After glassing the inside of the top skin I found that I must have measured wrong when I decided I needed extra on the leading edge. Now the foam was too long. I'd caught it at knife trim time, cut the 11 lines and let it cure in shape on the strake. Now I had to trim some glass and foam from the back, and redo the taper and sand the ribs a bit more to get a good fit. I had to reglass the trimmed foam at the back, so, while I was at it, I decided to squeege epoxy per plans. Then I realized that this epoxy is supposed to be added just before setting the top on the flox. I wasn't anywhere near ready to seal the tank, so I had to peel ply the whole top skin again.
I got the fuel vent pipe and researched where it should go. Carl Denk recommends a fancy system of two vent pipes for each tank. One per plans and one from the top at the inboard back of the tank. Both Mark Zeitlin and Brian DeFord used that approach. The pipes exit the tank at the rear inboard point, go over the spar, then up the forward face of the firewall. One pipe then does a U turn while the plans pipe carries over the top and back down the other side to exit under the cowl. Jeff Russel drills three 1/16 holes in the single plans vent pipe toward the back of the tank. The issue seems to be the need for a vent while grazing. I'd ordered 20 ft of pipe and that was just enough to do the plans vents up and over. I put the plans vents in, then drilled holes per Jeff. I think Carl's vents would need at least 35 feet of pipe. Most important issue seems to be getting the pipes high enough, then going back down to avoid fuel spill during climb and in case of inversion. I floxed and BID taped the pipes in place, then realized at knife trim time that one pipe was too high along the fuse side and would interfere with the skin. I cut the BID tape and lowered this pipe a little. My only problem now is that the pipes are supposed to go up the forward face of the firewall. I don't have the firewall in place yet, so the pipes stick up in the air.
I used the "top hat" (see archives) approach for the strake top skin. I put masking tape on the ribs sticky side up, ran my fingures along the bulkheads and ribs to make lines in the tape, then carefully (I thought) put the skin in place. The tape came with the skin when I lifted it up and I had my bulkhead lines. Now I transfered the lines to the foam, removed the masking tape and replaced it with strips of box sealing tape centered on the lines. I cut 1 inch strips of 1 ply BID and layed them over the lines. I put flox along all the bulkhead tops and Char & I lowered the skin into place once again. As I was about to walk away and leave it for cure I looked under the outboard end of the skin to confirm alignment of the tape against the rib. I have no idea how it happened, but somehow I'd missed. The sealing tape and BID was off to one side and the flox was against the skin glass. Good thing I checked! We lifted the skin off again. This time I had nice lines of flox to show me where the tape had to be. After "adjusting" the position of the tape, scraping the flox from the skin and putting it back on the bulkheads, we lowered the skin into place again.
I made the little wooden plugs for the ends of the pipes, per plans. Unfortunately one of my plugs was a little too little and it disappeared down the pipe. No problem. I'd made the pipe too long on purpose and I'll get the plug back when I cut the pipe to the 0.4 inch length. I floxed the pipe in place inside and out, then added 1 ply of BID over the gooseneck meeting itself in the middle. I didn't flox the fuel pipe in place along the fuse side yet. I thought it better to wait until the controls are in place to be sure it doesn't get in the way of anything. The last job before sealing the strakes was the fuel strainers. I floxed the 1 inch square al points in place and microed the corner to make sure the outlet is the lowest point. Now - whats a 1/8 NPT plug? I have no idea. Reading on I see that the 1/8 hole gets drilled and tapped to take the strainer valves. I have the strainer valves, but I didn't have a tap the right size so I just drilled the hole to 5/16 and moved on.
A kind reader wrote to tell me that NPT stand for something like National Pipe Thread, and is a standard in plumbing. He also mentioned that 1/8 NPT is really about 3/8 inches. Weird. I borrowed a 1/8 NPT tap, drilled the holes to the correct size and preceeded to tap the (already installed) 1 inch al squares. They both "popped" off the wings. Unphased, I finished the tapping in my vice, ruined one by tapping the thread too loose, made another al square, drilled and tapped it, test fitted the fuel drainers and floxed the squares back in place. This time I held them tightly in place for cure with hardware store bolts through the holes.
Just out of interest I closed off the fuel tap and tried blowing into the vent lines. The left (my) tank blew back. Not a real test, but definately a good sign. The right (Chars) side didn't. I shouldn't expect a good seal until the top skins are on, but it would be nice to know where the leak is. The method I came up with for finding the leak was a little different. I blew cigarette smoke into the vent line! After a few minutes I was able to see smoke coming out of a small gap at the aft corner. A close look showed a thin 2 inch gap along the edge. No problem. I'll flox this in when I glass the top.
One thing I learned from all this. I wish I'd cut out the baggage area's fully flush with the skins BEFORE BID taping the bulkheads to the skins inside the baggage areas. When I cut the fuse skin out of the way I found that my BID tapes were not as neat as I'd like them. I had to cut away some of the tape and redo it.
I've read recently of some chemical which goes purple when exposed to ammonia. Search the archives.
But who am I to give advise. I still have a leak!
Much later note: How about this for an idea? Fill the damn tank with gasoline at this stage. The leaks are easy to find when there's a brown stain all around them. I know. See fixing the leak below.
When Eddie came back to pick up his sniffer I mentioned that I still had leaks on one side, so Eddie set out to find them. His "sniffer" technique is much better than mine. He found three more leaks. Hopefully, when I've fixed these, the right tank will be fully sealed. While chatting with Eddie I found out that there is another level of sniffing used by A/C guys which would be perfect for this job. Apparantly they add some flurescent dye to the freon and use ultra-violet or "black" light to see the gas escaping. Eddie only charged me $35 for the freon I'd used. I promised Eddie a ride.
Rocket Rooter
I'd noticed a 3 inch plastic cap in the patio floor right under the nose of my Cozy, but never really thought about what it was for. I found out the hard way. Someone in the house flushed the toilet and, instead of going wherever it usually goes, the contents (all of it) came right out of the pipe. The plane is currently fixed in place since the wings are now glassed onto the strakes and I haven't cut them off again yet. No way I can move it. After three panic calls to "Rocket Rooter" the guy finally showed up. He'd never had to do his work under an airplane before. To add insult to injury, Mr. "Rooter" said the plane looked like an "overgrown go cart". He worked for 10 minutes, charged us $155 for an emergency call out and went on his mucky way. Mr. Rooter does NOT get a ride.
Leading edges and fairing blocks
Finally I got to the fun bit of carving urathane. I actually enjoy this, except for the grit that gets in your eyes when you sand urathane from below. After all the fuss (by me) about the strakes being too big I found that fairing the strakes to the wings was no problem. My strakes kink in a tad as they approach the wing, but who's to know it's not supposed to be like that. I added dry micro at the edges to get a smooth join, then glassed the leading edges and fairings. I carried the glass over onto the wings at the sides. After cure and primary finishing I cut the wings
back off. The cuts looked great, but I found that the fore / aft cut on one side was about 95 degrees. I had to redo the cut and build up the wing rib to meet the strake. I wish I'd glassed over the bottom spar gap. I had to put the wings on upside down to handle this later.
Compared to other parts of the plane the strakes took me a long time (almost 7 weeks) to complete. This was partly due to the fact that they took a lot of work and partly because I had to work on other things during that time.
I left the fuel pump for now. I'll do this when I know what engine and fuel system I'll be using.
Fuel System
I researched the various ways to modify the fuel system for Electronic Fuel Injection (EFI). The problem here is that the EFI pumps use fuel to cool themselves, and typically they have a bypass which returns excess fuel to the tanks. The problem is that you cant draw from one tank and return to another. (at least not without a lot of juggling and pilot intervention). Some talk of automating fuel transfer from one tank to another like the big boys do. Tracy Crook draws from one side only, and has a manually operated fuel transfer pump. I looked at Jeff Russell's AeroCanard system. He has 1/2 inch feeds going directly to a sump tank in the "hell hole" behind the rear seats. The EFI return line goes to this tank. As far as I know there is no easy way to isolate the fuel system for emergency landings using this method. Others recommend an (expensive $380) Andair fuel valve which switches the return as well as the feed, but then you have high pressure fuel lines in the cockpit. General consensus seems to be that there is no RIGHT way to handle all the issues here, but you'd damn well better end up with a system that works well, because fuel starvation and/or mismangement is probably THE biggest cause of GA accidents. Eventually I decided to go with a fuel sump or header tank behind the rear seats, fed from the plans fuel valve.
I removed the glass cover over the landing gear box and reglassed it in as low as I could get it. I drilled a hole in the forward top landing gear bulkhead, rear landing gear bulkhead and firewall for the heat duct and made a long 1 1/4 inch fiberglass tube by doing a layup on an old flourecent light. (nice smooth finish, but it was a bit.. to get the glass out). I cut off the heat duct transition piece, shortened it and glassed on my tube. The heat duct will now fit under the sump tank.
Sump tank
I made the sump tank to fit in the hole behind the rear seats, allowing 1/4 inch on all sides for mounting.
Once this was made I put it aside and got on with the rest of the plane. About a year later I thought about the fuel system again. I revised my plan and decided on a Tracy Crook type system but with full redundancy all the way to the fuel rail. The feed on the right side would be for take off and emergency use. Later still I thought about the andair fuel valve again. I'm going to have to make a decision eventually.
