- Secondary Failure: Background -
An Introductory Look at the Problem
In October, 2002, another RotorWay 162 crashed as the result of the in-flight fracture of the secondary transmission shaft. That ship was equipped with the two-bearing, 30mm secondary configuration, and the ProDrive™ toothbelt system. This failure was by no means the first, nor would it be the last.
In February, 2003, the owner of that ship sent EPI his broken 30mm secondary hardware and engaged us to (a) determine the cause(s) of the secondary shaft failures and (b) to devise a solution, preferably one which does not require modification of the airframe.
Later in 2003, EPI was the first (of many) to examine the broken 35mm secondary from the prize-winning 162 which crashed in July. That ship also had the ProDrive™ belt system, and the 35mm secondary shaft had only 55 hours on it.
EPI did a substantial amount of examination, inspection and analysis of the causes of this failure problem. As a result of that work, EPI developed a secondary transmission upgrade which solved the fatigue issues with the secondary. EPI rebuilt over 30 secondaries with that system. Because of rapidly-rising costs to produce that system and a substantial resistance to price increases, EPI has since withdrawn the product from the market.
However, the results of the analysis work on the secondary transmission are presented in the following few pages in order that the information will remain available to the community.
The fracture of the secondary transmission shaft in the RotorWay power transmission system results in an instantaneous loss of power to the main rotor. If the failed system is a ball-bearing configuration, the failure frees a substantial hunk of steel in the upper part of the dirve machinery to find its way down into other components and cause more havoc. If the failed system is a spherical bearing configuration, the ship can lose all tailrotor authority as well.
That failure generates the very definition of an in-flight emergency, which requires the flawless execution of a full-down autorotation (most likely the pilot's first full-down) on random terrain. The history of successful outcomes is regrettably short.
RotorWay's official position is that their 35mm secondary shaft has solved the secondary shaft breakage problem, and in Late 2003, they published a lengthy study to justify that position. However, it is important to understand that the factory 35mm secondary shaft is only claimed to fix the secondary breakage problem for ships equipped with the standard factory roller-chain main rotor drive. Further, even with a chain drive, the factory secondary shaft has a published life limit of 500 hours.
There have been breakages of the 35mm shaft, but so far (as of mid-2007), there has not been a KNOWN failure of a 35mm shaft in a ship equipped with the factory-supplied roller-chain main rotor drive, so for now, it appears as if the factory 35mm shaft is a solution for a chain-drive ship.
After lots of 35mm secondaries had been sold, the factory quietly replaced the press-fit upper bearing spacer with a clever slip-fit specer held in place by o-ring friction. That fix eliminates one of the sources of fretting from this problem-fraught design.
NOTE: The chain-drive 35mm shaft that failed at the factory in February of 2003 was a failure below the 8-sheave V-belt pulley, in the 1" diameter shaft section. It was possibly caused by damage to the shaft in at least two previous destructive tailrotor strikes. Another possibility is fretting under the lower bearing. Based on inspections of secondary shafts in service, fretting under the lower secondary bearing is a very real issue However, the bending loads there are substantially lower than at the top of the shaft.
Keep in mind that the factory has declared previous configurations to be "the solution" too. If the factory has finally solved the problem for their original equipment chain drive, that is a good thing, and it is long overdue. If you accept RWI’s claim of the 35-mm fix, you must keep in mind that their claim applies only to ships with the factory 35-mm, chain-drive.
The January 2004 issue of the Rotorway quarterly newsletter Sport Helicopter contains an article which restates the RWI arguments in defense of the 35mm secondary. That article concludes with this important statement, which provides the backdrop for the discussion of the secondary failure problem:
"It should be understood that any deviations from the kit’s standard design, as RWI supplies it, constitutes a redesign. IT IS NO LONGER A ROTORWAY DESIGN".
In other words, "If you’re flying with a main rotor drive system which is other than the factory chain drive on a factory 35mm shaft, then any drive problems which occur are your responsibility to deal with".
AND WE AGREE WITH ROTORWAY on that point. RotorWay is not responsible, and should not be considered responsible, for failures to anything except the systems which they supply. (Of course, there have been several failures of 30mm chain-drive configurations which RWI also declared to be "satisfactory".)
As the Failure Database suggests, the failure rate of the 30mm shaft seems fairly evenly distributed between chain-drive systems and ProDrive™ tooth-belt systems. It is our opinion that the 30mm design, in any of its versions, is inadequate for the task on a 162, and so the focus of this analysis is the 35mm shaft.
To begin our analysis of the problem, we conducted detailed, magnified visual inspection of several failed secondary shaft pieces. Those inspections strongly indicated that the failure was in bending fatigue mode, and that the fractures originated from areas of severe fretting on the surface of the secondary shaft.
EPI contracted with an independent failure analysis lab to conduct a metallurgical examination, a chemical analysis and a failure analysis of the broken secondary shaft sections. That lab confirmed our determination: bending fatigue failure, originating from a severely fretted portion of the shaft surface.
The lab inspections also determined that the factory shaft material is almost a high-grade chrome-moly steel (E-9310). The autorotation clutch journal has been hardened to approximately 60 HRc but in all other locations, is in a relatively soft condition (therefore not very strong). This material is customarily used for extreme-duty gears, but not commonly used for high-fatigue shaft applications.
The difference between the failure of the 30mm shaft and the 35mm shaft was the location of the fretting from which the failure originated. In the two-bearing, 30mm configuration which EPI examined, the fatigue crack originated from a severely-fretted area at the surface of the shaft, in a transverse plane roughly in the middle of the inner race of the upper bearing of the two-bearing set.
