William Wynne

"The Corvair Authority"
5000-18 HWY 17 #247
Orange Park, FL 32003

Important Research Update

January 15, 2006

A defining characteristic of my work with Corvairs is continuously working to improve the engine and its installation. Improvements fall in many categories: power output, ease of installation, simplification, even appearance. But all of these are secondary to my primary goal of building better engines, and teaching and encouraging others to do so. I've said many times that aviation will always involve risks, but our goal is to make you aware of these and show you as many ways as possible to reduce them.

While some companies build a product, send it out into the marketplace then consider their job done, our work is very different. The central axis of all we do is testing. This is what provides the information and drives the design of our products. Thus, our development and research is never done. It is an ongoing process that is integrated continuously in the schedule.

If our only product was manufactured engines, we would only be concerned with how people use them. The extensive scope of our teaching operations requires that we additionally be concerned with how people build and install them as well as how they operate them.

This past week marked the first phase of one of our most intensive investigative projects. The focal point of this work is crankshafts. Most everyone building a Corvair engine followed the crank issues in 2005. It has often been pointed out to me that few companies in this industry discuss service difficulties frankly. For example, virtually every type of engine that has been put in an airplane and flown extensively, be it Continental, Lycoming, VW, Subaru, Jabbiru, Rotax or Corvair, has experienced some type of crankshaft failure. Yet, you will not often find plain discussions on the subject prominently featured on the respective Web sites. Each business has their own approach, but in our case, relentless testing and frank discussion will always be our approach. If you're one of the rare customers who'd prefer the blissful illusion that there is such a thing as the perfect engine which can't be broken, perhaps you'd be more comfortable with some other engine company. While the general feeling among Corvair builders was that the case was closed on crankshaft issues in 2005, my position as the authority on Corvair flight requires that no case is ever closed, and no issue goes unmonitored. This week, we summoned four builders from the fleet of Corvair aircraft to bring their engines to us for teardown and inspection of their crankshafts. I selected these engines because of certain characteristics they possessed and theories we wanted to test. Three of these engines were torn down in house beside one of our own engines. The fifth crankshaft was shipped in by air. The owners of these engines voluntarily and without compensation did this at my request. While I was prepared to offer a lengthy discussion on the importance of the project, the four builders understood quickly and were glad to play a positive role in research that would potentially provide risk reduction to all Corvair engine builders. Keep this in mind when you meet one of these gentlemen later. You'll certainly want to thank them.

The engines torn down were:
1. Our original 601 engine with 200 hours, .010/.010 regrind, no nitride. This engine represents a standard installation, albeit one that was flown at its limits by Gus.
2. Steve Makish's KR-2 engine, a standard GM nitrided crankshaft with 181 hours on it. Notably, this time was flown with a mid-length prop extension and the crank suffered two prop strikes in this time. This crank was chosen because it is the highest time nitrided crank accessible.
3. Bill Clapp's KR-2S engine with 165 hours on a .010/.010 crank, no nitride. Bill accumulated this many hours since he fractured a crank 12 months ago. The damage from Bill's break was minimal; he only replaced his crank and cam gear. All of his subsequent hours were flown on the same parts, including the case.
4. Dan Weseman's 3,100cc Cleanex engine with 46 hours on a .010/.010 non-nitrided crank. Dan had always planned to be one of the first customers to fly a prototype of our fifth bearing crankshaft. Dan opted to fly off his 40 hour test period with a typical Corvair crank, and then have us convert his engine to a fifth bearing motor. Knowing that we'd inspect the engine at the conversion, and knowing Dan's skill as a pilot, his plane was cleared by us as a test case for limited aerobatics. Dan flew the engine with care, but at its full power output. He did not include any snap maneuvers in his flights.
5. Mark Jones' KR-2S standard non-nitrided crank. Mark flew this for 71 hours in 2005. With his airplane disassembled for painting and upgrades, I suggested that Mark exchange his crankshaft with us for a nitrided one. This gave us the opportunity to inspect his former crank.

The gameplan of the test was to disassemble the engines, clean the cranks and have a meticulous magnaflux inspection done on them. Fifteen miles north of my shop, inside my alma mater's Embry-Riddle FAA certified engine repair station, lives the mother of all magnaflux machines. In the hands of a skilled and experienced operator, it can detect stress fractures and cracks that would slip past virtually all machines and operators in the automotive world. To operate it, I called my old roommate Chris Welsh, who now works for Grumman in St. Augustine. When we were students, Chris worked for many years in the Embry Riddle repair station on the very same magnaflux machine. If you're new to aviation, you may not understand that the airworthiness of certified engines is ensured by the FAA's system of rigorous standards which are put in practice in a controlled environment called a repair station. While there are hundreds of repair stations in America, there are only a handful of famous ones. And among these, Embry Riddle's is unique because it has the dual mission of providing all the engines for Riddle's fleet of aircraft (the largest private fleet of light aircraft in the world) while training legions of A&Ps. The training nature of the repair station makes them extremely aware of how equipment is used and how standards are met. It produces a tighter standard than a situation where one guy operates the same machine alone for years. In short, we took the crankshafts to the right place with the right people.

Here are the results of the magnaflux testing. Keep in mind that we focused our attention on the area between the front main bearing and the number six rod bearing because in a standard Corvair installation, this is the most highly stressed part of the crank. Crankshaft #1 showed stress fractures on both sides of the area in question. Crank #2 showed absolutely no flaws, even with the circular magnetism set as high as 2,400 amps. Crank #3 also showed no flaws even under the highest power applied. Crank #4 had a very significant crack running around the edge of the fillet on the front side of the #6 bearing. This crack was apparent at any amperage above 2,000. Crank #5, at 2,400 amps, showed a stress fracture approximately 1/4" long on the front bearing fillet.

Concurrently with the teardowns and inspections, we also took the week to have a general review of all procedures involving crankshafts. We talked extensively with the grinder, heat treater, and a CNC shop with very good optical measuring equipment. The following conclusions are based on this information. Keep in mind this article is a small summary of what we learned. It's not possible nor necessary to present every word of conversation here. My conclusion from these observations can be grouped under two topics: nitriding and fillets.

Based on what we now know, I consider it an unacceptable risk to fly a crankshaft that is not nitrided in a Corvair. I am well aware that this statement will be received as heavy news by many people who already have an assembled engine with a non-nitrided crankshaft. Although this is not the news that many people want to hear, it's not my job to always make you feel good; it's to report the facts as we know them and provide you with guidance based on our experience. Keep in mind that in my heart I know exactly why people, even those on tight budgets, feel the tremendous call to create and build airplanes. And I balance this with the personal knowledge of the cost of aircraft accidents. There are many difficult judgement calls in aviation, but this isn't one of them. Nitriding is cheap, available, and it works. Granted, the test samples are not large, but the indications are not debatable. The only argument that an individual builder could have against the process would be the day to disassemble his engine, the two weeks of downtime in process, a few hundred dollars in expense for the process, shipping and gaskets, and perhaps two days for reassembly. Being unhappy reading this at first pass is understandable. Being reluctant to perform this after a day's consideration or seeking out someone, perhaps on the Net, who will give an answer (based on no experience) some builders want to hear, is a different story. Let me be blunt and say the minority of builders who continuously have this mindset when confronted with a clear issue should consider leaving aviation while they're still alive. Anyone who has owned a certified airplane knows that there have been countless ADs issued on these planes. As an A&P mechanic, I can't think of a single AD on engines that was this inexpensive to fix.

