Fuel Tank, Round 3 - A Blessing in Disguise?!

Okay, so here we go again.  After considerable discussion with various other builders, the consensus became, "Yes, we do have a potential problem here".  More importantly, it is just good conservative judgement to change things when your gut tells you so.  Such is the case of the "misplaced" location of the Rotax fuel return line.  All politic'ing aside, the tank shall be opened again (third time) with the original return line removed and a new return line installed in a different location, that should suit the Viking engine.

Same song, third verse: work a small break in the Proseal to remove the fuel sender plate and ground plate to gain access into the bowels of the tank.

Below is the original return line that is being removed.  A total of two rivets that held clips which restrained the line to the roof of the tank were drilled out, the AN fitting loosened and the assembly removed.  Relatively easy process requiring no more that about 20 minutes.

In order to minimize or eliminate any outlet flow disturbances in the sump or baffled section of the fuel tank, the flow of the returning fuel needs to be slowed down as much as possible.  To that end, I substituted a larger 3/8 inch AN fitting and upsized the new return line to 3/8 from 1/4 inch.  Shown below are the 1/4 inch fitting on the left and the new 3/8 fitting on the right.

Next, it was time to get creative and bend some tubing into the new shape required to bring the returning fuel back to the bottom of the sump area of the tank and as far away from the outlet drain as possible.  Furthermore I also attempted to have the outlet of the return line close to one of the baffle walls, again, to minimize any risk of disrupting or disturbing the outlet flow of fuel.  The resultant line is shown below.

Yet another attempt to slow the flow rate of the returning fuel was to drill 5 more #30 holes in the sides and bottom of the return line to more evenly and gradually disperse the returning fuel into the tank. The next two pictures attempt to show the auxiliary holes.

Break out the Locktite 567 thread sealer and wrench down the new return line.  It was quite a challenge to photograph the installation with one hand reaching the camera blindly into the fuel sender plate's hole and capture the image I was after.  Shot below shows a portion of the line's AN fitting and the tubing turning over and approaching the opposing baffle wall.

Yet another couple of poor quality shots of the new return line installation with and without the camera flash.

 

The last two pictures of the installation that give a bit better perspective of what I was trying to accomplish with the alternative location of the return line.

 

As I was beginning to put the tank back together, I took the opportunity to double-check the published spec's of the sender unit.  Van's states that the fuel sender's minimum resistance "should be approximately 30 ohms.  The maximum resistance should be approximately 240 ohms."

Fate pitied me for all of my fuel tank sorrows and smiled upon me.  As I tested the unit again, I found that the max resistance would never go past 198 ohms, but the other stop reached exactly 30 ohms.  Further investigation found that the last 3/16 inch of movement of the float arm resulted in a dead spot on the unit.  I never could get a full scale reading past the 198 ohms.  Thankfully, I had saved the Stewart Warner box, found the part number and went shopping on Ebay.  30-something bucks later I'm in business with a new unit on order.  I removed the old unit from the ground plate, cleaned it and the sender plate up and all is ready for the wicked Proseal yet again.

And yes, I will be testing the new unit for full resistive range before I install it.  So, this was my "blessing in disguise."

One final thought.  The Van's Pilot Operating Handbook, states in the Before Takeoff section on page iii, "Fuel Quantity Indication - CHECK (no take-off with less than 4 gallons fuel)."

....Fast forward a few days.....turns out my blessing in disguise was just an extraordinarily stupid mistake on my part.  When one checks resistance with an Ohm meter, one must use the proper scale.  Its pretty fundamental that when using a 200 scale setting on the meter and the measured entity exceeds 200 ohms, that the meter would stops displaying values!  Such ends the saga of the bad fuel sender.  Oh well, at least I now have a back up unit.  Also, it was quite satisfying to find that the tech people at the distributor from whom I purchased the sender unit reproduced my "error".  To their credit, they figured out the problem in about 15 minutes.