Fuel Caps
Once all the painting was finished and the strakes were beutifully polished I took out my hole cutter and, being careful not to go through the inner skin, I cut the holes for my Jack Wilhelmson fuel caps. Following the method I'd read about I cut away the foam in the hole to expose the inner skin, dug out a little more foam between the skins and floxed the edge to get rid of any foam debris. Now I cut the inner skin with a Stanley knife. I didnt drop so much as a crumb of glass or foam into the tank. Everything was going just fine until I hit the bulkhead. Yes, I'd cut the holes per plans as close as I could figure it. When you're working with a finished plane its hard to match the dimensions in the plans which refer to places that are now covered or extended. Anyway, the forward end of the holes ended up about level with the forward edge of the rear canopy window, and the front edge of the hole overlapped the forward bulkhead inside the strake by about 1/8 inch. It was the same on both sides, so at least I was accurate in my measurements. Now I was going to have to hacksaw and dremel to get things to fit.
At this point Char came out to look at the holes. "Y'know that new vacuum cleaner you just bought", I said tentatively. YES, she said emphatically. "uh...thinking quickly...do you think I could borrow the old one?" I said. She smiled and told me where the old one could be found. I got the old vacuum cleaner out and, holding the suction pipe right next to the hacksaw blade, I cut away the offending glass. When it came to the dremelling, I held the suction pipe right up against the dremel and all the dust got sucked out of the tank. I swear not a thing went in the tank.
Now it came time to fit the fuel caps I'd got from Jack Wilhelmson over a year before. I had a hell of a time getting the caps off. When I put the inner piece in the hole I decided I didnt like it. There is a piece of steel wire across the opening which the cap grabs as you turn the screw. I wont be able to get the filler nozzel into the tank, so I'll have to use a gas can or a funnel every time I fill. This may be a good thing, because I'll be mixing 2 stroke oil into the fuel, but it would also be a PITA. My other concern was removing the caps. They're a real tight fit with the airtight seal, and there's nothing to grab onto once the caps are fitted. The outer diameter of the flange - the bit that lays flush with the skin, is actually part of the cap. The part that's glassed into the skin has no horizontal surface to glass in place.
I considered making screwdriver slots front and back to make cap removal easier, but in the end I decided this would look ugly and what I really needed was different fuel caps. Besides, I'll probably be keeping the plane on the ramp, and I want locking caps to stop anyone stealing fuel, or worse - adding stuff to it. I called Buly, who'd just got his from AC Spruce, and decided to go and look at these before floxing Jack's caps in place and regretting it every time I fueled up for years to come. Don't get me wrong - Jacks caps are beutifully made and are probably just fine if you want to save a few bucks, but I knew I wasn't going to like them, so I eventually bit the bullet and ordered two FF36L caps from Wicks. When they arrived I realized that these were the fuel caps I should have bought in the first place. Made in UK, no less. Not cheap, but absolutely perfect for the task. My advice to other builders - don't even think about it - order these caps, and fit them just before you do the gloss paint, not after like I did. It really hurt to take a dremmel to the gloss paintwork and dig out down to the glass. I wonder if I'll ever get the new surface to match the rest of the strake. Probably not. Tomorrow I'll glass the flanges in place. In the meantime I left the holes covered to make sure no animals crawled in there and died during the night.
If you want to prepare for the FF36L caps while you're saving up for them, the flange is exactly 3.5 inches in diameter and the outside diameter of the neck (the size of the hole in the inner skin) is 2 3/8. You'll need a depression of 1/8 inch for the flange to sit down in, with room for one ply BID over the top overlapping the top of the outside skin for 3/4 inch all around. A little sticker on the plastic bag tells you to drill lots of large holes around the flange and glass it in with flox and one ply BID above and below the flange. Sounds simple enough. Oh, and you might want to make the holes about 1/2 inch aft of where the plans say to be sure of missing that baffel bulkhead. On the other hand, its no big deal if you hit it on the edge as I did.
A visit to the Pope
I saw Bulent talking about Swagelock fittings in a mail list and sent him a note suggesting that he come up and show me how it's done. His reply - "If the people wanna see the Pope, they go to Rome". OK. Fair enough. I took a Saturday afternoon and drove down to Bulent's hanger for a visit. The fuel fittings look good, but I think I'll stick to AN flare fittings now I've bought an expensive flaring tool.
Fuel Pipes
The first job was to remove the old (per plans) pipes. It felt like sacriledge dremelling away the BID tape and flox holding the pipes to the fuselage wall, but I got over it. I cut off the pipes about 1/2 inch from the fuselage and cut out a 2 inch circle of the outer skin around the pipe. Now I used my trusty little dremel sanding drum to remove the foam and de-micro the inner skin. I removed the flox from the remaining pipe by dremelling around it until it got quite hot. Then, with a pair of pliers I pulled on the pipe and out it came nice and clean. (Note - I've also found this tecnique quite useful when I've had to move an aluminum hard point. Just clean the al with a dremel sanding drum. It gets hot enough to melt the flox.) I was left with a 2 inch circle of inner skin with a very clean 3/8 hole in the middle. Next I cut out a little of the foam around the outside of the circle to give the flox something to bind on. I cut similar 2 inch circles at the top for the returns.
I bought four AN867-2 flanges, two 230106 in-line filters and various AN fittings from Wicks. I got the fuel pumps from Tracy Crook. The flanges have a tapered 1/4 NPT thread starting at the small (inside) side. I wanted a short 3/8 pipe to poke through the hole and make sure none of the flox went in the pipe. Nothing would screw into the outside end of the thread because it's the small end of a taper. After wandering around the shop for a while I discovered that those little aluminum things that slip over the pipe when you're fitting an AN female fitting... ok... alright... I'll look it up... I mean the MS20819-6D sleeve. A little "adjustment" with my dremel and this little guy fit tightly in the hole in the flange. A short length of 3/8 al pipe could now protrude from the flange into the tank. Having followed the plans, then retrofitted these aluminum flanges I can't imagine why plans call for floxing the pipe straight through the side of the tank, then bending it to run along the fuselage side. I don't plan to build again, but If I did, this one would be a no-brainer. I'd get four of these aluminum flanges and two of the fuel capacitor flanges and build them into the tank sides before sealing the strakes during construction. Actually, I'd probably go with the next size up for the feeds and add one of the $7 fuel finger strainers that Wicks sells. The flanges can be plugged for pressure testing and fitted with AN right angle fittings to avoid bending the aluminum pipe too much. The fuel system can then be plumbed per plans and, if necessary, easily retrofitted for EFI pumps etc. later.
Replumbing the fuel system
After occilating back and forth for almost three years on which way to do the fuel system I think I've settled on a plan. [later note: read on. this wasnt the final plan]. I'll plumb an electric pump and an in-line filter from the outlet of each tank direct to the rail with 3/8 al pipe and Aeroquip hose. This will give me two independant fuel sources. I'll run the returns to motorized valves, one on each side which will switch the return at the same time I switch the pumps. i.e. Left pump on will open the left valve. Right pump on will open the right valve. If a pump is off the valve will close. The valves will fail open. All I've got to do now is find the valves. I was looking around for suitable valves when another builder made a suggestion. The returns are very low pressure, max flow about 30 gal/hour. Why not T them and let the return fuel flow naturally to whichever tank has the lowest pressure. This OUGHT to be the tank from which we're drawing. Interesting idea, and very simple. I talked to Tracy Crook about it and he thought that the return direction would be unpredicatable based on minor variances in pipes and, perhaps vent pressure. I decided that the last thing I need is to run a pump dry while at idle (max return flow) during decent.
I considered a lot of alternatives before settling on the above plan. Many, including Greg Richter, Aerocad and Velocity, use a gravity drain into a header tank and return to the header. This is definately a proven approach, but I really didnt want a header in the "hell-hole". Also, I kinda fixated on having dual redundant fuel systems and the header introduces a single point of failure and something else that could leak. The header system has its advantages and disadvantages like anything else. Another alternative is the Andair duplex valve. This would mean having four long aluminum lines in the cockpit and it would force the pump to suck a long distance. The Andair valve sucks on the wallet pretty hard too. I could have used the plans fuel valve, with another weatherhead valve on the return. Same problem of long fuel lines. Then there's Tracy Crook's system of always returning to one tank and dealing with the fuel management. An addition to this idea is a facet pump to transfer fuel so you can access every last drop of the golden liquid. Again, you have fuel management issues I'd rather not deal with. I could add a facet pump transfer to my system if I wanted this benefit, but its not worth the comlication to me. I've always been in the habit of maintaining a good fuel reserve. Finally, there's the pump in the tank option. I considered this a bit late, and I couldn't really see a way to install the pump and make it easily accessible for maintenance. There are other variations on the themes, but I think I've covered the main options. If, reading this, you think I've missed something - please don't hesitate to tell me so. I'm not married to this plan, and I wont be offended - ego doesnt mix with aviation.
I continued occilating on the fuel system even as I got into installing it. After a lot of dicussion on the canard forum I'd settled on a variation of Tracy Crooks system of drawing and feeding from one tank and transferring fuel from the other with a facet pump. This effectively uses one tank as a big sump and the other as a reserve. When I came to install it I didnt like it. It just didnt feel right. I think this was because I was loosing my redundancy. The practical implementation was turning out to be complicated. I had no room for parallel fuel pumps on the firewall because all the space at the bottom was taken up by my radiator, oil coolers and fan. I read on the electric list about a guy who was using two infra red sensors and some electronics to transfer fuel automatically. Sounds cool and hi-tech, but the temptation there is to forget about fuel and let the system handle it. I prefer to keep fuel, like Georgia, always on my mind. (I'm not going to tell you why Georgia is always on my mind, except to say that the memory involves driving through it non-stop late one night). Anyway,,,one reason that there can never be a consensus on what is the "right" fuel system is that every pilot has his or her own personal preference. Some, like Jim Sower, are anal about having access to every last drop of fuel in the plane. His plane uses a sump. One tank failed to drain into it. Unfortunately this was the tank that had some fuel in it. Others want a "both" system, maybe because thats what they're used to or what their own experience has taught them is important. Jim, for example, had his "fuel starvation problem" over Canada, and had to GLIDE back to America. I guess a bad day like that will tend to influence you're choices.