There is a lot of talk about the size of radiuses on Corvair cranks. They are typical of automotive radiuses. By comparison, aircraft crankshafts have gigantic radiuses or fillets in the corners. Sitting in racks in the shop, I have more than 50 Corvair cranks. We can easily observe what a typical GM radius was, and compare it to the radius after the crankshaft is ground undersize. Cross referencing this with the tests at the magnaflux machine, which shows with absolute precision the exact point of propagation of a crack, allows me to make the following comments about radiuses. There are large variations in how GM ground the cranks. Many stock cranks had very square corners ground by GM. Once a crank is ground this way, it is very difficult to grind even a 3/32" radius in it. Without the benefit of two cranks side by side, radius gauges, and 10X magnification, it is difficult to see one in favor of the other. In magnafluxing the cranks and considering their experience, discussions of radiuses are deceiving. The #4 crank, with the worst crack, had some of the best looking radiuses. The #3 crank had sharper radiuses but exhibited no defects. Removing stress risers and unnecessarily sharp radiuses is important, but no one should feel that this is a substitute for nitriding. It should be regarded as a complementary procedure. Radiusing alone has proven to work in individual cases. We flew hundreds of hours on .010/.010 ground non-nitrided cranks without problem. This is similar to the experience with crank #3. This led to my previous recommendation that this was acceptable. However, the evidence now clearly shows that an individual builder faces an unacceptable risk operating a crankshaft this way. Nitriding is emerging as the proven deterrent to trouble, not radiusing. My recommendations on these issues have to serve everyone. An individual case that this worked once for one guy does the average builder no service. I regard dispensing conflicting information based on no flight experience or firsthand testing as amoral, even if it's couched in the usual disclaimers. I could easily go back in the Web archives and find calculations provided by Internet contributers showing that non-nitrided cranks would never suffer a fatigue failure. As well meaning as some of this was, it certainly points to the fallibility of theory. While I'm now tightening our own recommendations and rescinding my previous endorsement, keep in mind that we flew all this stuff personally for hundreds of hours, and I never suggested to anyone that they do anything that we had not already previously flown. In discussing this with Mark Langford, he pointed out that Page 1 of the Conversion Manual contains the sentence "I reserve the right to get smarter." In the long run, builders are obviously best served by our continuing testing. This testing and commentary on crankshafts is far more important to builders than any of our promotional, educational or installation efforts. Service Bulletins, Notes and ADs are part of how Lycoming and Continental maintain their safety standards. Some companies, like Rotax, do the same. Other experimental engine companies do not deal with issues in the same way and that's their call.

Further Testing
To further confirm that nitriding is the correct approach, crankshaft #2 is being reassembled into its engine to be flown several hundred more hours and then re-inspected. When you read a post from or meet in person Steve Makish, recognize that he has volunteered to be a critical part of risk reduction to anyone operating a Corvair engine. As a builder, this merits your respect. Crankshaft #3, despite having no flaws, is retired. I intend to submit it for X-ray inspection to confirm my belief that magnaflux is the front line of NDI on crankshafts. Bill Clapp's engine has already been reassembled with a .020/.020 nitrided crank that was put through our process. The airplane will likely be flying again in two or three days. Bill is going to install a turbo on this engine and fly it to Sun 'N Fun. He accumulates hours on his aircraft very quickly, and like Steve, is very willing to participate in any inspection that will yield useful information. We intend to work his engine very hard and inspect it at regular intervals. The other three crankshafts will obviously be retired. The #1 crank is being kept for an important project. Engine experts have told me that it's possible that a small accelerometer mounted on an engine can, if properly calibrated, detect the vibration caused by the onset of a crack such as the one in crank #1. By building and running an engine on our dynomometer with the #1 crank in it, we could potentially develop simple equipment to spot the onset of a problem. Please note that even if this eventually comes to pass, it will never be a substitute for nitriding. Engine #4 is being converted to a fifth bearing engine. In February, we'll have more updates and photos of this. We're working to have at least three flying examples of the fifth bearing at Sun 'N Fun April 4-10. Our in house testing of it will be done on either the Skycoupe or the Wagabond. Dan's airplane, the Cleanex, will also feature this and we plan to convert Mark Langford's 3,100cc engine to a fifth bearing also. Mark plans to re-install this engine in his aircraft as soon as possible. The crankshaft in engine #5 is being replaced with a nitrided one. This and the replacement of the crank in #3 means that all the flying Corvair powered KR-2s in North America have nitrided crankshafts. The most recent KR pilot to join the fleet is Joe Horton in Pennsylvania. On the eve of his first flight, he opted to get a nitrided crank from us and replace the one in his engine. The swap took Joe only one weekend. The Conversion Manual contains a paragraph about Aviator #1, Jimmy Doolittle, Master of the Calculated Risk. He survived by always stacking everything in his favor that he could control. Joe replacing his crankshaft is a sterling example of the Doolittle philosophy in action. Keep in mind our own 601 has about 100 hours on a nitrided crankshaft. Additionally, Mark Langford's current engine, a 2,700cc Corvair, has a reground nitrided crankshaft. In the coming months, both of these engines will offer additional confirmation of the benefits of nitriding. I intend to tear down the 601's engine for inspection at an appropriate time, and also, I've spoken with Mark about inspecting the crank in his current engine when he converts to his 3,100cc fifth bearing engine. As said before, our testing is continuous and ongoing.

Solutions for Individual Builders
If you have a Corvair crankshaft that you intend to fly, it should be magnafluxed and nitrided. We are currently lining up a second source of nitriding by a slightly different procedure which will allow freshly ground cranks to be nitrided and then directly re-installed in their engines. We will have a detailed Web site update on this the first week in February. Grace's trip and the organization of this will preclude having a Web update in the next 15 days. Builders should hold onto their crankshafts because ideally, they will be shipped directly to the nitrider and not to our hangar.

For builders who purchased completed engines from us, it's important to note we built engines with both nitrided and non-nitrided cranks in 2005. In the next two weeks, Kevin and I will finalize a procedure that will allow us to exchange the crankshafts on all the non-nitrided engines. The most efficient way to do this in some cases will be the direct swap of engines. In other cases, it will be having one of us replace it in the field.

Owners of our products should rest assured that we will find a fair, equitable and timely way to allow all builders to take advantage of what we have learned about crankshafts. Again, we'll announce this procedure the first week in February. For builders who assembled their engines in our shop during Colleges and at other times, we will likely offer a schedule of replacement workshops. However, the majority of builders who worked under our tutelage on the assembly of their engine in the first case are certainly qualified to repeat the procedure at home. Bill Clapp was in our hangar during the crank testing, and I discussed with him the option of utilizing him as a flying field rep to assist some builders with the crankshaft exchange.

Whatever the final procedure, keep in mind the goal is that everyone utilizing the Corvair engine should operate a magnafluxed and nitrided crankshaft. While in the first 40 years of flying Corvair engines this did not prove to be necessary, my testing has shown that with increases in power output and the type of use the engine sees today, it has become a requirement.

Further Drive Developments
In addition to getting everyone in the fleet to operate a nitrided crank, we are also closing in on two other approaches that in time will allow the use of heavy constant-speed props and the approval for aerobatics with the Corvair engine. These two programs are referred to in our shop as the Fifth Bearing and Spline Drive. The fifth bearing adds an extension on the Corvair's crank, made of CNC'd 4340 heat treated steel. The prototypes have been made to use the tapered shaft prop hub arrangement from a Continental C-85. The thrust bearings for these engines are rear mains for small block Chevys, and they're located only 2" behind the prop flange. Installing this on an engine requires starting from scratch with a case and a crank. By this process, the Corvair engine will be converted to a crankshaft strength standard associated with certified aircraft engines. It will not even be in the same category as previous Corvairs, VWs or Jabbiru engines. The aforementioned engines have previously served well with light props, but the fifth bearing is designed to work with props as heavy as 30 pounds, and absorb aerobatic loads with 15 pound props. We're working to have flying examples of this and static displays at Sun 'N Fun. The nature of this design precludes its installation in the field. It is not a bolt-on system. It will add in excess of $2,000 to the expense of building an engine, but even so, the Corvair will remain far less expensive than other engines like the Rotax or Jabbiru, and it will have the capability of absorbing loads that are well beyond the limits of other engines.

Now that the fifth bearing is being turned into metal parts and will shortly be entering the test phase, we're gearing up to build and test another drive method. The most important character difference of the spline drive method is that we're working to make it field installable as a bolt-on item to a standard Corvair engine. The prop hub is independently supported on ball bearings that absorb all thrust and flight loads. It is driven by a concentric spline shaft addressing female splines inside the prop shaft. The male spline shaft is bolted to the Corvair's crank. Involute splines allow the different bearing styles of the crank and the hub to work with each other. This is the proven drive method that appears in high quality aircraft gearboxes attached to plain bearing motors. This is very different from any system which would attempt to rigidly attach a ball or roller bearing on the outboard end of a plain bearing crank. While initially seeming simpler than a fifth bearing engine, the design and flight testing are far more extensive on a spline drive. It must be carefully investigated for harmonic resonance, torsional stress and wear. This will not be debugged and finalized overnight. It may eventually be restricted to propellers of 12 pounds or less. In February, we'll show photos on the Web site of its progress, and we'll have physical examples at Sun 'N Fun, but its flight testing will have to follow instrumented dynomometer runs and careful evaluation.

