While I was waiting for the spinner to arrive I got on with mounting the engine. I striped all the accessories off and found that I could manouver the engine around fairly easily. I didn't have a "block & tackle" or any other way of hoisting the engine and I didn't want to buy one. I considered making the plans recommended support table, and will later, but right now I needed to get the engine off the floor and mated up with the mount upside down. Hmmm. I searched around and found a hydrolic trolly jack. I put the engine on the jack (not too hard, even on my own) and gradually "walked" the engine up into place using bricks and blue foam. I got about 2 feet in the air and decided that someone, probably me, was going to get seriously hurt. I went down to NAPA and got one of their "come-along" things, then went to home depot to get anchor bolts to put through my patio beams. This made life a lot easier. An hour later I had the engine bolted in place. I put the redrive back on and stood back to admire this new dimension. Wow. My Cozy is now 41 inches longer! My baby has an engine, and I installed it on my daughter, Julia's, 27th birthday. That's gotta be a good vibe.
At this point I diverted from the engine for a week to make the lower cowling.
When it came to bolting the engine in place I noticed that I'm going to need some non standard bolts. The sump bolts are too short because they have to go through the 1/4 inch al plate, and the big engine mount bolts at the front are too long because they don't have a stock rubber mounting to go through. I think NAPA will be able to provide what I need.
I raided the spare room, brought out my redrive, turbo, alternator and water pump, and fitted them on the engine. I spent some time trying to see what shape the upper cowling would have to be to accomodate the various parts, then created a web page with pictures and asked for comments in the flyrotary mail list. I got various suggestions about how to move or lower the water pump. For now I think I'll just make the cowl the shape that looks right and let "function follow form". While trial fitting the lower cowl I noticed that the rubber bushings for the top engine mounts had pulled away from the firewall about 1/8 inch and the engine was sagging down under its own weight. The prop hub was about 1/2 inch lower than it had been when I made the lower cowl. This was because the thickness of the firewall and glass mount reinforcements was less than the 3/4 inch required by the Barry mounts, so the rubber bushings weren't gripping the firewall properly. I decided to thicken the upper mounts with 1/8 aluminum floxed and glassed in place, then drilled for the 1.5 inch rubber mountings. So, if you're planning on using the same Barry mounts as me, make sure you have 3/4 inch thickness at the mount points when you do the engine mount reinforcement layups.
With Char's help I fitted the wings. Again. This must be about 6 times I've done this. I've got the old hacksaw blade & duct tape trick down to a fine art for fitting the nuts. (In case you're not familiar with this - you bend the last inch of a hacksaw blade to about 20 degrees and stick duct tape to it, stick side out. Now you stick the nut to the duct tape, reach carefully into the hole and, using a flashlight, position the nut on the end of the bolt. Now you reach around the wing and gently turn the prepositioned socket wrench. Bingo! The nut catches a thread. Now pull off the hacksaw blade and use the sticky end to rotate the nut on the threads. Works like a charm. I got the nut threaded on the first attempt on one side, third on the other. As usual, I couldn't remember which bolt(s) get washers. This is one of the penalties of being a bit disorganized. I wrote this information down somewhere, but I can't remember where. You'd think I'd have the sense to write this in marker pen on the inside of the wing right next to the bolt. But no, I always forget when I take the wings off. Once the wings are on tight it's obvious where the washer goes because one wing is slightly out of line with the strake. Now I have to remove that wing, put the washer on and put it back. Damn. This is a note to me. WING WASHERS (search key). One thick washer goes on the starboard side upper outboard bolt. OK. Got it? Good.
I had a little "incident" while fitting the wings. Knowing that the weight was all in the back (you know where this is going don't you?) I'd put a heavy car battery and a lot of weights in the front. The winglets were supported on foam and Char went inside to get on with some work. I jiggled the wings into final position and tightened up the bolts. One winglet was hitting the roof of the patio, so I needed to move the plane forward an inch. I pulled the plane by the nose, and the foam fell out from under the winglet. No problem, the bolts are on. Now the weight of the wings and engine, with no canard or canopy, was enough to override my battery and the plane wanted to sit down. I had it by the nose, but it was late at night and I was alone on the patio.....