My next iteration of the plan was a seperate pump and filter for each tank switching the return with the plans weatherhead valve. I agonized about a gascolator for a long time, but then Jim Sower made a good point. Water will always drain from the front of the strakes. If I do get any crud through the screen in the fuel blisters, or water in the feed lines, there is so little fuel in the system that I'll be burning new fuel from the strakes well before I get to the runup. Also, I'll have a "spare" fuel system and the chances of both becoming contaminated at the same time is remote. I didnt like the idea of running all that pipe (again) to switch the return, and having to do two things when it came time to switch tanks, but I hadnt been able to find a suitable or reasonably priced solenoid valve.
A visit to the hydraulics store
I was down at my local hydraulics supply store (Hydraulic Supply Co, 954-507-9651) getting some AN fittings (for my "off-road vehicle"). Jim at the counter seemed to know everything and produced all the parts I needed in seconds, including special metric to AN flare fittings for both my oil coolers. He also had the Aeroquip thread guage kit which Greg Richter had told me to get months ago. I bought one, and Jim showed me how to check threads on the fittings I was buying. Anyway, in an attempt to stump the guy, I asked if he had a 12v, normally open, three way, continuous duty, fail open, lightweight, small, durable but inexpensive solenoid valve. "Sure, $76 ok?", he says, and plonks one down on the counter. I bought it without a second thought. The valve comes in three parts. A standard solenoid, an aluminum block with input and output ports, and various cartridge plungers depending on what you want it to do. Jim also had An flare fittings for it. Mine takes the input and feeds it to output port A when it's off, and to output port B when its energized. Worst case is that I'll be returning to one side if it fails. Maybe I can rig a warning light for that situation. I think (and hope) that this is the final iteration of my fuel system. Jim mentioned that the pump isnt "explosion proof". Hmmmm. I guess this is ok. There's a lot of stuff on my plane that isnt explosion proof, including me. Then again, does "explosion proof" mean it'll keep working if theres an explosion near it, or it wont cause an explosion? Ah well. I guess you cant have everything. The valve looks and feels solid. The solenoid body is physically seperate from the fuel body. When I got it home I gave it power and was rewarded with a small "click". Sure enough the default output was now blocked and the alternate output was open. Works for me.
For anyone who's interested, the valve is made by Vickers. Its available from Aero Hardware Supply tel. 800-507-9651. Part numbers are VICK SV4 10 3 0 00, VICK 30545 and VICK 20197AA8. The AN flare adapters are part number AQP ADP 6M ORB 6M JIC. When assembled with the flare fittings it weighs in at 33 1/2 oz. Don't forget to tell them its for you're "off-road" vehicle. Chatting with the guys over the counter, one of them asked me what I was building. I said "Its an off-road vehicle..... Wayyyyyy off the road".
I weighted my weatherhead valve, fittings and piping. This was 21 1/2 oz, so I just gave up another 11oz to the gravity demon as well as $76 to the dollar demon.
Now the question became "where do I put the pumps and filters, pipes and valves". I could squeeze the pumps on the firewall by the oil coolers, but this would mean messy piping or drilling near the lower engine mounts. I could put them inside the NACA plenum near the radiator "per plans". They'd have good cooling airflow here, but the pumps and pipes would be a bit too exposed for my liking. Any "foreign objects" that get sucked into the NACA would have a good chance of damaging them. I could put them higher up on the firewall, but pumps are supposed to be low so they arent sucking a head of fuel. Hmmm. I decided to test my pump by making it suck a three foot head of air out of one gas can into another. It pumped fuel in a couple of seconds. OK, I guess that would work. On the firewall by the spar the head would only be a few inches. But then I'll need that space for coils, water collection tank and other stuff.
In the end I decided to put the pumps in the triangles under the back seats. This gave me 4 inch pipes from the outlets. I made brackets to secure the pumps to the forward LG bulkhead, installed nutplates in the gear well, and used silicone baffelling to insulate the pumps from vibration. I decided to seal off this area by making a removable cover for the triangular space with vent holes in the top, and a fuel drain into the gear well at the bottom. I glassed a 3 BID V shape about 2 inches high onto the floor just behind the seat hinges to attach the cover to. This way I'll smell any leaks, and heat will convect out, but any fuel will run out of the drain holes rather than down the fuselage floor.
I discovered something while fitting the AL pipes. I'd used a cheap flaring tool when doing the original system and the pipes were scratched at the ends. I was uncomfortable about this, so I bought the Rolo-Flair tool from Wicks. When I used this the pipe ends were perfect. But, after making a few flares I decided something wasnt right. I hadnt been tightening the tool fully on each flare and the flares were too narrow. I tried winding the screw up all the way until I hit a significant resistance, then giving it another 1/4 turn. Now the flares were just the right size to fit inside the nuts. I redid all the connections I'd flared wrongly.
Moving the filters
I eventually decided I didn't like the filters being in the back seat area, so I moved them to the gear well. This became possible because I ripped out all the ac and heater pipes I'd put in that area and used bulkhead fittings to send them through the firewall instead. Now, and I hope this is the final solution, the fuel pipes go out of the pumps directly into the gear well. The filters are clamped to the "roof" of the gear well about 2 inches above the main gear. The output from the filters comes into the hell hole to bulkhead fittings in the firewall. I have drain holes for any fuel leaks which might occur either in my sealed triangular area where the pumps are, in the gear well or in the hell hole.
Fuel Pumps
I'd ordered one pump from Tracy way back, planning to get the second one when I was ready for final installation. I had two sets of special fittings made to fit the pump. When the second one arrived it was a different type (A Walbro 393). It has an all metal casing instead of a part plastic housing. Damn. The new pump is better, but the custom [read expensive] fittings arn't fitting. Maybe my hydraulics guy has AN-6 fittings that'll fitit. Maybe I should bite the bullet and get another of the all metal ones.....
Shut off valves
Since my fuel system plumbing doesnt come forward to the seatback, there's no easy way to shut the fuel off in an emergency. I accepted this limitation knowing that the pumps have check valves and wont pass fuel when they're powered off. I still had a niggling feeling that I'd like a way to shut off the fuel in flight in case of fire or forced landing.
I'm sure an experienced airplane mechanic would just look at the problem and just know the solution. Unfortunately novices like me have to reinvent the wheel every time. I'd added mechanical valves to the outlets for maintenance so I spent some time trying to devise a way to operate these from the cockpit. I drilled a small hole in the valve control arm and oriented the valve so it closed by moving the arm forward. Perhaps I could attach a pull cable from the armrest pocket. Once the valves were fixed in place I drilled a 1/4 hole in the back of the armrest pocket about an inch from the top. I continued the hole through the seatback and the rear armrest bulkead until I could push a wire through to the valve arm. The next step in my evolution process was to think that I'd really like to be able to change my mind about shutting off the fuel. With a wire I get to switch off, but not back on. OK, I need to be able to push as well as pull. What can I use that wont bend when I push? A wire coat-hanger (no laughing please) I'd been using to pull cables caught my eye. I pushed it through the hole and decided it would both pull and push the valve arm. But.... do I want a wire coathanger activated control in my plane? Not really. What do the professionals use to push and pull an control arm? A push-pull cable! I just happen to have one of these on my shelf. I think it's the only thing I every bought from Spruce. They sent me the wrong thing, but it wasnt worth the trouble to send back. I dug it out, cut if off to length and fitted it. Perfect. All I need now is the little thingys to hold the outer cable in place and to lock the inner cable to the arm. I'll also need another one for the other side. Its scarey how these things have to evolve, but I usually get there in the end.
But the "end" is often not really the end.....
Next day I needed to put the rear armrests in to take pictures for my upholstery guy. You guessed it. The armrests interfered with the shut-off valve. This is in the very back. By sitting in the rear seat I could I tell that this area wasnt really needed by the passenger. It's well behind where the right hip would be. I cut a circle out of the armrest to allow the fuel pipe and shut-off valve some space. Now the problem is that a bored kid (or a curious adult with a death wish) could get at the shut-off valve. Damn. I made it accessible. Now I have to hide it again. I need some kind of door. I fixed this with a piece of flat cured 2 BID. I bent the glass into a curved shape lapping around from the gear bulkhead to the armrest, then warmed it up with a heat gun. When it cooled I had a nicely shaped removable cover for the shut-off valves.
A few months later, down at the hangar, one of the many expert passers by pointed out that my push-pull shut-off cables would pull just fine, but they were at too much of an angle to the valve to push. They worked, but I could see the cables bending as they opened the valves. So, in an emergency I'd be able to shut off the fuel fine. I just wouldn't be sure of turning it back on again. To solve this I glassed 2 inch blocks of wood to the fuse side and mounted the cables on these to give them a straight line to the valves.