The crankshaft work is being done side by side with the normal production of parts in the shop. I encourage anyone to call the shop who has serious questions. I'd like to thank people in advance for the patience they'll show waiting for the procedural update the first week of February. Some of the builders we spoke with expressed the thoughts that this news might come as something of a setback for the Corvair engine movement, and potentially a large scale expenditure of resources for us. The bottom line is simple: Testing reveals the truth, and the truth dictates recommendations. Financial or business consequences have zero bearing on this issue. Any businessman who would hesitate to make a recommendation for reasons of finance, trouble or ego does not deserve your trust or business. Years ago, I set out to build the least expensive engine at a reasonable risk. The current round of testing has merely revealed that the engine will cost a few hundred dollars more than planned to stay within a reasonable risk. While I am sure this news will generate a handful of fairweather builders who will move to other engines, the vast majority of our builders will recognize it as a significant improvement at modest cost and they will remain glad to be part of an engine program founded on testing, improvement and rigid adherance to the truth in the face of all pressures. The serious builders who respect these values have always been the customers I've worked to serve.

Crank Update

August 5, 2005

It has been about 45 days since our last technical update on the Web on crankshaft issues. I had a chance to speak with thousands of people in person at Oshkosh, and also touch base with a few other professionals from industry. In the past 90 days, there's been a tremendous amount of Internet discussion regarding crankshafts and my post-Oshkosh impression is that some of the people reading the Internet do not understand the situation, and most of the people on the Net have little or no exposure to airplane crankshaft issues and their applications to all the engines in the alternative engine field. The following comments are intended to clarify and enlighten the average builder and dispel some of the misconceptions that have arisen from the great volume of discussion; but the talk, talk, talk and commentary is based mostly on speculation and very little testing.

A Quick Synopsis

Corvair engine first flys in 1960. It is continously flown from 1960 to today. Approximately 400 Corvair powered aircraft have taken to the air. The highest time examples are in the 1,200 hour range. Despite the fact that the engines are flown with a variety of handmade propellers, and hubs of mediocre quality, crankshaft issues are unknown on the engine prior to 2004. There is a popular misconception that all of these engines were being used as 2,500rpm, low power engines. While they are not as potent as engines built to my specs today, some of these engines are worked very hard. The Oshkosh 2005 Update on my Web site includes Bernie Pietenpol's personal Air Camper on display at Pioneer Airport. This aircraft flew hundreds of hours with a 66x30 prop. I have Bernie Pietenpol's notes on the desk, which include the use of rpms as high as 3,300. Its common cruise rpm was 3,000. I have many more examples of classic engines worked at higher rpm than this. All were without failures.

Since 2004, there have been four crankshaft failures in KRs. The average time to failure on these engines is about 60 hours. When you look at all the aircraft that are Corvair powered and have flown more than 500 hours without incident, you have to simply come to the conclusion that the four failures had something about them that made them different than engines that reliably logged ten or twenty times as many hours.

In my 15 years of working with flying Corvair engines, and 10 years of doing this commercially, no engine we have built or flown, nor anyone who has built a copy of the engines that we have built and flown, has had any type of trouble with the crankshaft. The four cranks in the KRs that fractured each had significant deviations from how I teach people to build engines. The airplanes we've owned or directly used as testbeds have logged more than a thousand hours on Corvair power without any type of crank trouble.

If you're a builder simply looking for a trouble free installation, common logic suggests that if you duplicate what works for us, it will work for you. The unspoken corollary to this is, of course, if you duplicate what has proven to break before, in all likelihood, you'll be rewarded with the same result. In all my years as an A&P mechanic, I've actually heard a number of people, after experiencing trouble with an airplane or hearing of an identical airplane's problems, say something like: "I'll just try it/run it that way until I have a problem with it." I carefully take these people aside and explain to them that the definition of insanity is running an identical experiment and expecting a different result. In the past four years, many people have argued that my efforts to encourage people to build replicas of proven engines hamper people's creativity. If you want to build a really creative airplane engine out of a Corvair, it's your choice. We don't recommend it. And, since the KR crank incidents of late, people who casually spoke of modifying the engine rethink their positions when they consider the consequences of untested modifications.

There has been some interesting and useful discussion on the Internet. I applaud the candor with which Mark Langford discusses his issues. Because of his job, he has access to a number of respected aerospace engineers. Some of these people have offered interesting commentary. John Kearney has embarked on an interesting data collection project to evaluate non-standard engines. However, there's been a fair amount of conclusion jumping from people who've never seen a running Corvair turning a prop. After speaking with thousands of people at Oshkosh, the detrimental effect of speculative chatter is that perhaps a third of the people we spoke with did not have a clear picture of any of the crank information. It was just a muddy morass of talk. I'm the only person who's seen all four crankshafts, built dozens of engines which have had no trouble, and have on hand all the historical data of all the other airplanes that flew trouble free. This doesn't mean that I possess the only valid information. But it does mean that I can shoot down virtually 95% of the commentary and conclusion jumping.

For example, at Oshkosh a builder told me he'd read on the Net someone's wise opinion that all 3,100cc engines would get the crankshaft in 10 hours. Obviously this isn't true, because Steve Makish flew his 3,100 about 80 hours on a .010/.010 non-nitrided crank with an extension on it (not recommended), Les Laidlaw has his 3,100 flying trouble free in his Dragonfly now, and of course, Gus and I flew the 601 to Oskhosh with my own 3,100 in it. Clearly, the person who wrote the original comment to the Net had no exposure to what's really flying out there. I had numerous people ask me if the power output from the engine was to blame. Many people had read this on the Net. Obviously, I do not believe that it is. My 3,100cc engine is the most powerful naturally aspirated Corvair powerplant flying. The turbo engine in the Skycoupe sitting outside my hangar is significantly more powerful than my 3,100. I could continue on with examples that would dispell most of the speculation and preclude the air being filled with such chatter. However, the Internet will always be a discussion group and people are going to discuss whatever they want even if it flys in the face of what's actually flying.

Many of the comments from people indicate that they have little or no knowledge of what's inside many engines. You can look up crankshafts on the NTSB Web site and there are page after page of accidents from crankshaft failures. Obviously, any crankshaft in an airplane can fail if it is operated outside the accepted set of specifications. In extremely rare circumstances, there have been manufacturing defects in certified crankshafts that led to trouble. But pretty much every accident involving a crankshaft is from operation outside the limits of what's proven to work.

Sticking with the subject of alternative engines, let's examine the Corvair, the VW and the Jabiru 3,300. All of these engines have maximum recommended prop hub lengths, propeller sizes and weights, and rpms. Quick Quiz for Internet Experts: How many of these three have broken crankshafts because they were operated outside the recommended limits? The Correct Answer Is: All Three. Most people know that VW engines have broken crankshafts in flight. Yet, when well built and operated with light, two-bladed wood props, VWs have an excellent track record. At Oshkosh, I spoke with our friend Jeremy Monnett. He told me he'd recently had someone ask about putting a heavy in-flight adjustable prop on one of their 2,180 Aerovee engines. In short, Jeremy told him that you could use a recommended propeller or buy somebody else's engine. Propeller extensions on Aerovee engines are not recommended, nor necessary. On the subject of Corvairs, clearly, anyone who owns my Conversion Manual understands that for years I have taught people not to put any kind of extension on the engine, nor run heavy props. I have taken great pains to explain to people gyroscopic forces and other factors people hadn't considered. Most people listened, some people did not. On the subject of Jabiru engines, most people are stunned to learn that there has ever been a broken crankshaft. One of the people I told this to accused me of making up the story for business purposes. I don't tolerate allegations like that well. Especially because the person who told me that a crank has broken in a Jabiru engine from excessive prop loads (not recommended by Jabiru) was Pete Krotie. Pete told me this when Grace Ellen and I walked over to his booth at Oshkosh this year. Perhaps the guy who didn't believe it might consider what Pete does for a job: He is, of course, the CEO of Jabiru USA. Before anybody jumps off on a tangent, I think the Jabiru 3,300 is an excellent engine. And although it's expensive, it's a very fine and high quality product represented by good people. We have friends who own them who have flown hundreds of hours and think they're good powerplants. My point is merely that the Corvair, VW and the Jabiru are restricted to light weight props, and no extensions because the common thread in all three designs is that they do not have an extended, long bearing in the front of the engine common to Lycomings and Continentals. Many Lycomings and Continentals can handle metal props and fly aerobatic routines. Fortunately for us, wood and composite props are much more economical, and are available in an infinite variety of applications, plus, very few pilots are interested in serious aerobatics, and those pilots are interested in serious aerobatic airplanes, serious training, etc. Thus, Corvair engines and the other alternative engines can serve the purposes of the majority of homebuilders. Pete showed us that the maximum allowable prop hub length for a Jabiru is virtually the same difference from the last bearing on the crankshaft on the 3,300 as is our prop hub face from the last bearing on a Corvair crankshaft. Pete additionally told us that Warp Drive props were acceptable, but only in diameters up to 60". Considering the amount of time that I have on 66" and 68" two-blade Warp Drives on Corvair engines, you could conlcude that the Corvair's crankshaft is at least as strong as the Jabiru's.

