My first mistake was to install a used stock '89 turbo. I'd seen that Mistral were using it, and Paul Lamar was recommending it. This was in the days when I didn't know any better than to listen to PL. Quite a few decisions bring a whole lot of baggage with them. Once the turbo and stock manifold were choosen the cowl and heat shield took shape around them, and the prop was made to match the power.
Then the turbo suffered overspeed damage. Changing to a larger turbo at this stage would have meant remaking the cowl or adding a large bump on the existing cowl and redesigning the heat shield and exhaust. I didnt want to do that, so I persevered. Hoping and wishing that my original turbo was just weak or unbalanced, I installed a replacement of the same type. It turned out that it was me that was weak and unbalanced. The second turbo blew up too, taking an apex seal with it. Another turbo rotary flyer had a similar problem, and Mistral later confirmed that they too had blown a couple of stock turbos from overspeeding at altitude. What with the delay in getting the modified replacement turbo from Australia and the damage done to the engine when the second turbo blew, this whole learning experience cost me around 7 months and $2500.
No new rotary flyer needs to go through this expense or delay again. The solution is simple - Install either a stock turbo modified by Max Heywood of ATS in AU or an approropriate T04 from Turbonetics, thereby saving yourself 7 months and $2500.
An alternative some have considered is to start out normally aspirated, then add a turbo later. The only problem I have with this is that a good prop costs upwards of $2000, and, unless you use an adjustable prop, you'll need to buy two. You'll also need to work out a muffler solution which then will be removed when the turbo is added.
Why bother with a turbo in the first place? Simple. The Cozy is predominantly a travel machine. The higher you fly, the faster you go and the easier it is to fly over the weather. The turbo gets you full power at cruise altitude, and, as a bonus, it saves you having to mess with the expense and heat from a muffler.
Being something of a novice at electronics when I started, this part of the build was a challenge for me. When things started to misbehave I blamed my wiring. Over a period of many months I rewired the entire ignition and injection control system twice and sent the engine control computer back four times. While I found and fixed a few weak areas along the way, none of these were responsible for my woes. Finally I discovered that the problems were caused by a badly seated connector with an intermittent connection. One way or another this cost me a year of frustration, and at least 5 months of delay. Are others likely to experience similar problems? No. Of all the 100 or so rotary flyers who have installed Tracy Crook's EC2 only a handful have experienced problems taking a week or so to solve. Usually these are things like bad injectors, incorrect wiring, and poor grounding. My experience was mostly just bad luck. I've been told that an intermittent connection is one of the hardest problems to track down. I can verify that rumor. They ain't kidding. You end up stabbing at the problem from random directions and suspecting and replacing everything. It's very unlikely that future builders will experience anything like the problems I encountered in this area.
It just didn't seem practical for me to do my flight testing at an airport ideal for the task. In retrospect this may have been a mistake. My local field has short runways, Class C airspace overhead, and absolutely no off-field landing opportunities. Having been through the experience, and with the stains to prove it, I highly recommend that flight testing NOT be done from a field such as mine. The critical phase of flight is, obviously, the take off. If you can't land straight ahead on the runway and there's no reasonable off-field landing site ahead you're stuck between a rock and a hard place. With a 6000' runway this problem goes away. By the time it's too late to land straight ahead, you have enough height to turn back or make it to a cross runway. This takes pretty much the pucker factor away completely. Add to this the need to either deal with approach clearance on every flight, or trundle out from under the class C airspace at 1000' with nowhere to go if there's a problem, and you have potential for a bad situation. With these issues in mind I was extremely picky about weather conditions and traffic volume, and the slightest hiccup from the engine caused an immediate abort. If I'd had a long runway out in the country somewhere the flight testing would have gone much, much quicker.
Building on what I learned there's no reason why a good Cozy rotary installation can't be tested and flown off in 14 days, never mind 14 months.
Despite all the issues I've had to deal with over the past 14+ months, I'm still convinced that the rotary is an excellent power-plant for a Cozy. A Cozy, however, which likes to land fast and doesn't like fields, isn't the ideal test bed for an one-off engine installation of any type, and there you have the dichotomy. Testing the engine on a dyno, the back of a truck, tied to a tree, on a boat or on a piper cub might help eliminate some issues, but not many. It's the final configuration of the engine and, especially, the configuration of supporting hardware in the target airplane that matters. What works on the boat might be negatively affected by airflow or cooling differences on the Cozy. The solution, in my opinion, is redundancy. However redundancy brings its own baggage along for the ride.