Another retrofit job was the fuel pipes and filters. I'd installed the filters behind the back seats, but decided that I'd rather have them in the gear well. You can see pictures of how the pumps and filters end up in Chapter 23
Fuel Guages
After searching around and considering the homemade capacitance senders I ordered a set of 24 inch bendable probes from Jim Ratte at Recreational Mobility. These probes can be cut down to 12.5 inches as needed. The probes come with aluminum flanges which can be floxed to the inside wall of the tank either during construction or later. Note that the flanges (at least on my set) are not assymetric - i.e the holes are not equally spaced around the perimeter, so you have to be careful to get them oriented correctly if you want the small air outlet at the top of the probe to be in the up position. Since I'm mounting the probes sideways this air outlet will be about 1.5 inches below the top surface of the tank. Fitted as they come, they'll read full for a while before dropping down. I think I can get around this by bending the probe up, then down to point diagonally down into the tank. If I install a small 1.5 inch tube in the air hole taking the outlet to the top of the tank I should get a full scalle reading. At least - that's my theory. (Note- theory failed in practice. The tubes didnt look like they'd bend this much so I didnt do it.) When it came to fitting these I had a dilema. I think the aluminum flange is designed to go inside of the tank. Unfortunately I don't have that option. The screws protrude a couple of threads past the flange, so they'll be pushing the flange off the tank wall if it's on the outside of the tank. I was planning to strip off the outer skin and mount the flanges against the inner skin. I decided that this wouldnt work unless I shortened the screws. The least invasive method was to cut away a 3/4 circle of the outer skin, clean up the rest of the outer skin in an area big enough for the flange plus an inch or so, then dremel out the foam in the center hole and clean up the inner skin. Once I had the inner skin clean I cut away 1/4 inch of foam from between the skins and then cut out the inner skin. I filled the gap between the skins with flox and pushed a socket wrench wrapped in duct tape into the hole. Now I floxed the outer skin where the flange would sit, filled the screw holes with a dab of grease, floxed the flange, slid it over the socket and pressed it into place. Five inch long beads of flox, with a dab of grease on the end of each, flowed through the screw holes. Duh! Its this kind of event that makes building so much fun. I think its also why I keep getting epoxy on my face. I keep slapping my forehead. Hmmm. How do I get out of this one? I wiped off the flox, dipped the ends of the screws in grease, then screwed them into the holes until I felt them hit the glass, then backed them off a bit. Next morning I was able to remove the screws by hand. Phew! As for the 1/16 of an inch of screw that protrudes through the flange - I'm guessing that the flox pad is about that depth, so I'll use a drill to remove some flox and give the screws some room. Worst case, I'll have to drill through the outer skin which wont hurt anything. I put 2 BID tape around the edge of the flange to hold it in place.
Once the flange was cured I inserted the full 24inch probe and, with a flashlight, looked to see how far it reached. It hit the inboard fore-aft bulkhead after about 20 inches. So, if I bend the probe a little to the left (left tank) and downwards, I should be able to use the entire probe diagonally in the tank. I took the probe out, carefully bent it down and put it back in. It now reached all the way in without interfering with anything, but it wasnt touching the bottom of the tank. I took it out, bent it some more, and broke the aluminum probe tube out of the plastic housing. Damn. I sent an email to Jim Ratte asking if he'd had any other idiots do this, and asked if the probes could be returned for repair. He replied that no, no other "customers" had managed to break the probes. His supplier is a warehouse that imports them for overseas, but he offered to take a look at the damage. I decided to repair the probe. If it works, fine. If it doesnt I'll buy a new one. At least they're not $100+ to replace like the ones from Wicks. So... a word of advice to anyone installing capacitance guages - The al tube is quite weak where it connects to housing because it's drilled for the air hole and the terminal connection. Support this area when bending, otherwise you're in big danger of joining the club I just initiated.
How to fix a broken fuel capacitance probe
Here is my repaired probe. I measured the resistance as the probe dipped into fuel with an ohm meter and it tests exactly the same as the other one. I think it's stronger than the original. In case some other "idiot" does the same, this is how I fixed it....
1. Drill out the rivet holding the tube and terminal to what's left of the plastic.
2. Slide the outer aluminum tube off inner terminal wire.
3. Cut off 1/2 inch of aluminum tube and redrill a small hole through both sides
4. Tap the hole for a small screw to hold the terminal
5. Slide the al tube back over the inner terminal
6. Attach the terminal with small screw
7. Wrap a small drill bit or similar with duct tape
8. Insert this into the air hole opposite the terminal
9. Pot the whole lot in wet flox
10. Remove drill bit to reveal the air outlet hole
Grounding
After reading some more about grounding issues, I decided I wanted to ground my tank and filler to the common electrical ground. This way a line-boy can hook his clip to the exhaust and grounding will occur as normal. I couldnt find copper plate anywhere, so I got a piece of 3/4 copper pipe, sawed it lengthwise, hammered it flat and cut triangle out of it. I linked thie triangle to the filler neck and the cap with copper wire. OK, so how do I get a ground wire out of the tank? I've found that I've become fairly good at figuring things like this out - unfortunately I often figure them out in the wrong order. So it was today. After looking around for a good place to drill the tank (again) it occured to me that the fuel capacitance probe has a ground. I tested the probe and found that grounding the terminal at the top of the tube didnt affect the performance of the probe in any way. Great. All I have to do now is run a copper wire from the filler ring, through the fuel to the terminal on the probe (which will be inside the tank) and I don't have to cut another hole for the ground wire. So, what's the problem??? Last night I potted the terminal in flox On one of my probes. A few minutes careful dremel work and I uncovered the thick wire running from the terminal into the probe body and soldered my copper wire to it. Using a coat hanger wire I threaded the copper wire under one of the baffle holes and picked it up from the filler cap. I'd just connected the wire to the copper triangle when I realized that the seal for the probe was still on the bench. Duh! I undid everything, put the seal on and rethreaded the wire. I do wish I could start figuring things out in the correct order!
Speaking of doing things in the correct order, test you're gauge and probes before bending or modifying the probes. When my guage arrived I tested it and found that I could only get 1/4 full reading no matter what I did. I removed my grounding wires. No difference. After a bunch of experiments I called Jim Ratte and explained the problem. He contacted the manufacturers with the symptoms and they said that it sounded like the gauge. I hope so, because I can't see anyone taking these probes back after I've bent and butchered them. Jim didnt have another dual gauge in stock, so he sent a new single gauge (at his own expense) for me to test. We figured that, if this worked, I could finalize the probe fittings and wiring while he got a replacement dual gauge for me. Fedex arrived with the single gauge and it worked perfectly with both probes, and with my extra ground wire attached. Excellent! I returned both gauges. Based on this experience I can highly recommend Jim Ratte at Recreational Mobility. He's knowledgeable, concerned and responsive.
Sacriledge
A general note about this whole retrofit job. When we build the strakes we accumulate a kind of reverence for them. They have to be sealed perfectly and must never be breached. Here I am, 2 years later digging a total of 6 holes in them. The job has been in my mind for two years... whats the best way to do it? Am I going to screw up the tanks? How hard will it be? Guess what.... it was really no big deal. It's the sort of job I'd do on the ramp on a Saturday afternoon. What I learned from all this is that making a few changes down the road is pretty simple - much simpler than agonizing about doing it the right way first time around. So... if you're in doubt about the best way to proceed during construction, don't sweat it. Just build the sucker per plans and move on. If you decide to modify something down the road it won't be as hard as you think. On the other side of the coin, another thought I've probably mentioned before.... once you veer off the straight and narrow road the plans provide, be prepared for things to get complicated quickly. For example, the plans fuel system is a piece of cake. You order the parts specified, fit them where shown and move to the next chapter. For someone who isnt trained in all this and, like me, is building a plane for the first time, designing you're own, even a simple fuel system, is a whole new ball game. You study the pictures in a dozen catalogs and try to figure out what parts you need. You order them and find that they don't fit quite right, so you return things and order something else. You theorize and discuss it with others and hope that it'll work like you planned. Of course, no-one has done it just the same way, so you wont be sure. Maybe you didnt think of something. Maybe you're solution will be better. Who knows. You sure don't. Of course you learn a lot along the way and that's half the fun, but if you want a simple life and just want a Cozy that'll fly as advertised... stick to the plans. Changing things around a little is easy when you reach the finishing stage.
Time to fill the tanks
Once the fuel system was plumbed and the pumps were wired up I decided it was time to fill the tanks with gasoline and find any leaks before I do the final upholstery inside the cabin. Avid readers may recall that I wasnt at all sure about the right (Char's side) tank, so I filled this one first. There's a 24hr gas station at the end of my block. I considered pushing the plane there at 3am, rolling up to the pumps and saying "fill'er up", just to see the look on the gas attendant's face. This would also have made a great picture. Common sense got the better of me this time. I filled a 5 gallon gas container and poured it into the right filler. I'd have thought that 5 gallons would be enough to show on the sight gauge, but it didnt. Hmmm. I plumbed the fuel feed direct to the return solenoid and switched on the right pump. Gas happily gurgled through the system and back into the tank. Cool. My plan was to fill the tank to the brim to see how much it took. I also wanted to let the plane sit for a few days to ensure that there were no leaks. I put another 5 gallons in the tank. Still no reading on the sight gauge. Damn. The little hole at the bottom must be blocked. I guess I'll be doing the retrofit of the little floating balls and lights. I let the plane sit overnight. That evening in the hot tub I thought I could smell gasonine, but there were no signs of leaks anywhere.
