Fuel Tanks

All fuel system pages

Roger Mellema

Fuel cell leakage

The BD-4 has more than it’s share of fuel cell problems but don’t feel like the Lone Ranger – many airplanes have the same problem. The best way to insure good cells is to prepare the glass panels before wing assembly. There are many people who think. sloshing material is a lot better than it really is and hope it will fill all the cracks and pin holes. The panels should be inspected and the best ones chosen for fuel. These should be checked carefully for flaws and these should be fixed on the inside of the wing using fiberglass (some builders lay another layer of glass on the two end ribs of the tank). The inside of the cells should be sanded until the entire surface is roughed up (don’t sand into the cloth). The surfaces should be cleaned more than once with acetone or MEK – not just wiped but flow it on with one cloth and wipe it dry with another. I then painted four coats of thinned slosh material on all surfaces except the 1.5 inch edge to be glued. Make sure all the pin holes get filled. After the wing has been glued together, the tanks are sloshed with unthinned slosh. I like to keep turning the wing while a vacuum cleaner, hooked backwards, blows warm air into the fuel gauge hole.

This helps the slosh dry evenly on all surfaces. Be very careful to keep sparks from getting anywhere near the fumes as they are very explosive. I know of one person who blew up a wing.

Some people coat the entire outside of the tank area with a light layer of fiberglass. This works very well but the last time I just put 2 inch strips of glass over the cell joints. The 2216 used on the wings does not have very good peel strength. I had a couple joints come loose on my first BD-4. Av-gas eats asphalt beautifully so I quickly flew away before they decided to charge me for it. Make sure you fuel test your wings before you go to fly. Fill them up to the top and let them set for at least two days. It eaves a lot of embarrassment come test flight time.

Roger Mellema
Newsletter #13 (March 86)

Paul Wood from Lincoln, NE recently purchased a BD-4 and is having some trouble with leaky fuel tanks and other things. Sorry I haven’t written back sooner, Paul, but I will answer your questions here as maybe other readers have the same problem.

The fuel tank problem is a very common one and many people who buy a used Br) are told by the seller that “it never leaked when I had it”. This could be true but knowing BD-4’s, I rather doubt it. The leaks that occur in the outside skin are usually very easy to correct by putting some sealant over the leaky spot on the inside of the tank or by putting a small fiberglass patch on the outside.

Leaks that occur along the rear of the wing are usually caused by not sealing the flap/aileron attachment brackets properly. These again are not too difficult to fix if you don’t mind taking the flaps and ailerons off and mettind in and using some Pro-Seal to fix the oroblerii. By the way, I have seen builders pour unbelievable ariic-urits of slosh into the wine in the hopes of sealing a leak such as this and never oettina it done. Slosh is really only good for coating surfaces so oas can’t attack therii and sealing piri holes. The hardest leaks I’ve seen to seal are the ones that Occur in the ribs at each end of the fuel tank. The inboard rib seems to be the worst problem. This may be because of the slight flexing of the tubular spar which will be worse in the vicinity of the inboard rib than by the outboard rib. When you have a seep in one of these ribs, it usually requires cutting a hole in the last cell of the fuel tank, cleaning all the gunk (sloshing material and 2216 glue, etc.) off of the rib and then painting it with brushable Pro-Seal.

Before doing anything drastic, check the hose fittings and the fuel sender gasket to be sure they are not leaking. it sometimes helps to clean the area with solvent (or soap and water, or diesel fuel) and blow a thin layer of white flc-ur or powder oritc, the surfaces to make it easier, to see where the leak actually is. It is very difficult to tighten the fuel lines to the fittings that go into the tank so this is a frequent source of leaks. I like to coat the rotarig surfaces of the fittings with a little Pro-Seal to make sure they don’t leak.. Par-t of the problem here is that Bede supplied 2000 lb hydraulic lines to be used as fuel lines as they are extremely stiff. It would be a very good idea to get rid of these and put in something reasonable. The fuel pressures in this area are extremely small. When working with the hoses be careful to riot turn the fittings that go iritc- the tank as that will certainly cause leaks – it is very difficult to work on the aft fitting, sometimes requiring griridirig down wrenches etc..

