Egg Gearbox Failure

DanH said:

The subject is an Egg gearbox failure. Does anyone have a dimensioned drawing of the actual components? Same for the flywheel assembly, with whatever sort of "damper" it might have? Third, mass moment of inertia for the propeller?

Without these things at minimum, (1) serious failure analysis is impossible, and (2) neither fans nor critics know the truth.

Click to expand...

First off, I am no engineer, but I can make these observations about the EGG design.

Starting with the engine with the PSRU removed - the original fly wheel has been replaced with a spring loaded devise that weighs about 20 pounds. It is a fly wheel but not connected directed to the crank shaft but rather the load is carried by a series of springs in compression. If you move the prop, you can feel the springs compressing before the crankshaft moves. Its purpose is to dampen the load transferred forward to the PSRU.

The engine is mounted to a half inch aluminum plate, aproximately 30"x18" in size. The PSRU is bolted to the forward side of the plate and connected to the engine with a splined shaft about 8 inches long which slides into a splined plate bolted to the forward side of the fly wheel devise. Fore and aft movement is controlled by 2 large snap rings. Alignment is controlled by plugs in the aluminum plate and in the engine similar to the arrangment with auto transmissions except the plate is aligned to the engine with one set of plugs and the PSRU is aligned to the plate with another. An alignment tool was created to check the total alignment before the PRSU is installed. The aluminum plate is attached to the engine mount structure with bolts through bushings.

The torsion load of the prop is absorbed by the PSRU and the aluminum plate but not the engine crankshaft. The PSRU is a straight through gear reduction (planetary?) resulting in the prop being aligned with the crank shaft and turning in the same direction.

The PSRU failures to date IMHO have been caused by misalignment with the engine resulting in side loads on the aft bearing or torsion loads beyond the design limit of the bearings within the unit. Thats why GEN1 has been quite successful with the original 2.5 engine but not so with later, more powerful engines. GEN2 is simply an upgraded GEN1 with different gears and bearings. It too did not meet the load requirements or was misaligned in the process of field installation. The alignment tool may have come too late as some were installed without it. The tool is necessary to install GEN3.

As previously stated, GEN3 is a totally new beast. The gears and bearings are stronger, it has a different ratio (2.02:1 vrs 1.87:1) and has a machined case instead of cast. It is considerably more expensive to manufacture and to date has been available to upgrading customers at cost or less. The EGG factory is trying to be fair and not go broke in the process.

Actually, no. Tracy's first few production drives had a bearing problem which he uncovered and corrected before any customers had accumulated any flight time. This is from Tracy's extensive article in Contact #87, page 23.

I'd consider his boxes to be among the most successful to date.
He has many units flying now successfully and has captured a large majority of the Wankel market. Tracy is the man here.

I've seen the Egg drives being assembled. The gears/shafts are off the shelf components from one of the big OEMs and sizes are in the 5 ton truck range. The gears will never break in this application IMO. There have already been at least one bearing revision on the Gen 3. The case looks well designed and machined and very strong. It is a layshaft setup using helical gears. There are some welded assemblies inside which should be fine if proper stress relief and heat treatment is done afterwards. (We'd do a welded mod to the W50 boxes for racing to retain 5th gear at high hp. This worked fine for us).

I am not aware of any rubber TV damper being used but a very heavy dual mass flywheel is always fitted. There seems to be a TV period at about 250 rpm and perhaps another more minor one at around 900 rpm by seat of the pants feel. Engine idle is generally kept just above 1200 where it is very, very smooth.

It is interesting to me that the 3 very different engine/ drive/ prop combos I've been associated with all have a TV period right around 600-950 rpm. Certainly my Sube and the Rotax 912 are a blur down there. Coincidence? Any comment on that Dan?

Ross-
You are probably correct regarding bearing failure- Tracy's original Ross drive did not fail catastrophically, but it was a bad design that drove his desire for something better.

Tracy's "bearing problem" that you refer to involved the ball bearing upgrade I mentioned in my 4th paragraph, which Tracy implemented to improve the 300-hr TBO inherent with the original needle bearing (from Tracy's historical evolution articles in his website). That failed bearing turned out to be a poor quality, cheap Chinese bearing- he fixed that problem with a better quality ball bearing, but eventually upgraded to the German roller bearing now used which has a TBO>2500 hrs. Like you said, no defective units were ever released to the public.

Tracy also mentioned an early bolt failure at the attachment between his drive and flywheel on one early RD-1C. He also fixed that problem with a minor alteration and recall at his own expense (unlike others in the certified aircraft world who will remain nameless). Tracy is definitely a class act.

I agree with you that the RWS RD-1B and 1C are the best affordable redrives available at this time. It is hard to beat a good quality planetary gearset for hard-use durability, compact size, and light weight.

The 2.85 ratio works great within the rotaries' 5000-8000 rpm peak torque/power band (1700-2800 prop rpm). The 2.18 ratio has been effective for fixed pitch economy in cruise, but it has also been reported that it is too tall to deliver peak rpm/power.

