Vehicle: Car: Engine: Rotary
Needle Bearing Wankel Apex Seals   (+2, -3)  [vote for, against]
roll, don't rub

Wankel rotary engines currently use spring-loaded cast iron apex seals to close the gaps between the rotors and the housing. When these seals wear out, the engine typically goes kablooey.

To prevent this problem, and to lower the amount of lubricating oil that must be mixed with the fuel, I propose using spring-loaded needle bearings for the apex seals. Needle bearings typically allow for lower friction than a sliding surface and can be made of harder material for less wear.

These needle rollers (one per rotor apex) would be contained within a half-cylindrical journal made from brass or babbit metal and supported from behind by a flat spring similar to those already used in Wankel engines.

Lubricating oil could be supplied either by mixing with the fuel or by feeding it through passages through the rotor. The roller would then ideally float on top of a thin film of oil, much as a car tire hydroplanes on wet pavement.

The only problem I can forsee is if these needle bearings fail to seal between the combustion chambers effectively and end up leaking combustion gasses.
-- discontinuuity, Mar 15 2010

Wankel Engine http://en.wikipedia.../wiki/Wankel_engine
[discontinuuity, Mar 15 2010]

Needle Roller Bearing http://en.wikipedia...wiki/Needle_bearing
[discontinuuity, Mar 15 2010]

the motion of the rotor is hardly linear or circular enough to satisfy a needle bearing, it would have to be more like a ballpoint pen roller bearing to truly support this motion rolling without rubbing. In addition you are underestimating the speed at which the tips are moving when that motor is reved up to 9k (or perhaps instead overestimating how well the needle bearings will perform). The apex seals DO float on a layer of oil, much like typical piston/crank bearings are little more than smooth sleeves with oil. The sealing on this to the wall is akin to the piston ring, if it allows pressure to pass through then the motor sufferes a loss of compression and power. I can't seem a visualize a scenerio where utilizing bearings in this manner would not result in increased complication, reduced performance, and reduced life. A traditional seal may benefit from a better design to feed oil and float on it, better high-performance materials, etc. Just not feelin the bearings here, sorry.
-- AutoMcDonough, Mar 15 2010


I don't see the advantage here. I get the desire to replace sliding seals with rolling seals, but that ignores the extra complication involved in sealing the ends of a roller bearing AND sealing a roller bearing thru it's motion. The present seal probably doesn't have that much leakage behind the seal, but this probably would as the tolerances cannot be tight as that would stop the bearing from rotating.

I think what [auto] is referencing when mentioning ball point pens is the idea of almost enclosing the circumference of the bearing leaving only a small section exposed that contacts the cylinder surface, which is how a pen works. This would extend that idea in a linear fashion.

In any case, added complexity with lower efficiency (-).
-- MisterQED, Mar 15 2010


BTW, I've changed my username from rasberry re-tart to discontinuuity in an effort to integrate my various online personas.
-- discontinuuity, Mar 15 2010


If the rotor spun like a wheel it would work because you could align the needle with the circle's radius, but it does not move this way. It's more like an inverted spirograph. In fact there is an animation on the wiki page! You can see just from the shape of the combustion chamber that the rotor has quite a bit of lateral movement. It's easiest to see on the right side where the spark plugs are, the seal is sliding almost totally along it's length there.

This is where I was going with the ballpoint, but i agree that the ballpoint is totally unsuitible in shape to work as a seal.
-- AutoMcDonough, Mar 15 2010


[Auto], I'm still not sure what you mean. The roller would contact the housing at a single point (well, line actually). This line might move across the roller as the rotor turns (ie the angle between the rotor and the housing will change), but if the roller is of a large enough diameter this should be no problem.
-- discontinuuity, Mar 15 2010


It seems some of my comments got destroyed in the account transfer. Here they are again:

#1: [Ian], I think that would be called the Infinite Recursion Needle Bearing Wankel.

[Auto], you may be right about the rpm's being too much, but I'm not sure what you mean by "the motion of the rotor is hardly linear or circular enough to satisfy a needle bearing, it would have to be more like a ballpoint pen roller bearing to truly support this motion rolling without rubbing." My understanding is that the motion would be akin to a wheel rolling along a road, with the springs accounting for any variations in the surface. And ball bearings would not work at all in this situation.

[MisterQED], the rollers would seal on the ends with the same journal bearing material (babbit, etc.) that supports them from behind. It would also be possible to have rollers that were thinner on the ends, allowing them to slot into holes in the corner seals.

#2: Tolerances would be about the same as current Wankel apex seals, since during engine break-in the bearing material would be worn down until it conformed to the needle rollers. Thus the only gaps between sliding surfaces would be filled with an oil film, as with a conventional journal bearing. As I said before, the tricky part would be in getting this oil film to stay put and provide a good seal.
-- discontinuuity, Mar 15 2010


[+] Forget supporting the bearings on an axle or anything of the sort. Instead, put half-cylinder indentations in the rotors for them to ride in. Pump oil through the eccentric shaft, up through the rotors, and out underneath the apex bearings as is currently done in IC engines (where oil gets pumped through the crankshaft to reach the rod journals). The extra oil would burn, but Wankel engines already burn oil by design. You might also consider injecting fuel if the timing can be controlled (using your fuel as a lubricant). I'll bet the epitrochoid housing could be tweaked to accommodate the moving line of contact (to where it wouldn't be a true epitrochoid any more).
-- kevinthenerd, Sep 27 2012


The apex seal is exposed to combustion byproducts and coked oil. Whatever style of seal you use must be able to scrape these contaminants away from the sealing surface or it will eventually fail to seal due to their accumulation.
-- WcW, Sep 27 2012



random, halfbakery