Part 2, the Snowmobile Engine
The History of the OMC Wankel Rotary Program
by John Sheldon
former OMC project engineer on rotary engines
Part 2; the Snowmobile Engine
With the demise of the oil cooled rotor engine program I was reassigned to Harry Ward who was working on an air-cooled engine for snowmobiles. Harry knew from his experience at CW, an oil cooled engine could not be hand started at –40. Ficthel and Sachs from Germany had developed a charged cooled rotor engine where the incoming air-fuel mix passed thru the rotor to cool the internal parts. Roller bearings were used instead of the babitt bearings of an oil cool rotor, which required oil to be mixed with the gas ala 2-cycles. Unlike a piston engine, the rotor is a captive heat source with very little heat rejected to the housing walls. As such most of the cooling had to come from the incoming charge to cool the rotor. The result is a significant reduction in volumetric efficiency, the actual amount of air going thru the engine vs its theoretical capacity. This is due to the significant heating of the charge as it passes thru the engine. A rotary engine is unique in that only one section of the engine is exposed to combustion. The down side is that it always hot, not like a piston engine, as combustion occurs almost continuously in that section. Harry had done some work at Curtis Wright with partially air-cooled housings. Unfortunately he had designed a 2 stage axial blower running at 30,000 RPM and took 35 hp to drive to cool the CW design. Ficthel had used an axial fan on their engines, but they had cooling fins all around the engine and had reduced performance to be able to cool the engine. Harry decided a partially cooled engine was possible if an adequate fan could be developed. I developed a computer program to predict airflow required vs. fin design. We produced rotor housings without any fins and machined various fin designs into them. We then had an external variable speed blower to blow air thru the fins measuring air velocity at each fin. Harry developed an ingenious method of measuring heat flux thru the engine. A pair of accurately spaced spring loaded thermocouples measured temperatures allowing us to calculate heat flux. Testing various fin configurations and measuring the air velocity required to cool the engine, we verified my computer predictions. Using the computer program and manufacturing constraints for casting fins, we were able to predict how much air and at what pressure would be required to cool this engine. One of the design constraints for the engine was it had to run at +90 in addition to -40. With the fin design fairly well set due to die casting restrictions, the challenge became to design a fan that would provide sufficient air flow at a given pressure to cool the engine. Very little has been written on centrifugal fans. We knew an axial fan could not generate sufficient pressure that would result in the velocity required to cool the engine. We ended up with a two-piece fan that was die castable in magnesium including a pre-swirl inlet section. Research had developed a sprayed tungsten carbide from Metco that would meet the durability requirements when coupled with tool steel apex seals. A durability cycle of 55 minutes WOT, followed by 5 min idle, for 100 hrs. was established as the minimum life requirement. More details on this engine can be found in the SAE paper written by Harry Ward. The snowmobile engine was released for production and in the winter of 1972. 150 engines were built on a production line in Milwaukee. 100 of these engines were installed in the new snowmobile designed for this engine. The 100 units were given to various dealers throughout the snow belt with instructions to loan them to their customers for no more than a week at a time. No service was to be performed on these engines at the dealer level. They were instructed to call OMC if they had any problems and OMC would send a service technician with a new engine. For the entire season, no one reported any problem with an engine. Not even a sparkplug was replaced. The units were returned in the spring with the plan being to tear all the engines apart for inspection. The 6 units with the most hours were torn down and the engines looked so good that the remainder were never disassembled. Production started for the ’73 model year. Approximately 15,000 engines were produced for the 73 & 74 season. OMC went out of the snowmobile business after the ’74 season. Because of the high maintenance of the die cast die fin sections, one of the last changes was to change the rotor housing from high pressure die cast 380 alloy to permanent molded 356 alloy. This not only reduced the costly die maintenance, but also gave the added benefit of better thermal conductivity, lowering temperatures. Many of these engines are still running today, 35 years later. With no home for an air cooled engine, attention turned to water cooled versions of this engine.
Because of the availability of production parts and the severe environment of the air cooled engine, the snowmobile engine continued to be the workhorse for component development. One of my responsibilities at the time was all the rotating components of this engine, bearings, seals and rotors. The rotor bearing was the weak link of this engine. Being inside the rotor, it is completely dependent on the incoming charge for cooling and lubrication. The bearing was already pretty special. It had a silver plated retainer and a high temperature temper of the outer race. Steel grows; get larger, as it gets hotter. That is until it exceeds the temper temperature where in it gets smaller. As I figured out after 100’s of tests, there was a lubrication breakdown between the retainer and the outer race. The result was steel running on steel at very high speeds generating extreme heat. The result was the bearing shrinking onto the crank eccentric stalling the engine. It took a hydraulic press or a sledge hammer to get them apart. I took several samples to SKF Research and talked to the chief metallurgist. He told me this was impossible to do. When I suggested he come to Waukegan and see for himself, he said the bearing had to be reaching 1800 F for this to happen. This type failure was self-destructing. When the bearing got near its thermal breakdown temperature it would self generate additional heat causing further oil breakdown, causing additional heat, etc, etc,etc. The end result was a shrunk bearing and a stalled engine. Dozens of bearing designs and modifications were tried before a successful solution was found. Interesting enough, none of the snowmobile engines in use showed any bearing problems. Part of the reason is it was very difficult to maintain WOT for extended periods of time and by nature, snowmobiles ran at much colder ambients. The apex seals were a second area where thermal breakdown of the oil caused problems. Similar to the bearing, it was a snowballing failure; generating addition heat as failure started. Wear was so severe the seals would look like horseshoe nails after only a few hours of running. Many, many, many, iterations were tested with varying results before a better seal was developed. Less than .0005 in wear was measured in 100 hrs WOT testing. Unfortunately, these never saw production. One might ask why all this testing and development when there was no home for the air-cooled engine. Part 3 is on the water-cooled engines based on the snowmobile design.