Friday, October 25, 2019

Page 37: Fuel tank -- There's gotta be a better way

The tank is made up of four large pieces of aluminum sheet which comprise the walls, plus lots of other bits and pieces -- structural stuff which goes on the outside and various baffles which go inside. All of this has to be riveted together with different rivet types plus a few machine screws.
There are a couple of hundred rivets providing a couple of hundred sites for potential fuel leaks.

The piece shown at left (one marked scrap that I screwed up and the replacement) has seven different kinds of rivets or machine screws, some counter sunk, some not.  It's a beefy piece supporting two of the three bolts which actually attach the tank to the airplane.  The bolts are frangible (hollow) designed to shear in a crash or super-hard landing so the tank won't distort and spill fuel in the cockpit (something which could ruin your whole day).  All the rivets in this piece, along with all the rest, pierce the inside of the tank (leak possibilities).

Here's the bad thing about all these rivets: Each one has to be washed in naphtha and coated with ProSeal before being set.  ProSeal has a working life of two hours from the time it's mixed with the
catalyst until it hardens and can't be used.  This means that the process has to be carefully choreographed, with the appropriate rivets laid out, washed, and ready to be set.  As shown, I washed the rivets using a strainer belonging to the Spousal Unit (who is more than ready for the airplane to be finished).

A major mistake I made was ordering the ProSeal in the small (2 oz?) tubes which contain the sealant, the catalyst, and a built-in plunger to mix the two, thinking I'd break the process into multiple events, wasting less sealant.  Wrong.  A much better plan is to buy a quart and do the whole thing within the shelf-life window (90 days).  Turns out, one of the 2-oz batches I got was bad, causing indescribable anguish (more below).





Over three riveting sessions I installed all the aforementioned pieces along with the fuel-sender plate (actually two circular plates), which doubles as an access panel should a need arise to get back into the tank after it's sealed, a horrifying prospect which came true twice.

The assembly is to be leak tested after allowing the ProSeal to set up for several days by attaching a ballon to the tank, pressurizing the tank until the balloon inflates, then spraying soapy water all over the tank and looking for bubbles.  The balloon limits the pressure to something which won't cause the sealant to fail, supposedly less than one psi.  Being a curious sort, I decided to measure the pressure required to inflate a balloon by constructing a manometer. Turns out that every balloon I inflated needed about 1/3 psi (8 inches of water).  It took more to initially inflate it, but once stretched, 1/3 pis did it.  The Mothership says pressurize to one psi, but as long as I have a big enough delta P to blow bubbles, I think I'm good.  And blow bubbles I did.

With bated breath I pressurized the tank and proceeded to squirt the soapy water all over it.  Leaks appeared in six places.  After the misery of building the tank, it was as if someone had stepped on my soul. As Tom Cruise's sidekick said in Risky Business, sometimes you just have to say WTF?  I had
followed instructions, had done each procedure carefully, and this!  WTF!  At this point I had no choice but to open the tank, which involves inserting a putty knife between the access cover and the tank, removing the cover, then re-sealing all the offending sites from the inside.  The group wisdom of the Van's Air Force forums (and the Mothership) said don't be tempted to try it from the outside!

Turns out, contrary to what other blogs said, opening the tank was easy.  This should have been my first clue that the batch of ProSeal was bad, but it didn't occur to me yet.  I opened the tank, applied more ProSeal to the offending areas, and repeated the balloon test.  More leaks (picturing now a razor blade making lengthwise cuts of the arteries in my wrists).  I resigned myself to once again opening the tank and proceeded to do so.

This time it's almost impossible to do, far harder than the first time.  Bad sealant the first time!  I know I was within the shelf life, but it was bad.  I now bought a quart of ProSeal from the Mothership
and a digital scale from the aviation aisle at Harbor Freight, and did it all again.  Incredibly, there was another leak in the center of one of the rivets that was supposed to be "solid."  Rivet failure!  No way would I open the tank a third time.  I applied a suction to the tank, filled a cut-off syringe with ProSeal (needle removed, of course), and applied as much pressure as I could to the center of the rivet from the outside.  Success!

My rig for pressurizing the tank is shown at left: a 12-volt emergency pump for cars.


After letting the ProSeal cure for a couple of weeks, I put seven gallons (limiting the weight to about 50 pounds) of ethanol-free gas in the tank.  Over the course of a few days I turned the tank so that five of the six sides were down (didn't put it on its top), letting it sit this way for a day or so, and observed no leaks.  Of course, in the airplane the tank will be flexing somewhat, but I can't simulate that.

Boats don't have this problem.  The tank should be made of nylon or plastic or anything but riveted aluminum.  Horror stories about leaking tanks are plentiful on the forums, even the factory-made ones.  This now resides at the top of my Worst Jobs of the Build list, above the landing light and the tail cone fiberglass fairing.


Off Topic

This past summer I made my 31st pilgrimage to Mecca. Hard to believe I've been to Oshkosh that many times.  Can't wait for next year!  While there, I spoke with an insurance broker who told me that due to my age (hard to believe I'm 55....Oh, wait....) I should be insured and flying before another birthday happens.  Gotta be in the air before next September.  I'm on a waiting list for a hangar in Colorado and NC.