Trials and tribulations of a fuel tank

A bit of fun from today and last week.

The LS1-powered MG was showing signs of a fuel pickup problem. I'd been considering building a new tank, but the stock one fit so well under the car I figured I'd see if I could do something with it. So I pulled the tank out and opened it up.

To keep from blowing myself up, I filled the tank with water before using a sawzall to cut the top off. And was I surprised at what I found inside! This is the top of the tank, placed upside down on the driveway.

Check out that big canister...thing. I popped the four spot welds holding it to the top of the tank, here's what it looked like in place.

The MG is a 1972 model that I got out of California. It had a charcoal canister setup, and that's the line to the charcoal canister you can see in the first picture. The line ran between the canister and the top of the tank and breathed through a really small orifice at the top of the canister. There's a round hole in the bottom you can see. A dent in the bottom of the tank (I'm thinking someone tried to use it as a jacking point) was perfectly placed to block this lower hole off. So it was basically a giant empty cylinder in my tank.

The problem? I'd been using that charcoal canister line as my fuel return line for the fuel injection system. It's a wonder the poor pump managed to push any fuel back into the tank at all. It would have been squeezing it out of the grooves on each side of that canister at the top of the tank.

I decided to address all of this by building new pickup and return lines, using the big canister as a sump of sorts.

I cut the top off the canister and enlarged the bottom hole considerably. I flipped it over, cut a few slits in the (now) bottom and bent the resulting "doors" inwards a bit. The new pickup line was bent up out of 3/8" brake line with a 5/16" return - they also dodge around the fuel level sender, thus the odd routing. The return dumps back into the sump, aimed at the side of the cylinder so it'll swirl into place. The previous charcoal canister line was cut off and capped. I also beat the big dents out of the bottom of the tank.

There's an added trick in here. There's a baffle that runs from front to back on the tank. The float for the fuel sender is on the side away from the pickup. But the hole in the baffle is about 2" off the lowest point of the tank. So when you get down to your last gallon or two of fuel, it no longer gets transferred to the other side of that baffle...where the pickup is. At a certain point, the fuel gauge would stop going down as you continued to suck the fuel out of the tank. There's no good final result here. It's like they wanted you to run out of gas.

A few holes at the bottom of the baffle avoided the sneak "out of fuel" attack. Seriously, who thought that was a good idea?

Welded it all back together and voila, a new fuel tank! I ended up having to go back into it later because of a pinhole in the feed line that had the pump sucking air instead of fuel - should have pressure-tested that before assembly. The tank is upside down here.

I also had a problem of spitting a bit of fuel out of the filler neck on track. The filler neck is right above one of the mufflers, which led to some exciting moments.

The solution was to put in a flapper from a modern filler neck. I cut one out of a Miata filler and welded it into the neck for the MG. I'm hoping that will solve the problem. The future will hopefully bring a cool Aston filler neck setup, but not yet.

Thanks to all this I have an enlarged tank capacity (in large part due to the removal of the big sealed canister taking up space), an improved pickup, better baffling, fewer fireballs and (hopefully) no weird fuel delivery problems. Time will tell!

My mind just exploded (no pun intended) at the awesomeness of this thread, carry on

Nice work. Where did you source the Miata filler neck flapper?

Raze wrote: My mind just exploded (no pun intended) at the awesomeness of this thread, carry on

Indeed. A fuel tank is one of those scary bits I don't like to play with.

"Seriously, who thought that was a good idea? "

It's an MG there wasn't much "thinking" at all when that car was designed.

I wonder what that canister was designed to do.. seems odd to limit the amount of fuel you could carry, unless it was done for some arcane tax law or to make the tank smaller during the fuel crisis so you fill up with less gas

Was the fuel pickup line in that canister ? It may have been an aid to fuel pickup when cornering.

The pickup was beside the canister - the second picture has everything in its stock location. My theory is that the canister was somehow supposed to collect fuel vapor without collecting any fuel, as that's what is supposed to go into the charcoal canister. Exactly how, I'm not sure. Without the big dent on the bottom of the tank closing it off, the 2" hole in the bottom of the canister would have been able to keep the fuel level inside approximately level with the fuel level in the rest of the tank but with reduced sloshing. Since I switched to a vented cap, I don't need the charcoal canister system in order to regulate air pressure inside the tank.

Of course, with the 2" hole in the bottom of the canister sealed off by the big dent and the vapor line for the charcoal canister repurposed for use as a fuel return line, this was not working correctly in a bunch of ways.

Fuel tanks aren't that scary. I've heard of people filling them with exhaust to cut down on the explosiveness, but the water trick worked perfectly. Once the top was off, there was nothing to contain any vapors so the tank didn't even smell like fuel after a few hours of fiddling around. I didn't have to do anything special while welding it up because there were no explosive vapors left.

I've also been inside a Miata fuel tank. Not a lot of baffling, but with a small change in the routing of the return line we had that thing maintaining a good fuel flow through turns 1-2 at Thunderhill with a quarter of a gallon in the tank. It's pretty clever. It's kind of fun to put a bit of fuel in and then lift the car on one side to simulate light cornering. I had some fun playing with water in the MG tank after I modified the pickup too.

Here's a question for the carburetor guys - is fuel tank baffling as important as it is for fuel injected cars? The bowl in the carb might be able to act as a buffer against momentary lapses in fuel delivery. Not a huge one, of course, but you'll feel every hiccup in fuel pressure in an EFI car. I would think this would especially be the case in a low-power car such as an MG which isn't consuming mass quantities of fuel at full throttle.

The float bowl will act as a sort of tank to make up for slosh but lower the float bowl the leaner things get in a hurry due to less "head" pressure pushing the fuel through the jets. They used to sell a gas saving device that was nothing but a in line fuel pressure regulator with vacuum reference when you were at high vacuum is lowered fuel pressure thus drooped bowl height and leaned out but step on the go pedal it would go to full pressure and fill up the bowl. JC whittney sold a million of them...

I too weld on fuel tanks, a big thing to remember is that it's hard to say exactly what's inside. Some have chambers like the tank above, and it's hard to tell if they are completely empty or not. Filling with water usually does the trick, but if a chamber was accidentally sealed off by some random occurance like Keith's tank, it would be hard to detect. I always like to empty the tank after the water flush and shake/listen (in a quiet setting) for any liquid that might be trapped somewhere inside

Good suggestion. I actually picked up the water idea from someone on this forum, so thanks to whoever that was In my case, since I had the tank wide open after being cut, I was sure there were no surprises inside when I got the welder out. I used the Sawzall to cut it open partly because of the minimal spark production when compared to a rotary tool.

I love that car. If it's ever for sale please let me know!

I bet the canister's purpose was to keep fuel from sloshing into the charcoal canister line. As a secondary benefit it was also a baffle. That would be a lot easier to manufacture/install compared to a bunch of straight steel baffles going every which way.

Also, fuel tanks expand/contract with changes in fuel vaporization. This can cause problems, GM J cars had a recall where the tank baffles would, due to the flex, cause the tank to crack. That canister dealie would be much less likely to cause something like that. Pretty smart engineering, when you think about it.