The odd thing is, I think few people understand that the Jabiru's crank contains no magic. Like the Corvair and the VW, it has limits on hub length, prop weight and rpm. A gentleman at Oshkosh told me that he was sure that a Jabiru had a stronger crankshaft than a Corvair, and his idea was based solely on the idea that a Jabiru has a main bearing between each rod throw. I kindly pointed out that the Corvair's crankshaft is drop forged, while the Jabiru's is machined from a billet. It is universally accepted among engine designers that forged cranks are stronger than billet cranks, and that perhaps having a main bearing between each rod throw was dictated by having a billet crank (horizontally opposed six-cylinder drop forged cranks are immensely expensive to tool up for and generally out of reach of short production run engines like the Jabiru). Although this man had not been shy about sharing his analysis of engine design on the Net, his qualifications obviously did not extend beyond the ability to count the number of main bearings in an engine.

If anyone thinks that my comments are a knock at Jabiru or VW engines, you're quite mistaken. These engines have served people when operated in their recommended form, they will both continue to be around a long time. And, no one should use any of my comments here for any purpose other than to understand that all engines have prop and operating limits. I think the majority of people who spend $15K for a Jabiru engine tend to follow the recommendations closer. But the $15,000 you have in an engine is not the most valuable thing you have in your airplane. The most valuable thing you put in your airplane is your passenger, and therefore, it doesn't matter how much you spent on your engine: You should fly it within accepted limits. If you want to fly outside recommended limits, perhaps you should do it alone over unpopulated areas, like the ocean. You most certainly shouldn't recommend to other people that they, or their passengers, should fly outside limits you've personally tested at length yourself.

The other day, it was brought to my attention that a small outfit in Australia named "The VW Center" is trying to cash in on the disinformation. They are working on a bolt-on external bearing prop flange unit for the Corvair. Seems like a great idea until you consider the following facts:

  • They've never flown one.

  • They don't have a running Corvair engine.

  • I'm pretty sure they've never tested it.

  • Its bearing style conflicts severely with the Corvair's main bearing design.

  • Its length would exclude use on any existing cowling.

  • Although they want your $1,500 today, plus $100 in shipping, they don't have any of them made yet.

    Sounds like a great bargain, huh? Most people who know me understand that the instantaneous way to get on my bad side is to offer a product that has never been tested before, nor will you likely fly yourself, yet you offer it to people for sale complete with the assumption that you know what you're doing.

  • Their Web site contains commentary on oil pressure that clearly shows they have no idea how plain bearings work, what the thrust load on a crank from a clutch is, etc.

  • The guy in Australia promoting this, hoping to see a flood of checks for $1,500 for this coming out of North America, is named Phil Matheson. I wrote Phil an e-mail explaining my concerns about selling untested products. Here's how well researched and well versed Phil is on the world of flying Corvairs: He had no idea who I was. You'd like to think before you sent a guy $1,500 perhaps he'd done a little more homework than that. (Here's a little tip to anybody buying an expensive product from overseas: Find out who their USA distributor is, make sure he's incorporated in the U.S. and not the Cayman Islands, and insist that you buy it through the distributor. That way, you'll have someone to hold personally responsible who lives in your hemisphere. The USA distributor will always have to show up at Oshkosh, and if you have trouble, you'll be able to explain it to him in person. The VW Center has a USA distributor, but my bet is he'll balk at handling a product with zero flight testing on it.)

    I could write a few thousand words explaining technically why their bearing idea is bogus. Let me just say that when I worked with Jim Rahm on the V-8 Lancair IV-P project, the toughest part of the gearbox puzzle to crack was simply how to make a plain bearing crankshaft work with a ball or roller bearing on the end of the shaft attached to it. The answer is, they don't work well together, and Jim had the privilege of spending $100,000 or so to learn that. Although many people think of me as an advocate of simple engines, I only became that way after exposure to the highest levels of engineering brought to bear on the most complex of aircraft engine installations. I'm sure Phil's going to come back with a technical argument, but to everything he says, my response will always be, that's your theory, based on 0.0 hours of operation. People can talk all day long about their VW experience, but you're not going to fly a VW engine. You're going to fly a Corvair. Even with all my experience owning, building and flying all types of engines, I only market products for the Corvair because those are the only ones I'm actively flight testing, and I only have respect for aviation businesses who adhere to this principal.

    And just to be really blunt, the VW Center Web site is loaded with gearboxes and other products for VWs. Lately, they've tried to adapt these to the Corvair engine. You have to ask yourself if their products for VWs were so great, and therefore in theoretically great demand, why don't they sell like hotcakes already, with no need ever to be bolted on Corvairs? My Web site is filled with photos of running engines, flying planes, in-flight photos and movies, plus trips to places a thousand miles away year after year. The VW Center Web site is missing these things, but it does make a cool airplane noise when it loads up.

    If Phil wants to market this stuff to Americans, it is not unfair of me to ask him to send us an air to air photo of him flying it, with one of his children sitting next to him. Builders should know, in case anyone else has ideas about front bearings, that there can be no bolt-on solution to an outer bearing on the front end of a Corvair engine because the Corvair's case is not accurately enough made for an external bearing housing to simply be bolted on and have it align with the crankshaft's axis. In the Corvair car, the transmission is driven by a two-foot long, flexible driveshaft, which is splined to the clutch disk, which, of course, is mounted on the Corvair's unique flexible flywheel. That system allows misalignment that occured in manufacturing to be a non-issue. However, anyone suggesting bolting on a rigid bearing structure as an aftermarket piece, without any type of alignment other than the dowel pins, has a severe misunderstanding of the Corvair engine. And if it fit on one, it could not be exchanged to another engine. Yes, the bell housings will interchange between engines, but the fit of a seal is far less demanding than the fit of a bearing.

    Remember, I grew up in Asia, I met a lot of Australians in my youth, I have great respect for our two dozen Australian customers, and anybody who wonders what Australians are made out of should go down to Blockbuster tonight and rent the Mel Gibson film Galipoli or the Edward Woodward film Breaker Morant. My problem resides just with the two "gentlemen" in Australia, and their alleged product.

    In conclusion, if you follow the proven path that we've forged, you can expect the same trouble free service we've gotten from the engine. Among 601 builders, I've noticed that they tend to follow our exact installation. The same one we flew 200 hours in the first year, that was a pure evolution of all our previous fligh testing. The opposite end of the spectrum was represented by the independent thinking and building KR community. One of the positive aspects of the KR crank situation is the fact that many people who'd previously deviated from our recommendations are re-fitting their aircraft to bring them closer to our standard installation. Steve Makish, who has had no trouble, has decided to remove his prop extension and utilize one of our new cowls. Mark Langford has also decided to shorten his installation, down to our recommended 3" length. Bill Clapp's airplane also will be re-fitted with one of our cowlings and a Standard Front Starter Installation. In the past month, a number of Corvair/KR builders across the U.S. have come to the conclusion that they too will build an engine that will conform as closely as possible to our flight proven work. These installations will provide the same type of service as we have always had. Even Bob Lester, whose aircraft was substantially damaged, has absolutely sworn off any type of installation that deviates at all from my recommendations and flight testing.

    I believe that the only Corvair powered aircraft that could use a properly designed and built front bearing would be turbocharged engines using heavy in-flight adjustable props, aerobatic installations that impose severe bending loads, and propellers that operate in the wake of other props, i.e. push pulls and overlapping props on installations like tri-motors. To this point, such applications are rare, and no one's close to flying one. If you're building a standard KR, 601, Pietenpol, Davis, Wagabond, etc., my recommendation is to operate the engine within our proven limits. Keep in mind that all of my recommendations are based on real world experience and flight testing, and my overwhelming desire not to see anyone get hurt in an airplane.