Testing the system
I didnt have my intake yet, so I plumbed the feeds to a T and plumbed the T directly to the return input. I wired the pumps from the battery bus on their respective sides and ran the return solenoid from the "endurance" bus. I had three switches in a row on the panel. I may wire this differently later, powering the solenoid from the left tank switch, but this would remove the ability to transfer fuel from left to right. With the system wired this way I was able to move the fuel around at will. With the right pump on and the solenoid off the fuel flows out of the right tank and back in again through the return. I ran this for a while to check for leaks in the system and filter out any crud. Next I energized the solenoid with the right pump running and sent all the fuel from the right tank to the left tank. Leaving the solenoid energized I turned the right pump off and the left pump on to swish the fuel through the left tank system for a bit. Finally I turned off the solenoid and let all the fuel go back to the right tank. It took about 15 minutes to transfer 10 gallons. I checked all the connections for leaks and found none. I also noticed that the pumps were stone cold after (and while) transferring 10 gallons of fuel from side to side. Cool! I measured the amp draw of the pumps at about 1.3 amps. The solenoid takes about 2 amps. I left the caps on while the system was pumping to check the vents. There were no signs of any vacuum build up when I removed the caps at the end of each run. It looks like everything is working nicely. One nice advantage I hadnt really thought through is that I can get at every last drop of fuel in the airplane. Even if the solenoid fails I can run the left pump only until all fuel has been returned to the right tank. All I need now is LED warnings for crossfeed, and tanks that dont leak.
Time to fix the tanks
Next morning the gasoline smell was VERY strong, almost overpowering. New rule. No smoking on the patio. It didnt take me long to find the problem. The right baggage area had a brown gasoline stain centered on the lower front corner, where the baggage area joins the tank, right by the drain tap. I must have missed a spot right in the acute angle at the front. The gas stain extended through the baggage area floor and into the seatback for a couple of inches. OK. At least I can get at the problem. It isnt going to be fun, but its doable. Step one is to get all the fuel out of that strake before it spreads any further. How do I do that quickly? I had finished all the plumbing and connected the right feed to the return, but hadn't wired the return solenoid yet. With the solenoid energized, return will go to the left tank. I grabbed my wire stripper, crimper, some wire and some crimp fittings and had the solenoid wired in about three minutes. The wires didnt go through the electical conduits or go via a switch or anything fancy like that - they just went around the outside of the fuselage to the firewall. One wire to the battery bus, and one to the ground bus. Click. With the solenoid energized I switched on the right pump and watched as 10 gallons of fuel went from the right tank to the left tank in about 15 minutes flat.
I set to with a hacksaw blade and a very carefully used dremel. I cut away an area of inner skin about 3 inches by 2 inches right around the joint between the strake bulkhead, the baggage area floor and the strake leading edge. Once the foam was exposed the culprit was obvious. There was a sheetrock screw sized hole right through the foam and, sure enough, once I removed the foam there was a matching hole through the inner skin. Damn Hicks and his sheetrock screws. I blame him.
Pleased that I found the hole so easily, I cleaned up the glass, painted it with epoxy, dabbed flox into the hole and layed up 2 ply BID over the hole. Once this has cured for a few days I'll pump the fuel back over to the right tank and see if the leak has gone. It took me all of a couple of hours to discover the mess, remove the skin and fix the hole. Last year I spent four days trying to sniff it out. Perhaps I've discovered the best leak detector of all. Gasoline.
Next day my gasoline detector system worked again. I had a small stain in the baggage area below and to the right of the repair. I cut away some more of the inside skin on the baggage area floor and R33 until I'd exposed the wet area. This time there was no obvious indication of where the gas was coming from. I was looking at the inboard side of the flox holding the drain valve in place. The gas was either coming around this from the front side, or seaping through the join between the vertical and horizontal skins. I'm beginning to wonder if I'm going to have to remove the outer skin around the drain hole. I cleaned up the whole area, forced flox into the corner as far as I could get it, covered it all with three ply BID and hoped for the best. Three days later the gasoline smell on the patio was gone and the right tank was holding 20 gallons of gas with no leaks. I added another 5 gallons. It seemed that there was still room for a lot more gas. Strange.
The next issue was what to do about the non working sight gauge. I called Vance Atkinson and had a good discussion with his wife. I've been admiring the new lighted sight guages with little colored balls inside, and Vance's web site says the retrofit is fairly easy. The upgrade kit is $45, and have to "pop off" the front of the gauge to do it. I can redrill the hole while I do the upgrade. Excellent. I ordered the upgrade.
Houston. We have a problem
I spent the next week or working on wiring and air conditioning giving the right tank a chance to sit with 25 gallons of gasoline. While working around the plane I noticed a "suspition" of a gas smell, but couldnt track it down. I thought perhaps it was coming from the vent, so I installed my vent covers. After a few days of working on other stuff and occassionally sniffing a gas smell, I decided it was somewhere near the spar. I could smell it from inside the spar at the back, and from the outboard wing bolt hole. Looking with a torch (flashlight) inside the spar revealed no problems. There was still plenty of dust in there and it was all dry.
After a week or so, I decided to top off the tank all the way to see if gas would seep out of the top seam somewhere. I added another 5 gallons of gas and still had room to spare. Wow. That's 30 gallons. The tanks are only supposed to hold 26. I remember shifting the bulkheads a little to try and expand the capacity, but 30+ gallons seemed like a helluva lot. I added another 2 gallons before the tank was full to the brim. (Still reading almost zero on the sight gauge.) Pleased that I'd somehow increased my fuel capacity by over 20% (can you see where this is going?) I carried on with some wire rerouting (see chapter 22). Walking around the plane that evening I noticed the gas smell again. It wasnt much, but it was definately there. I searched all around, inside and out with a flashlight and eventually spotted a small brown stain on the outboard right strake leading edge. (Char's side) On close examination I could see that this was a gas stain eminating from the corner of the layup at the end of the strake - the one that seals in the eurathane foam used to fill the corner. I cut out a one inch square of glass with the dremel and, sure enough, the eurathane behind it was a little damp with gas at the bottom. OK. I cut out the entire layup and started removing the foam. As I dug, the foam got wetter and wetter until I had to get a cloth to mop up the gas dripping out of the hole. This is not looking good. I decided now would be a good time to pump all the gas to the left tank.
Pumping a full tank from one side to the other takes about 50 minutes. My curiosity couldnt wait that long. I dug out some more foam using a kitchen knife until I reached bulkhead OD. I know it's OD because it has OD written on it. As I scraped the foam away from OD my knife suddenly disappeared through the bulkhead and now fuel started to pour through the hole. It felt as though I'd just put my knife through a big hole into the body of the tank. Uttering the full name of a well know deity, I pushed the foam back in and stuffed the hole with a couple of rags. While rushing about wringing out the rags into a container and replacing them, I tried to figure out what could have happened. Had I left a big hole in the tank by mistake? How could I be that stupid. Nah! Couldnt be.
Eventually the pump drained the tank, the deluge slowed and eventually stopped. I removed the rest of the brown foam and discovered a large (I mean large - 4inches * 3inches) oval hole in the side of OD. Looking through the hole I could see fuel sloshing around in what I thought was the tank. But the tank is dry. I just drained it......then I realized that OD doesnt form the side of the tank. I was looking at the cavity you create to keep the fuel forward and in the correct C of G area. Only my cavity was full of fuel. This must be where my additional capacity had come from. If I'd waited I could probably have got another ten gallons in as the spar filled up too. SHeeesh! I bet this area holds at least a gallon. By now it was down to about a quarter full. I mopped out the last of the gas and looked around with my flashlight. No obvious signs of a problem with the tank seam, but there must be a serious leak somewhere in there. I wonder if I'll be able to fix this without destroying the nicely finished, painted and polished outer strake skin. I left it to dry off for a few hours while I pondered the situation. Char is sleeping. I wont tell her yet. Maybe I can get it fixed before she wakes up.
After pondering for about 30 minutes I went back out to the patio and set out to find the leak. I dremelled away what was left of the OD bulkhead where it met the lower skin, but couldnt find any obvious holes. Time to use my tried and tested gas detector system again. I pumped 10 gallons from the left tank into the right tank while watching for any appearance of wetness. After an hour a suspect wet area developed about 4 inches aft of the strake leading edge right where OD meets the the strake floor and the leading edge bulkhead. I dremelled the leak. It was like following a river upstream. As I dremelled the wet area became dry, the wet when I stopped. I was able to follow the reappearing wet area to the tank. Just as before in the baggage area, the leak seemed to come from the foam under the tank. I could see a line of micro just where the leak appeared under the floor of the tank, and realized that this was one of the cuts we make in the strake skin to bend it. I decided that my original guess had been right. The fuel was coming from the outboard side of the fuel strainer and tracking along the microed cut all the way to the end of the strake. OK. Time to bite the bullet. I made a 2 inch cut in the strake about 3 inches back from the leading edge about an inch and a half outboard of the fuel strainer. Fuel immediately began dripping out of the cut. Bingo! I pumped all the fuel back into the left tank again, and cut out a three inch square of the outer skin around the fuel strainer. I removed the fuel strainer (I wish I'd done this during finishing) and dremelled away the foam until I came to the inner skin where it met the aluminum square. The foam went all the way up to the aluminum square. It looks like I hadn't done a very good job of floxing the aluminum square in place between the skins as clearly shown in the plans. Ah well. Everything can be fixed. No one will ever know.
