In writing an article, especially one that comes across as critical of a product or business, an author takes a chance that a number of readers will have their feathers ruffled. This is especially true in an industry such as ours where individuals often form strong opinions and loyalties in and of the marketplace. In the first article (Reduction Drives: An Engineering Perspective, November '92) I came out somewhat critical of the reduction drive industry and the lack of engineering savvy our company encountered in searching for a reduction drive for the engine development project we're involved in. Having said that, we expected a less than friendly response to the article from manufacturers and loyal customers.
Imagine our surprise then when out of nearly sixty letters and phone calls only one was negative, one was a little of both, and the rest were very positive and reaffirming. The respondents included a couple of reduction drive manufacturers, pilots and aircraft owners, a couple of retired transmission designers for Ford and GM, several aircraft mechanics, doctors, and attorneys, among others. The most frequent comment was that often these individuals found suppliers (airframe, reduction drive, engine, etc.) to our industry who seem to hide behind the "Experimental" label, as if it was an excuse for not doing all that was reasonably possible to assure the safest product.
Now does this mean that people expect suppliers to have large design staffs, testing their product(s) for years before bringing them to market? Of course not, nor was anyone suggesting that. Neither was any form of certification or regulation suggested - let's face it, that's why we're in this industry in the first place; to get away from all that bureaucracy. But in the same breath it is clear that many folks would like to see more responsible design work and documentation in producing aircraft and their components.
In writing this second article my purpose is fivefold. First I would like to answer one or two comments made about my statements regarding engineering and manufacturing. Second, there were many question asked regarding the rotary engine's torsional vibration characteristics, and third, a few individuals asked for more information about the types of testing which may be required to substantiate a reduction drive.
Finally, the last two items will deal with the progress on our prototype box and provide an introduction to Hayes Rotary, a small northwest company dedicated to the rotary engine and its application to aircraft, marine and of course automotive vehicles.
First, engineering and the production floor. Two of the respondents suggested that I placed too much emphasis on the engineering of the product and did not give enough credit to those individuals who have extensive production experience and design drives (or other components) based on their years of dealing hands-on with the materials and processes. To this I say true and partially apologize for the oversight. Having been a machinist for a number of years before getting my engineering education and degree, and having on many occasions been in charge of engineering design staffs, I have from time-to-time said that I would prefer to have working for me a machinist who knows the design process rather than a staff of engineers who have little or no fabrication experience. Many industries are plagued by problems associated with over-engineered items which cannot be made due to the lack of understanding and foresight of the manufacturing process during the design stages of the products' development.
But at the same time I would never wish to depend on a product that had no engineering design or analysis to back it up. The experienced fabricator or machinist may have the expertise to design a functional reduction drive but in general he does not have the background or experience to prove the drive's reliability, its life expectancy and its functionality at other than ideal conditions. Manufacturing know-how is important and critical to the development process, but is only a small part of the products' design, analysis and substantiation. Imagine an individual who designs an outwardly good looking airplane but neglects to take into account the effects of gust loading on the wing structure, or the effects of an extended fairing on the stability envelope - the airplane may be functional but at some point down the line could provide the owner with a nasty surprise.
The same argument holds for the smaller components also. Responsible engineering needs to be a primary part of the development process regardless of the product. The machinist may have an applicable knowledge of the materials he works with but generally knows little of the design requirements, engineering mechanical properties, behavior of components under load, etc. Practical fabrication experience usually provides insight to some design steps but rarely the entire process.
Well, enough of that, now let's talk a bit about the rotary engine and torsional vibration. Some individuals seem to be of the opinion that if an engine can sit on a test stand or on an airplane, and idle without any vibration, that it does not have a torsional resonance problem - as if just because you can't feel it, it isn't there. Nothing could be further from the truth.
In fact, every piece of rotating machinery has a natural frequency where torsional feedback can become a serious problem. It's just that in some systems (such as Lycomings) the vibration is more obvious than in others. If the natural frequency of the system is low, as it is on most conventional aircraft engines, the effects torsional feedback can be felt and observed as the magnitude of the reaction force starts to build. If however the natural frequency is higher, the resonance of torsional feedback may not be felt until its too late. Having a high natural frequency means that things happen very quickly per a given amount of time so the effect of a torsional vibration may build faster than an individual can feel or react to. Now let's briefly examine how this can affect the parts.
