1919 Eaton Restoration – Part 3 of 3
Initially, it was thought that little work would be required on the mixer apart from stripping and cleaning it. The check ball was a good fit in the inlet pipe, and the needle and what could be seen of the seat appeared satisfactory. The mesh filter to the inlet tube was cleaned along with the rest of the mixer and left to the side so attention could turn to the missing fuel tank.
It was only later, after assembly and trying to start the engine, that it was realized all was not well. Priming the engine, the needle was opened 1/2 turn, the choke closed and the engine turned over once. This resulted in an enormous amount of fuel gushing out of the mixer.
The seat for the needle was lightly cleaned up with a taper reamer before it was coated in engineer’s blue and the needle was re-fitted in the closed position. This revealed that when fully closed there was little contact with the seat, so a new needle was made from brass.
To do this, some 0.375-inch-diameter brass rod was held in the lathe chuck and the tip turned to a 60-degree included angle taper before narrowing the first 0.75 inch of the needle to 0.1875 inch diameter. Next, a 0.375 inch x 28 TPI thread was cut on the body of the needle using the screw-cutting feature of the lathe as the appropriate thread die wasn’t on hand.
The retaining pin was removed from the old adjusting head and the replacement needle fitted, a new cross hole drilled and finally a new retaining pin fitted.
With this done, the needle had only to be open 1/8 of a turn for the engine to run smoothly, giving rise to the suspicion that the hole in the needle seat was badly worn and enlarged. However, the mixer worked, so it was left “as is” rather than drilling it out and fitting a new needle seat.
The tank needed was round, approximately 8 inches in diameter and 2 inches deep. It would require four 0.25-inch-diameter studs to bolt the mixer to it and thus support the tank under the cylinder head, as well as a filler cap. The filler cap would be 1.5 inches in diameter.
To start, some 20-gauge steel sheet was cut into a strip 2 inches wide and 26 inches long. This was an inch longer than needed, but allowed for trimming and an overlap of 0.75 inch where the ends meet.
To make the base and top, two 8.75-inch-diameter steel discs were cut. A 90-degree lip would be formed to a depth of 0.375 inches around the edge on each piece. To mark these out, engineer’s blue was first painted on the steel then lines were scribed using a compass. The lip would be bent from the disc. To make the lip, a pulley 8 inches in diameter was found in the shed that would be used as a former.
One of the steel discs was centered and then clamped to this pulley, ready for the hammer to be used to form the lip. This was done by hitting the thin sheet around the edge resting on the pulley, using light and frequent taps rather than big hits. It took a bit of time to get the lips square, and then a file was used to tidy up any irregularities to the edge.
The finished top and bottom pieces were used as a guide to check the length of the center section that would form the body of the tank, allowing 0.75 inches for an overlap where the ends meet. To form this overlap, the last 0.75 inches was bent over some 0.0625-inch-thick steel to form a shoulder and recess so the outer side of the side would be completely smooth.
Where the ends of the center section would meet the faces were cleaned with a small grinding wheel to remove the galvanized finish, ready for fluxing prior to soldering. The inside of the top and bottom edges were also cleaned and fluxed. The first step was to join the base to the center section.
First the bottom edges were tinned, old, heavy soldering irons being used rather than bare flame so there was no chance of overheating and bending the thin sheet metal. The middle part of the center section was clamped to the base and heat applied to melt the tinned solder. Then, by working around the base, the center section was gradually joined to the base, finishing at the recessed joint in the center section, which was also soldered.
The top piece required cutting holes for the fuel filler cap, fuel feed pipe and the four studs to secure the mixer. The filler cap was cut from the top of an old can, then cleaned up and soldered in place. The four studs were made from brass rod threaded 0.25-inch UNC, with a flange at the end to fit on the inside of the tank. Using the mixer as a template to hold them, these four studs were soldered in place.
The overlap joint of the center section might spring open if too much heat was accidentally applied when soldering the top lid. To counter this, two 0.0625-inch holes were drilled through at this juncture, with two brass rivets fitted after cleaning the area around the holes and applying flux. These rivets were then soldered and once the lid had been fitted the outside heads were filed smooth.