As for Subaru applications, it is questionable whether the two available planetary gear ratios with the Ford 4/6-gear carrier present an acceptable fit to deliver peak hp to the propeller. Not knowing the Subi's torque/hp graph, Id have to guess a desire to keep the engine in the 3500-5500 rpm range, which implies approx 2:1 gearing given RV props. Didn't Jan try the 2.178 planetary setup early on? It proved to be less than ideal in delivering peak hp to a fixed prop and finding a better ratio has been nothing but grief to users since. The only solutions I anticipate would involve the use of a lightweight adjustable prop (like the IVO) to work within the heavier-than-Lyc engine weight balance envelope, or possibly a simple 2-speed transmission.

It looks to me like an engine and redrive mismatch problem is in play here: the Subi engines cannot reliably rev high enough for the 2.85 ratio, and the engines cannot produce enough rpm (to deliver peak horsepower) with the 2.18 ratio.

I Think Tracy has been so successful in the field because he is smart, tests a lot, gives the straight goods and makes things right if something isn't. All things that any business should be doing.

Ratios of 1.9 to about 2.2 are suitable for most Subaru applications allowing these engines to reach close to power peak rpm at 2700 prop rpm. You are right, The EZ30s would perform better with about a 2.4 ratio due to their higher power peak rpm. Prop blade angle may not be well optimized at lower rpm cruise settings unfortunately and frictional losses at high engine rpm conspire to raise SFCs. It is a careful balancing act.

Forced induction can change torque curves considerably so in all cases, drive ratios should be selected according to each engine's torque and power curves. This was one reason I chose the EG33 over the more modern EZ30. The power and torque peak rpms are only 1000 rpm apart and power peak rpm is 1100 rpm lower as well.

With turbos, there is far less worry about whether you can turn the prop at low engine revs and make any power- usually you can and the engine is far more efficient at low revs, high manifold pressure. Unfortunately then you have the added weight and complexity and cost of the turbo system.

Jan has never used planetary gear sets to my knowledge.

Tracy's RWS & Wankel have much to recommend

rv6ejguy said:

I Think Tracy has been so successful in the field because he is smart, tests a lot, gives the straight goods and makes things right if something isn't. All things that any business should be doing.

With turbos, there is far less worry about whether you can turn the prop at low engine revs and make any power- usually you can and the engine is far more efficient at low revs, high manifold pressure. Unfortunately then you have the added weight and complexity and cost of the turbo system.

Click to expand...

I agree if going Alternative I would go Rotary and RWS stuff, not as a vote against a Subaru, just that the Rotary is a true alternative engine, it's nothing like a Lyc. A big plus is a true roll-your own deal with RWS parts, so you can make a FWF package for less than a new Lycoming. Rebuild a 13B core yourself, you can come in cheaper than a new Lyc, no doubt by many thousands. The FWF Eggs and Power Sports kits are more expensive than a Lyc. The negatives I mentioned before like higher fuel burn fuel, noise and mixing fuel/gas come with the package, however there are ways to mitigate these, TURBO CHARGER, which is particularly effective on Wankel's.

With a rotary, a turbo is almost a MUST in my opinion. The Rotary wastes or expends some wonderfully high temp, high velocity exhaust gases, which is wasted energy (thus fuel and noise issues). This is why it's perfect for turbocharging, to recover all that energy. It also tones down the exhaust bark, while increasing efficiency, particularly at altitude. When you get a rotary up high into the teens, it comes into its own and MPG is actually good. Also a rotary can be lighter than a Subie package but still not as light as a Lyc.

My flying preference does not included a lot of high altitude long rang cruising, sucking 02. It's just not that fun for me or something I do much of anymore. If I go X-C, 8k to 12k feet or lower is plenty high. Turbo's in general are a pain, at least on Lycs. Lyc powered RV's don't need them because the engines are powerful enough as an ATMO engine. Also I don't want to fly at rarefied altitudes sucking on an Oh-2 mask. There are also Vne issues. Still a turboed Wankel and flying high is a cool deal. To be fair Jets suck gas like crazy down low and only come into their own above FL250.

This is probably not the best place to discuss this, hijacking a Subi thread, but since you started it...

"When you get a rotary up high into the teens, it comes into its own and MPG is actually good. Also a rotary can be lighter than a Subie package but still not as light as a Lyc."

Simply not true, let us see your biased data George! Let's put some relevant facts on the table and compare. Not dissing btw, I believe the Lyc is a very worth opponent here, the standard for comparisons.

A rotary does run very well with a turbo, but turbos are certainly is NOT required to meet or outperform a 360 class Lyc power to weight with the correct intake. The 13B class rotary fwf is very close to the o-320/360 in weight with more power potential, closer to O-540 output, assumed range 200-250hp.

We've already mentioned the recent p-port Renesis, naturally aspirated test using 87-octane gas produced 265hp @ 7200rpm ON A DYNOMETER. Ive never seen Lyc output measured on a dyno..., guessing posted numbers are adhoc or estimated. P-porting is not particularly difficult- drill a 2" hole and glue or weld in a pipe; it greatly simplifies the intake system, with only 2 intake runners vs 6.

