December 16, 2019

Fuel tank leak test passed

After the required minimum 3 days elapsed to allow the sealant to cure, I was able to perform leak tests on my left fuel tank this past weekend. I closed off the vent and fuel supply fittings, and then used a bicycle pump tethered to the fuel return fitting via 1/2″ ID tubing to pressurize the tank. The instructions say no more than 1 bar of pressure, which is a little less than 15psi – pretty low to register on the pressure gauge on the pump, so I just gave it a few pumps. You can see the skins start ballooning up right away, so I knew I was getting air into the tank!

Typical floor pump with built-in pressure gauge
No bubbles!
Drats – lots of bubbles

All of the surfaces along the rivet lines were bubble-free, but my first attempt to close the fuel level sender cover didn’t work. I really didn’t want to apply sealant to the bolts for this cover, so that it will be easier to service in the future. My first attempt involved using nylon washers on the M4 bolts in place of stainless washers, but it just wasn’t enough of a seal.

I tried adding small Viton o-rings on the screw, at the end of the threaded shaft. In theory this will work well – but the problem with this idea is that the o-ring gets squished too much as the screw is torqued down, and eventually causes the o-ring to tear. I tried chamfering both the washer and the cover surface, to give the ring some place to expand into as the bolt is tightened, but there was still too much tearing.

My third idea was to make flat washers out of the same material as the custom gasket between the cover and the rib, but this material is also pretty squishy (75A durometer) and wants to tear before getting to the nominal 50 in-lbs of torque. There is a company that makes sealing washers called Seeloc, but they don’t appear to be fuel resistant, and at $2 a piece they are quite expensive. Their solution appears to work because the washer and seal are designed correctly so that the seal can squish down and form a tight seal without tearing.

My fourth and final idea was to apply a short section of heat-shrink tubing to the last third of the bolt shaft closest to the head, so that a seal is formed as the bolt is tightened. The heat shrink I’m using is made from polyolefin, which I learned is resistant to fuel. The idea is similar to using Teflon tape, but without the worry of bits of tape tearing off and potentially clogging something in the fuel injection system.

Small bit of 3.7mm diameter heat shrink tubing on the M4 bolts for the cover

The heat shrink tubing seems to have done the trick – I can feel the resistance of the tubing as it starts biting into the threading of the rivnut, and it torques down fine. I backed a screw out to make sure it doesn’t tear apart and leave debris, and all looks fine. Since this little piece of tubing does not extend past the end of the rivnut and into the fuel tank, I feel confident that I won’t have debris entering the fuel system.

One more pressure test, and this time no bubbles! So it looks like my heat shrink tubing solution will work.

No bubbles!

I’m on another business trip this week to Germany, so I won’t be able to make any more progress this week until the weekend. I had hoped to finish up the other tank before my sealant expires at the end of the month, but I don’t think I’ll be able to. I wonder if it’s still safe to use the sealant? I’m using the Flamemaster CS3204 sealant recommended by TAF, and it’s been in my cool dry garage / basement all winter.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.