In the spherical-bearing 35mm configuration, the fatigue crack originated from a severely-fretted area at the surface of the shaft, in a transverse plane under the lip of the press-fit seal-sleeve below the upper bearing.
The high cantilevered load on the upper end of the secondary shaft produces a measurable deflection of the secondary shaft toward the main rotor shaft. When the shaft rotates, the deflection at any given spot on the shaft cycles from zero to full-tensile to zero to full compressive back to zero in one rotation.
The fretting is caused by microscopic cyclic deflections and the resulting relative motion between the secondary shaft and the inner race of the upper bearing. In the case of the 35mm shaft, the pressed-on sleeve below the upper bearing provides an additional area for destructive fretting. Later factory 35mm systems introduced a slip-fit spacer held in place by o-ring friction which cleverly eliminates this source of fretting.
The severely fretted areas on the shaft surface cause the fatigue strength of the secondary shaft to deteriorate to a level which is well below the applied cyclic stresses.
On the 30mm shaft, the fretting is severe on both chain and toothbelt drive systems. On 35mm shafts, the fretting seems to be worse on toothbelt drive systems.
In addition to the high cantilevered loads, there is a significant amount of vibration energy present in the entire drive which aggravates the stress and deflection problem on the secondary. The primary source of that vibration energy is the engine, which operates around 4250 RPM. At that speed, the engine torsional excitation (141.7 hz.) is close to one of the drive system's four major resonant frequencies.
The combination of (a) the high cyclic bending load and the cyclic deflection it produces, together with (b) the torsional vibration produced by the proximity of the engine excitation frequency to one system resonant frequency increases the fretting rate and the resulting loss of strength of the shaft.
Because the only known 35mm failures have been on ships using the ProDrive™ belt system, and since there are so many ships equipped with that system, we think it is important to make a clear explanation of the substantially greater fatigue loads and deflections that the ProDrive™ belt system imposes on the secondary shaft and bearings. When that overlaod is combined with the fretting-prone shaft design and the high level of torsional vibration in the system, it can produce an alarming fretting-rate, leading to rapid shaft failure. The documented failure rate is sufficient evidence of that fact. The analysis presented later in this section make it unquestionable.
Although the Factory claims otherwise, it is our opinion that the factory 35mm shaft design is susceptible to fretting in all applications. Clearly the rate is slower with the chain drive, but we think it is unquestionable, given the design, that the 35mm shaft will fret on a chain drive. We have no hard evidence to support that opinion, but as soon as we do, it will be published immediately. Note that Later factory 35mm systems introduced a slip-fit spacer hald in place by o-ring friction which cleverly eliminates this source of fretting.
As of May 2005, only 18 documented failures had occurred out of an estimated population of more than 400 helicopters flying. That means there are lots of shafts that haven't failed yet.
The nagging question is, however: "How much time can safely be put on the shafts currently in service"? We think that question is impossible to answer because of the large number of variables which effect the cyclic loading applied to the shaft and the rate at which fretting corrosion occurs. Here are a few of those variables.
The ProDrive™ toothbelt main rotor drive system significantly increases the cyclic bending loads applied to the secondary shaft (this fact is documented in detail later in this section).
Tooth Belt Preload: If the ProDrive system is installed, the level of preload in the belt system dramatically alters the cyclic bending stress applied to the shaft as well as the dynamic deflection which generates the destructive fretting corrosion. The preload is set by a shimming operation, at whatever ambient temperature prevails when the installation occurs. The existing preload is measured by an imprecise "belt-deflection / fish-scale" test, and therefore can vary significantly from one installer to another. Then, the temperature sensitivity of the ProDrive™ tooth belt system significantly increases the preload value as the drive heats up in operation. (As temperature rises, the the belt shrinks quite a bit, the large aluminum sprocket expands quite a bit, the iron sprocket expands a bit, and the airframe expands a bit.) Now, if the installation takes place in a cool hangar in winter, at, say 50°F, and is then operated in a summer ambient of, say, 85°F, the increase in preload is downright scary.
The 30mm shafts are more susceptible to failure because of the higher stress levels they encounter for the same applied loads.
Secondary Shaft Manufacturing: The tightness of the bearing press fit alters the rate of fretting, which effects the loss of fatigue strength of the shaft. In addition, The factory assembly techniques leave a lot to be desired. For example, every factory system we have taken apart has substantial scoring damage to the shaft surface in the area under the bearing inner race, suggesting the hard bearing was forced into place on the soft shaft..
Secondary Shaft Metallurgy: The allowable variation in the chemical content of the shaft material causes significant differences in shaft hardness after heat-treatment. The shafts with lower surface hardness will fret faster than harder ones. Add to that the fact that the factory material seems to be deficient in one element which is a critical component of the material the factory claims to be using.
Secondary Shaft Alignment: alters the bending stress applied to the secondary shaft, especially in the tooth-belt drive system.
V-Belt Pulley Eccentricity: A few builders claim to have detected substantial eccentricity in the secondary shaft 8-sheave V-belt / autorotation clutch pulley. We think that core-shift combined with machining errors can be the cause of the sheave eccentricity, which can produce a substantial cyclic level in the upper and lower secondary bearings, as well as an additional fatigue load on the shaft.
One owner complained to us that he had installed a new 35mm secondary system from RotorWay and failed the upper secondary bearing in 5 hours of hover testing. Post failure examination revealed nearly 1/4 inch of eccentricity in the pulley.
Airframe Flexing: We think it would be revealing to instrument an airframe and measure the bearing-support deflections under the influence of the dynamic drive loads.