[Schurkey]

Originally Posted by Kenth

all mechanical fuel pumps have a check valve that prevents fuel from "flowing back into the tank".

Some pumps have two check valves--one inlet, one outlet.

Some pumps have six check valves--multiple inlet, multiple outlet for higher flow volume. No doubt over the course of automotive production, there's pumps with other numbers of check-valves.

It's entirely possible for them to not seal completely. Even slight leakage of an inlet and outlet will drain everything from the needle valve to the level of the fuel in the tank, given enough time.

Thus my suggestion to connect a vacuum pump to the inlet of the fuel pump, to see if--or how badly--the inlet check valves leak.

[Shiny]

Originally Posted by Schurkey

....
It's entirely possible for them to not seal completely. Even slight leakage of an inlet and outlet will drain everything from the needle valve to the level of the fuel in the tank, given enough time.

Good point. I have never torn a pump apart but every online drawing I find suggests the valves are simple - like a disc or flap that may not have a backing spring to hold it closed. If this is the case, and since there is no significant back-pressure or residual pressure other than the weight of the fuel, I can see how these valves could leak a little.

I think these valves are primarily designed to seal against the internal pressure as the diaphragm moves so maybe not reliable for preventing slow drain-back? They don't have to seal very well to allow the diaphragm to push fuel through the outlet.

So how to know how much they leak?

Would be interesting to know the "specs" on reverse flow rate for these valves.

If, for example, there were 3 oz of fuel in the bowl and line, then a leak would have to be 3 oz/24 hr to drain it in a day. This is over 3.5 cc/hr, which I guess is feasible but sounds like a lot....maybe not?

A 1/2cc/hr leak would drain 3 oz in about a week. That sounds feasible to me, but just guessing.....

Originally Posted by Schurkey

... Thus my suggestion to connect a vacuum pump to the inlet of the fuel pump, to see if--or how badly--the inlet check valves leak.

A vacuum pump might apply enough differential pressure to close a valve that leaks a little under just the weight of fuel above. IMO, this could fail to detect a small leak but if the outlet check valve leaked a lot, this would detect it.

Mike

[grivera]

Originally Posted by Shiny

Good point. I have never torn a pump apart but every online drawing I find suggests the valves are simple - like a disc or flap that may not have a backing spring to hold it closed. If this is the case, and since there is no significant back-pressure or residual pressure other than the weight of the fuel, I can see how these valves could leak a little.

Mike

The Carter 6907 has springs for the fuel inlet/outlet valves. The rebuild kit included new springs, pins, diaphragm, and sealing discs.

[Shiny]

Originally Posted by grivera

The Carter 6907 has springs for the fuel inlet/outlet valves. The rebuild kit included new springs, pins, diaphragm, and sealing discs.

Just a guess, but seems unlikely to leak a lot. Maybe others are different?

[chrisp]

Originally Posted by grivera

The Carter 6907 has springs for the fuel inlet/outlet valves. The rebuild kit included new springs, pins, diaphragm, and sealing discs.

Have a 6907 & rebuild kit sealed if any in need.

[Kenth]

Originally Posted by Schurkey

Even slight leakage of an inlet and outlet will drain everything from the needle valve to the level of the fuel in the tank, given enough time.

This assumes that the inlet valve in the carburetor is at the bottom. I haven't seen any carburetors like that in the 40 that I have serviced carburetors.