Looking at a few mathematical terms, let's let "g" represent the natural frequency of the system and "w" the frequency of the applied load. The term (w/g) is then the frequency ratio. The next term is called the "Magnification Factor" (MF), and is defined as the ratio of the amplitude of steady-state vibration to the static deflection which is caused by the amplitude of the periodic force. Mathematically this is represented by:
MF = 1/[1-(w/g)2]
In slightly more plain English this means that the "Magnification Factor" can be envisioned as a force multiplier or load factor, and that its value gets very large as the energizing force-pulse-rate starts approaching the natural, or resonance, frequency of the rotating system. This factor needs to be taken into account when components are designed to operate at or near this frequency. Although in theory the value of this multiplier can be infinite, in practical applications it is much smaller since most rotating systems have some form of "internal damping" which absorbs and dissipates a portion of the energy.
While conventional piston aircraft engines seem to encounter their natural frequency near idle, according to available tests and published data, the rotary engine encounters its natural frequency in the neighborhood of 4,500 rpm. At this speed if a problem occurs, the reaction is so fast that it can't be detected by sound or feel, only by sensitive measuring instrumentation. The results however can be very noticeable.
Recently we were testing the instrumentation of a new dyno installed at the facilities of Hayes Rotary Engineering. The engine we used was an old and tired "13B" which for some reason couldn't rev up past 5,000 rpm. We estimated that at best the engine was developing 115 hp.
In order to interface better with the existing hardware, we mounted a gear box from a sprint car (rated for 700 hp by the manufacturer) between the engine and the dyno's drive-line. To keep the installation simple for these early runs, a coupling similar to a standard clutch plate (similar to some forms of marine couplings) was used to join the reduction drive to the engine's crankshaft.
The early runs were done at relatively low revs and as observed by many in other applications, the engine ran smooth with not even a hint of vibration. Once the early adjustments were completed we started running at a higher rpm and at a somewhat higher load. As we brought the speed above 4,000 rpm there was a sudden grinding noise, a bang, and a growing oil puddle on the ground - the case of the gear box split, revealing support bearings and gears. Analysis of the failure showed that the breakage was beyond question a function of the torsional feedback. There was no initial vibration, no unusual noise, no warning of any kind - the failure was sudden and catastrophic.
Others developing rotary engine applications have had similar incidents, damaging everything from gear boxes to drive shafts, even the dynamometers themselves. In short, the rotary engine does have a torsional vibration problem and failure to address it in the reduction drive and coupling design may, sooner or later, result in failure.
So, assuming you have a gear-box design and a prototype, how do you assure yourself and your customers that it will hold together? Well first of all, a "static test" by itself is not enough. The first article listed a number of the considerations and drivers that have to be accounted for in the design of an aircraft engine's gear system and its housing - most are dynamic in nature so simple static testing will not be capable of generating the necessary loads to substantiate the hardware. The best place to start your testing is probably on a dynamometer where the procedures should at least include running at length at the predicted service RPMs and torques; variables which cannot be simulated on a static test fixture, especially with a fixed pitch prop.
The next procedures should include a test stand, but one which can be pivoted about an axis perpendicular to the axis of the prop rotation. While running the prop at full RPM, the stand should be cycled around this second axis thus testing the reduction drive case's capability of absorbing the gyroscopic load of the prop. This is an important test for all cases but especially those which will be subject to aerobatic flight or rough operating conditions.
The static stand can also examine the torsional response of the system so if any problems appear they can be detected before installation on an airplane. In order to do this it is necessary that the engine is run not only at idle but also up to full RPM with a prop identical (or at least very similar) to one which will be used in service. This may necessitate a lower pitched or constant speed propeller so the engine can reach the extents of its operating envelope. Ideally, these tests should include a number of propeller combinations so different diameters, masses and applications can be examined.
A limit of ground based testing is that it cannot simulate thrust. The static thrust of most propellers is about one fourth (or less in some cases) of the prop's capability at speed so unless the stand is mounted on a truck or a rail so it can move, there is no practical way to simulate this force. Fortunately though, it is a minor consideration as compared to the magnitude of the other loads.
In the case of our own reduction drive, we are currently undergoing dyno testing and fine tuning of the design. By the time you read this we should have at least one hundred hours of running time, much of it at rated load. No, the program has not been without difficulties but the few we've had have been minor and almost predictable. So far the major problems had to do with weight and with cooling.