The basic RENESIS engine weighs about 180 lbs, a little less than older 13B engines- I can lift my engine off the ground (w/o mount, exhaust manifold, cooling system, or redrive). The Lyc 360 weighs close to 300 lbs., the IO360 a bit more. Add to that the weight for mount, exhaust system, ducting, etc. If you have more exact numbers, Id like to see and compare them.

I can show you similar carburetted dynographs of the OEM Renesis and p-ported 13B's = 200-240hp depending on engine configuration. FWIW, you can add an IVO composite magnum adjustable 3 blade prop and add only another 27 lbs complete (and 1/3 the cost), a bit lighter than the Lyc c/s choices I believe.

rtry9a said:

Ive never seen Lyc output measured on a dyno..., guessing posted numbers are adhoc or estimated.

Click to expand...

Barrett Precision Engines has ran a lot of unmodified, and modified Lycomings on their dynamometer. Rhonda Barrett-Bewley reports that O-360s typically produce 180 hp, but the nominally 200 hp IO-360-A series engines typically only produce 196 hp.

<<It is interesting to me that the 3 very different engine/ drive/ prop combos I've been associated with all have a TV period right around 600-950 rpm. Certainly my Sube and the Rotax 912 are a blur down there. Coincidence? Any comment on that Dan?>>

Your 4-cyl w/ Marcotte hammers hard at an RPM a little further up the scale...a peak at about 1350, which you push through as quickly as possible, yes? The 912 rattles the ramp and dog system somewhere not far below 2000 if I remember correctly from the last time I did a carb sync. Neither has a soft element suitable for lowering the natural frequency of the system. Yeah, I know, your Marcotte has urethane bushings in the flywheel. Trust me, they are near useless for the above (full disclosure; my opinion, but based on measured spring rates plugged into TV prediction software runs).

David described soft springs in the flywheel of the Egg system. Springs are a reasonable way to lower system frequencies, if they allow enough angular displacement without coil bind. At this time we do not know the Egg's F1 and F2 frequencies, but if your observation is correct, it matches what I would expect for a system with a very soft torsional soft element.

BTW David, springs are not dampers. I've explained the difference before, but I'll try another tack. It may seem hard to believe, but any shaft is also a torsional spring, as all materials are elastic. If the shaft is short and has a large diameter, the spring rate will be very high. If it is long and has a small diameter, the torsional spring rate will be low. The springs in your Egg system have exactly the same effect as a very long, very thin shaft. Now, would you consider a shaft to be a damper?

Ross, back to your question. Concerning outselves only with the system's F1 frequency, the prop drive systems I've observed seem to be between 20 and 50 hertz. The difference is due to design variations, specifically variations in the torsional stiffness of connecting members, and variations in inertias. You'll recall the Powersport design example also, which was to shoot for a very stiff system and thus an F1 above the operating range. The three subjects of your experience merely happen to be at the soft end of the scale.

Let's assume we have a system (you do) with an F1 (fundemental natural frequency) of 45hz. If we bolt it to a 4-cyl 4-stroke, it will resonate at 1350 RPM because that is where natural frequency is matched by the major exciting frequency, firing events. If we bolt it to a 6-cyl 4-stroke (assume the same crank-flywheel stiffness and inertia for this example), it will resonate at 900 RPM. The equation for exciting frequency due to firing events with a 4-stroke is (RPM x # of cyls)/120 = hertz, or turn it around if you know the hertz, ie, (hertz x 120)/ # of cyls = RPM. Point is, just bolting the same drive on a different engine configuration will give you a new resonant RPM.

I ran a 912 on a test stand a lot last year and it is very unhappy from about 600 (min idle) to almost 1200 rpm. The frame was very flexible so you could see (and hear) the effects of TV quite graphically. Pretty darn scary actually. Super smooth at 1400.

On my EJ22 setup it is pretty bad from 600 to 950, ok at 1000-1050 (idle) and really bad from 1100 to about 1600 so the 1350 figure is right in the middle of that second point all right.

On the EG33, I'm now installing an 8 pound steel ring on the outer part of the aluminum flywheel to see what effect it will have with and without.

Marcotte has changed the bushing style in my later drive to a stiffer, smaller diameter. Both are neoprene.

It will be interesting to see how bad the EG33 shakes at 900 rpm and how wide that bad range is.

David, do you notice a vibration in your EZ30 around 900 or do you ever idle it that low?

I'm guessing that that death shake some Lycomings do as they are shut down just before the prop stops is TV as well?

Yea sounds Good

rtry9a said:

"When you get a rotary up high into the teens, it comes into its own and MPG is actually good. Also a rotary can be lighter than a Subie package but still not as light as a Lyc." The 13B class rotary fwf is very close to the o-320/360 in weight with more power potential, closer to O-540 output, assumed range 200-250hp.

Click to expand...

Of all the alternates the rotary does the best weight wise, but show me some W&B's. I've never seen "light" but some respeciable empty weights. If you push a 13B to 250hp, I think it will die soon and not be advised for a daily flyer, going on what I recall Tracy told me, conservative wise. It physically can make HP for sure, just how long is the question. Radical racing Wankel's do blow up spectacularly.