So. How did I get out of this one? I managed to reach a cement block. I snagged it with one hand and lifted it up onto the canard support. The nose wheel came down again and all was well. Maybe this incident will help me remember weight and balance later when I have a prop to damage.
Once the wings were on properly I fitted the lower cowling again. I taped supported it along the sides to keep the edges even with the wing roots. I layed up a reinforcement rib of 5 ply BID, 3 inches wide along the edge of the cowling parallel with each wing root. Why 5 ply? Well, I could say I wanted to give the cowl extra strength. The truth is I misread the plans. The reinforcement should have been 2 ply. Anyway, I covered this layup with plastic and layed up the wing root lips on top of it. When everything cured I drilled and bolted the cowl to the wing roots temporarily. I was looking forward to working on the upper cowl, but I needed the lower cowl properly fitted first. I needed to add the 2 ply reinforcement layup along the rear edges, but the edges weren't "straight as an arrow". I came up with an idea to fix this. I clamped an aluminum extrusion above and below the edge, then heated the aluminum with a heat gun for about 10 minutes. The hot straight edges warmed the glass. When everything cooled down I had perfect straight edges. Now I had a straight base, I layed up the 2 ply BID and clamped the aluminum back on for cure.
The next item that needed attention was the spinner area. I'd tried to curve the back of the cowl around the spinner, but it needed some adjustment.
Finally, it was time to make the upper cowling.
Plans say to put a wedge shaped piece of wood at the top of the firewall to support the upper cowl as it goes on. Instead of using wood, I found an old aluminum rudder pulley bracket which had been replaced by steel. I bent this and screwed it to the firewall.
Wayne suggested I consider the fancy cowl fitting method which uses removable hinge pins, but I decided to use nut plates per plans for now. Later I'd like to make large cowling doors.
I found that the cowl lips were inexplicably too high in a couple of places. This made the cowl too low and would have required quite a bit of micro to correct the curve between fuselage and cowling. So I moved the lips. How do you move a fiberglass layup? Simple. You glass another lip on the underneath, then grind away the original lip. Getting the nutplates to line up was a bit tricky. After a couple of false starts, and a few extra holes I had to glass over, I found that the best way was to install the nutplates two or three at a time. I started with one at the top and one on either side. With these aligned and screwed up tight, I drilled the holes for a couple more etc. etc. Futzing with cowl lips and nutplates took a few days intersperse with some finishing work. I'm not sure how the plans prefab cowl comes, but I suspect most people have to jiggle with the fitting. Perhaps mine was easier - at least the cowl was made in place to match the fuselage fairings.
Given that there are about 26 screws (make that 30) holding the upper cowling in place, I'm considering making large doors on either side to make inspection easier. I'm holding off on this until the cowls are finished and painted and I see how well the floxed nutplates and stainless screws work. So far they seem to work very well. I've had the top cowl on and off a couple of times without any problems. With a power screwdriver and a good fitting phillips attachment I can whip the screws out in a couple of minutes. I considered using hinges and/or camlocks, but the screws are quick and easy (now).
Tracy says: The igniter is the module that drives the coils in responce to control signals from the ECU (or EC2). These are built in to the 86 - 92 RX7 coil modules but are separate parts on the 13BREW coils. The 93 - 96 igniter is a module about 3" x 5" with a finned aluminum heat sink on it mounted on the driver side of the engine compartment in the RX7.No. I don't have an igniter module. I talked to Bruce Turrentine and he tells me that igniter modules are pretty hard to come by and cost a ridiculous amount new from Mazda. He suggests I use the GM-LS1 coils which have the igniter module built in. Tracy is using these in his new engine. They're AC Delco part # 12558948. Tracy says he wouldn't use them if they were actually made by Delco but they are, in fact, made by Nippon Denso and are very high quality. Apparantly they were used on 1998 LS1 Corvette & others. Tracy mentioned that another aggravating factor is the connector used on these coils. It's only available at GM dealers and is expensive. It's sold in the 4 connector harness for the corvette. However, he says it's possible to solder wires to the connector pins and pot the connector cavity with epoxy to avoid this cost. I decided to get the LS1 coils and asked Tracy to set the EC2 up accordingly.
