New Limits vs. Service Limits
Differences Between Aircraft Engines and Auto Conversions – from William Wynne’s Speaking Notes
By William Wynne, EAA 331351, WilliamTCA@aol.com
The title of this article contains two terms that aircraft mechanics use all the time but aren’t widely understood in the automotive conversion world. They can be casually tossed around in hangar flying sessions. But A&P mechanics use them very carefully because what they define has a very strict legal meaning when you write it in the logbook of a certified airplane.
"New limits" on an engine is loosely defined as being the factory tolerance for the range of acceptability of a part or assembly to go out the door as being brand new. All certified engines have this data published. It’s generally referred to as the “table of limits” on an engine.
Many people mistakenly think that when they read “overhauled” or “zero (hours) since major overhaul” (0-SMOH) that the engine they’re looking at is as good as new. It's not! All the parts inside haven’t been returned to new limits. They’ve actually been returned to a more liberal set of specifications called the "service limits." These are also known and published for certified engines. To write the word “overhauled” in a logbook, the engine must meet only service limits. This is a very serious thing that neither mechanics nor the FAA takes lightly. 0-SMOH is often mistaken for the term "zero timed," which actually means to return to new limits, but isn’t the case.
If a Lycoming or Continental engine has a published 2,400-hour lifespan, a “new” engine or a “zero timed” one will almost always make this mark without trouble or further work. However, the same model engine that is just “overhauled” is far less likely to go the full time, because it’s starting out with an acceptable amount of wear that the new engine doesn't have.
It’s very important to note that none of the engines are expected to break or be less reliable than the others. We’re only speaking of wear, not failure. With aircraft engines, getting "tired" early is okay; breaking isn’t.
In the land of automobile conversions, most engines don’t have a hard and fast set of service limits. We have some indication from the limits in the back of many service manuals, but this usually doesn't cover the whole engine. To really know, we would have to tear down a lot of engines from the fleet and look at a lot of statistics. In the Corvair engine world, I’ve seen more of this than just about anyone else, but my training in aviation and statistics tells me that our sample size is still too small to have very accurate service limit numbers. We can have close estimates, though.
How does this affect your engine? For the most part, anyone overhauling his auto conversion, replacing virtually every moving part with a new one, will have a "zero timed" engine by default – it will be returned to new limits. The fastest way to return any part to new limits is to replace it with a new part; if the part is correctly made, it’s within new limits, job done. Pistons, rings, lifters, valves and springs, bearings, etc., all fall in this category. A lot of the reworked items (such as cylinder bores, valve seats, and connecting rods) are also returned to new limits with a proper overhaul. When cylinders are bored oversize, the clearance is set to new limits. Same goes for connecting rods, crankshafts, reground cams, etc.
The one item that we use in engines again and again that needs to be examined is the case or block. Just because the case or block has 100,000 land miles on it and has had nothing done to it, that doesn’t mean it’s worn. It has been shown that a very high percentage of Corvair cases, all of them at least 40 years old (and well used), will still meet new limits, as we see them in the original General Motors (GM) drawings. The only cases that are often out of round on the #2 and #3 mains are 140-hp engines (as compared to the 110-hp engines that share the same case), as they were often run much harder in cars.
Editor Pat Panzera checking the cam gear for run-out while zero timing his Corvair engine.
Since my forte is with the Corvair engine, the balance of this article will be specific to the Corvair. It may not be 100 percent applicable to your engine of choice, but I hope you continue reading and can apply some of what I write to your project.
Horizontally opposed, air-cooled aircraft engines are absolutely designed with far looser tolerances on crankcase fits than automobile conversions, because the case is designed to have some degree of flexing in operation. But what are the service limits of the Corvair? I have a feeling that if we took five engines – each with more than 500 flight hours – and measured them, the majority, if not all, wouldn’t meet new specifications. Yet all this means is that the flying engines, because they work and we aren’t seeing failures like spun bearings, are operating within an acceptable limit, which we can consider the service limit.
I applaud and support the research into the fit of parts in the bottom end of the Corvair engine that's being done by Roy's Garage. I think it will yield valuable data. However, I don’t want builders to misunderstand Roy Szarafinski's personal limits as the only way to build an engine.
Before anyone thinks that I’m advocating sloppy building, let me say that there’s a big difference between accuracy and close tolerance. I would like everything to be as accurate as possible, but I’m not always a fan of super tight tolerances. Example: When I designed a front bearing for the Corvair engine to take the propeller loads off the crank, in round one, I picked a very tight tolerance to the crank. The nature of the design indicates the bearing is perfectly round and stays that way as it heats up. In the middle of the first winter's operation, my beta tester called to say that the engine wouldn’t rotate on the ground. Tight tolerance had led to a cold seizure once the ambient temperature dropped enough. Round two, slightly looser, now works like a charm. Notice how this two-step process began to define how wide the limits for this part should be.
The test run of William's fifth bearing, the obviously billet aluminum part just behind the flywheel.
A lot of people reading this may think, if all we do is tighten tolerances, the engine is going to be like a certified engine. This notion is off the mark. Air-cooled certified engines tend to be very loose inside. The center main in a Lycoming is allowed close to 0.005 inch out. Keep in mind the difference between accuracy and tolerance.
I don't know what the spec on the GM drawing is for main bearing bores. I’m pretty sure that Roy was looking for 0.0005 inch out of round and he was rejecting cases that were more than 0.001 inch. This is a personal choice, but consider this: I’ve seen a great number of very beaten 140-hp cases that had 0.002-inch and 0.003-inch out-of-round #2 and #3 bearing bores. Yet these were running engines when we took them apart. None of them showed any sign of spinning a bearing. I’m not advocating building an engine from such a case – I'm just pointing out that a case with a 0.0012-inch tolerance isn't likely to break in five minutes in a plane.
