length of 0.500-inch hex brass, reducing it to 0.405 inch round for the 0.125-inch NPT thread in the igniter, and to 0.375 inch inside the cup. A 0.250-inch hole was drilled through this brass, except for the top 0.250 inch, which was threaded 10-inch-by-32-inch for the priming cup sealing bolt. A 0.093-inch hole was drilled through both sides of the tube near the base of the cup to allow fuel to drain into the tube.
The cup was turned from some scrap brass, with a 0.375-inch hole to match the end of the tube, and was then silver soldered in position, taking care not to block the drain holes.
The stem of the sealing screw, made from a 2.750-inch length of steel rod, was threaded 10-inch-by-32-inch to within 1.250 inch of the top, and some brass was tapered at 45 degrees to act as the plug. The threaded rod was then trial fitted to identify the point at which the handle should be bent, the position being marked by nicking it with a 3-corner file. After heating and bending, the screw was fitted and the plug brazed in position and cleaned.
Testing and fitting
The magneto and bracket were then assembled and wired up. The electrode arm screw was adjusted so that it barely touched the push finger when the igniter points were closed. The bottom of the igniter was held in a vise and a screwdriver was used to trip the magneto by levering the studs on the spring arm and magnet bar. There was a healthy blue spark from the points when it tripped.
Knowing that the new bracket now worked, the battery and coil driven igniter were removed from the Bull Dog, along with the trip lever and old valve rod clamp.
The new valve rod clamp was fitted and the length of the pushrod adjusted so that the exhaust valve started to open 30 degrees before bottom dead center, and was fully closed at or just beyond top dead center.
The length of the pushrod and position of the wedge was then adjusted so that it tripped the igniter 30 degrees before top dead center. This was earlier than my battery and coil setting of 15 degrees, and I assumed it was because the Webster is slower to spark with the time taken for the trip finger to move then knock the contacts open to generate the spark.
Setting this was a little bit of trial and error as I did not have a starting lever to set an accurate firing position. Instead, I used a piece of bar between the roller pins and my eye! The correct way to set up a Webster is to cock it with the starting lever when the engine is in the firing position (30 degrees before top dead center). The timing lever needs to be in the normal running position. Adjust the length of the trip rod until the end of it just touches the push finger, the wedge being further back so that it does not touch the roller. Move the wedge toward the trip finger until the bottom edge of the pushrod clears the push finger and tighten securely. Remove the starting lever and turn the engine over slowly until it trips to verify the firing position.
The starting lever is useful as it shows the exact position at which the magneto should fire the engine.
After some fine tuning the engine fired up and with some further adjustments ran evenly. This completed the final step in the lengthiest and the most complicated engine restoration I have ever undertaken.