When cleaning the ribs, be sure to pay special attention to the area j,ust around the spar. This is a very difficult area to get at but is by far the riiost iriiportant. The wing cell should be cut into so that you have good access to this area. If you think you are going to out a 6 inch hole in the top of the wing and work throumh it, you are very riiistakeri. What it really requires is to cut a hole in the BOTTOM of the cell that noes 12 or more inches forward of the spar and 12 or more inches aft of the spar. It should be cut to withir, 1.5 inches of each rib. The panel can be cut in an ‘HI shape with the center of the “H parallel with the spar. You will then have to find a way to hold the ‘HI open so that you car, work freely. It really is a lot easier to work if the whole panel is cut out but it is a little easier to close up the hole again if everything is not cut away. Next you should use acetone or MEK and/or a flat scraper (looks like a wood chisel) to scrape away the slosh around the hole and from the entire rib. Don’t get lazy and quit before the job is done – set everything up so you can sit down and take your time while doing this. It helps to use a trouble light held on the other side of the rib you are cleaning to see where you aren’t getting it right. This is a lot easier on the inboard rib where both sides are accessable.

The next job is to coat the rib with brushable Pro-Seal. A couple of good even coats should do it. The light will come in handy again here to see areas that have been missed. The fiberglass skin car, be closed up again by getting the cutout part to lay right and then covering the cut with about three layers of 8 oz fiberglass cloth (at least 2 inches wide) and polyester, r,esiri. The very early, very white opaque wing panels were made with epoxy and you will have to experiment to see what will work with these. If you think you need rilore strength than the fiberglass patch provides you can put manner- to the gas cap installation.

I guess the one point I want to make clear is that you CANNOT fix a leaking tank by dumping more slosh into it. I have seen a couple thousand dollars in slosh go down the drain with out doing much r3c-od. A couple of times I have seen the tanks stay sealed for a month or two and then take up right where they left off.

Fuel System

Roger Mellema
Newsletter #2 (Jul 81)

Some builders have had fuel starvation when they have a high angle of attack. The fix, according to some builders, is a three gallon header tank on the firewall. My ED-4 is built according to plans and I’ve never had a problem. The problem I have had with fuel is that with large tanks (5 cells on each side) a slight “off ball” condition will cause the fuel in one tank to be inboard and outboard in the other. This is not much of a problem if you have a both position on the fuel selector. If I’m low on fuel, I watch the ball very carefully!

Remember that if a header tank is used, it must be vented up to the tops of the wing tanks. The fuel selector can be modified to have a both position but the FAA requires that the tanks then must be vented to each other. I had a both position on my first BD and it worked alright without a common vent.

George Deveault
Newsletter #18

The other day my insurance man told me I couldn’t be covered anymore unless my BD-4 has a Header Tank. I wrote him the following letter, and much to my amazement, he called a few days later and said the underwriters (Eastern) agree with me!! – So now I’m insured for at least another year.

Dear Sirs,
Back in the’70’s, when I was building my BD-4, I learned of a take-off accident at Long Beach caused by low fuel/ steep climb-out/ venting the forward fuel pickup port. I considered this very sedously, built simulator models, and discussed the problem with many knowledgeable people.

Finally, in July 1977, while in final assembly stage at Mojave, I had the opportunity to “brainstorm” the problem with Burt Rutan, who wasn’t too busy in those days and showed an interest in my project during his lunch hour. Burt and I simultaneously came to the conclusion that the forward port should be eliminated. A most simple and obvious solution!
I don’t remember any specific conversations with the FAA regarding this fuel system, but they did oversee my BD-4’s construction and OK’d the first flight of Aug 3, ’77. FAA inspector Bill Daniel issued my Airworthiness Certificate Dec 1, ’77 after 50.5 hours of flight- testina.

I have intentionally run a tank dry in flight with no problem, immediately re-starting by switching tanks. I’ve flown many skidding maneuvers and unusual attitudes with never a burp from the engine/fuel system. The only conceivable fuel-starvation would be at a negative angle of attack with very low fuel supply. And as Burt Rutan said to me” “Who cares- you’re going down-hill anyway.” But to be safe, I’m happy to placard my panel with “This Aircraft not to be operated with tanks less than 1/4 full”.

