Real World Solutions, Inc

Home of RA Classic Apex Seals and RA Super Seals

[ Home ] [ Search ] [ Recent Updates ] [ Contact Us ] [ Directions ] [ Site Map ] [ How to Order ] [ Guestbook ]

RD-1x Gear Design Philosophy
Zen and the Art of Gear Drive Design : Part 1.
The most common question about our new gear drive is "What's the difference between your drive and the Ross?" It's a fair question given that they both use the same gear set (RD-1A only). The gears come from the C6, which is Ford's biggest automotive automatic transmission and is made in 3 varieties: normal, and heavy-duty. Other than clutch pack differences, the main difference in the heavy-duty version is the use of a 4-planet carrier in place of a 3-planet set. The 4-planet version is used on the 428, 429 (Cobra Jet), and 460 cubic inch car and truck engines as well as 7+ liter Power stroke diesel trucks. I use only the 4-planet set, and I think the same is true of Ross. Except for the overall length, this is where the similarity with the Ross ends. The RD-1B and RD-1C use gears from a later version of the transmission (Super Duty) which have 6 pinion gears

Why So Long?
Several builders have complained about the length of the Ross drive and asked if I could make mine shorter. There were different reasons for this request. Some wanted to mount the engine on the stock Lycoming mount (with an adapter) and use the cowl that came with their kit. The 14.750 inch length of my redrive put the prop flange too far forward to do this. Others were concerned with CG problems. They had been told that the rotary was so much lighter than the Lycoming that the engine needed to be mounted as far forward as possible to avoid having the CG too far aft. This, in turn, would call for a short redrive to keep the prop hub in a reasonable position.

I can understand the desire to keep the stock mount in the hope of minimizing the work needed to mount the engine. The CG argument makes sense, as well, if it were true that the Mazda rotary was lighter than the Lycoming O - 320. I wish both of these things were true, but, as the saying goes, if wishes were horses everybody would ride. Unfortunately, neither is true. Adapting the 13B to a Dynafocal or conical mount is more work than starting from scratch, and if you are careful about your installation choices, the rotary weighs about the same as the Lycoming. If you're not careful, it will weigh more.

Other factors which favor the longer-length redrive are: 1. It eliminates the need for a prop extension; 2. Prop efficiency is improved when it is moved away from the cowl into cleaner air; and 3, the greater distance between bearings on the propeller shaft reduces the stress on the bearings.

Bearing Choices
One of my main design objectives was to eliminate the radial- and end-play of the propeller shaft that caused several problems in my Ross Drive. In addition to actual part failures (thrust bearing), this play caused a lot of resonance and vibration at various rpm settings. Control of this play in my design is accomplished mainly by the front prop-shaft bearing. This is a single row deep-groove ball bearing, which takes most of the radial loads, and the entire thrust load, from the prop.

Some people have objected to this choice pointing out that ball bearings are not designed to take thrust loads. It is true that they are primarily designed for radial loads, but the idea that thrust loads are forbidden is simply not true. There are very specific guidelines for how much thrust they can take. For instance, the maximum radial load for the bearing I use is 3484 lb static or 5732 lb dynamic load. The manufacturers' factor for calculating maximum axial load (thrust) is .5. Using the static radial load rating, this gives us an axial load rating of 1742 lb, which is much more than the actual thrust load will be. Note that bearing tables give loading in Newtons. I have converted to pounds in these examples.

Note: These figures are for the early RD-1 drive. The RD-1A/B/C drives use a larger bearing rated about double the figures quoted above.

The other thing we would like to know is the expected life of the bearing. The bearing industry has developed very reliable formulas to calculate bearing life, which is referred to as the L10 life of the bearing. L10 is defined as the number of revolutions a group of bearings will run under the specified conditions before 10% of them have failed. I used SKF's interactive design guide (available free on their website) to calculate the L10 life of the front bearing under the conditions it would be under at full throttle and maximum aircraft speed (I assumed 200 hp, 210 mph, and a prop speed of 3000 rpm). This resulted in a calculated bearing life of over 15,000 hours. My hope is to fly long enough to wear one out.