I think I'll install the center console with screws rather than glassing it in, at lest until everything's finalized in terms of control cables and wiring. The back of the console meets the seatback just above my plans fuel valve. No problem. This valve is coming out anyway. I haven't made the final decision on how the fuel system is going to work yet, but I'll need to do something with switching the return and the seatback valve isnt going to figure into the final solution.
I took my center console with me on a visit to Bulent's project, just to show it off. I placed it in his fuselage. It was about 8 inches too short. If Bulent ever offers you a back seat ride, be prepared for a tight fit. Buly is making good progress, but visiting a project that's behind you is always a good thing to do. It makes you feel better about how far you have to go.
I told you I had some thinking about your engine setup.
What worries me is having your engine bolted on the engine mount. From my
understanding both rotors would send explosions pulses through the engine
case to the mount. These pulses will be "Torque" pulses ie tending to rotate
the engine counterwise vs the propeller rotation.
I have the feeling that you have to damp those pulses by some means . Otherwise you might get your engine mount to crack after some time as it is bolted strongly on the engine.The mount absorbers you have on the firewall to hold the engine mount do not seem to be doing the job because the torque pulses would react as a shear action on them. Are they designed for shear vibrations?
Also as you add power to the engine, the torque send to the propeller means that the engine will want to slightly torque the other way, so you need to allow that movement (as it is done with the dynafocal mount used with Lycoming engines).
These thinking are only "thinking" and of course I only have the expertise of the fact that I made engineering studies and that is from a "Physics" stand point that I am reasonning.
I talked to a guy (Homebuilder) who professionaly was involved in a rotary engine for cars and he kind of agreed with me. He also told me that fuel burning was a bit high with that type of engine. Any idea on that? Have you talked with experts about your engine mount?
My reply was:
This is a very astute observation, Benoît. The mount was designed by a trusted aviation engineer based on input from other engineers familiar with the issues. At that time there were two schools of thought about damping - you can do it either at the engine, or the firewall. If you do it at the engine there can be too much movement (I forget the other negatives), but if you do it at the firewall you have the problem of torque being transmitted through the mount. Notice that the mount is not square. It has triangles from the forward and aft support points which support each other. Also there are braces which join three of the four mount points vertically and horizontally, so it should be a good "box" structure which will transmit the drive torque to the airframe. The explosion pulses are the main concern for causing fatigue cracks. In a rotary these pulses are MUCH less than from a 4 cyl piston engine like a lycoming. In general there is much less vibration and smaller sudden torque pulses because the large mass is simply rotating rather than reciprocating. There is still the overall torque, however, which is more gradual and should be handled OK by the stucture I have.
Yes, the Barry mounts are supposed to handle the sheer action, (They have steel inserts surrounded by delrin) and a number of trusted engineers have given the design their blessing. However, I am still a little concerned about this issue. There is now a "standard" Cozy IV / Mazda engine mount (See http://conversionconcepts.com/page24.html ) which uses isolators at the engine. This has not been tested yet either, but it looks good and I wish it had been available when I needed mine. I will keep a VERY close eye on the mount during the first few hours and will abandon it if there are any signs of a problem. I'd rather not, of course, since this will require a redesign of the cowls and a $1000 cost. To facilitate inspections I will be making cowl doors on either side so I can study the mount (and everything else) carefully during every preflight.
As for the fuel burn. Tracy Crook (flying an RV with a rotary for over 1000 hours) and a few other 13B flyers report similar fuel burn to a Lycoming doing the same work. It can definately be a little worse at high power settings (take off and climb), but it can be lower at cruise because you can lean the engine to the performance limit without worrying about burned exhaust valves.
Thank you for your concern and your well thought out points. Unfortunately there is no proven standard to follow in the Cozy 13B mount, so I am at the edge of the experimental envelope. I will be careful.
A few weeks later I returned to Pompono Beach to pick my various parts. Ed had cut the top off the water pump, made a new cover and added -16 AN fittings to the main outlets. While he was at it he glass beaded the pump inside and out. In the picture you can also see my radiator, oil cooler and fuel pump with the custom AN fittings attached. The price for all this work, including the revised motor mount, caught my breath. Hell - that's being generous. I almost needed resusitation. It seems that Ed's prices depend on how well his day has gone. He must have had a really NASTY day. I won't be going back to Pompono much.