We've already mentioned the recent p-port Renesis, naturally aspirated test using 87-octane gas produced 265hp @ 7200rpm ON A DYNOMETER. Ive never seen Lyc output measured on a dyno..., guessing posted numbers are adhoc or estimated.

Click to expand...

Displacement is twice as much as a 13B right, so that is expected. I am guessing they are not laying around cheap? Nope Lyc and Superior Dyno there engines. What Keven said, the 180's are getting about 185. Don't know about the IO360angle valve being/under or over nominally on the Dyno. From the fact the 200HP was 9.5 mph shower than the 180HP Factory prototype in the PowerSport fly off baffles me. I still can't figure out what that was about. I can only assume the 180HP was making 185HP and the 200HP was tired.

The basic RENESIS engine weighs about 180 lbs, a little less than older 13B engines- I can lift my engine off the ground (w/o mount, exhaust manifold, cooling system, or redrive). The Lyc 360 weighs close to 300 lbs., the IO360 a bit more. Add to that the weight for mount, exhaust system, ducting, etc. If you have more exact numbers, Id like to see and compare them.

Click to expand...

I would love to see that. Are RENESIS engines laying around cheap? I think that single lower scoop like Sam James cowl is probably as good as it gets, verses trying to make a stock Vans cowl with to cheek inlets work. Just saw DOG Fights on the History channel with the P51. That lower belly scoop seems to be the ticket? (Like what Ross is doing on the RV10)

I can show you similar carburetted dynographs of the OEM Renesis and p-ported 13B's = 200-240hp depending on engine configuration. FWIW, you can add an IVO composite magnum adjustable 3 blade prop and add only another 27 lbs complete (and 1/3 the cost), a bit lighter than the Lyc c/s choices I believe.

Click to expand...

Lighter weight, cool, whats the empty weight? I watch finished RV weights and find water cooled engines are at least 50-100lbs more. Clearly some weight savings can be had with prop and being smart about outfitting the plane. The belted power guys have a reasonably light V6 RV demo plane (not as light as a Lyc plane), but its bare bone, which is fine. That is what you have to do, keep it light. Add weight FWF you better look to save it aft of the FW. Save weight you must. Cost wise, OK I'll quit arguing but 1/3rd? I have seen the math and you're looking at mid-high teens for a RWS Rotary setup. With a NEW Lyc & prop mid-high 20k something. I'm sceptical, may be 33% less but not 66% less. It all depends on what you buy and make with sweat equity. Now being a cheap skate I got my O360A1A for $2,500 and overhauled it for a total of $12,500. I have to admit the used Lyc engine market has dried up, but it's still not impossible to find a used deal. Frankly apples and apples you are talking about a used rotary engine. Still you can beat the Lyc price, I'll grant you. You probably can do it with a Subie but you have to make it all up yourself.

<<I ran a 912 on a test stand a lot last year and it is very unhappy from about 600 (min idle) to almost 1200 rpm>>

Whoops. It's been a while since I fooled with a 912. I must be thinking of a 582.

<<the 1350 figure is right in the middle of that second point all right.>>

We talked about that in a previous thread.

<<I'm now installing an 8 pound steel ring on the outer part of the aluminum flywheel to see what effect it will have with and without.>>

It will help. How much would require modeling.

<<Marcotte has changed the bushing style in my later drive to a stiffer, smaller diameter. >>

That will worsen shake at the F1 intersection, but may help ensure the F2 intersection is above the operating range, a real concern with the switch from four to six cylinders. To illustrate, consider an F2 of 200 hz and play with the equations from my previous note.

<<It will be interesting to see how bad the EG33 shakes at 900 rpm and how wide that bad range is.>>

It will shake less than it does on the 4-cyl because exciting frequency amplitude is reduced. Expect the range to be the usual 0.8 to 1.2 times peak RPM, give or take a bit.

Subaru Options

FWIW Maxwell Propulsion Systems have taken over the NSI baton and are producing Subaru EJ25 FWF packages which are a pretty complete option.
The sprague clutch in their 2.12:1 gearbox seems to overcome the torsional vibration problem but there have been failures, apparently associated with the former ECU related backfires. (At least the failures involve the clutch locking up so the a/c is still flyable.)
Their Wirlwind fabricated electric adjustable pitch propellor allows the engine to develop its full power at 5700 RPM or cruise at 4000 - 4800 RPM WOT giving reasonable performance and very good fuel burn.
I have an NSI package in a 9A and without having addressed the cooling drag issue can achieve 155 KTAS at a DA of 8,500' or coming back to 4500 RPM, 145 KTAS at a fuel burn of 4.6 gal/hr.

Rupert Clarke
Melbourne Australia

rv6ejguy said:

I ran a 912 on a test stand a lot last year and it is very unhappy from about 600 (min idle) to almost 1200 rpm. The frame was very flexible so you could see (and hear) the effects of TV quite graphically. Pretty darn scary actually. Super smooth at 1400.

Click to expand...