In an effort to increase safety, we add redundancy and thereby increase complexity. Compared to an air cooled aircraft engine with a carburettor and magneto ignition, my installation is incredibly complex. Where the Lycoming has a simple carb, I have two sets of two electronically controlled injectors. Where the Lycoming has engine driven magnetos I have two batteries, two ignition computers, two crank angle sensors and four coils driving two independant sets of two plugs. On top of all this, I have a liquid cooling system with all its associated hoses and connectors, and a turbo with its own water and oil cooling, and a pressurized intake system and intercooler. Finally, where the plans Lycoming has a simple gravity fed fuel system, I have duplicate high pressure fuel systems, each with their own pumps and filters, each of which require fuel to be returned to the correct tank. Sounds ridiculous, doesn't it? What I've just described is becoming the standard for flying rotary installations. Yes, the redundancy breeds complexity, but it also breeds safety. Combine the rotary's internal robustness and propensity for continuing to run under almost all circumstances with all this redundancy and you have the energizer bunny of aircraft engines. It just won't stop. If you've read my log you'll know that, while getting the bugs out of my system I suffered coil failure, fuel pump blockage, catastrophic turbo failure (twice), intermittent ignition, mixture and fuel injection problems. Not once did the engine stop running. On every occasion I simply flew back to base, taxied back to the hangar and fixed the problem.
Back at the hangar there's a VariEZ with it's engine removed for a $12,000 rebuilt. There are a TOTAL of 5 wires hanging from the firewall. Five. I'd bet I have more than 50.... but, then, you could snip 25 of them and the engine would still run.
What are the advantages? Reliability - Other rotary flyers have shown that these engines just keep on running. If things go wrong, as they can with any mechanical device, the rotary usually gets you to an airfield while the certified engines that do quit have a nasty habit of doing so catastrophically and immediately.
Power - That extra 50+ HP is nice to have. Having it at 15,000 feet is even nicer.
Maintenance Cost - Any engine can fail. If my engine fails internally the worst it can be is a $6000 brand new replacement. If a Lycoming fails the cost can be $17k or more. Normal preventative maintenance is also inexpensive.
Vibration - Often not considered, this is a big issue for me. I've flown a Lycoming Cozy and a rotary. Rotary is a lot easier on the bones. I watched Al start his LongEZ today. You could see the vibration all the way to the winglets. By contrast the rotary purrs. At the end of a 4 hour flight in a Lycoming powered plane I'm tired and usually have a headache, just from the vibration. You have to experience the difference to appreciate how important this factor is.
The jury's still out on my fuel system. I love the total redundancy, but the pilot workload is heavy. I need better strainers before the pumps. I'll probably add a transfer pump and operate it a la Tracy Crook in normal use. I still think the overall concept is better than the alternatives. I don't like sump tanks and all their potential for vapor lock, air lock etc. and I'm not keen on the long runs of fuel pipe needed for duplex fuel valves which can also give rise to vapor lock.
I'd improve the exhaust augmentation, maybe using the cowl boat tail for the exit. My cowl / heat shield interface isn't ideal. It needs an air gap or ceramic coating to keep the exhaust air from discoloring the paint. Firing the exhaust directly at the prop was a concern, but it seems to work out fine. The alternative of an angled exhaust is draggy. I plan to send the prop back for finishing soon, and will ask Clark Lydick for his thoughts on whether the heat and/or pulses are damaging it's structural integrity. I'm confident the answer will be no.
The cowls themselves are, I think, a success. It IS possible to get everything under them, there's no bump in the airsteam and there seems to be sufficient air exit. According to the involutary oil trail tests I've done, the boat tail does its job admirably. Building the plenum into the lower cowl is effective, but it's a tremendous pain to remove the cowl. I'd install the plenum with nutplates or hinges so the lower cowl can come off without disconnecting the rad and oil coolers. If there's enough interest I may pull molds from my cowls and offer cowls for sale. They're not perfect, but they're a pretty good start.