Finally, the top and center section were cleaned, fluxed and tinned before being fitted and then soldered together. Once cool, the tank was filled with boiling water to wash out any flux residue and to test for leaks. Although the metal of the tank was galvanized, there were a few scratches. These were touched up with fuel-proof silver paint.
Once more, the Internet proved useful, providing pictures of reproduction splashguards to help in designing the correct style of guard for this engine.
It was then a case of looking through the workshop to see what plate steel was available in order to decide how the guard could be fabricated. The main part was to be 0.1875 inches thick, with a strip across the bottom of 0.25-inch depth increasing to 0.50-inch thick for the area of the securing bolt tab. The main body was cut 8.50 inches long from some 0.1875-inch-thick steel plate. This was an oversize 5.25 inches wide at the base, decreasing to 2.75 inches wide at the top. The “extra” width at the bottom end was to allow the edges to the bottom third to be curved before welding on side pieces.
To start, a thick piece for the bolt tab (0.50 inch) was machined on the end of some 1-inch-wide steel. The machining operations were completed before this piece was sawn off the bar to a length of 1.50 inches, with the slot for a 0.3125-inch bolt being cut first. Next, a recess was cut at the opposite end to the slot so that the inside facing edge would sit flush against the 4.25-inch wide piece of 0.25-inch thick bar that formed the base of the guard.
These two parts were then bolted in place on the engine so they could be used to provide a reference point to check the shape and position of the main body of the guard. Before starting to bend the main part, a center line was marked on both sides as well as the 4.25-inch-wide edges for the sides ready for bending. The vice was used with a bending jig (three pieces of round bar) to create the curve, progress being checked by holding the guard in place and then turning the crankshaft to check clearance with the big end.
Once satisfied with the main curve, the edges at the bottom end were bent in the vice to create a curved edge to this section of the guard. To make bending easier, a channel was cut using a disc grinder along the bend inside the guard. To complete the sides of the guard, cardboard templates were cut to fit the gaps between the guard and the base of the engine casting. It was then a case of cutting steel plate to match these templates, checking their fit and adjusting as necessary. These were then held against the guard when it was fitted to the engine and the sides were tack-welded in place.
It was then a case of removing the guard and fully welding all the joints before grinding off surplus weld and shaping the edges with a file. All that remained was to then check the fit, making some minor adjustments with a file, and the guard was ready for painting.
Skid and primer cup
The catalog pictures I found showed the engines fitted to two rails, but having recently acquired some 1.25-inch-thick mahogany I decided to use a plank, cutting a piece 29 inches long and 8.50 inches wide. This was sanded and then finished with two coats of clear varnish before drilling with four 0.375-inch-diameter holes for the bolts to secure the engine.
There was no priming cup fitted to the magneto bracket, only a square-headed plug sealing the hole. For completeness a priming cup would be made, mirroring one fitted to another engine in the shed.
To start, a length of 0.50-inch-thick hexagonal brass bar was chucked on the lathe and the first 0.75 inch turned down to a diameter of 0.40 inch so that the outside could then be threaded 0.125 inch NPT. This was then center drilled with a No. 24 drill bit ready for a 0.1875 x 32 TPI internal thread. This hole was then opened up to 0.25 inch for 0.375 of its depth and the start point finished with a 45-degree countersink. This end was then ready for the fitting of the sealing plug to shut off the cup.
This brass stem was then parted off to a length of 1.35 inches and reversed in the chuck. From this end the first 0.35 inch was turned down to a diameter of 0.3125 inch, leaving a shoulder with the remaining hexagonal body being 0.25 inches long. Next, the inside hole of this end (top) was then threaded 0.1875 inch x 32 TPI and a 0.125-inch hole drilled through the side wall 0.20 inch from the end, so that it came out in the larger diameter center hole, clear of the threaded portion. This was to enable fuel in the cup to flow down the tube.
To finish the cup, a piece of 0.80-inch-diameter brass bar was center drilled 0.3125 inch to fit over the top end of the stem. Then both the inside and outside were profiled to produce a bowl. This profile was cut by setting the compound slide to an angle of 10 degrees. The finished bowl was then brazed to the core part of the cup.