John Brecher
Newsletter #14 (May 87)

I’m getting down to final assembly now. I’ll be taking the BD to the airport next weekend. There I should have 3 or 4 weeks of readying for flight test. Tinsman is to do the test +lying and he is having some knee surgery next week so plane and test pilot should be ready by mid Dec.

I just went through a month of Avemco, EAA, and chapter designee drill trying to get insurance coverage. I knew there had been some problems with fuel starvation but didn’t know the magnitude. Avemco says they insured 9 BD-4’s before my request and all 9 crashed with what they felt was a fuel flow problem. Avemco asked me to contact Ben Owen at EAA headquarters to discuss my fuel system. I told Owen that I had run the fuel lines down the front and rear door posts like Cessna. He asked that I call a local designee +or a visit. I got Jim Miller from Chapter 91 in Kansas City to come and look over my system. Miller, Tinsman, and I went over everything. Miller got to the point where the two lines from each wing “T” together. He felt this was totally unacceptable. He was sure the air from an unported line could win out over fuel flow, particularly with just a “right” or “left” fuel selector. Tinsman and I argued that nost BD’s were plumbed that way. Miller also didn’t like the rear fuel pickup running forward to just 3 inches from the front pickup. He felt that would cause head pressure difference to be very low and thus allowing air in the system.

First, Miller suggested a “flop tube” and one line per wing. I also said I wanted to make a “both” position as per Jim Kerr’s letter in BD-4 newsletter in the mid 70’s. Miller agreed. The flop tube could get into the fuel sender unit but naturally I thought of that after Miller had machined two “work of art” rolling -Flop tubes at a cost of $100.00. I told Miller I’d heard that the insurance company would accept a BD-4 if it had a header tank. I didn’t want to put it under the panel as that would be difficult with the panel complete. I suggested a pair of tanks behing the seat would be easier and better for weight and balance. Jim said that would work fine. We came up with the size due to Wickes tube size available 4 inches in diameter and 20 inches tall fits nicely. This holds 5 quarts each tank. Miller feels very strongly that this mod will cure the fuel +low problem and thinks it should be mandatory. Tinsman and I feel there are other fixes too. I think the both position is a major +actor. The system the fellow in PA did with outboard bay sumps is good.

The fuel selector mod takes less than one hour and costs nothing. I’d like to know if any BD had fuel flow problems with the both selector. Miller is sending the attached article to EAA +or publication. I installed the system as drawn. It would take someone about a weekend to retrofit at a cost of almost $120.00. Miller welded up the tanks from 6061-T6 tubing. Nylo-seal was used from bulkhead at wing to the fuel selector. As I’ve said, this is a good fix, but not the only one. it is going to get me insured with Avemcc. Rates with me having 300 hours logged ‘is:

Liability – 500,000 each accident, 50,000 each person = $540.00
Ground and Taxi – 16,000 value = $824.00
Liability, ground, and in flight = $2603.00

At meetings I’ve gotten the impression that insurance doesn’t rate very high with others but it does with me.


The following letter was sent to the EAA.

The BD-4 Aircraft has apparently earned a reputation for fuel -Flow problems. One insurance company reports that 100 % of the BD-4 aircraft that they have insured have been damaged as a result of fuel +low problems. Consequently, this insurance company will not insure a BD-4 with the original design fuel system. After being refused insurance, a local builder sought assistance in evaluating his fuel system. This article is the result of that evaluation. The BD-4 utilizes fiberglass wing cells +or fuel tanks, with the majority of the fuel space being located behind the tubular spar. Feed lines from each tank port join together near the fuel selector valve that is located below the instrument panel. The fuel selector valve has 3 positions: Right Tank, Left Tank, and 0++. This arrangement appears prone to experience fuel flow problems with anything less than a full fuel load, because one port can be uncovered during high angle of attack operations. The uncovered port then allows air to enter the line, near the selector valve, and reach the carburetor or fuel pump, resulting in complete loss of power. Secondly, the lack of a “both” position at the -Fuel selector valve subjects the aircraft to total loss of fuel +low in a slightly uncoordinated or climbing turn if fuel is being drawn from the tank in the lower wing. The obvious solution for the fuel selector valve is to modify it to provide a “both tanks” position or, to obtain a valve incorporating this provision. It is also important to placard the aircraft for use of the “both tanks” position during take-off and landing operations.

