Prop Extension Failure
October 1994
(Click on images to enlarge)
At Oshkosh this year, we were shown
photographs of a prop extension that had failed, catastrophically, resulting in the loss
of the prop and a forced landing that seriously damaged the E-Racer which, while not an
RAF design, is a similar pushet. The pilot and passenger were not hurt.
The engine was a Lycoming 0-360, 180hp. The prop was a B&T prop and it was driven by a Brock prop extension 6" long with a 7" diameter flange at the prop end. The fracture started right in the radius between the barrel and the aft flange and propagated across the extension. This fracture has been characterized by experts as being a high cycle, fatigue failure. The total time on this prop extension (and on the aircraft) was 72 hours. What caused this failure? Is it something we should be worried about?
A little history may be helpful here. Several years ago, a good friend who was an excellent engineer and VariEze builder, Bob Beard, experienced a large vibration while in flight, shut it down and glided to a safe landing. He discovered that his 8" long prop extension had an enormous crack in it. (See photo). He analyzed the prop extension and found that it was machined from 6061-T6 aluminum instead of the required 2024-T351 aluminum. This happened on his original design aircraft, the Two-EZ , a large four-place similar to a Long-EZ. He had a Lycoming 0-360, 180hp engine and a wood prop.
About the same time, Danny Meyer was flight testing his Velocity, also a pusher, when he had almost exactly the same experience Bob Beard had. It turned out that both prop extensions had come from the same source. Both were made from 6061-T6, both were 8" long and both were using Lycoming 0-360s.
The difference in strength between 6061-T6 and 2024-T351 is only 18-20% so although the wrong material may have been the cause of this problem, at least on an 8" long prop extension, 18-20% is not much margin of safety. Bob Beard designed a prop extension that had a 4.5" diameter in the middle as compared to a 3.25" diameter on his previous extension. He sent this editor a letter and a drawing of his new design and said that this 8" long extension would be just as stiff as a standard Brock 4" long extension and that its natural frequency should occur above 4000 rpm.
We purchased a billet of aluminum (2024-T351) and machined a 9" long modified Beard design which has a 5" diameter in the middle and which has been tested, in flight, to show that peak stress occurs at an rpm that is out of the normal operating range of the engine. The problem with this design is that it does not lend itself to economic manufacture.
We have borrowed a torsional order analyzer. This is a magic box that has a built-in x-y plotter and receives a signal from a magnetic pickup which is mounted close to the teeth on the starter ring gear. Basically, this machine measures the speed of each tooth passing by the magnetic sensor. As the engine drives the prop, it speeds up and slows down with each firing stroke and each compression stroke, this causes the crankshaft, prop extension and prop assembly to twist like a spring. This "spring" winds up and unwinds many times per second as the engine drives the prop.
Now, obviously, the magnitude of this windup/unwind action is very small. In fact, this machine measures the rotational displacement in milidegrees, that is to say, thousandths of degrees. One of the uses of this machine is to determine if an engine/prop combination should have a "yellow arc" on the tach. A Grumman Tiger, for example, has a "yellow arc" from 1500 - 18OOrpm. This means that the pilot should not operate within this yellow arc. He may pass through it in either direction but must not fly within the yellow arc.
We are concerned that there may be a yellow arc on some of our RAF designs and we have spent many hours flying several airplanes and a bunch of different engine/prop extension combinations. We have talked to experts in this field and the consensus is that a light weight, low inertia, wood prop simply cannot damage a Lycoming aircraft engine - good news! Introduce a prop extension, particularly an aluminum, spool-type, prop extension, and maybe you can have a problem! It turns out that a spool-type, aluminum extension is relatively soft, torsionally. It also turns out that a crankshaft, prop extension, prop assembly is what is called a first mode shape. This means there is only one node (a node is a point where there is no action or movement - if you grab a spring with one hand at each end of the spring and twist it, someplace in the spring, there is no movement - this is the node). It further turns out that the node in this assembly usually occurs between the crankshaft flange and the propeller. That is to say, most, if not all, of the twisting we are measuring takes place within the prop extension.