12558948 Coil 12569111 Coil harness 12102754 7 pin connector for aboveAccording to Ed, the best plugs are supposed to be Leading B9EV or B9EGV, Trailing B10EV or B10EGV Gold Iridium. A web search turned up boxes of 4 for $19. I got the EGV ones.
At Ed's suggestion I got the plug leads custom made by Magnecor. When Roger from Japtrix saw the leads installed he mentioned that these are the best leads available, and he uses them for his race car.
It would have been nice to have had help, but I didnt want to wait, and I guess it became a bit of a challenge. Can I get the engine in the car, out of the car and onto the airplane on my own. Answer is yes, in about 6 hours. Would I want to do it again? Only if I really had to. I wish I hadn't said that.
If you're wondering why there's blue foam under then engine in the picture, that's to stop the plane tipping over. It very nearly went over when I took the canard off to work on the electrics, so I put the foam there to catch it in case I forget about weight and balance again.
About 4 weeks later I forgot about weight and balance again. :) Unfortunately I'd moved the plane and my support structure was elsewhere in the hangar. I'd decided to thread a couple of wires for the landing brake warning. I took the pilot seat cusion out and dumped it on the strake. Up she came. Before I knew quite what was happening the nose was 8 feet in the air and I was hanging onto the longerons for grim death. I kinda climbed my way down the lonerons toward the nose and, with some effort, pulled it back down. Figuring it out later I realized I'd set myself up. The engine was fully installed, but the canopy and canard were off. I guess this kind of experience is part of the training, so you don't forget weight and balance issues later.
I followed Tracy's instructions and had no trouble getting a good alignment. I bought a brand new torque wrench for this very event, and was ready to torque down the main adapter flange bolts. That was when I realized that the tortque wrench had a 1/2 drive, and all my sockets were 3/8. After antother trip to Pepboys I was able tp proceed. I'd set the wrench for 30lbs, but it didnt seem to be breaking, even when the built-in torque wrench in my arm was telling me I was way past 30lb. Hmmm. I adjusted the wrench to 10 lbs, and discovered that it doesnt "break" like the one I'd used before... it clicks. You have to listen for the click and stop. I hadnt heard the click and I hadnt stopped. All the bolts were now clicking just fine at 10 lb except one. I torqued it some more, and it just didnt seem to get any tighter. uh oh. I took the bolt out and discovered that it now had a "waist". A 12mm bolt, and I'd twisted it without even noticing. It seems that a two foot wrench can put a lot of torque on a bolt. I drove over to NAPA and picked up a replacement bolt. Same markings on the head and everything. Excellent.
Now to install the drive housing. I'd just added the RTV to the flange faces when I had a visitor who wanted to look around the plane. I didn't have time to torque all the nuts down before the RTV was starting to set, so I took everything apart, cleaned it all up and started again.
The second time I had no interruptions and the drive housing went on fine, until one of the AN3 nuts started to turn without tightening. Damn. I'd done it again. This time I'd stripped the threads on the nut. Not wanting to remove the housing again I dremelled the nut and managed to remove it without damaging the bolt. OK. I had some spare nuts of the same type, so I was back in business. I tightened all the nuts crosswise and was getting to about 25 ft pounds torque all around when there was a snap, my knuckles hit the ring gear, and one of the nuts came off, complete with the end of the bolt.
Next day, after I'd calmed down and my hand had stopped bleeding, I picked up another AN4-47A bolt from the FBO and tried again. I removed that adapter plate, inseted the new bolt and torqued away. 10lbs ok, 15 lbs ok, 20lbs pk, 25lbs ok....snap. Another nut came off with the end of its bolt still attached. Sometime I get the feeling I'm not working my way through this project, I'm stumbling my way through it. Oh. Thanks for you're agreement.
