David,

Thanks for that suggestion and sorry for the delay with my reply. No I haven't performed a fuel quantity test yet (mainly due to the lack of a hose with the required fitting to fit in place of the return line from the fuel distributor). Unfortunately, since the problem really only shows when the car is driven, a fuel quantity test may not have shown a problem as will become evident as I describe how I eventually found the cause of the problem.

When I previously drained and flushed the tank, it was still fitted to the car. To flush it I removed the strainer and poured fuel and methylated spirits in through the filler and collected it as it drained from the opening the strainer fits into. I used methylated spirits to hopefully absorb any water that may have been present in the tank. I should have noticed that no matter how quickly I poured liquid into the filler, the rate at which it drained from the bottom was slow as this was actually a clue to cause of the problem, as was the fact that the car had been in storage for many years.

On Saturday morning I fitted a new fuel pump hoping this would fix the problem. For a while it ran fine but after about 20 minutes of driving the engine again stumbled and I noticed the new fuel pump had also become noisy. Obviously there was a restriction, presumably at the strainer, and the pump noise was due to cavitation. It was time for the tank to come out for a better look.

With the tank removed and the fuel guage sender (transmitter) and strainer removed from the tank I could have a look inside. Firstly I noticed the strainer was indeed blocked with what looked almost like mud. The bottom of the tank was also not clean. Obviously with the car stationary all this dirt would settle and allow it to run fine. When driving the car it would get stirred up, eventually blocking the strainer.

I plugged the opening in the bottom of the tank and added a few litres of fuel and gave the tank a vigorous shake. By then draining it quickly through the filler I was amazed at how much dirt washed out. The dirt included fine muddy silt and a small amount of rust. There was also still a small amount of water present. I repeatedly shook and flushed the tank with both fuel and methylated spirits until it looked fairly clean inside and no more dirt could be flushed out. By filtering the fuel and methylated spirits and allowing time for the dirt to settle, I was able to use much of it for several flushing attempts. Flushing and shaking the tank took an hour or more so I felt like I'd had a good workout at the end of it all!

I was also now able to see and understand how the swirl pot which sits on the bottom of the tank surrounding the strainer functions and understand why the strainer was drawing air with the tank less than about a quarter full. The fuel return line (from the engine) enters the bottom of the swirl pot via a venturi which is designed to draw fuel into the pot from the tank. This way the swirl pot is kept full and overflowing even when the fuel level in the tank is below the top of the swirl pot. The dirt in the tank was blocking this venturi such that the swirl pot could only get fuel from the tank if the level was above the top of the pot. It did of course still also receive fuel from the return line, but this had to have entered the swirl pot from the tank to have got to the return line in the first place.

With the strainer cleaned again and the tank refitted, I finally had the car running well that night just in time for it to attend its first club event on Sunday morning. Unfortunately I did not get to clean the rest of the car as much as I would have liked as I had spent all my time sorting out the fuel problems. The car drove to and from the event without a problem although I did notice that by the time I got home the pump was again a bit noisy. Presumably there is still some dirt present in the tank so I may have to look for another tank, not a bad idea considering there must also be some rust present.

Thanks David and Joseph for your suggestions. Hopefully this thread may help someone with similar problems one day.

Greg

__________________
107.023: 350SLC, 3-speed auto, icon gold, parchment MBtex (sold 2012 after 29 years ownership).
107.026: 500SLC, 4-speed auto, thistle green, green velour.
124.090: 300TE, 4-speed auto, arctic white, cream-beige MBtex.
201.028: 190E 2.3 Sportline, 5-speed manual, arctic white, blue leather.
201.028: 190E 2.3, 4-speed auto, blue-black, grey MBtex.
201.034: 190E 2.3-16, 5-speed manual, blue-black, black leather.