I cleaned up the inner skin and floxed a 3 BID pad over and around the aluminum square. Once this cured I dremelled the glass off the aluminum, replaced the fuel strainer and filled the tank yet again. After an hour or so the edge of my repair was wet with fuel. Damn. I cut away another chunk of finished, painted and polished outer skin to chase the leak. "chasing" is pretty much what I did. As the very lightly used dremel removed foam, the foam would go dry, then the gas would begin to wet it again. It was easy to see the origin point because that's where the wetness was starting from. Once I'd dremelled away all the foam I could see that there were two tiny nicks in the inner skin which were the source of my leak. Judging by their location its a pretty fair assumption that I caused them myself while cleaning up the inner skin for the previous repair. I pumped the fuel into the left tank again. Once the tank had dried out completely and no more wetness was showing I roughed up the area and did another flox / BID repair.
The story goes on. I filled the tank again, and soon saw that there was another leak somewhere. The edge of the foam near my repair got wet after about an hour. This time I had no idea where it was coming from. Somewhere along the lower front seam of the strake. I decided to use a technique called "binary chop" which is often used in computer program searches. The idea is that you go to the half way point and decide if the item you're looking for is above or below this. Then you go to the half way point of the remainder. You continue with this until you find the culprit. Theoretically you have to find the item in 10 chops or less. I found my leak in three chops. i.e. four holes. Once again I floxed and glassed the inner skin. After cure, the gas stayed put for over a week with no smell of gas, so I got on with some other stuff. (but see below).
Perspective
To a new or prospective builder the above might sound like an absolute nightmare. Let's get this in perspective. I probably spent a total of four hours finding and fixing these leaks. True, I now have to replace the OD bulkhead and repair the strake skin, but this wont take me long. Compared with the other work I've been doing lately - wiring, planning the panel, engine configuration etc., it's almost a relief to revert to something I already know how to do. While I'd obviously rather not have to do all this, compared to the overall building effort and experience, this fuel leak problem was no big deal.
A sympathetic builder sent me an email wondering if ANYONE had EVER managed to build an EZ fuel system that was totally leak-free first time. I answered yes. Me! Unfortunately I only built one. I need two.
Later note: The above perspective is getting a little stretched. See below.
Vance Atkinson's Balls
Vance Atkinson's sight gauge upgrade kit arrived. It included two tiny orange balls and two very small but very powerful prewired LEDs. At first glance you might think that Vance has a lot of balls to charge $45 for this upgrade. In a way you'd be right. The balls are special order, made in Spain, and the minimum lot size is around 1500 units. Vance tells me the balls are made out of cork, then painted with special gasoline resistant paint. A lot of research and effort went into chosing the materials and suppliers. I figure the balls and lights are well worth the money.
I started in on the right tank since it was already empty. The trick now was to "pop off" the clear cover to install the orange ball and, in my case, clear out any flox from the two holes so that the fuel could flow in and out of the gauge. I carefully dremelled at the edge until I could see the join between the clear cover and the white base. I'd done a really good job of installing the sight gauges during Chaper 9. I tried prizing the cover off with a sharp blade per Vance's instructions. The covers didnt want to "pop" anywhere. I resorted to the heat gun and discovered that the melting point of flox (made with MGS) is a little higher than the melting point of the plastic used to make the clear cover. Right. You guessed it. I got the cover off, but totally ruined it in the process. I called Vance and ordered new glass covers. While I was waiting for the new covers I
played with the LEDs in my panel.
When the new glasses arrived I trimmed one to fit the white backplate and floxed it in place. Once this cured I installed the LED and covered the edges with a little more flox and 2 ply BID. The instructions for installing the LED say to put it 3/8 away from the edge of the bubble. I don't have 3/8 of white plastic below the bubble, so I made do with about 3/16 or 1/4 clearance. I wired the LED's in with my panel light dimmer (from Bob Nuckolls) and they dim down nicely.
Another damned leak
Once the new sight glass was installed in the right tank I pumped all the gas back into it so I could work on the left one. I climbed into the back with the dremel and stopped. Sniff. What's that smell? Gasoline? I searched around and found a small brown stain at the inboard rear edge of the baggage area. Since I had the dremel in hand I began to dig. I removed a small chunk of inner skin. The foam was dry, and there was no sign of gasoline - and the tank was still full to the brim. Hmmm. Maybe it just tracked over from one of the other leaks. Next day everything was still dry. Cool. It wasnt a leak after all.... or was it?
I fitted the other sight glass, with its orange ball and light, then got on with the repair to the front of the strake. I made a new OD bulkhead and floxed and taped it in place. A new supply of pour foam arrived. I taped over the oval hole in the new OD with duct tape to stop the pour foam going into the cavity. Now - how do I fill the hole with pour foam without leaving any cavities? If I lower the nose all the way it will be downhill going into the hole and I can just pout it in. I lowered the nose, mixed my pour foam and poured it in the hole. Of course it didnt run down the way I expected and I had a bit of a hole at the back. I mixed some more pour foam and used syringe to quickly squirt it down the back of the hole. I got about 7 squirts in before the foam expanded inside the syringe, but it was enough. I watched the foam as it expanded and filled up the hole. Cool. But what's that smell?
Gasoline. I looked around and couldnt see any spillage until I climbed in the back. The chunk of inner skin I'd removed at the inboard rear of the baggage area was leaking badly. Liquid gasoline had filled the hole in the outer skin and was running out of the foam. It didnt take me more than a minute or two (say, one cup of gas) to figure out what had happened. I'd lowered the nose. The fuel level must have been just below where ever it was now leaking from. When I lowered the nose the fuel tipped forward and, presumably, over the top edge of the tank. This is starting to get tedious.
I raised the nose and the leak stopped. I dried the area off then gradually lowered the nose again while watching the baggage area. When the nose reached about 12 inches from the ground fuel immediately started running out of the foam. I raised the nose, and it stopped.
Filter Check
Before pumping the fuel again I checked both fuel filters which had now passed about 120 gallons each. There was very minor accumulation of crud / dust on one - maybe 5% of the filter area. The other was totally clean, so my tanks have now been effectively flushed of contaminates.
The last leak
Once again I pumped 30 gallons of fuel to the other side and began chasing the leak with the dremel by removing a strip of inner skin. I ended up at the top inboard side of the capacitance sender. The fuel was coming out of the corner between the backplate and the top seam of the tank. When I installed these backplates I'd stuck them on the outer skin and filled the gap between the skins around the hole with flox. In retrospect this wasnt too bright. I removed the sender and, sure enough, there was a small indentation in the flox surrounding the hole. It didnt look as though there was a hole, but it Occam would surely tell me that this was the problem. I tried to remove the backplate for the sender so I reinstall it with a glass to glass join, but it was very firmly stuck in place and I think I'd have destroyed the inner skin if I removed it. The safest way seemed to be to seal off the area. I dug a "moat" all the way around the sender backplate and did a flox / BID repair on the inner skin. I also added more flox to the indentation inside the hole. If fuel leaks in this area now It has nowhere to go beyond the tiny bit of foam behind the backplate. It'll meet a barrier of flox and BID at the edge of the backplate.
Once the repairs cured I refilled the tank. There was no sign of any gas. I lowered the nose. Immediately the familiar smell of gas filled the air. Dejected raised the nose and hunted around for the leak. The baggage area was dry. No problem there. Nothing at any of the previous leaks points. Then I spotted it - the fuel cap. I'd attached my ground wire to the nut on the underside of the cap, and hadn't retightened it enough to get a proper seal when the cap lever was snapped over. A few more turns on the nut and the seal did its job perfectly. I lowered the nose again. This time there was no sign of gas. I left it overnight with the tank full and the nose down. Next morning - still nothing. After two more days there was not a sign of gas anywhere. No leaks. No smell. Nothing.By George. I think he's got it!". Finally my battle with the strake was over. At last I could finish the repairs inside the baggage area, install my upholstery and move on. I was happy. The happy feeling didnt last long.
EFI Fuel Pumps Suck
I struggled to come up with a title for this entry. "Total Fuc.ing Disaster" came to mind, but the above was the most appropriate and informational. The good news is that I now know that my right tank is TOTALLY sealed. Here's the bad news:
On December 21, 2002 I was repairing the area inside the right strake baggage area. I'd microed foam in place and sanded it down flush with the skin. I needed to pump the fuel out of the tank before I could remove the capacitance sender and finish the repair in that area. I dont like to run the pumps unattended (as it turns out, this is a really good policy), and it takes about 40 minutes to completely empty a tank. I was on "puppy watch" making sure that our seven-week old Golden Retriever puppies didn't get into too much trouble while romping in the backyard. This would be an ideal time to run the fuel transfer. I flipped on the right pump and the return solenoid so the fuel would pump out of the right tank and into the left one, and then I continued playing with the puppies on the other side of the pool. After a couple of minutes I heard a sharp "bang" from over by the plane. At first I thought one of the puppies must have got under the cowling which was on the concrete and banged it on the patio floor, but no. All seven puppies were at my feet. One of the cats, perhaps. I went over to the plane to investigate and heard another sharp cracking sound. This time I was sure it wasnt a puppy or a cat. My plane was making bad noises all by itself. I quickly turned off the fuel pump and the master switch. What the hell was going on?
I looked around inside. Nothing. Underneath. Nothing. I walked around the outside of the plane to the right side, and then I saw it. My right strake had imploded.
