In the summer of '88 I bought a 5 HP Maynard kerosene gasoline engine to be restored during the winter months. This engine was stuck, cracked, rusty and missing the carburetor and igniter. I decided to design and build a low voltage homemade igniter to use until I could find an authentic igniter. The first step was to purchase some Mica washers, Mica tubes and Tungsten igniter points from one of the GEM engine parts advertisers. Next came some studying of Fairbanks-Morse and Stover igniter specifications to obtain low voltage igniter construction ideas.
Pictures 1 and 2 show the housing that resulted from a lot of cutting, welding, grinding and drilling. The length of the housing was chosen to place the points approximately flush with the side of the cylinder bore. This length also serves to fill most of the entrance tunnel void, thereby keeping the compression ratio high.
The stationary point was brazed to a small steel piece that was welded to the head of a 1/4 x 28 machine bolt, as shown in picture 3. The other point was brazed to a 1/4" diameter rod that was bent 90° to form the movable arm and shaft, shown in picture 4. Picture 5 shows the bolt and shaft in their respective igniter positions. Note that a 1/4 x 20 hex nut was brazed to the shaft to form a bearing surface and gas seal at the point where the shaft enters the housing block.
Picture 6 shows the spring that holds the points open and the stop pin that determines the distance between the points when fully open. This distance should be no more than 1/16", otherwise there may be a problem obtaining sufficient dwell. Some igniter designs put the stop pin on the inside near the points, but I like to keep the number of components in the combustion area to a minimum so I located the stop pin on the outside.
Now we come to the more complicated part of the design. The length of the igniter trip finger is important to achieve proper igniter operation. If it is too long, the side bar may reach the end of its travel without the igniter reaching the trip point, or the square igniter push arm may not drop down past the trip finger during the igniter intake stroke reset operation. Conversely, if the trip finger is too short, the achievement of proper spark advance/retard range may be difficult. The cam on the half speed gear will impart a fixed amount of the side bar motion for each cycle of the engine. I measured the side bar motion with respect to the crankshaft rotation and plotted the travel distance vs. rotation on a graph. It isn't necessary to make a timing graph to build an igniter but for me it helped put various details in perspective. Side bar (and push arm) travel was very close to 5/8" total before the cam reached the power stroke plateau. In that 5/8" the following three events must occur:
First, the push arm must move a short distance, approximately 1/8"; to take up the space between the push arm and the finger. This space is necessary to allow the push arm to drop down past the trip finger on the igniter reset cycle.
Second, after the push arm makes contact with the trip finger, further motion is utilized to rotate the igniter shaft and the movable contact point toward the fixed contact point. It takes 1/4" of push arm motion to rotate the points closed.
Third, after the points close, a small amount of side bar motion must be utilized to hold the points closed long enough for the current to build up in the inductance coil. This "dwell time" does not have to be very long with slow running engines, but there must be a finite, reliable amount of time before the trip finger trips and lets the points snap open. It is during the closed point dwell time that the igniter over-travel spring is wound up. We have about 3/16" of push arm movement left before the trip point and that amount will be sufficient for the dwell time, with a little to spare. If there is a satisfactory amount of dwell time, with the timing in the advance settings, there will be more than enough with the timing set at the retard setting.
Picture 7 shows the trip finger. A trip finger length of 1 1/8", measured from the shaft center to knife edge was used to implement the three aforementioned events.
The next step was a trip to the local hardware store to see what could be used for an over-travel spring. This spring must have a stronger rotational force than the spring being used to hold the points open. The design of the over-travel spring mounting must have a minimum of lost motion. Stated another way, keep the spring stiffness high and the tolerances close to avoid excessive lost motion. Picture 8 shows the spring I used and the mounting arrangement. With an igniter, it is a good idea to design all parts for ease of disassembly, e.g. set screws vs. pins.
The last piece to design and fabricate is the advance/retard eccentric spool. Picture 9 shows this spool. Note that the shaft that the spool rotates about is displaced from the center of the spool by a small amount. This offset is very small and is used to provide the spark advance and retard function. I made the center offset just enough to provide a range of 32° before top dead center to about 10° after top dead center. I put a 60° angle on the sides of the spool to aid in keeping the square igniter push arm centered in the spool.
Picture 10 shows the assembled igniter and also shows the wing nut that I used to lock the advance/retard spool in place. This may be an unorthodox spool locking concept, but I like it because it affords me the opportunity to set spark timing any place within the 32°BTDC to 10°ATDC range. Many igniters have only two timing positions, start and run.
With regard to spool diameter, I recommend turning two or three spools of different dimensions on the lathe at the same time. Then try each on the igniter to see which one gives the desired results.
I enjoyed making this igniter and it gave me some insight into the ingenuity early engine designers used to solve problems. For the photography enthusiasts out there, I used a +4 diopter close-up lens for pictures 1 through 9. Fast film is needed to keep the depth of field to a reasonable value.