On the specific aircraft that was the basis of this article, the unported tank outlets that occur during high or low angle of attack operations were accomodated by installing a pair of 5 quart header tanks behind the rear seat. Each wing tank port +eeds directly into the associated header tank via an individual 3/8″ minimum diameter line. Each tank port is protected by a finger strainer. A balancing line to each header tank was installed. This is absolutely necessary to avoid an air locked condition and insure rapid refill of the header tank when normal aircraft attitude is resumed. The normal main tank vents must not be directly connected to this balancing line. The header tanks were constructed of 20 inch lengths of 4 inch diameter 6061-T6 tube with 0.050 inch walls. The tanks were located vertically just a+t of the rear seat with the lower end secured to the aircraft’s lower skin via the drain valve fitting. The upper end was secured to the seatback frame. Fuel is +ed to the selector valve at a point one inch above the bottom of each tank. This provides a sump for water collection and removal. The general arrangement is as shown in the drawing. Total empty weight addition was approximately 4 pounds +or both the header tanks and the additional lines. The positive aspects of these changes are that the header tanks not only add 2 1/2 gallons of additional fuel, but also allow reliable use of almost all fuel on board. Take-of+ acceleration forces will tend to insure that the header tanks are full. Also, full header tanks will provide at least 10 to 15 minutes of usable fuel after almost every drop of fuel is used from the wing tanks.

The changes described appear to overcome the problems inherent with the original BD-4 system. It is suggested that equivalent changes be considered mandatory modifications for all BD-4 aircraft (see Figure 2).

Jim Miller, EAA Technical Counselor #529


There is merit to the above fuel fix. I guess the fuel in the cockpit bothers me some, but it is in a better place than under the instrument panel. I think that a single tank would be much better than two as far as complexity and probably weight.

The thing that really bothers me is sending such a letter to the EAA and claiming that it should be mandatory on all BD-4’s. When this was sent, there had been no physical testing of any sort done to prove his theory. Avemco’5 story bothers me also in that I have called several times over the past 5 years and they always say that they cannot cover BD-4s because the +laps and ailerons rip off in +light!

Joe Gauthier, Roger Mellema
Newsletter #18

Engine Driven Fuel Pump Failure/Gravity Flow

A recent incident involving a fellow BD-4 pilot and an off airport landing caused me to modify my fuel pump installation. I have operated this, and one other BD-4, for many years (800+ hours) without an electrical backup to the mechanical pump. The theory being that; if the mechanical pump failed, gravity will take over and continue to deliver fuel. My friend and his BD-4 proved that theory wrong! His mechanical pump failed and gravity alone did not provide the fuel to keep his engine running.

It appears that even if gravity delivers enough fuel for takeoff through an idle mechanical pump, a backup is still necessary. The failure modes for mechanical pumps obviously include situations where the fuel simply will not flow adequately. The failed pump would flow all the fuel necessary when sifting idle on the bench. The operational test, in our local engine shop test cell, revealed that it would run the engine at idle RPM with about 1 PSI at the carb inlet. However, it would not deliver enough pressure or volume to allow the engine to run higher than idle. When the electrical boost pump that the test cell is equipped with was turned on, the engine produced all the power it should have.