Somethings wrong here. I can twist a big 12mm bolt at 30lbs, but I'm supposed to torque these little AN4 guys to 60lbs. And, Tracy's instructions make a point of saying not to be tempted to overtighten them. Something don't jibe. I reread the instructions. Sure enough - the final torque on the main adapter bolts is 30 pounds, and the final torque on the 12 AN4 bolts is 60lbs. Then I noticed something. A four letter word. "inch". The big bolts get 30 FOOT pounds, the little ones get 60 INCH pounds. Damned engineers - he changed units on me on the fly. I'm no engineer, but I'd hazard a guess that an inch pound is 12 twelth of a foot pound. (PS - Yes, I do understand moments).
Since all 11 remaining bolts had been torqued to 25 FOOT lb or more, I ordered 12 replacement bolts. When they arrived I didnt want to disassemble the redrive, so I unbolted the flange plate from the engine, replaced the bolts one by one, tightend them to 5 ft lb, then reinstalled the flange plate. Once everything was back together I tested for binding by rotating the flywheel and feeling the end play. The end play was constant all the way around, so it looks like my alignment is ok.
The hose shop had a fitting which went in the "oil to oil cooler" port show The engine port identification page. From there we ran dash 10 AN fittings and stainless braided hose across to the inlet of the oil cooler on the starboard side of the engine, out of the oil cooler outlet and over to the inlet of the port side oil cooler. The pipe from the outlet of this cooler has to return oil to the engine. Unfortunately, don't ask me how, my engine doesn't have an oil return port. Apparantly the 2nd gen has a return just below the oil filter, but the 3rd gen doesnt. I wonder how they get the oil back. It seems that my oil filter is going to have to move, and we'll need an adapter for this. The current plan is to drill and tap this adapter to take the oil return.
We're done with the main oil hoses, except for the oil return which needs a custom adapter plate to go where the oil filter used to be. Next week we'll do the water, fuel, turbo and redrive oil, turbo water, heater and AC. Add the part finished intake and we should be ready to start the engine any time soon.
I installed the big radiator hose and fittings. Tightening the fittings was tricky. I couldnt find a 1 3/4 wrench, or even an adjustable one that wide. I didnt want to use a pipe wrench and damage the nice colors so I used one of those rubber band type wrenches they make for oil filters. This worked, up to a point, but I had to use the pipe wrench for the final tightening. You cant see the scratches unless you look close. The radiator connections were pretty easy. I'd had 37 degree fittings welded to the rad, so all I had to do was make a short pipe from the water pump to the front of the rad, and a long one around to the back.
The redrive oil fittings were ok, but the drain pipes were much too long. I plan to take these back and have them redone to get a neater and lighter result. The two drains need to be T'd together. I decided that the T could be a fitting in the forward drain. One pipe would go to the other drain and other back to the engine. Much simpler. I'm waiting for Roger to make the oil adapter plate which goes on the oil filter flange. This will have a feed for the remote filter, an oil return, and a fitting for feeding oil to the redrive.
When we're at the bottom of the mountain the ambient pressure is around 30 inches of mercury, and the turbo can put out around another 8psi. For the sake of discussion, lets use inches of mercury as the unit. 8psi is about 16 inches, so full boost gives you a Manifold Absolute Pressure (MAP) of 46 inches and you're engine puts out all the power it can muster.
Now, as you climb the mountain the boost stays roughly the same, and the ambient goes down about 1 inch / thousand feet. So, by the time you reach 10,000 ft, the MAP is down from 46 to 36. So we have a 10 inch map gap and the engine is going to run at say 25% less than optimum power. Not too bad, but the real trouble is that the turbo wastegate is controlled by the difference between it's output and input pressures. It doesn't know sqat about the change in outside pressure. The turbo trys to fill the map gap by boosting more, but just as the boost starts to go up the pressure sensor tells it it's overboosting the engine, and opens the wastegate. The system is designed to stop the engine getting more than, say, 46 inches of MAP, but it thinks it's overboosting when the MAP is only 36 or less. So, how do we fool it? One simple way is to override it by closing the wastegate manually. The vacuum tube that controls the wastegate has a T which goes to a solenoid in the car. Connect that T to a needle valve and you can modify the pressure difference. Open the valve and the pressure difference goes down. Close it and the pressure difference goes up. Fully closed, the system will work as in the stock car. Fully open and the wastegate will never open. With the wastegate shut all you're boost is available at altitude. Make sense? Good.