It would be easy for a builder who hasn’t seen a lot of engines and doesn’t have a panoramic view of flying planes to take the new limits and falsely conclude engines that aren’t to that spec may be prone to failure or somehow substandard.
The other part of the engine that Roy has spent a lot of time looking at is the fit between the cam and crank gears. On the drawings, it specifies 0.002-inch to 0.004-inch clearance. This is obviously the tolerance on a new part, one that GM was willing to send out the door in a new Corvair that was going to have a warranty and be expected to see 100,000 road miles at an average cruise rpm of 3,500 or so. Obviously it was a good number because I’ve taken apart 300 or so cores; if an engine had oil in it, it didn't have a broken cam gear. Again, we don't know what the service limit is. It’s very likely a lot more liberal than the new specification. Gear clearance is a hard number to check, but it has been shown by flying examples that as long as there is backlash, there is clearance and it will work.
An aluminum cam gear tooth can be seen mating with the steel crank gear tooth, looking through one of six bolt holes in the crank flange.
The primary reason that an engine wouldn’t have the proper cam gear clearance might be that the crank had been incorrectly ground. A local respected car guy whom Roy knows said that when reading the 0.002 to 0.004 new limits on the drawing, he would only use 0.003. He may very well be a nice guy, but right off the bat I dislike people like that. He probably has never owned nor built a Corvair, yet he’s somehow saying that he’s smarter than the people who designed the engine. Just picking the number in the middle doesn't mean anything. Careful engineering analysis might show that 0.004 inch was by far the best number for our application. As an A&P mechanic, I’ve seen countless incidents that whenever a published acceptable range is from, say, 5 to 10, there will always be a guy who wants to tout himself as a better human because he spent several hours making the item 7.5 instead of spending two minutes ensuring that the parts are inside the correct 5 to 10 range. Often, people who engage in this are prone to getting 'creative' in other ways around engine to express something about their personality. There is no valid reason for seeking the middle of an established limit. Anywhere in the range is acceptable.
An article recently brought to my attention referenced the topic of "infant mortality" in aircraft parts. The article devoted very little space to a topic that was covered in many classes I attended while a student at Embry-Riddle Aeronautical University. Infant mortality is a descriptive term for a part breaking in the first stages of its service life. Note that it’s describing a part breaking, not wearing out. It’s generally understood that such a part wasn’t made to specification (defective) or was installed wrong. The life cycle analysis assumes that the design is good and proven, otherwise the failures wouldn’t distribute themselves in a normal manner, and no predictions could be gathered. The simplified chart may falsely give the image that infant mortality on a 2,000-hour engine is a serious possibility out to 500 hours. Everyone who has studied this subject in depth knows this isn't accurate. In reality, infant mortality on engine parts almost always happens in less than one million cycles, which is well under 40 hours at typical aircraft engine rpm.
This is a big part of why the FAA puts a 40-hour flight restriction on your experimental aircraft, not a 500-hour one. The topic is a good one; the article can be considered a good general introduction, but statistical analysis of aircraft part failures is a very complex field. The average novice engine builder can spend his time with a far greater result just working to improve his basic workmanship. A person who is concerned about infant mortality issues who doesn't know how to use a timing light will drastically improve his personal risk management by first mastering the latter.
As a final thought: One of the first jobs I had was working in New Jersey for a man named Werner Habberman. When I first met him, he seemed to know a lot about machines for a guy who put aluminum siding on houses for a living. After I got to know him better, he showed me photos of what he did in 1941-1945. He had several hundred missions in Me-109Es and Gs on the Eastern Front. He didn’t speak of it much – he just let me look at the black and white photos in his hallway and ask a few questions. He had me change the head gasket in a Chevy Vega he owned, and it sparked some conversation on engines. One of the things he spoke of was that the Soviet engines would operate at minus 40 degrees, but Daimler-Benz 601s wouldn’t. The Germans put some effort into recovering Soviet engines to send them back to Germany to discover their secret. Before being a pilot, he was a mechanic, and he was involved in this effort. The big secret turned out to be nothing more than internal clearance. Thirty-five years after the fact, he still derided the crude Soviet machine work. It was hard for him to accept that the Soviet equipment had functioned better. Lost in the memory, he went on for a minute about how the cam lobes weren’t even polished. He stopped suddenly and just looked at me. Today, 30 years later, I know that he had just caught himself sharing something with an idiot teenager in suburbia who's understanding of the Eastern Front was building models of machines that he had placed his very life in. I worked for him for a while after that, but he never spoke of anything to do with WW II ever again.
William Wynne (WilliamTCA@aol.com, http://FlyCorvair.com) is best known as an engine guy and specifically known as the Corvair Authority. In the last 20 years, he has potentially helped more people toward their goal of building and flying an affordable homebuilt than anyone you may know personally. In addition to pioneering the concept of a group-learning environment specifically aimed at engine rebuilding for the novice and specifically known as a “Corvair College,” the most recent of which ended a month ago, William is a prolific forums and workshop presenter at EAA AirVenture Oshkosh and Sun ’n Fun Fly-In at Lakeland, Florida. William comes to us by way of CONTACT! Magazine.
Permission to reprint is granted to all. Please credit William Wynne and CONTACT! Magazine. This article was intended for CONTACT!, but I felt it deserved a broader audience. ~Pat