Before the fuel pump situation developed, I decided to look very closely at the fuel delivery rates in our BD-4. I am concerned about the number of unexplained fuel problems. Originally I was interested in the effects of ram pressure on the tank vents. The fuel vent diameters, lengths and configurations were all areas that needed study. Recognizing that measurement of pressures so small would be difficult to read, I decided to temporarily install PYREX glass inline filters so I could actually look at the flow. The glass filters were installed so they could be observed at all times during flight.
Other than my glass filters, the fuel system that I tested is per Bede plans. It includes 3/8 inch aluminum plumbing, flush fuel filler caps and a selector valve that allows a choice between the right or left tank and not both, as well as 1/4 inch vent lines that faced down and slightly aft. Either tank is capable of holding 28 gallons of fuel and each has two pickups, one at the low point in climb. All four pickups are equipped with standard mesh screens to keep any debris out of the plumbing. This plane is equipped with a 180 hp Lycoming engine with a mechanical fuel pump with no backup during these tests.

I began by documenting the performance of the system as originally installed. As one could expect, the fuel level in the glass goes down to about 3/4 full with application of full power for the takeoff run. Reaching cruising altitude I reduced power and switched tanks from right to left and back, no change. Both tanks are nearly full. Fifteen minutes into my flight I began to notice that the fuel level in my PYREX filter is now down to about 113. After I returned to the traffic pattern and reduced power, the filter level increased back to 3/4 full. After landing the glass is 7/8 full.

My next test was to see if repositioning the vent lines would have any material effect on the fuel level in the filters. I bent them just slightly forward and filed a 45 degree angle facing into the slipstream. I flew it this way for a few months and no real change in filter fuel level.

The third item I tested was fuel caps. With the decrease in flow not occurring until after a fifteen minutes or so of flight, I began to suspect that I might be loosing some of my vent ram pressure through leaking fuel caps. I replaced the old “o’ rings and increased the tension on the lock to be certain they didn’t leak any air. This made my most significant change. The glass remained about 213 full all the time. It’s hard to believe but, there appears to be enough air drawn out of the tank through the cap to significantly affect the flow. I expect that the combination of leaking fuel caps, and a long, (72 inches) small diameter vent line, (with a poor forward facing angle) caused the decrease in fuel flow that I observed. The faster and higher I went, the lower the fuel delivery rate became. The dynamics of aircraft speed and altitude, cap position, the condition of it’s seals and a small diameter, poorly placed vent lead to the significant reduction in the pressure difference which I observed in my PYREX looking glass.

In addition to tightly sealed fuel caps and forward facing vents, our BD-4 now has an electrical back-up pump as well as a mechanical fuel pressure gauge. The BD-4 that landed off-airport did not have forward facing vents or an electrical backup pump. The BD-4 airframe did a great job of protecting the occupants in this incident. The landing was in small trees and heavy bushes in a swampy area. The airframe remained intact and there were only minor injuries and bruises to the occupants. The plane is repairable.

ed. note: Joe didn’t say just where the PYREX filters were located, but it is obvious that he did not purge his fuel lines of air after the modifications. I say this because of the different amounts of air appearing in his filters. I feel that the amount of air in the filter did not indicate just fuel pressure but also rate of fuel flow. With the size of fuel line we use, fuel will flow past an air bubble without pulling the bubble all the way through the system (self purge). What will happen is that the bubble will move down the fuel line in proportion to the amount of fuel flow. If the tanks are pressurized more, the air bubble in the line will be compressed some and this will result in less air in the filters.
We believe it is always dangerous to have air in the lines because it makes the loss of siphon happen quicker. After running your lines dry for any reason, the fuel lines should be purged by opening the gascolator drain and blowing on the fuel filler hole or the vent line to pressurize the tank slightly. Ray Ward and I learned this lesson at the CAFE 400 race a couple of years back!!

The size of the vent lines should not be a problem as long as the air flowing through them is very slow. Our fuel usage is so low that our vent lines are big enough. 20 gal/hour is .7 ounce per second, or 1/3 gallon per minute.

I think that pressurizing the fuel tanks is helpful. Of course this does not help as much during an 80 mph takeoff as it does during 200 mph cruise when you don’t need it as bad (lower fuel flow). The pressure on the vent tube is from two sources: the higher pressure under your wing due to lift (0.16 lb/sq in) and the ram pressure due to facing the vent into the airstream (0.7 lb/sq in @ 200 mph). As you can see, the pressurization is very small, but helpful. Remember that you do not get much ram if you are inside the boundary layer. Make sure the ram port is below the wing by 1/2 inch or so. The fuel cap seal is something that several people have mentioned to me. This is the first time it was concerning ram pressure loss. Usually it is water leakage due to an old cracked “o” ring.