    What makes our work with Corvairs one of the greatest jobs you could ever have is the fact that we're having a lot of fun with our friends. The day anyone gets hurt in a Corvair powered airplane, I'm sure a lot of the fun will disappear. I'm merely trying to teach people what we have learned, and provide proven recommendations, so that no one ever has to call me and tell me that you or your passenger had an accident.

  • Crank Issues, June 2005


    The following notes are technical commentary on the subject of crankshaft breaks in direct drive Corvair aircraft powerplants. This dissertation will provide readers with the available data and a discussion of aggravating factors.


    In the past 12 months, there have been four crankshaft failures in these engines. Notably, all of these engines had significant deviations from the guidelines established in my Conversion Manual. Engines built to the specifications in the Conversion Manual have flown many times more hours total than the four engines in the three aircraft under discussion - without any type of crank failure. While reading this, keep in mind that I am the only person who has seen all four of the engines in person, and conducted inspections on all four.

    History 1960-2005

    The Corvair engine has been flying since 1960, almost half the time that there have been powered aircraft in the world. My extensive research indicates that during this period, there have been probably 350-400 Corvair powered airplanes. The highest documented times flown on a single engine are approximately 1,200 hours. This has been done by both Tom Brown and Virl Deal. There are many pilots who have accumulated more than 500 hours. While most of this time was flown in Pietenpols, there are still a number of higher rpm installations such as Jim Ballew's DA-2 with close to 500 hours. In all the years of my work with Corvair engines, no one has shared with me a single story of a broken crank in a direct drive engine. There was a rumor of one in the 1970s in a Pietenpol, but no one could provide a name, state, N-number or year. It should be noted that no one, to my knowledge, has been fatally injured in a Corvair powered airplane. The sole instance of a broken crank prior to June 2004 was done by Rich Dietrich with a belt reduction on a Corvair that had no outer bearing on the crank sprocket. All of my work is with direct drive Corvairs, and more than 98% of the estimated 30,000 hours that have been flown on Corvair engines have been done behind direct drive engines.

    Description of the Four Engines In Question

    Bob Lester's KR-2 engine, first installation (hereafter referred to as BL1): The longblock of this 2,700cc Corvair engine was built by me in my shop at the Spruce Creek airport about four years ago. Bob did his own installation on the engine, including a Mark Langford style rear starter and a 4" extension on top of my 3" Prop Hub. Bob made both of these changes to adapt the Corvair to an existing KR-2 cowling. I considered this a mistake, and told him so at the time. All of our published literature warns people against prop extensions. Bob took about a year of part time work to finish his installation. He ran the engine for 45 minutes on the ground once without any type of cooling baffles or cowling, which I also warn against. The engine ran hot enough to detonate, destroy both heads, and require the replacement of several pistons and cylinders. Once flying, Bob installed his own carburetor, and had significant problems getting it to work correctly. He had the float set incorrectly on this Stromberg carb and ran it to the point of lean detonation. With the float corrected, the aircraft flew about 100 hours with a variety of five pound, two-blade wood propellers operating in the 3,000-3,500rpm range. About 10 hours before the crank broke in his plane, on a visit to our airport, Bob had a hard landing which was severe enough to grind the bleeder screws off the Cleveland calipers. This was done after Bob installed a 3-blade, 58" diameter Warp Drive prop. This prop had blades modified by Gary Hunter. This modification changes the shape of the blade face, and requires extra care when setting the blade pitch with a protractor. A mistake in pitch setting by Bob led to a flight with this 10 pound propeller turning in excess of 4,000rpm. About 10 hours later, Bob felt a vibration in flight significant enough to warrant a landing and inspection. He flew 10 minutes to an airport, landed, and discovered that his crank was fractured. I personally did the teardown of the engine, and the fracture occurred between the front bearing and the number 6 rod. This engine had a polished, standard crankshaft in it that was never ground. After the teardown, I discussed this event publicly on the CorvairCraft discussion group. The archives should be checked on this for comments at the time.

    The second crank issue was in Bill Clapp's KR-2S in February 2005, hereafter referred to as BC. Bill had built his own engine in his shop in Valdosta, Georgia. It had a Mark Langford style rear starter, a WW Prop Hub, and a 2 3/8" spacer between the prop and the hub. This combination was flown approximately 110 hours with a Prince 54x50 prop that was originally made for Bill's 2180 turbo VW. His Corvair could turn this prop more than 3,900rpm in level flight. Most of Bill's flight time was in the 3,400-3,600rpm range. The engine was of average smoothness. Bill had experienced two stripped sparkplugs, and replaced them with 18mm aircraft plugs of unknown heat range. On a cross country flight with his daughter, Bill detected a vibration and diverted to Keystone airport in Florida. The engine continued to run for the duration of the 10 minute deviation. Upon landing, inspection revealed the crankshaft to be fractured. The engine was torn down in my shop that night. Internal damage was very light. During the rebuild of his engine, Bill opted to use the same case, pistons, rods, heads, etc., replacing only the bearings, cam gear and crankshaft. Close inspection of his .010/.010 reground crank showed that a local Valdosta automotive shop ground it without performing the standard practice of leaving a radius in the corner. The fracture occurred between the front bearing and the number 6 rod. The engine was reassembled, installed in the airplane and flown again in only five days. It has flown trouble free since then with a properly radiused crankshaft, logging 60 more hours. Bill's plane is now being converted to use a Standard 3" Hub, and a Front Starter Assembly. It now has a 54x54 Sensenich propeller installed, which provides peak performance while only turning the engine 3,500rpm. The cylinder head with the 18mm sparkplugs was replaced with a standard unit.

    Following his first engine, I built another longblock for Bob Lester. This engine included a new case and a crankshaft that was reground with proper radiuses and had been subjected to a first class magnaflux test in an aircraft repair station done while I watched. Bob felt that the culprit had been the three-blade prop, so he reverted to a Prince 2-blade prop as the only significant change from his previous installation. This prop was a handmade, carbon fiber coated wood prop, about 54" in diameter and of unknown pitch. It would only turn 3,200-3,300rpm on Bob's aircraft. Bob flew this combination, herafter referred to as BL2, about 60 hours. About two weeks before his second crank break, Bob flew in to visit our shop. While at our place, several of us told Bob that his engine was not running smooth. Bob was focused on solving an idling problem with his engine. While running it briefly uncowled, I noted that the vibration on Bob's engine was bad enough to have loosened his entire rear starter. A dial indicator was put on Bob's prop hub, and it was shown to be running 50/1000" out. This is an enormous deviation. Upon disassembly and inspection, it was revealed that Bob had allowed a significant amount of Loctite to get between the crank and the propeller hub upon assembly. On a visit with Grace Ellen to Bob's airport years earlier, I had advised him to be wary of this, showing him Steve Makish's installation as an example. We worked several hours, and reduced the runout in Bob's assembly below 20/1000". Still very high, but a large improvement. This could have been cut in half by removing the 4" extension. Our friend, A&P Arnold Holmes, brought over vibration testing equipment and showed Bob that his installation was running rough enough that some of the values were off the scale. Arnold volunteered to work with Bob, and over a period of two hours, made a very large improvement in the smoothness of Bob's installation by installing numerous balance weights on his spinner bulkhead. Bob's front spinner bulkhead was a homemade piece of very poor quality. It had globs of epoxy and flox, which was a large part of his out of balance situation. While the engine did not run perfectly smooth, the balancing work made a drastic improvement. Bob opted to fly home to get to work on time, but called to say the airplane actually turned 100 more rpm on the flight home. About 10 hours later, while on a cross country flight, Bob experienced an engine vibration. A check on the GPS showed an airport below him, but Bob opted to try to fly to a bigger airport with more facilities, approximately 20 miles farther away. He did not make it to this airport. Instead, he made a forced landing on a road short of the airport. He did extensive damage to the airframe, and injured his back. He called us, and we drove 300 miles and inspected the airplane with the FAA. The gentleman from the Atlanta FSDO carefully noted the remains of the front spinner bulkhead and commented on the great length of the prop extension. The engine was torn down at my shop about a week later. The break in the crankshaft had occurred between the number 5 and 6 rods. There were signs of oil starvation to the bearings on the front end of the engine. The aircraft is currently under reconstruction, and Bob has declared that he will not fly any type of prop extension again.