With all of the above in mind, we set out to run in-flight tests on Long-EZs with Lycoming engines, 6" long prop extensions and wood props. An 0-235 powered Long-EZ categorically does not have any measurable problem with a 6" aluminum spool-type prop extension. The same is true of a pusher, such as a Defiant, with an 0-320 and a 6" prop extension. It may not, however, be true that an 0-360 with a 6" aluminum spool-type extension on a pusher is as free of problems. (A Long-EZ with any engine larger than a Lycoming 0-235 is not approved by RAF).
We have not fully analyzed all of the data and we plan on generating a finite element model to help with this analysis. At this time, we are unable to say (as we can with the 0-235 and the 0-320) that a Lycoming 0-360 with a 6" or longer prop extension on a pusher-type aircraft is completely safe. Some facts: If you are in the market to buy a Lycoming 0-360 (for your new Defiant), we strongly advise that you purchase one that is equipped with a 6tb order damped crankshaft. We have designed, and are testing, a couple of prop extensions that show promise to eliminate this problem, however, there has been only one failure of a Brock prop extension with many hundreds out there in the field accumulating hundreds and, in some cases, thousands of hours. We will continue to test and evaluate and keep the builders and flyers informed.
The prop extension that did fail had three strikes against it. First of all, the radius between the flange and the barrel of the spool-type prop extension was too small. Other prop extensions we have examined, including several other Brock extensions, have 1/4" radii. The failed extension had only a 3/32" radius (less than half the normal radius). Also, in this radius, there were machine marks, tool "chatter" marks, in fact. Expert opinion says that chatter marks in a highly stressed part are bad news. These chatter tool marks are longitudinal "ridges" and are torsional stress risers. Also, the forward face of the prop flange was in the same plane as a change of inside diameter and this area had a sharp radius.
None of these features are good news - all of them in one prop extension are probably bad news. Add to that the possibility of a slightly out of balance prop and then throw in the possibility that the engine/prop extension/prop may have been running in resonance causing maximum stress in the aluminum prop extension.
The torsional order analyzer shows the rpm at which peak stress occurs, if there is such a point. We tested a long-service Brock 6" long prop extension on a Lycoming 0-360-A4A, 180hp engine (with no 6th order dampers) and measured a peak torsional displacement (windup) of 20 milidegrees at 2770rpm, yet at 25OOrpm, the peak displacement was only 3 milidegrees. Running continuously at 2770rpm in this pusher aircraft would probably fail this prop extension. This same test was done using all the same parts, but with a Lycoming 0-360 with 6th order dampers installed, and the peak displacement was only 3-1/2 milidegrees! With our modified Beard prop extension, these numbers changed significantly even with no 6th order dampers. Peak displacement is only 12.5 milidegrees at 2870rmp! (At 2770rpm, maximum displacement is only 6 milidegrees). This data is all the more impressive when you consider that this prop extension is 50% longer than the 6" Brock extension.
We have designed, and are having made, a 6" long prop extension that we believe will eliminate any problem associated with the 0-360 Lycoming. It has not been tested yet and is not available at this time. We will report on its performance in the next CP. We would like to state that a correctly designed prop extension should run virtually indefinitely because peak stress would be below the maximum allowable stress. This is the key to the whole problem - the maximum allowable continuous stress must = be exceeded.
We have had the cooperation, not only of Ken Brock Mfg. in this endeavor, but also of Woofter Manufacturing (formally Woofter Custom Metal Fabrication) of Pembroke Pines, FL. We would like to thank Judith Saber of Woofter Mfg. for all of her help. She has machined and sent to us for testing five different prop extensions and she is currently machining a proprietary design which we hope to test soon. If you have not seen a Woofter prop extension, you are in for a treat! Judith runs the CNC lathe and machines all of the various prop extensions sold by Woofter Mfg.and all of them correct the problems mentioned in this article. The radii are at least 1/4", there are absolutely no machine marks of any kind, and the I.D. has a really clever "S" curve transition from the smallest diameter to the diameter that fits your crankshaft. The workmanship is absolutely first-class and we are very happy to report that just as we were going to press with this CP, Ken Brock Mfg. has decided to order prop extension from Woofter Manufacturing. Stay tuned!