The factory minimum idle RPM for a Rotax 912 / 912S is just under 1400 RPM. If you are really good at adjusting them you can get them to run smooth around 1,000 RPM. They aren't meant to idle below that. If I remember right they have to be turning at least 6-700 RPM just to make the dual Ducati ignitions fire. If you had a rough engine below 1,200 RPM you needed to balance the carbs. Idling a Rotax 912 lower than 1,000 sets up pulses & harmonics from the prop through the gear box that can damage the gear box. If you "ran it alot on a test stand" trying to get it to idle at 600 RPM I suggest you rebuilt the gear box and replace the "spring" washers / spacers. You may have damaged them.

They love to run 5K RPM all day long. Some have over 4,000 hour on the original engine & gear box (properly maintained) and are still in specs.

The Rotax 912 series engines are about as bullet proof an aircraft engine as you can get. There are more Rotax engines flying than ANY other engine.

<<The sprague clutch in their 2.12:1 gearbox seems to overcome the torsional vibration problem but there have been failures, apparently associated with the former ECU related backfires.>>

I think not (an opinion <g>).

Torsional theory says a sprague will be great. Unfortunately, the engineers at the sprague company recoil in horror at the idea (I called, oh, six years ago). As previously noted, most of our systems are resonant in the 25 to 50 hz range. I suppose there is a special sprague out there somewhere, but spragues sourced from automatic transmissions were never intended to lock and unlock at 25 to 50 times per second.

So, an example. Temporarily lock the clutch by some means, run the system, and assume you observe an obvious evident F1 intersection at 1000 RPM. It starts making noise and shaking the airplane at about 800 RPM, rises to a peak at 1000, and tapers off by 1200. Unlock the clutch, and presto, everything is smooth as glass. Wonderful result, but to do it the clutch is cycling at 33 hz.

Customer A has a light airframe and keeps his tires inflated, so he tends to taxi at 1000 RPM. Customer B has a heavier airframe and his tires are always a little flat, so he tends to taxi at 1300 RPM. Both report excellent system "feel", but poor Customer A will suffer a sprague failure, while customer B will be happy as a clam. "A" racks up a very great number of lock-unlock cycles in not many total operating hours, while "B" may not reach the same number of cycles until he has flown the classic "1000's of hours of testing!".

The catch for sprague flyers should be evident; they have no idea what RPM to avoid.

Sprague clutch numbers

Thanks Dan. Good info.
The conventional NSI wisdom (probably an oxymoron) is to keep the idle speed at 2000 RPM or better but I think that the last line of your post suggests that the problem resonance could be anywhere ?
I know that you were only offering an example but if heavy aircraft were an answer the Subaru drivers might be OK because the package is not light. Including header tank with 1.3 gal of gas, coolant, oil, battery, engine mounts (but not front gear), dual coils, pumps, radiator, oil/water heat exchanger, exhaust pipes and muffler, cowls and prop, plus a heap of wiring my little lot weigh in at more than 450 lbs.

Rupert

rv6ejguy said:

David, do you notice a vibration in your EZ30 around 900 or do you ever idle it that low?

Click to expand...

Idle is set at 800-900 but I never leave it there except for starting. I have not reset it to the factory recommended 1400 because the engine does not like a partially open throttle for start. It starts better the way it is which usually is without hesitation. As soon as the ECU has found its brain, I move the rpm up to 1200-1400. At 800-900 there is what sounds like gear back lash and general vibration throughout. When the engine is warm, it will idle at 700 with quite a bit of gear noise. Is that a separate issue from the low end vibration you and Dan are discussing?

I have run the engine without the PSRU and prop and it appears to be quite smooth at idle but there may be stuff going on I can not detect without vibration gages.

It would be good to know if there are any rpm ranges above 1400 where these plots are converging or whatever happens with vibration. It would be easy to avoid those areas with the CS prop.

Re: rotory discussion- GMCjetpilot

If you push a 13B to 250hp, I think it will die soon

No evidence of that- the power levels depend on the rpm selected- the power curves are linear up to 8000 rpm. Rotaries do not break and they seldom show measurable wear in hard service (up to 3000 hrs so far in a gyrocopter, Tracy has also nearly reached that point in his RV).

That said, most rotary drivers keep rpms in the 5000-6000 range for cruise (abt 160-180hp) where the rotary operates at its best efficiency. Because the envelope is so wide, there are no problems expected by adding more power if wanted for whatever purpose, and, that extra margin is available at higher altitudes as power drops. The only concern is burning more fuel. There is no reason to ever exceed 8000 rpm- that is where the peaks reside and represent the OEM's design red line. The auto racers see rpms in excess of 12K and often under big boost levels where detonation is a primary concern- that is where things get dicy.

Displacement is twice as much as a 13B right, so that is expected. I am guessing they are not laying around cheap

No way- the Renesis is a 13B motor; the rotors and housings are the same size and interchangeable, the 1300cc displacement is identical. The only real differences in the Renesis vs older designs (quite a few iterations involving port size and orientation) is the exclusive use of large side intake and exhaust ports on the Renesis that improve breathing and eliminate cross contamination of the intake charge over the other stock engines. The rotors are slightly higher compression and lighter. The Renesis is ~10 lbs lighter than older motors. The internal gears are hardened (allow higher speeds- a common racing upgrade).