There was a line about three feet long, parallel to the leading edge where the upper strake skin simply bent at a 45 angle for 2 inches, then bent back again. The forward part of the tank was lower than it should be by about 2 inches tapering to nothing at the leading edge. A small crack in the micro along the leading edge showed where the skin went back to the proper shape. A crack just forward of the filler cap showed where the damage started. Images of remakng the entire strake ran through my head as I went inside to get Char and show her what I'd done. I also got the camera to take this shot. As we examined the damage I thought about what must have happened. I've definately fixed all the leaks. The strake is now completey air tight. Unfortunately this must also include the vent line. I hadn't hit the problem before because the filler cap had been leaking air. Once I fixed that, air had no way to enter the tank. As I removed fuel the tank just collapsed in on itself.
I tried to undo the cap to release the pressure but it wouldnt budge. OK, so I'll pump fuel back over to equalize the pressure. I flipped on the left pump and watched as the strake magically returned to its proper shape like a rubber bladder. Now I was able to remove the cap. I put the cap back on and pumped a bit more fuel. The tank pressurized, the cracks disappeared, and the strake looked as though nothing had happened.
But, of course, WE know that's not the end of the story. Obviously the composite sandwich had been compromised and a repair was going to be needed. I figured that the top skin would have to come off a few inches past the area that had been damaged. I'd fix any voids with micro and re-skin the top of the strake. As you can imagine, I was pretty depressed about the whole thing. I got a couple of tylanol and called Nat.
I described what had happened and Nat said "that's nothing!" He told me about repairs he'd done on his Cozy's which were much worse than this. All I had to do was drill 3/32 holes all along the areas where the skin had bent, inject pure epoxy with a syringe, then cover the repair with a ply of fiberglass. There would be no need to take the top skin off. Nat cheered me up a lot. I emailed my picture to him and he sent the following note:
Thanks for the picture. It doesn't look as bad as I visualized from your
description, and easier to repair than I visualized. It you have the
collapsed section raised again to the original contour, just drill a bunch
of holes along the 2 bend lines where the foam is most certainly collapsed.
Add enough epoxy to fill the void, and ad one more layer of fiberglass,
probably extending 2 inches past the collapsed area. Use UNI with fibers
perpendicular to the fault lines, and you will be home free. Happy holidays!
Best regards,
Nat
No time like the present
It was 8pm. I decided I'd sleep better if the repair was started immediately. I put a small drill bit in the dremel and fitted the depth guide to be sure I didnt go through the inner skin. It was hard to tell where the "fault lines" were, even when I took a copy of the collapsed picture outside for comparison. I decided to collapse the tank a bit to see where it bent. If I removed some fuel it would bend at the weak point before breaking somewhere else. Right? Right! 30 seconds of pumping and the bends began to reappear. As soon as I could tell exactly where they were I stopped the pump and used the dremel to drill a row of holes along each of the bends. I spaced the holes about 2 inches apart, and drilled a few extra ones for good measure. I could tell when the drill met a delaminated area because the feel was different. Once the holes were drilled I released the pressure to get the strake back to its correct shape, pumped a couple of squirts of epoxy, filled a syringe, and squirted epoxy in all the holes. It didnt seem like much epoxy went in the holes. I decided that the skins and foam were butted against each other, so a bright idea dawned on me. If I suck some pressure and collapse the tank a bit, this will expose the fractures. If I squirt epoxy in the holes with the tank collapsed it will fill the gaps. Then if I repressurize the tank it will force the epoxy along all the ridges and out of the holes. This approach worked much better, and I had to pump two more squirts of epoxy to fill all the holes. I got to the point that squirting epoxy in one hole would force epoxy out of all the others. This meant that the voids were all full. Once I was sure I'd got all the epoxy in that I could, I let the pressure back in the tank, and all the excess epoxy came out of the holes as the tank returned to its proper shape once again. I cleaned off the excess and decided that the job was done.
Once back in the house I decided that there was a better way. I went back out, collapsed the tank a bit, squirted even more epoxy in the holes and then taped over all the holes with some metal duct tape. This time, when I gradually released the pressure, the holes were blocked, so (my theory) the "excess" epoxy was forced to track everywhere else it could before being forced out of the holes. A bit came out, but a lot more stayed in. I left it to cure overnight with a little bit of positive pressure in the tank to force the skins together. Once the epoxy reached "thick putty" stage, I removed the aluminum tape and scraped off any excess. Next morning the strake looked normal. I walked around the plane and noticed that my vent covers were still on. Could it be that my vent cover made such a good seal that it blocked any air from entering? I doubt it. The covered were made of a short piece of tygon tubing with a bolt in the end. They weren't a tight fit. Even so - this whole episode has proved something relevant to my fuel system arrangement, and making changes in general. You never quite know what set of circumstances will get you to a failure mode you haven't thought of. These fuel pumps Suck! If there's no way for the air to get in to replace the fuel removed they'll keep on sucking until the tank can't hold the negative pressure. I had executed the worst possible senario - 35 gals / minute pumping with the return going somewhere else. However, taking off with sealed vent covers could be VERY BAD. Not only can the engine face possible fuel starvation, but also you're strakes (read airfoils) can get smaller as you go. Not a good combination, I think. I'm glad I discovered this one on the ground. As well as checking what's wrong with the vents, I think I'll drill some small holes in my vent covers, put red flags on them, allocate a place in the cockpit for them and the pitot cover to be stored, and add verifying their presence to the pre-flight check list. I still dont believe that the vent cover caused the blockage. The fit is too loose.
[Late note: I made permanent vent covers by putting some aluminum window screen over the pipe, sticking it with epoxy and sliding a 3/8 al pipe over the top.
A little Physics goes a long way
I taped a latex glove over the vent pipe, put a shop air line in the filler cap, covered the opening with my hand as best I could, and pressurized the tank. After a few seconds I looked at the glove and it was a one foot ballon. OK. I added a bit more pressure until I had a two foot ballon. I removed my hand from the filler hole expecting the glove to deflate. It did.... very slowly over about 45 seconds. Hmmm. The vent line must be partially blocked. I poked a piece of rudder cable up the vent line and met no restrictions except at the tight bends. Thinking now, I got a 2 foot piece of 1/4 al tube and taped a glove on the end. I blew up the glove by putting the air hose on the end of the al tube, then removed the hose. The glove deflated.... very slowly. I think I've figured it out. When you pressurize a large tank the pressure exerts equally over the entire area. The 3/16 ID vent tube outlet only gets a very small portion of the pressure and the vent can only equalize the pressure at a certain rate which depends on how fast air can flow through the tube. This rate is further constrained by the length of the pipe and the twists and turns in it. OK, put you're finger over the end of the vent. You can stop the air flow with ease. So... my loose fitting vent covers were enough to stop air entering the tank, and the pressure built up too fast. The pump was evacuating the tank at its full capacity since I was returning fuel to the other tank. The question remains - are the uncovered vents big enough to allow air in fast enough such that the tank wont collapse in this senario. Another question hovers in my mind - how do I find out? If anyone reading this remembers enough fluid dynamics to do the calculations, I'd appreciate an email. What about you, Hicks - you're the damn rocket scientist!
Repairing the strake - again
I cleaned up the damaged area and prepared it with 33 grit, then duct taped around the outside to prevent (further) damage to the paint. Per Nat I added 3 ply BID over the entire damaged area, then I covered it with plastic and squeeged with a hair drier to remove all bubbles.
A little sanding around the edges, a couple of sessions with micro and the repaired area was almost ready for primer. I used red oxide spot filler under the smoothprime to correct any minor valleys and gave it a couple of coats of Top Gloss. Another day of color sanding and the repair was done.
Crossfeed
I like having the ability to transfer fuel from one side to the other. [I'm not entirely sure why], but I'd like to have some kind of warning when this is happening. I wired the pumps on seperate busses, each with it's own on switch. I wired the transfer solenoid with its own switch also. If the solenoid is on, return fuel goes to the left tank. If it's off, return fuel goes to the right tank. I could power the solenoid from the left pump switch, but this would remove the ability to transfer fuel from left to right tanks. I asked for a wiring suggestion on the Aeroelectric list but was told by Bob that my system was too complex. He gave me a link to a carberated system with only three components. My system is EFI, which requires returns, and I only have three components. I can't see how you could get simpler than that. Anyway, another builder called Bernie came back with a nice simple wiring diagram for my warning lights which uses a relay to switch on an LED whenever a crossfeed situation occurs. Works for me.
Draining the tanks
Want to know how a simple 20 minute job like draining the tank can take 3 hours, and cost you $35,000?
The tanks had been full for well over 3 months. I decided that the gas was probably deteriorating, and I'd like fresh gas when it comes time to start her up. I removed the right tank return line and attached a long 3/8 pipe instead. I pumped a couple of gallons into a can, then filled the lawn mower. Before putting the gas in the car I wanted to be sure it was OK, so I mowed the lawn. One hour gone. The mower worked fine, so I backed Char's car around the side of the patio. As I neared the end of the tube there was a crack as the right side door mirror collided with a bamboo tree. Damn. The mirror wasn't broken, but it had come out of it's socket and the quarter panel holding it to the door was badly bent. I filled the car, then went inside to tell Char the good news (the car is full of gas) and the bad news (the mirror's broken). She was pleased about the good news and even more pleased about the bad news - I need a new car anyway, she said, smiling. How about one of those new BMW SUV's? Aghhhhh!
I went back outside and spend a couple of hours straightening the panel and reinstalling the mirror. No-one will ever know. Next day I pumped the rest of the gas into my Saab being very careful not to hit the bamboo tree in the process.