    In June 2005, Mark Langford of Alabama fractured the crankshaft in his engine with approximately 50 hours on the engine, which we'll refer to as the ML. The engine had his rear starter setup and a 5" prop hub of his own design. The propeller was a heavily reworked Sterba prop, approximately 56x60. This propeller turned about 3,200rpm on Mark's installation. The break in Mark's crankshaft occurred between the front bearing and the number 6 rod. His crankshaft had been reground by a crank grinder with radiuses about five years ago. Mark had the crank magnafluxed before installation. We were in the area and diverted to Mark's house the day after the break. We were able to see the installation in person in Mark's shop. The number 6 rod was removed to allow inspection of the radius. Mark's 5" hub was removed and a small amount of Loctite was found between his hub and the crank. At the time, my primary focus was on the Sterba prop, which appeared to be particularly rough. Mark stated that the engine had run smoothly and was run without a spinner, although the balance of the propeller was not checked with sophisticated equipment. The first sign of vibration in flight to extremely rough operation was only a few seconds. Mark opted to glide engine off to the nearest airport, the Redstone Arsenal in Huntsville. Twice on the glide he restarted the engine to check what power would be available if needed. He landed smoothly, and rolled to the first turnoff. His engine, a 3,100cc Corvair, had experienced some teething problems in early operation. During a ground run prior to flight, the engine had run hot enough to damage a cylinder head, which he replaced. In the week before the crank failure, Mark also reported having experienced detonation after fueling with low grade gasoline. Mark has said he intends to run one of our Standard Prop Hubs and a CNC manufactured propeller. He is currently focused on examining the role of the rear starter assembly as a potential factor and a common thread in the failures.

    Concurrent Perspective

    During the same period of time, our own aircraft logged more hours in the air than any other Corvair powered airplane. Numerous other installations without extensions or rear starters logged signifcant amounts of time without any type of failure. Additionally, Steve Makish's KR-2 has flown about 170 hours with a rear starter and a 3" extension on top of my standard Prop Hub. It should be noted that Steve is an excellent and careful mechanic and operates his aircraft conservatively. His rear starter setup is slightly different than Mark Langford's. Although he runs automotive fuel, he has carefully tuned the engine not to detonate. It is natural for builders to look for the common thread between these engines, but it is my observation that they probably broke for different reasons. BL2 was clearly the runout and out of balance; BC was the sharp crank radius. All observations have to be put in the perspective of the thousands of trouble free hours that have been flown with the engine. Few builders have the big picture on this issue. Example: John Martindale's KR-2S in Australia has logged about 125 hours on a Front Starter Standard Hub engine built in our shop about five years ago. Prior to the sale to John, this engine had logged about 100 hours flying my Pietenpol. The same crank and case in this engine had been flying in Pietenpol N1777W since 1968. Thus, that crank has in excess of 500 flight hours on it. For any observation on the crank issues to be valid, it must fit all of the available data, not just the four specific instances mentioned above. Below is a list of specific topics which I consider to be factors worth discussing on the issue.


    I have always been against extensions on Corvairs. My Prop Hub is 3" long, the same as all the Bernie Pietenpol hubs. This is plenty of length to effectively streamline the engine installation. I have never installed an extension on any Prop Hub. While Lycomings and Continentals can use extensions, they have purpose built and engineered aircraft crankshafts. Customers of mine have, in several cases, put on extensions of up to 4" long. It's a free world, and once I sell something, it's out of my control. But everyone who owns any Conversion Manual of mine can easily look up my arguments against extensions. To date, I have no evidence that any Corvair with a 3" propeller hub length has ever broken a crank in flight. Three inch prop hubs have logged at least 98% of the time flown on direct drive Corvairs. The handful of people who have flown longer extensions is a small number. I believe that Steve Makish, with my 3" Hub and his 3" extension on it, is the highest time pilot at 170 hours flying with an extension. Last year on the Corvaircraft Internet discussion group, one person publicly made the false argument that a propeller extension 4 feet long would have no more effect than one 4 inches long and plenty of people jumped right on that bandwagon. Notably, he had never flown behind a Corvair engine, and I don't believe he's ever seen one run, nor does he have the intention of using one for himself nor his family. To the extent that he convinced anyone that this was acceptable, I hold him responsible for the effects. His comments are on the record in the Corvaircraft archives, as are mine. People offering advice to others should be cognizant that the archives are a permanent record of all advice, ethical and unethical.

    Out of Track Hubs

    The largest factor in BL2 was the out of track hub. We tried several times to get the combination of prop hub and extension to track accurately on Bob's crankshaft when he stopped by my hangar. The first issue was the installation of Hybrid Studs using red Loctite with excess Loctite between the crank and the prop hub. When Steve Makish was first flying his plane on Corvair power, we visited his hangar because he'd found his hub was loose. This was about two years earlier, and it was Grace Ellen who suggested we take off Steve's hub and check it for Loctite interfering with mating to the crank. She was right, and Steve cleaned up the hub and re-installed it cleanly. Subsequently, we put numerous warnings in our Corvair Flyer newsletter, e-mails, and other communications. The Hybrid Stud installation instructions specifically address this issue. We had built the BL2 longblock in my shop, but the red Loctite was a giveaway that Bob had installed the hub himself using the wrong Loctite, as we specify and use Loctite 620, which is green, for Stud installations. When we were looking at the ML engine, Mark's engine utilized metric Allen bolts to secure his 5" hub. One of the issues when using bolts to install a prop hub is the fact that it is impossible to very difficult to ensure that the Loctite used on the bolts does not creep in between the crank and the hub. Prior to 1999, we used bolts to install our Prop Hubs and faced this same issue. That's why we knew to look for it on Mark's installation. If you have bolts holding on your hub, they must all be installed and torqued in place, and the hub checked for runout. Then, they should be removed one at a time and reinstalled with sealer on them. This will minimize the chance of any of the sealer getting between the hub and the crank. Hybrid Studs eliminate this issue. In either case, all hub installations should be checked with a dial indicator for run out. Acceptable run out measured on the perimeter of the prop hub is 1-2/1000". This is well within the capability of a CNC Prop Hub and a GM crank. Mark told us that he did not dial indicate his hub, but I do not suspect it to be a major factor, as with the BL2 engine.

    Handmade vs. CNC Props

    This is a big issue. If the two blades are different, one is going to pull harder than the other all the time. While propeller blades see different angles of attack in normal flight, the differences are not as great as those generated by hand carved propellers. While many hand carved props have an excellent safety record on Lycomings, the Corvair's automotive origins now lead me to discourage the use of hand carved props on the engine. Especially if any other aggravating factor, such as higher rpm, is considered. BL2, BC and ML had Prince, Prince and Sterba props, respectively. While these prop names have been around for a long time, and some of them have logged a fair amount of hours on Corvairs, I am now advocating only the use of symmetrical props such as Warp Drive, and CNC machined wooden props like Sensenichs and Aymar-Demuths. I own a large number of props, and have spent a lot of time working with them, and can assure builders that no hand carved prop can be made symmetrical without care which would require a price far above what people are willing to pay. In the BL1 incident, I suspect that the highly modified blades made setting the pitch difficult, and it may have been asymmetric. Standard Warp Drive blades have a perfectly flat blade face, and it's very easy to set a two-blade version identically.

    Proper Crank Radiuses

    Journals on crankshafts have a radius in the corner. If this radius is removed by improper grinding techniques, the crank is drastically weakened. In an automotive application, the crank sees fairly low bending loads. Sharp corners that would last in a car are completely unacceptable in an airplane, especially on the number six rod journal. I would put the primary reason for failure in the BC accident as the tight to non-existent radius that the automotive grinder put on his crank. Mark has posted photos of the radius his local crank grinder left on his crank, in comparison to one ground by our crank grinder, and they are comparable. The BC crank had almost no radius when compared to either the ML crank or one of ours. We've used the same crank grinder for the past eight years. And engines built with those cranks have flown hundreds of hours without failure. I don't consider the radius on a properly ground crankshaft to be the primary issue. The proper conclusion to draw from the ML and BL2 engines is that a good radius that will perform flawlessly under a typical set of circumstances can still be broken if enough aggravating factors are present in a sufficient magnitude.