Here comes the other issue. How do we avoid blowing up the engine, the turbo or both? First lets look at what it would take to do that. We climb to 10,000 ft and shut the wastegate to get our full 46 inches of MAP. Now we decend with the wastegate shut, do a go around with a lean mixture and the MAP goes up to 56 inches. Not good. We could get a thing called overflash, which is the equivelent to detonation in a piston engine, where the fuel in the next chamber gets ignited before its time. Apparantly this can destroy the engine "in a hurry", so we have to avoid doing that. One way is to install a pop-off valve in the intake somewhere. This is basically a flap with a spring. When the pressure gets above a certain level it overpowers the spring and the flap opens releasing the pressure. You can get these popoff valves adjustable for 6 - 10 PSI (or 12 - 20 inches) of boost. The other way, of course, is pilot awareness. The decent checklist should include removing the boost override. Given both levels of protection we should be safe from "overflash". There are two other issues which become important here. The initial damage from overflash is usually to the seals. The seals then damage everything else they touch as the bits fly around inside the engine. Uugh! I had 3mm seals installed instead of the stock 2mm ones, so hopefully this will make my seals much more resistant to getting blown to bits. Also, I note that Tracy Crook now offers a new type of seal which might be very good for this application. I'll probably use these on my "next" engine. I hope I dont need a "next" engine.
The other bad possibility is overspeeding the turbo. Apparantly this is very rare and tends to result in a fairly benign failure where the turbo simply stops spinning and the engine sucks air through the broken turbo to become normally aspirated (NA). Roger says the blades may break off the exhaust side of the turbo, but they usually wont go into the intake. They may come out of the exhaust, but that's very unlikely because of the wastegate constrictions. Typically they'll get stuck in the turbo. The engine won't sound too good, and it'll be way down on power, but it'll probably get you home... after it gets your attention. We avoid overspeed by being conservative with boost at altitude.
Disclaimer: The above is my laymans understanding of the basic principles. I have much to learn, and there's a LOT more to turbos, such as sizing and compressor maps which might lead you to want a larger turbine, inconel wheels etc. etc. Turbonetics seem to be the company to contact if you want get it really right, and spend about $4k in the process. Right now I'm concentrating on the stock turbo and what's called "turbo normallization" i.e. getting full NA power at altitude by adding a little boost to make up the map gap. Roger mentioned that he can send my turbo away and have it fully reworked and reconditioned, and have a different compressor installed for about $800. This process would optimize it for low rpm torque and make it much more resistant to failure. Tempting...
I hope this all makes sense and helps someone ask better questions. If you know about turbos, read the above and feel it is incorrect, please send me an email at sladerj at bellsouth dot net.
Speaking of green grass at home, I'd promised to clean up the patio after building an airplane on it for 4 years. The best way to cover the epoxy stains was some green outside "grass" carpet. Before installing this I needed to take down the wooden wall which had supported my shelves. Behind the wood were two old windows. "We" decided to remove one of these and replace it with patio doors into the storeroom. But - lets put the new doors in the kitchen and move the old doors to the storeroom. Two days with a mallet, hammer and chisle and I had sore arms, correctly sized holes in the walls and a LOT of rubble. Two more days and I had the patio somewhat back together and was able to get back to the plane. Now.... where did those crank angle sensor wires go....?
The above was just one day in the life of an engine installer and probably represents 5% of the details we had to cover. I say "we" because Roger led me through this, with the occasional diversion while I led him on aircraft safety issues.
I cut the top off a 1/8 NPT brass vacuum connector, drilled out the hole a little and used JB weld to embed Tracy's intake temp sensor. This will now screw into a 1/8NPT hole in the manifold.
Oil sensors pressure are a bit of a problem. I have three of them, all 1/4 npt. One for the gauge, one for the Hobbs and one for the voice annunciator. Seems like I overdid this. There's a 1/4 NPT hole in the block, so that handled one of them. I'll need to make an Aluminum pad for the other two. The vacuum feed for the boost guage is a 1/8 npt fitting drilled and tapped into the manifold. I need two more of these for the EC2 computer and another for the EC2 air temp sensor. The CHT needs smaller sized plug washers and the EGT and air fuel mixture need to go in the exhaust. I think that covers them all.