Chris Radek's stuff

Today I repaired the fuel pump on my '52 Chevy. I had replaced the pump once before and this replacement failed in the same way. A fuel pump on an old car is simple - it has a plunger and two one-way valves. When the plunger goes one way, it pulls fuel from the tank through the first valve. When it goes the other way, it pushes fuel into the carburetor through the second valve. The plunger is operated by an arm that sits against a lobe on the camshaft, so it is constantly pumping while the engine turns. When the carburetor is full enough, it simply plugs its fuel intake so no more fuel comes through. The fuel pump plunger is flexible enough to tolerate pumping to nowhere.

This pump originally used rubber washers for the valves. Above the washer is a metal "umbrella" which holds a spring that presses the rubber washer down against the valve's openings. The failure mode here is that the spring still works, and still presses the rubber down, but the rubber is no longer flat. It is curled up at the edges, letting fuel through in the wrong direction. If the umbrella was removable, it might be possible to repair the valve, maybe just by flipping the washer over. But, no.

This pump would still work well enough, when primed, for the car to run. But if the car would sit for a few days, the fuel would drain back through the pump the wrong way and into the tank. Once there was no fuel remaining in the pump, the car would not start because the pump could not pump air to prime itself.

I had planned to make ball valves (a ball held against a seat with a spring) but there was not sufficient clearance above the plunger to switch to this kind of valve. I needed something flat so I settled on a reed valve. The design is very simple. The body of the valve is a 0.2 inch thick piece of 1 inch diameter brass with one face carefully turned flat. The reed is a piece of 0.005 inch thick brass shim stock. After much practice I was able to cut out the reeds without distortion using sharp scissors. I filed very slight chamfers on all edges of the reeds, and then soldered them at one end to the valve body using a propane torch. It took care to not solder the entire valve shut (one of them took two tries.) The old valves were pressed in, so I only had to measure the seat and make mine the measured diameter plus about 0.002 inches so they would press into place. I used brass because I figure gasoline contains enough absorbed water to rust steel reeds. Being able to solder it easily was a nice bonus, though.

Upon reinstalling the pump dry, it pulled fuel from the tank and primed itself in about three strokes, and filled the carburetor to allow the car to start in about ten. I think the problem will stayed fixed this time.