As far as I can tell the fuel system is done. I'm just a little unhappy about the excess noise from the right pump. I'll probably replace it before the bird flys.
Fuel regulator
I'd always thought that the fuel regulator went on the rail. The one recommended by my engine guy, Roger, was a firewall mount model. It's input comes from the rail, and the output goes to my return solenoid. I wanted it on the cool (left) side, but that didn't work well with where the return solenoid was mounted. After a few experiments I moved the return solenoid so that both it and the regulator are on the left side. While moving things around I replaced all the aluminum pipes with steel braided teflon.
The remaining component of the fuel system is the fuel rail. See the intake system for details.
<<< Back | Index | Next >>>
The pour foam was no problem. I took extra care to use a 50/50 mix this time since I'd had problems with shrinkage earlier. I trimmed it flat and sanded the top of the R57 bulkhead a little to get the top skin flush with the spar. I had to mess with the contour where the skin meets the fuselage. Once I was done sanding the contour I found that the skins were about 1/2 inch too short. I added a 1/2 strip of foam to the leading edge before glassing. My measurements may have been a little out, but I still think it was better to have the skins in one piece of foam rather than two.
On the second strake I was extra careful placing the masking tape and the top hats lined up fairly well. Finally I used flox to radius the corners and 1 ply BID to tape the hats in place around the tank edges. I roughed up the top of the tapes and sanded the edges straight. When the top comes down on these floxed top hats, I should get a good seal. Using the skin as a form gets you the correct curve at the front of the ribs, but everything is flat at the back. If I was doing this again I think I might try to do everything but the forward 12 inches on the ribs using flat cured 1 BID tapes, then use the top hat method described above for the curved bits. (its a complex curve here and just curving flat tapes wouldnt work). When you squish the flox down with the top skin there's no way to smooth it out and the result can be a bit lumpy. Using cured tapes would allow you to radius the corners nicely with flox. I guess aesthetics don't really matter inside the tank, but a smooth corner would make taping the "hats" to the ribs much easier and ensure a good seal.
I got out the aluminum pipe and fittings and figured out how everything fitted together. Plans say these are "flared" fittings. I haven't worked with this type of fitting for many years and it took a while for me to realize that I needed some special tools for this job. It just so happened that my buddy, Bill, needed the same tools and was planning to buy them that day. So I waited till the afternoon, then borrowed Bills brand new pipe cutter, pipe bending spring and 3/8 pipe flaring tool. It took me a few hours, but I finally got the two fuel pipes in place. Rather than have to remove everything again later, I made a pipe to bring the engine supply line out from under the seat brace and attached this to the fuel valve. While doing this I noticed that the plans fuel valve has a third input, presumably for an auxilliary tank. Handy. I've been wondering about the possibility of fitting auxiliary tanks, but didnt really want tanks in the back seat area. I like the way Mike Mellvile put nacelles on the wings for the round their world trip. I even considered adding extra fuel lines in advance, but decided that the lines could go down the inside of the spar later if necessary.
Have you noticed how Nat sometimes gets maximum value for minimum words in the plans. The next day we got to do the most work ever for four words of instruction. Turn the fuselage over. I removed the screen from the side of the patio and removed the wings and canard. Bill, Jason and I then shuffled the plane sideways, over the hot tub and onto the grass. Char took the classic "space shuttle" shot as we flipped her over, then we shuffled "Juliet" back over the hot tub and onto a couple of trestles. The name Juliet was chosen yesterday - from her reserved registration number N386JS. We'll christen her formally when she flys, but I think we'll hold off on the champagne bottle thing. I microed the joins between the bottom skin, the spar and the fuselage and installed the OD bulkheads. I had to cut away part of the top skin on both sides to get the OD's vertical. While the plane is upside down I'll be doing some finishing work on the bottom. I was hoping that I would be able to get the bottom in primer at this point, but then realized that you can't do this until the leading edges and fairings are done, and these come after the top skin. For now I just "slapped" some micro on the lower nose.
Yep. You saw it coming, didn't you. The heat light had rotated in its stand and was pointing directly down at the foam 3 inches away. A 4 inch by 6 inch oval area of foam was melted and burned. For those who follow this site regularly... guess whose side of the plane it was!
Isn't it wonderful how easily you can fix composites! 10 months ago, this would have been a disaster. Panic calls to Nat and very distressing. Now it was just a little cursing, and an hour of recovery. I used a small plate to draw a circle, then carefully cut out the damaged foam. I scraped and sanded all the way down to the glass and was delighted to find that the glass itself was undamaged. The foam had taken all the heat. I hereby attest that H45 Divinicell foam is non-flammable. I painted the glass with a good coat of epoxy, then used my plate to cut a circular piece of BID and a matching circle of new foam. I laid the BID, wet it up, microed the circle of foam and pressed it in place. A little sanding to contour and my fix was complete. No one will ever know!
I glassed the second bottom skin and added the fuel sumps with no problems. Now it was time to turn the plane over again, but there were no helpers in sight. I decided to spend a couple of days working on finishing the bottom until enough people showed up. It turned out that my local EAA chapter was having a fly-in picnic just a few miles away on Saturday morning. Char & I went over and enjoyed the good company, food, planes and hanger talk. As we were getting ready to leave I "just happened" to mention that we lived only a few miles away and before ya know it - I have a car full of builders (including one Cozy builder, Todd Silver) following me home to take a look at my project. How convenient! We flipped the plane in no time. We were chatting around the plane a few minutes later when my buddy, Bill, spotted this large group of people and wandered over to get help pushing his stuck trailer.
This was to be a "hospitality" weekend. I think we'll call it the First Annual West Palm Beach Cozy builder's meeting. Mid january is a great time for the northerners to come down for a spot of warmth. Next day I was visited by two more Cozy Builders, Norm Muzzy (IO) and Bulent Aliev (FL). Bulent is currently rebuilding a Mazda 13B cosmo engine for his Cozy IV, and Norm is going gangbusters with his project. He's had the plans less than three months and is done through chapter 8 and Chap 14. Amazing. Norm also has the equipment and skills to build handy things like engine mounts. We're all serious about the Mazda 13B engine and much good conversation was had about engine mounts, exhausts, cooling and cowlings. The following day, Monday, Pace Harmon (CA) dropped in with his family for yet another Cozy visit.
I'd prepared the top foam for glassing, but held off because of all the socializing.
Bulent took one look at my left strake and asked for a long straight edge. He frowned. Hmmm. Where's you're spline sander. Before you know it, Bulent is sanding away at my top skin trying to fair it into the wing. We measured, sanded and straight edged for an hour or two. No doubt about it - the strakes are both too far forward. I'd completely omitted to check that the strakes angles correctly to meet the wings at the join. Later, I triple checked and tried to decide on a cure. I was considering the chainsaw, or perhaps cutting off the front 2 feet of skin, reshaping and shortening the ribs and then floxing everything back together.
I asked for opinions on the Cozy mail list and the unilateral opinion of the resident experts was to fair the strake straight to the wing and not worry. Seems that many builders have this problem. Some cut the end of the wing to match the strake, others just "fair it in". I also called Jeff Russell and asked his opinion. General consensus seemed to be to fair it in and carry on. No-one will be able to see it from 10,000 feet anyway. Looking at the problem in this light, I decided that I'd tape the wings for release and fair the strakes to the wing about an inch outboard of the cut line. This way I'll have a "cuff" which the wing will fit neatly into. Much neater than that ugly cut, and probably less draggy too. Cool. I have a improvement! (Later note: Wayne Hicks later persuaded me to abandon this idea because the wings move relative to the strakes).
I waited a few days for warmer weather before glassing the tops. On Sunday, January 30th. 2000 I did the second top strake skin. I put urathane foam down the gaps between the strakes, the spar and the wings and carried the glass over the gaps. This was the LAST large layup on the airframe. Lot's of fiddly bits to do yet, of course, but this was the last of the big stuff.
After all this I still had a very small leak on each side. I was going dizzy from blowing into the tank, so I called Eddie, my friendly local air conditioning guy. His company name is Elite Eagle and every time I think of him I think of the English guy who entered the Olympic ski jump one year. Anyway, Eddie came round and lent me a bottle of (environmentally safe) freon and his freon sniffer for a few days. I set up a 30 inch water column on the fuel line to avoid overpressurizing the tanks and added enough freon pressure to sent the column to the top. After a little experimentation to get the hang of the sniffer I discovered leaks at the baggage area where the strake skin meets the fuselage. I hadnt taped there yet, and I hadn't been checking there with the soap. I wasn't expecting a leak so far from the tank. I guess there must be some pinholes which let the air into the skin foam. I also found a leak behind the front of OD. A small nail hole. I carefully floxed and BID taped all the baggage area joins. The good news is that there are no leaks at all in my fuel sumps, the inside of the spar was clear of freon (I don't see how you could check there without a sniffer), and all the top and bottom skins and LE bulkheads are good. Next day I found another tiny (and I mean tiny) leak in the bottom skin inside the rear right baggage area. I painted this area with epoxy and waited yet another night. Next day the left side tank was totally sealed. It held pressure for 3 days before I let the pressure out. After 3 days searching for leaks with the sniffer the right (Char's) tank still leaks about 1/4 inch of water per hour. The leak is so tiny that even the sniffer can't find it. I suspect the fuel drain valve, but there's not enough freon to be sure. After hunting everywhere I finally gave up on this. I figure that this leak will either go away, show itself so I can fix it, or bite me. One chance out of three ain't bad. [Later note: It showed itself, so I could fix it, and bit me a little]