Mazda is talking about bringing out a new higher displacement motor in the next year or two, the 16B (or 1600cc displacement vs 1300cc). IMHO, the biggest plus of the new motor resides in its aluminum end housings that will drop another 50 lbs off of the 13B family in addition to added power/torque.
Powerwise, the Renesis and any of the older p-ported 13B's are close, with a slight edge to the p-ports because of better charge flow, but, they tend to idle rough because of exhaust contamination of the intake air. In addition, there are 3 rotor (20B) engines and a smaller older design (12B) occasionally seen.

FWIW, I bought a nearly new 4-port Renesis for $1800 + 200 shipping. A big lot of motors were sold at the time in Australia (Mazda used these for emission testing, I think). Used 13B typically costs anywhere from $200 up depending on accessories (attached to a sports car...) A complete rotary overhaul costs less than $1000. The 3 moving parts and housings seldom show wear unless they are grossly dirty inside, and even then, generally usable after a good cleanup.

I think that single lower scoop like Sam James cowl is probably as good as it gets

The rotaries are quite compact- the radiators and oil cooler can be mounted horizontally under the motor, where the James "P-40" cowl works well. Another possibility is to use an offset shertz beam mount that allows a vertical radiator to be mounted on the side, parallel to the motor- the Vans cowl works fine for that arrangement.

Cost wise, OK I'll quit arguing but 1/3rd? I have seen the math and you're looking at mid-high teens for a RWS Rotary setup. With a NEW Lyc & prop mid-high 20k something.

It all depends on the specific installation, and level of sweat equity, as you said. On the low end, figure approx $2000 for motor+ 3500 for redrive + 1000 for engine controls + 1500 more for cooling, exhaust, intake, and motor mount materials- nothing particularly exotic, mostly welding and assembly. The motor could be less, the incidental materials a little more depending on what you select.

Premanufactured fwf's have been quite expensive, like the Eggs- Im guessing their pricing has been based more on the prices of the competetion (Lyc motors and clones) than on actual costs and labor.

Geico266 said:

The factory minimum idle RPM for a Rotax 912 / 912S is just under 1400 RPM. If you are really good at adjusting them you can get them to run smooth around 1,000 RPM. They aren't meant to idle below that. If I remember right they have to be turning at least 6-700 RPM just to make the dual Ducati ignitions fire. If you had a rough engine below 1,200 RPM you needed to balance the carbs. Idling a Rotax 912 lower than 1,000 sets up pulses & harmonics from the prop through the gear box that can damage the gear box. If you "ran it alot on a test stand" trying to get it to idle at 600 RPM I suggest you rebuilt the gear box and replace the thrust washers / spacers. You may have damaged them.

They love to run 5K RPM all day long. Some have over 4,000 hour on the original engine & gear box (properly maintained) and are still in specs.

The Rotax 912 series engines are about as bullet proof an aircraft engine as you can get. There are more Rotax engines flying than ANY other engine.

Click to expand...

We were running with EFI and both OE and our ignition systems on a test stand. The OE ignition requires around 300 rpm to fire, not 600-700 or the engine would never start. The EFI negates any carb synching and improves mixture distribution.

We'd generally establish an idle of 1400 but during R&D would have temporary dips well below 900 while adjusting things. The engine did go back to Rotax after we were done with it.

Having the engine on the stand really allows you to see the severity of low frequency TV. You could feel how hard these points were on the engine and drive.

Yes, these engines are very light and tough and show what a good geared engine can do if operated as per manufacturers guidelines. I think Van made a good choice here to use it on the RV12.

I enjoyed working with the 912. Cool little engine. BTW we will be offering an EFI setup for the 912UL and 912ULS engines in early 2008 after flight testing is complete.

rv6ejguy said:

I enjoyed working with the 912. Cool little engine. BTW we will be offering an EFI setup for the 912UL and 912ULS engines in early 2008 after flight testing is complete.

Click to expand...

Cool. The market for FI for Rotax is probably huge. If you can make it cheaper than these folks:
http://www.silent-hektik.com/UL_912_1.htm

They also make gears, but max 150 HP and probably wrong ratio for Subaru?:
http://www.silent-hektik.com/UL_R1200_1.htm

David,
<<At 800-900 there is what sounds like gear back lash and general vibration throughout. When the engine is warm, it will idle at 700 with quite a bit of gear noise. Is that a separate issue from the low end vibration you and Dan are discussing? >>

It is exactly the issue we are discussing, and don't do that!

<<I have run the engine without the PSRU and prop and it appears to be quite smooth at idle but there may be stuff going on I can not detect without vibration gages.>>

Engine without psru and prop is an engineered-by-Subaru system. In that configuration you get no torsional resonance, guaranteed.