    Starter Locations

    Mark Langford has proposed that the one thing that all three airframes had in common was a rear starter which was either his design or a copy. He feels that it may have been a factor, and has professionals who agree with his assessment. Personally, I don't know enough about the real world of that type of engine analysis to comment. The fact that all three airframes and four cranks had this in common is worth examining closely, but obviously a builder using a no radius crank and an 8" extension should not feel safe because he has a front starter on it. Two other rear starter arrangements have logged a lot of time. Steve Makish has about 170 trouble free hours on his rear starter arrangement, which hugs the ring gear much closer to the balancer. Between 1999 and 2001, my Pietenpol flew approximately 350 hours with a unique rear starter assembly that also mounted the ring gear very close to the balancer. Both Steve's and my setup drove the alternator off the balancer groove. Since 2000, I have only recommended front starters to people. The primary reason why people chose rear starters was the erroneous belief that it would aid the weight and balance situation on the engine, and because they were reluctant to modify existing cowlings. Over the years, I've always advocated modifying the cowling instead of the engine.

    Oil System Issues

    One of the major problems with rear starter assemblies is the modifications to the oil system required to install one. Few builders understand how the system works, including the typical operation of all the bypasses. The Corvaircraft archives contain a very lengthy dissertation from me under the title Subject: "The Mother of All Oil Posts." The BL1 and BL2 engines had the same oil system. It was a three hose setup with the inlet and outlet on the top of the rear cover. Although this may look externally similar to the two hose remote filter setup that we use on the 601, it is radically different in operation. The BL system contains the fundamental flaw that it has no cooler bypass, and thus, in cool weather, will have enormous drag through the cooler. It is possible to see zero oil pressure at the bearings while seeing 20 pounds regulated oil pressure at idle. The three hose system that I used on the Pietenpol's rear starter had the oil entering and exiting the engine at the stock cooler locations. This system utilized the stock cooler bypass, and at most would see a 7psi reduction in oil pressure at the bearings. Most of the people who modified the oil system for a rear starter had a mistake in it like the BL engines. Additionally, the BL engines had seven hard 90 AN fittings in series. This is an enormous amount of drag on cold oil. I suspect that it worked as long as it did because the airplane's based in South Florida. On the BL2 teardown, it was apparent that the 5 and 6 rods and the front main had occasionally seen an oil shortage. I think this is a contributing factor. The last bearing in the engine to see oil pressure is the number 6 rod. Obviously, you woudn't want lack of oil pressure to gouge this bearing journal in the engine. Front Starter Systems or rear starter systems with proper design eliminate these issues. Evidence shows that very few homebuilders understand the intricacies of the oil system. These people are always better off with front starters. The Conversion Manual cautions that any welding in the oil system, for rear starters or otherwise, is a bad idea. Most homebuilders and many "professional welders" don't possess the welding skills to keep slag out of the oil system when they're attempting to modify it. It should be noted that at least 85% of all the time flown on Corvair engines has been done with hand prop engines. The two highest time installations are both hand prop engines: Tom Brown and Virl Deal. So are the majority in the 500-700 hour range. Although many of these engines had cruise rpms in the 2,800 range, they virtually all used propeller hubs in the three inch range. And of course, we have no evidence that any of these engines broke a crank in flight.


    Throughout all my writing, I have taken great measures to warn people about detonation. I still believe that very few builders have an intrinsic understanding of how bad detonation is on an engine. Over the years, I've successfully advocated forged pistons, but the negative effects of detonation go far beyond the destructive effects on cast pistons. Each detonation provides a very harsh shock wave traveling through the engine. Consider that one of the few times I flew a Corvair that let out a few knocks from detonation on climb out was in my Pietenpol during an ignition advance test. Imagine how strong of a blow is struck on the inside of an engine that you can hear it in an open cockpit airplane while climbing out on full throttle. Similarly, consider this observation: When removing the flange and gear on a Corvair crankshaft, I frequently can load it to the full 20 ton limit of the press, and it won't budge. However, you can lower the pressure to 10 tons and take a 12 ounce ballpeen hammer, tap on one of the crankshaft throws with a blow so light it would hardly bury a four penny nail in a piece of soft pine, and the gear will jump 3/8" on its way toward coming off. I suspect that detonating engines have similar vibrations running through the crankshaft. One common thread in all four installations is detonation. Mark wrote about his engine experiencing detonation, from being fueled with low grade fuel, briefly before his incident. Bill Clapp had experienced several detonation issues that I believed were related to his use of two 18mm plugs in a head that should have been removed from the aircraft and replaced or helicoiled. The head has since been replaced. Bob Lester told me that he frequently leaned the engine in cross country flight. When leaning, he would lean until he saw a decrease in airspeed, and then slightly enrichen it. The evidence at teardown of BL2 showed that either this leaning technique or his Ellison EFS-3A installation was the cause of lean operation and evidence of detonation. In many installations, especially aircraft without mufflers or flown by pilots with headsets, detonation in the plane cannot be heard. It is important that pilots trained in flying Cessna 150s to lean the aircraft until the engine runs slightly rough and then enrichen it slightly, should not use this technique on a Corvair engine, especially below 8,000 feet. The difference is simple: a Continental or Lycoming with a compression ratio in the 7:1 range, when excessively leaned from cruise power settings, will experience a lean misfire in the cylinder. A lean misfire is a harmless event compared to detonation. Here, the air/fuel mixture has reached a point where it will not ignite. Conversely, leaning an engine with a 9:1 compression ratio like a Corvair has the potential to detonate the engine long before it lean misfires. This is aggravated by high available density, as in low altitudes or throttle openings yielding MAPs more than 24". Corvair engines can be leaned, it just requires an EGT and common sense. The contribution of detonation to crank failure will be the hardest factor to replicate, model or evaluate. But, successful builders will avoid detonation for a multitude of reasons, not just its potential contribution to crank failures.

    Black Misfiring Cylinders

    On both the BL1 engine and the BC engine, teardown revealed one of the six cylinders to be completely blackened with dry sooty carbon. This indicated to me that these cylinders were experiencing intermittent misfiring. Most builders accustomed to the operation of four cylinder engines are stunned at how smooth a correctly running Corvair aircraft engine is. The corollary to this is that many builders do not recognize when an engine is running on five or five and a half cylinders. Part of the reason why we have always taken the time to fly as many builders as possible in our own aircraft is that I want them to experience what a perfect running Corvair feels like so that their senses will tell them to be suspicious if they experience something less. This also goes back to the case for CNC props, which allows me to tell someone what the correct static rpm is for any known prop combination, something impossible to do with props that are handbuilt one at a time. On the BL1 engine, the black cylinder was directly traced to a poor ignition wire contact in the distributor cap. On the BC engine, the black cylinder was in the same hole as one of the 18mm sparkplugs. 18mm aircraft sparkplugs are incompatible with distributor ignition systems because they are intended for electrode gaps in the .015 to .018 range. Plugs like this produce a symptom we associate with magneto engines: plug fouling. Although it may turn out to be nothing, intermittently missing a cylinder in the firing order may have some harsh effect on the engine; at the very least, it's completely avoidable and not desirable.

    Overtemped Engines

    In ground runs, the BL1 and ML engines were inadvertently run hot enough to extrude aluminum from the head gasket area. I believe that this takes in excess of 700 degrees as a localized temp. I have a substantial amount of time running Corvair engines in excess of 500F CHT, and as long as they do not detonate, they will not extrude aluminum in the head gasket area. Thus, I believe that both of these engines were detonating during their ground runs. I replaced the cylinder heads on Bob's BL1 engine, and Mark Langford replaced a head on his ML engine. What I did not consider at the time, but find worthy of consideration now, is that this previous detonation may have had a detrimental effect on the engine. As discussed above, the BC and BL2 engines also had significant evidence of detonating previously.

    Bad Spinners Out of Balance

    I believe that very few handmade props are perfectly mass balanced. Over the years, I've worked with propellers from many different companies. At M-T Propeller, I personally did the balancing on all models of propellers. What few homebuilders appreciate is how small a weight makes a prop out of balance. Most of the repairs I've ever seen done to homebuilt spinners are several times the size of the weights used to balance 50 pound props. BL1 and BC were originally equipped with aluminum spinners. Both of these spinners cracked, and needed repairs. It's impossible to repair the spinners symmetrically. Additionally, handmade props frequently are most different down near the root section, requiring different size cutouts in the spinner, making the spinner unbalanced even if the prop has been balanced. When Bill Clapp put his replacement engine together, it was checked for dynamic balance by Arnold Holmes. He was able to make a significant seat of the pants difference. At certain rpms prior to Arnold's work, the flap handle would buzz in Bill's plane. Afterward, the handle remained still at all rpms. The computer data from the prop balancing printout told a very positive story that your senses could not feel. Arnold had cut the vibration to a third of its previous levels. Experience from Lycomings and Continentals does not apply here. If someone you knew happily flew around an out of balance wood prop on an O-320 for years without trouble, this means nothing in the land of Corvairs. An O-320 can perform a snap roll with a 50-pound metal constant speed prop on it. A Corvair, like a VW, Subaru, Jabbaru or most other alternative engines, could not even perform this trick once.