The inner stem was made from a 2.75-inch length of 0.1875-inch-diameter steel rod. This was turned down to a diameter of 0.13 inch for 1.25 inches of its length, and an 0.1875-inch thread cut for the next 0.375 inch. A brass cone was made, with a 0.1875-inch hole through the middle to fit the end of this rod and brazed in place. The rod was then passed through the thread and a right angle bend made to form the handle, thus completing the primer cup.
Painting and decal
As was noted earlier, the engine had been repainted at some stage, with no traces of the original paint visible. However, fragments of an old, thick layer of paint were found under the name tag when it was removed. This paint was nearly black, so it was assumed that this engine was originally painted Stover dark green. The finished color would be made by mixing green, brown and black enamel paint in a general ratio of 2:1:1.
The green paint that was on the engine had been applied without a primer onto bare cast iron, which generally looked as though it had been cleaned first given there was next to no rust. The engine was wire brushed and sanded to get a clean surface, the thin layer of “new” paint being easily cleaned off. The bare casting was immediately given two coats of bare metal primer. This is a high zinc content paint that although thick to brush on gives excellent resistance against rust.
Several coats of gloss enamel were then added and the whole engine was left in a hot greenhouse for a couple of weeks for the paint to thoroughly cure before assembly.
The Eaton decal was of the earlier square type, later engines using rectangular-shaped decals. The catalog pictures for the period show the core layout of the letter “E” and the words “Eaton Engines.” However, the scrollwork was a little unclear and the background color could not be identified. Several requests were made for help on Smokstak, but these failed to come up with a definitive answer.
As the engines from T. Eaton were actually manufactured by Stover, the decision was made to mimic the scrollwork on the Stover decal of that same period. First the Eaton logo was created using a drafting program before copying this onto the scroll pattern from a scanned image of a Stover 3-inch decal. It was then re-colored to give black letters and scrolls on a gold background. This was printed onto water decal paper and later fixed to the engine using clear spray varnish.
There were various items to take care of for the final assembly. One was the rocker arm spring retainer. To anchor the return spring for the rocker arm, a small plug was fitted in the hole in the casting that was used as the trip bar support for models without a magneto. Before tightening the clamp bolt for the gear-wheel pivot I ensured the pushrod assembly moved freely and traveled back to the rest position. If it stops short the magneto trip rod will not engage the push finger.
Next was the use of copper gaskets. There is only a narrow ridge for the mating surfaces for the cylinder head and the igniter. Normal gasket material would not work, so gaskets were made from sheet copper. However, after assembling the engine, compression was poor, despite the new valves, good bore and piston rings.
Attention turned to the igniter port, the mating part of the bracket being coated with engineer’s blue before offering it up. This revealed two areas of poor contact. The passage to the cylinder was filled with a rag and a scraper was used to remove the high spots.
Before fitting the flywheels the slots for the gib keys were in need of some attention. First a piece of 0.375-inch-square high speed steel was used to measure the slots as this steel gives an accurate and consistent measurement. A small scraper was then used to true up the groove so the steel was a snug fit, yet moved easily when all nicks and scrapes were removed. The gib keys themselves were loose in the slots, and on measurement were found to be 0.01 inch under size. New keys were made, 0.375 inches wide with a taper of 0.01-inch per inch.
On assembly there were no visible timing marks on the cam gear. The timing was set using the general rule that the exhaust valve should open some 35-40 degrees before bottom-dead-center and close at or just after top-dead-center. Once this was set attention turned to setting the igniter trip arm. There was a timing mark on the flywheel and this was aligned with the pushrod to set the timing.
Once the engine was running the rpm was measured and the governor springs were adjusted to give a speed of 450rpm, suitable for running without a load at a show. I have found that the smaller, lighter engines do not run well at really slow speeds without a lot of work replacing springs, etc.
This completed another full restoration, resulting in an engine returning to its original factory condition.
Crossley KB Illuminating Gas Engine Restored
Follow the acquisition and restoration of a Crossley illuminating gas engine from shattered pieces to a restored masterpiece.
Franklin Engine Modification
Follow Ronald McClellan through building the carburetor and replacing the air intake for the restoration of his Franklin gas engine.
Gas Engine Buzz Coil How-To
Learn how to make a simple, all-electronic buzz coil with David Cave as he walks through his process of building one.