Rupert,
<<The conventional NSI wisdom (probably an oxymoron) is to keep the idle speed at 2000 RPM or better but I think that the last line of your post suggests that the problem resonance could be anywhere ?>>

I do not know for sure, but I think the F1 is below 2000, for two reasons. First, the overall system is conventional (for our application) and likely falls in the 25 to 50 hz group I mentioned in a previous post. 50 hz in this case would be 1500 RPM. Two, I'm pretty sure NSI knew where the F1 was located, because they hired a Phd to review the system. Only part of the good doctor's report got released, that being the math proving a sprague would eliminate resonance. I have a copy somewhere in my files; all the mechanical review is missing.

In fairness, I have no idea what brand or type of sprague is currently installed in a Maxwell drive. That takes us back to a previous statement; without engineering data, neither fans nor critics know the truth.

Excellent Smithers

rtry9a said:

Re: rotory discussion- GMCjetpilot

No evidence of that-

No way-

The rotaries are quite compact-

It all depends on the specific installation

Click to expand...

I have been schooled. Thanks for taking the time to educate my rusty brain. Very interesting. Even though I don't have plans to build or fly a plane with out a Lyc, I enjoy following the progress. Who knows someday I'll come to the dark side. ha ha.

Egg gearbox failures seem to be entirely random, but perhaps there is a connection. I was scrubbing in the shower this AM (a great place to think), and a detail came to mind. It is just a suspicion, a theory without proof, but maybe worth investigating.

Most of them are bearing failures, yes? The only one I've seen personally seemed to be so; the propshaft had worked it's way forward in the case.

I've previously explained how pilots can run a PSRU system, by luck or on purpose, in operating ranges that don't have torsional issues and thus get very good life. Others may operate in less happy ranges, by ignoring warnings, by lack of warning, by circumstance, or just plain bad luck. Based on observations reported here by Ross and David, let's assume some percentage of operators are cranking and idling less than 1000 RPM. The system seems to be resonant around 800 RPM, meaning it sees high gear loading and likely gear load reversal....which is why it makes noise.

So how does it translate into a bearing failure?

I had a private note mentioning helical gears in the Egg box. A plain helical gear on a shaft generates axial thrust under load, not a problem under normal circumstances; bearings, even plain ball bearings, are rated for some axial thrust. However, a gearbox with load reversal would axially drive the shaft first one way and then the other (fore and aft in this case), as gear tooth contact moved from the front to the back of each tooth. There is some freeplay in the gear mesh, thus during the torque reversal there is some degree of impact when the teeth swap contact faces. The gears are probably as described in some other posts, quite heavy duty for the application, and the tooth impact is no bother to them. However, the impact would drive the gear and shaft axially against the bearing, with impact. The principle would be identical to that of an impact screwdriver, a tool well known to guys who worked on motorcycles in the days of phillips case screws.

Bearings do not survive a lot of impact. Microscopic dents in the balls and races shorten life a lot. In this case, assuming an 800 RPM gear clatter, the axial impact would be happening at 26 times per second.

The degree of impact would be hard to quantify, but it is safe to say axial shaft freeplay would make it worse. I have no idea if any effort is made to precisely control axial shaft freeplay, or even axially preload the bearings at assembly. It should be noted that there isn't any great reason for the designer to worry about precise control of axial freeplay if he doesn't expect load reversal.

Ross, you've seen the components (and you've worked with geaboxes). Are these plain helical gears, and do you think there a possibility of axial impact as installed?

Everyone remember the above is a theory, not a fact.

Dan,

Is that why Rotax uses the Sprague clutch?

Sprague Clutch

Yukon said:

Dan,

Is that why Rotax uses the Sprague clutch?

Click to expand...

I did not know what Sprague clutch was exactly so I Googled, coming up with this interesting technical artical on PSRU's. (May be old news but new to me.)

http://www.prime-mover.org/Engines/Torsional/sport_av92/sport_av92.html

Yes its longer than my post and detailed. Lots of good info for those who want to plow through it. He discusses different couplings, belt, torsional, flex, sprague and fluid. Here's the excerpt on the sprague clutch:

"The fifth option, a sprague (or overrunning) clutch, is not necessarily a coupling for the reduction drive; the few applications of this I'm familiar with use the clutch on the output shaft as a preventative measure. The configuration of this clutch is something similar to a ball bearing in that it has an inner race and an outer race, separated by the inner members(spragues). When the input shaft rotates one way, the spragues (which look something like out-of-round cylinders) jam between the inner and outer race, rigidly connecting the input and output shafts together.

In the case of torsional feedback, when the output shaft wants to turn ahead of the input shaft, the spragues unlock and allow the motion. Since the clutch lets go, the vibration isn't allowed to build past the first cycle. Since no energy is generated, none has to be dissipated, allowing the system to operate in a continuous relatively benign environment. I know of at least two manufacturers that are using this system with good success, although both applications are below 100 hp. (Rotax?)

Will any sprague clutch work? No. The system has to be designed for the load and for the frequency at which the system is expected to vibrate. Surprisingly enough though, some off-the-shelf clutches are applicable to a wide range of variables. The best positioning of the clutch seems to be on the output shaft, the slower rate of rotation resulting in lower hoop and locking stresses and allowing the separation of the prop flight loads from the gear train.