    Harsh Flight

    This is discussed at length in the Conversion Manual. Let it suffice to say here that this is an aggravating factor. We have shown people numerous times in videotapes and in person, our own Corvair powered airplane performing non-aerobatic maneuvers like wingovers, hammerheads, chandelles, falling leafs, etc. What few people realize is that the rate of angular change can be much harsher on a very hard landing, or with extremely jerky flying. While I don't think it's the major factor in any of these crank incidents, it's obviously a force to be reckoned with, and an excellent reason not to use crank extensions or any other aggravating factors in combination with it. In the case of BL1, you have an engine with a 4" extension, and a 10 pound prop turning more than 4,000rpm occassionally. To add a harsh maneuver to a combination like this is a very large force multiplier. It's important that people understand that severe, rapid, PIO in pitch or yaw can produce forces that exceed coordinated aerobatic maneuvers.

    Overweight Prop

    Not a week goes by without someone asking me if they can use some 25 pound prop on a Corvair. Again, my Conversion Manual covers this, but the heavier the prop, the greater the load on the crankshaft. Fortunately, our flight testing shows that the Corvair works extremely well with light, efficient propellers. The only one of the four that I consider the weight of the prop to be a factor in is the BL1 incident. Although the crank failed while in steady 3,200rpm cruise flight, the fact that it had seen far more rpm than this in the previous 10 hours gave it plenty of chances to start a crack that finished later.

    Prop Orientation

    The subject of how the prop is oriented on the crank has come up. I have not studied this in detail. In a quick view, it makes sense, but when you look at the past 45 years of flying Corvairs, no one has paid any attention to this, so it is a fair assumption that if there is a better and a worse way to install the prop, it has been installed both ways with equal frequency. Also, not all prop hubs made in the past 45 years are drilled in a way which allows the prop to be installed in the theoretically proper orientation. I have a lot of respect for the guy who first proposed this idea. But I think it has a long way to go before I would call it reality because the first computer model of the stresses showed the failure point at a different location than where the breaks have happened, and the fact that we would have seen other broken cranks earlier is a valid question that the theory has not addressed. I'm open minded to learn more; there's just a lot of work to move this theory from the computer screen to reality in your hangar.

    Excessive RPM

    The rpm that the engine is turning greatly influences the amount of bending stress put on the crankshaft. The BL1 engine had exceeded 4,000rpm. The BC engine turned nearly this much on numerous occassions. In general, it's not a case of having a specific redline. It's a balancing act of all the factors. The BC engine turned 3,900rpm at 180mph with a Prince 54x50 prop. It would do exactly the same speed with a Prince 54x54 while turning only 3,600. Currently, Bill's airplane has an even stronger rate of climb, and will turn the same speed at 3,500rpm with a Sensenich 54x54. Thus, the extra 400rpm of the original prop is unneccessary. The ML and BL2 engines both turned 3,100 to 3,200rpm. This is not excessive rpm. Many of the original Corvair installations on Pietenpols could turn rpm like this wide open; typical props of the day were 66x30s and 66x28s. Even on an 85mph Pietenpol, these props yielded very strong rates of climb, and flat out rpms of above 3,000. Bernie Pietenpol's writings from the '60s indicate that he occassionally saw 3,300rpm. Yet, no crank failures were reported in this era. Our own 601 turns 3,100rpm at a high cruise setting. Other than the BL1 case, most high rpm work is done with smaller diameter, lighter props. This works to offset the effect of increasing the rpm. Still, builders should ask themselves what is necessary, and use the least amount of rpm that will achieve their goals. I want to emphasize that there are other Corvair powered airplanes that have logged hundreds of hours trouble free with small diameter props. What these successful installations were missing were many of the contributing factors discussed here.

    Standard vs. .010 Under Cranks

    If properly ground, a .010 under crank is no different than a standard one. The BL1 engine had a standard crank in it. Yet every hour we have flown since 1998 has been behind a .010 under crank. A combination of aggravating forces can break any Corvair crank, standard or reground. Additionally, a properly ground crank has proven to be a reliable performer in hundreds of hours of service in our own personal aircraft. The BC crank was improperly ground by his local shop. The ML crank has not undergone a close enough inspection to draw a conclusion. The BL2 was magnafluxed before I installed it in Bob's longblock.


    Magnafluxing is the approved method of crack detection on aircraft cranks. The Corvaircraft archives contain a lot of comments I have made on magnafluxing in detail. Most magnaflux jobs that cost less than $100 simply provide a false sense of security. A properly done test takes very expensive equipment, and takes a trained technician more than an hour. The BL2 and ML cranks were both magnafluxed prior to their installation. I am absolutely positive that no detectable crack existed in the BL2 crank prior to its installation. No one should come to the conclusion that these cranks had pre-existing cracks, nor should you believe that having a crank pass a proper magnaflux test is a license to use a number of aggravating factors without consequence. The proper perspective is that the aggravating factors are enough to crack even a perfectly good crankshaft. A good magnaflux inspection will assure you that your crank does not have an existing flaw, but it will do nothing to prevent a failure in your aircraft if you employ a number of aggravating factors.

    A Note About Internet Experts

    As I pointed out earlier in this piece, there were "experts" on Internet working to convince builders to use insanely long extensions. I readily agree that there are many great things about the Internet as a source of information. The case of Corvair crankshaft issues would be a textbook example of what I dislike abouut the Internet. The owners of the aircraft discussed here are not the problem in my Internet story. These builders simply looked at the available information and decided to build their own plane, conforming to what they felt was a reasonable combination. They had the freedom to build their airplane their way. The people I have the greatest issue with are those who actively worked to discredit my warnings while having no intention of flying the systems they advocated. Mark Langford designed his own rear starter simply to serve his own needs. He used every opportunity to explain to people that it was not flight proven, and was truly the essence of experimental. Its popularity was driven not by Mark's promoting it, but by endless Internet dissertations about the idea of it, and how it theoretically improved the CG of the engine and made effective streamlining possible. This commentary largely came from people who had no running engine, nor any idea of the complexities of the oil systems that would be required. If you copied ML's system, carefully consider whose promotion and logic talked you into it.

    The downtime on the four crank failures varies from four days for Bill Clapp's plane to perhaps a year's worth of solid work it will take to make Bob's airplane airworthy again. I am sure that these incidents will lead to builders following more closely what has been proven to work before. In the past, I've been accused of hampering other people's creativity by suggesting they use what's proven. It's a fair observation that my loudest critics haven't built or flown anything. Hopefully, people who choose to read the Internet will be far less tolerant of these types of people in the future.

    A Few Closing Thoughts

    Bill Clapp's airplane is currently being converted to one of our Front Starter Setups, a system which utilizes one of our 3" Prop Hubs. He is currently flying a Sensenich 54x54 prop. With one of our new KR cowlings, his installation will then conform very closely to how I have always advocated engines be built: Simply, with front starters and short prop hubs. Bill is a very active pilot and will put more hours on the combination in the coming year than most pilots will in five. Steve Makish also telephoned to say that he is planning on removing his extension and switching to the new cowling. Bob Lester has said that when his airplane is rebuilt, it will never again use an extension. Mark Langford has plans to do further investigations, but already has the parts on hand to build a standard 2,700cc engine. Hopefully builders will learn from the discussion and have greater appreciation of how no aspect of building and operation is independent from any other. I have always carefully considered how people are going to use the information and products we offer, and attempted to reasonably gauge builders' capabilities and experience. We'll continue to do so in the future. We'll see you at Oshkosh.

    HOME  Zenair 601  At The Hangar  Turbocharging  Thrust Testing  Engine 4 Sale  Corvair College #8  Corvair College #7  Corvair College #6  Oshkosh 2004  Sun 'N Fun 2004  Corvair College #5  Daily Q&A  Flying Corvair Planes!  News  Mission Statement  Hangar Gang  Conversion Manual  Online Catalog  Maximum Horsepower  All About Corvairs  Corvair History  Engine Specs  Sonex Installation  RV Application  Why Fly Corvair?  FAQs  Liability Statement  Carb Ice  E-mail William 

    Copyright 2005 William Wynne Web Design by Aviatrix