Since our initial design constraints were for the box to be applicable to the horsepower range of the rotary engine (160 hp to 300 hp +), we had to use the worst case scenarios for the design numbers. A number of individuals thought that I was too conservative or pessimistic in the values listed in the previous article, but to cover that wide a range we not only had to address the low performance application with the wooden prop, but also the high speed airplane with the three blade constant speed metal propeller, and the flight envelopes for both cases
Historically, flight testing has shown that operations in moderate turbulence can result in instantaneous pitch changes of about 300 degrees per second. Granted these are very momentary impulses but the reactions due to the motion must be accounted for in designing all the rotating components. Considering more severe flight conditions and/or aerobatic flight we chose to use 360 deg/sec. for the gyroscopic load design criteria. In short, experience and testing has shown that the values used in the article were accurate if albeit somewhat conservative for some applications.
Since money was somewhat short we also took shortcuts in the machining of the case and some of the components. As a result we did not remove nearly as much material as will be removed in the final configuration. This, coupled with the conservative design, has resulted in an overall weight in excess of ninety pounds. With further refinement and more complete machining, our current estimates for the final configuration are in the neighborhood of sixty five to seventy pounds. Yes, it's still pretty heavy but to cover the range of applications we had to stay conservative.
The most recent stumbling block had to do with heat. In doing our initial dyno runs we have been seeing oil temperatures rise to well over 230 degrees, even at low RPMs - the initial configuration did not have a circulating oil system but one is being added now. To get better data, the reduction drive's oil system will be separate from the engine's however on the production package, the oil circulation will be part of the engine's oil sump. (Subsequent tests determined that the heat rise was due to a misalignment problem in the gears caused by a loose fixture at the machine shop.)
Like I said, so far the problems have been minimal - we hope they'll stay that way. If all goes well, by the time this article is published we should be running full power endurance tests.
So what's the eventual goal? By the time we're finished, Hayes Rotary Engineering would like to supply the homebuilt aircraft market with a turn-key firewall forward engine/reduction drive system which is dependable, reasonably priced and flexible for upgrade. Looking at the complete package, a Hayes aircraft engine, producing 190 hp at 6,000 rpm, will weigh about 185 pounds dry. With reduction drive, alternator and cooling system the overall installed weight is expected to reach about 275 pounds. With turbocharger and inter-cooler, producing approximately 275 hp at 6,000 rpm, the installation weight will be something over 300 pounds.
So who is Hayes? Hayes Rotary Engineering was established 1975, providing engine services to rotary powered car owners and racing teams. The company has been around for over seventeen years, all that time gaining significant experience with the Wankel engine. While others today are advertising selling conversion plans and a few engines over the past six years or so, Dennis (the owner) has been shipping on an average over sixty remanufactured engines per every month for each of those seventeen years. Adding that up that's over twelve thousand engines!
The outgoing engines are remanufactured, not rebuilt, to as good if not better than factory new specification. Experience has allowed Hayes Rotary to incorporate many patented improvements and embellishments to the design, all of which increase reliability, longevity, performance and economy, yet still allow engine prices to remain at very reasonable levels. My engine, which I bought several years ago for $3,500, has only moderate modifications yet is capable of delivering a bit over 200 hp at 6,000 rpm. Their racing engines have shown remarkable performance and reliability in track, off-road and marine applications.
For the aircraft market the buyer will be able to purchase the package complete with cooling system, ignition and exhaust, all ready to bolt in. Although the final price has not as of yet been set, preliminary estimates for the complete assembly run from about $5,300 for the standard 160 hp configuration and $6,300 for the improved 190 hp version, to around $7,500 for the turbocharged 275 hp version. The first sets should be going out the door by the first part of 1994. (To date the engines are avaialable without the reduction drive system.)
If you wish to purchase just an engine, these and other configurations are available now, less of course the reduction drive. Prices for engines alone are $2,800 for the standard configuration, $3,800 for the modified but still normally aspirated engine, and $4,900 for the turbocharged version.
Other combinations will be available, some capable of delivering in excess of the 275 hp currently foreseen. If development goes well, a three rotor engine will be available in about three years, producing as much as 450 hp (turbocharged), weighing about 440 pounds installed.
All in all we feel that the rotary engine is an ideal aircraft powerplant, capable of providing dependable and economic service to a wide array of airframes. For more information you can contact Dennis Hayes at Hayes Rotary Engineering at (206) 881 - 3604.