Roger Mellema, Jim McCord
Newsletter #19

Jim McCord pointed out an interesting fuel system pick-up that was featured in Sport Aviation. It is a valve that effectively closes off the pick-up (front or rear) that could suck air. It uses as ball bearing and will switch pick-ups when the aircraft accelerates or pitches.



Care must be taken to make the valve so that the ball can never be sucked tight over the exit hole. This can be done by having a drain valley between the lower ramped rails as shown on the right.

This valve looks like it could really help us solve any problems associated with pitch or acceleration. Remember that it still does not solve the “ball off center, or slip/skid” problems that force the fuel away from the pickups.

Jim also suggested that “flap valves” be used between the ribs so that fuel will drain away from the inboard bay much slower. I suppose these valves could be made so that they are pretty tight. The way you would use them would be to “pump” the fuel into the inboard bay by flying with the ball off center to the opposite side of the tank you wanted to pump (i.e. ball to the left, you would pump the right tank). Of course turbulence and other natural rocking while flying would also “pump” the system. The ribs could have holes through them in the upper 1/3 so that the tanks could be easily filled.

Another suggestion was that the bottoms of the outer fuel cells be filled with increasingly thicker layers of foam and thus force the fuel to the inboard rib. The only thing wrong with this idea is that changing altitudes can “pump” fuel into the foam if anything isn’t done just right.

An idea for pressurizing the fuel tanks is to use a NASA vent on the bottom of the wing. Maybe this way they could be pressurized with nothing protruding below the wing.

Carlos Serodio

The best fuel senders/gauges possible for a BD4… they never go wrong and they display clear and precise indications. I even use them to trim my rudder for a perfect attitude of the longitudinal axis. The front clear hose shows the fuel in the main wing tank (cells 2-6), the rear one shows the fuel of the auxiliary wing tank (cells 7-9). If you’re interested I can draft you a sketch for the installation of these gauges and the complete fuel system I have in my BD4.

My outlets from the tanks connect to fittings on the cabin wall just above the doors (see photo).

Main tanks (cells 2-6): the fittings on the cabin wall closer to the cabin spar (the lower one for fuel, the upper one for venting) are “t” type (i made them myself but stock ones could as well do). One arm of the fuel “t” is connected to one arm of the vent “t” with a piece of clear fuel hose. Inside the arm of the fuel “t” i put a small piston of fuel resistant rubber with 2 pieces of syringe needle driven across it. This reduces the freedom of flow of the fuel inside this bypass and stabilizes the movement of the fuel level inside the clear hose. I like it this way, the rubber piston is indeed optional. The other arm of the fuel “t” is a fuel line. The other arm of the vent “t” on one side connects to the other side for cross venting between the tanks.

Auxiliary tanks (cells 7-9): there are a fuel pick-up and a vent pick-up on the wall between cells 6 and 7, just by the wing spar and in a position similar to the corresponding main tank pick-ups in cell 2. I added a free-flow stop valve between cells 6 and 7, for safety. Fuel can flow from cell 7 to cell 6. An alu tube goes across all the cells of the main tank connecting the vent pick-up of both the main and the auxiliary tanks. Another alu tube connects the fuel pick-up to a fitting on the wall of the main tank (between cells 1 and 2) this latter becoming the new fuel pick-up for the auxiliary tank. I attach a drawing showing how this new fuel pick-up is used to run a fuel gage for the auxiliary tank and to have fuel pumped from the auxiliary tank to the main one. Please note that there is also a rubber piston with syringe needle in this system allowing for the fuel to always flow to the gage. When the facet pump is working the fuel is pumped through the gage (useful feature because makes the operation visible).

K. Skaggs: The Grumman Yankees (also known as the BD-1) had an AD requiring a floating red ball in the clear tube. Having had the misfortune of reading one wrong, I completely agreed with the AD and suggest you consider one here. Otherwise, they are failproof as long as you’re in balanced flight.