Fuel Tank Leak Testing, Round 2 and a Very Unpleasant Surprise!

After the completion of the second service bulletin on the fuel tank, I allowed the Proseal to set up and cure for a few weeks.  Today, with spring-like temperatures in the 70's, it was time to check the leak status of my service bulletin modifications as well as the optional fuel tank vent installation.  Following picture shows the setup, much like the Round 1 version performed in May 2013.

Notice the upgrade from green colored water to blue!  In this shot above, the tank is pressured up to almost 27 inches water column, resulting in 1 psi pressure to the tank.  There is a leak!

Turns out it is one of the two screws that Van's insisted needed a lock washer underneath, I believe, to insure proper grounding of the electronic fuel sender unit.  You will see bubbles around the screw in the 10 o'clock position below.

There were the other usual culprits causing leaks during the pressurization, like the fuel tank filler neck, where it adjoins the tank with the large diameter radiator hose.

 Another source of an imperfect seal was the tank vent's 45 degree bulkhead fitting, which had been firmly potted in a nice glob of Proseal.  The leak is better seen in the pictures below, with bubbles forming at the base of the cap - not the Proseal!

Those leaks aside, the 10'oclock screw on the sender unit was the only legitimate leak that was observed.  Note the tightened screw and the distinct lack of bubbles while under the second pressure test of the day.  I would say we are good to go!  But stay tuned.....the plot is about to thicken.

Not 30 minutes after I left the garage and put everything away, basking in the sweet glow of victory of a tight, nonleaking fuel tank, I got a phone call from Dick Gossen, another RV-12 Viking builder from the great state of Texas.

The conversation went something like this:  "Tom, I've think I've found a problem with our fuel tanks and the way they might not function properly with the stock fuel system as constructed for the Rotax engine...".  It was too funny to get upset about.

The long and short of it is that the fuel return line is probably in the wrong location for fuel tanks that operate with a Viking engine.  The fuel injected Viking is of course, a pressurized system that utilizes the stock return line, originally designed to minimize vapor lock with a Rotax engine.  The sump, where the fuel outlet is located is at the front of the tank.  The fuel return line terminates at the rear of the tank, pointing upward into the center of the filler neck.  This location of the return line is the best way to cool warmed fuel in order to prevent vapor lock.  Unfortunately this characteristic works completely against the Viking's requirement for returning fuel to the sump as quickly as possible.

The Viking engine with its automotive fuel pumps delivers fuel at the rate of about 30 gallons per hour, with a return rate of about 25 gallons per hour, netting a nominal consumption of about 5 gallons per hour.  Under this assumption, once the fuel level reaches the 5-6 gallons remaining in the tank, fuel is mostly contained in the baffled, sump area.  Fuel goes in and out of this smaller (approximately 1 quart) area via a series of 4-5 small holes at the base of the baffles, where they contact the floor of the tank.  Perhaps you can see the potential problem.  Fuel returned to the tank at the opposite end cannot "trickle" its way into the sump area fast enough to overcome the rapidly dwindling fuel supply...and a fuel starvation situation could exist.  Combine this with the possibility of a go around, where what little fuel remains is flowing out of the sump area and you could have a major problem on your hands.

If Dick's assumption is correct, and I have no reason to believe that it is not, it will mean opening the tank one more time to greatly shorten, relocate and secure the termination of the return line.  It only makes sense with our high pressure fuel injected engine to return unused fuel as close to the sump/fuel pickup area as practical, without creating other problems.

Dick has initially suggested shortening the return line such that as it enters the tank, adjacent to the fuel pickup fitting and travels vertically toward the top of the tank, that it stop before exceeding the height of the baffle wall and immediately take a 90 degree turn and point directly to one of the four baffle walls.  This configuration would keep returned fuel in the immediate area of the sump and hopefully minimize any flow disturbances as fuel leaves the outlet.

Onward through the fog!