Dell’Orto carb floats for the VHB 29 carburetors on loopframes sometimes develop pinhole leaks. The floats then fill with fuel, sink down a bit, and therefore do not completely shut off the flow of fuel to the float bowl when they should. This sounds complicated perhaps, but it’s really the same principle as a toilet tank valve. Archimedes would have grokked it in a second; fluid displacement, levers, valves, etc. In this case the floats rise on the level of fuel in the float bowl and push a conical rubber tip upward into an orifice in the carburetor body and stop the flow of fuel. When the fuel level goes down –as you wick it up to pass Myrna Tuttle doddling along in her Buick– the floats drop and the fuel flows to refill the float bowl.

When the floats don’t float, the valve doesn’t shut and fuel keeps coming through. The fuel then fills the float bowl, overflows into the carburetor, and runs into the intake manifold and cylinder (if the intake valve is open) and the air intake, and eventually out where it might cause larger issues. This is one reason why most owners of old bikes shut off the petcocks when not running.

Replacement floats are available, but I wanted to try repairing a set. If you decide to try, you’re also taking responsibility for checking in on them to make sure they are functioning as intended, and I’d still recommend closing the petcocks when the bike is parked.

First I drained my leaking set and let them sit in the sun outside for a while to dry out. Then I cleaned them thoroughly with alcohol and let them dry again. A bit of scrubbing is in order, and some carb cleaner might help as well. You want a good squeaky-dry bonding surface.

When I was certain they were both clean and dry, I lightly coated the entire float with West System’s G/flex, which is an epoxy that bonds to plastics. I gently mounted them in a vise upside-down after coating them, and put a bit of paper towel underneath to catch any drips. When I checked them after the G/flex had dried, some of it had run down to the lowest spot, forming an extra tip –on the top of the floats if they are dried upside-down– which can optionally be sanded off, but if on top won’t affect displacement as much, although obviously the coating adds a few ounces.

The float set that I coated is the greenish heavier variety. Even with the additional weight of the coating it appears to work well, and after over 17 months submerged in fuel does not leak or sink. If one can locate the exact spot of the leak, another way that might work is to enlarge that spot with a pin and pack some G/flex or other sealant in there and on the surrounding area. I have not tried doing it that way. I preferred to know that the whole float was covered, in case other leaks might develop over time.

As with all projects of this kind, the process is an open-ended experiment. I may find out that the G/flex eventually degrades in some way. At this point that does not appear to be so, but I am not claiming to be a chemist.

I am noting what appear to be spots on the float where the coating is either thin or has worn off, you can see one in the last photo, but I see no sign of leakage or failure so far.

Update: I had an opportunity to weigh the mended float and an intact one. A hanging postal scale that I have around the shop is probably not accurate and does not zero, so this is about the difference the Gflex makes, also taking into account that they are two different floats so may not have weighed the same originally. The intact float weighed 14g, the patched 15g. i had sanded off the “drips” at some point before weighing, now that I see the difference in weight I would do it again and also consider just patching only the area with the hole or holes.

Newly coated…


After 7 months of use…


After 17 months…