The rear prop-shaft bearing is a plain (bronze) bearing which is lubricated by pressurized oil from the engine. Using a plain bearing here eliminated the requirement of precisely locating it relative to the front bearing, which made machining much simpler. It also eliminated the problem of differential expansion between the aluminum gear housing and the steel prop shaft. The third reason was that the groove in the center of this bearing is used to feed pressurized oil to the prop shaft (which is hollow in this area). This is the key to lubricating the planetary gear set, which I felt was a weak point in the Ross drive.

Lubrication
The following drawing illustrates the general layout, dimensions, and oiling scheme of the RD-1 gear drive. Notice the oil port in the sun gear, which is a small hole between two teeth. The pilot bearing of the sun gear rides on the end of the prop-shaft, which supplies it with oil. Oil also flows out the end of the prop-shaft and sprays out of the oil port in the sun gear, which insures a constant supply of oil to the entire gear set. I'd like to take credit for this ingenious scheme but this is exactly the way Ford designed it to work in the transmission.

Maintainability
I place a high priority on ease of maintenance, which was another point I felt needed improvement on the Ross Drive. It was virtually impossible for the user to disassemble the main drive on it. Only one part is replaceable by the user if required (the planet carrier) because all other parts are altered and welded into an assembly, which makes even the bearings inaccessible. The RD-1 can be disassembled with ordinary hand tools and many parts are off-the-shelf and user replaceable.

Torsional Resonance... One More Time...
This did require a compromise. There are two additional splines in the RD-1 where the Ross Drive has welded assemblies. This increases the system lash and results in about 3 times the amount of free rotation at the prop tip as compared to the Ross. I was concerned about this, but after analyzing the situation and seeing the test results I have a new perspective.

As you may recall, Everett Hatch at Powersport went to great lengths to minimize the lash in their drive. The goal was to raise the resonant frequency of the driveline to a point that was higher than the engine impulse frequency at the highest expected rpm. They accomplished this by using an internal tooth ring and pinion gear set with a system lash of about .040 inches. This is certainly a valid approach and appears to have worked, but this is a fairly expensive approach and had other side effects that we will get further into in part two of this article.

The point I want to make now is that driveline resonant frequency goes down very quickly with increased lash. Think of dribbling a basketball at varying distances from the floor. If you dribble the ball a foot or so off the floor the ball bounces back and forth between hand and floor very rapidly. The speed of the dribble goes down rapidly with increasing distances between hand and floor. It probably was not by design, but the resonant frequency of the Ross Drive, which has about 6 1/2 times more lash than the Powersport, appears to coincide with an engine speed of about 1500 rpm. At this speed or less, the engine, gear drive, and propeller start to rattle quite noticeably. Fortunately, this is below the normal operational range of the engine in aircraft service, and as long as the engine is idled above this point all is well. The only aviators that this might adversely affect are float and amphibious plane operators who would like to idle the engine as low as possible to improve handling on the water. So, we see that the Powersport and Ross drives both avoid problems with torsional resonance problems, but they do it by using opposite tactics.

This puts my original plan (which was to reduce the lash of my drive compared to the Ross) in a whole new light. This would have very likely raised the resonant frequency into the useful speed range of the engine. When using the Ford gears, there is no way to get the lash as low as Powersport did , so that approach was out of the question. This realization caused me to change design strategy. I definitely did not want to have less lash than the Ross, so in spite of the conventional wisdom that dictates that less is better, I began to consider the ramifications of having more. This ultimately led to the design I ended up with and had the fringe benefits I mentioned before (using more stock Ford parts, improved maintainability, etc).

As you already know, the gamble did pay off and I have had no adverse side effects from this arrangement. I would have been happy with the same critical rpm as the Ross, but the actual results mirrored the theoretical predictions. The RD-1 gear drive idles smoothly down to 800 rpm (367 at the propeller) with no shaking or rattling. In part 2 of this article we will discuss the reasons behind other design choices including elastomeric vs. spring dampers, plate-and-spacer vs. cast bell-housing adapters, and thrust bearing alternatives.