The sprague clutch, however, adds complexity to the reduction drive, driving up parts count and cost. Furthermore, since the mechanism depends on metal-to-metal friction contact, the potential exists for wear and contamination of the gear-train by small metal particles.

Operationally, there is also concern since the propeller freewheels when the throttle is pulled back, causing a significant increase in drag and reduction of the glide ratio. If you're on final and take this drag increase into consideration the consequences are minimal, however, in the case where the engine fails enroute, the reduced glide performance can be the difference between a safe landing and one not quite so. One manufacturer has an option for a disk brake but this, of course, adds weight and further complexity to the system."

Excellent article, George. I love the last sentence.

John,
<<Is that why Rotax uses the Sprague clutch? >>

Friction clutch with limited rotation. Nice design, a true damper.

I think I finally got it -

The Cessna 150 0200 starter had a clutch with a zillion little grabbers that jammed on a flat shaft surface when engaged and then they slid over that surface when the engine started and exceeded the speed of the starter.

Was that a sprauge clutch? I remember rebuilding that little sucker every 50 hours or so.

<<The Cessna 150 0200 starter....Was that a sprauge clutch?>>

Yep.

DanH said:

Egg gearbox failures seem to be entirely random, but perhaps there is a connection. I was scrubbing in the shower this AM (a great place to think), and a detail came to mind. It is just a suspicion, a theory without proof, but maybe worth investigating.

Most of them are bearing failures, yes? The only one I've seen personally seemed to be so; the propshaft had worked it's way forward in the case.

I've previously explained how pilots can run a PSRU system, by luck or on purpose, in operating ranges that don't have torsional issues and thus get very good life. Others may operate in less happy ranges, by ignoring warnings, by lack of warning, by circumstance, or just plain bad luck. Based on observations reported here by Ross and David, let's assume some percentage of operators are cranking and idling less than 1000 RPM. The system seems to be resonant around 800 RPM, meaning it sees high gear loading and likely gear load reversal....which is why it makes noise.

So how does it translate into a bearing failure?

I had a private note mentioning helical gears in the Egg box. A plain helical gear on a shaft generates axial thrust under load, not a problem under normal circumstances; bearings, even plain ball bearings, are rated for some axial thrust. However, a gearbox with load reversal would axially drive the shaft first one way and then the other (fore and aft in this case), as gear tooth contact moved from the front to the back of each tooth. There is some freeplay in the gear mesh, thus during the torque reversal there is some degree of impact when the teeth swap contact faces. The gears are probably as described in some other posts, quite heavy duty for the application, and the tooth impact is no bother to them. However, the impact would drive the gear and shaft axially against the bearing, with impact. The principle would be identical to that of an impact screwdriver, a tool well known to guys who worked on motorcycles in the days of phillips case screws.

Bearings do not survive a lot of impact. Microscopic dents in the balls and races shorten life a lot. In this case, assuming an 800 RPM gear clatter, the axial impact would be happening at 26 times per second.

The degree of impact would be hard to quantify, but it is safe to say axial shaft freeplay would make it worse. I have no idea if any effort is made to precisely control axial shaft freeplay, or even axially preload the bearings at assembly. It should be noted that there isn't any great reason for the designer to worry about precise control of axial freeplay if he doesn't expect load reversal.

Ross, you've seen the components (and you've worked with geaboxes). Are these plain helical gears, and do you think there a possibility of axial impact as installed?

Everyone remember the above is a theory, not a fact.

Click to expand...

Yes, these are plain helical gears and as usual Dan you bring up good points. Most automotive manual gearboxes do not use ball bearings to take up thrust loads on these gears, they run the gears on glass hard shafts with flanges. Only radial loads are imposed on ball and roller bearings in most designs unless angular contact bearings are fitted. High frequency impact loadings is very bad for bearings in my experience but sustained TV is rarely encountered in automotives because real engineering and testing has been done by the OEM in most cases. However we can do stupid things like trying to pull from 500 rpm in top gear and experience TV outside the normal operating range.

Most problems that I've read about do involve bearing issues. If left uncorrected of course, bearing problems quickly cause gear and case problems leading to disintegration.

The one thing I don't like about the Marcotte box is that they have one ball and one roller bearing bearing supporting the drive gear- a bit short on thrust load capability possibly. Twin tapered roller bearings support the cantilevered driven gear so that part is fine. Maybe the fact that everything is so massive in these boxes makes this a non-issue. We shall see.

The thrust loading on helical gears is fairly substantial. If you try to grab neutral at WOT in your car without depressing the clutch, it is very hard to move the shifter. Try it with your foot off the throttle and it glides right out of gear. (Use your beater for this experiment).

Interestingly EPI discusses plain bearings in certain places being superior to ball or roller bearings. I tend to agree. I don't agree with their views that high pressure engine oil is superior to lubricate PSRU gearsets. While this is proven and makes sense in engines like the Merlin with integrated gearboxes, bath lubrication is lighter, simpler and less expensive and extremely well proven in automotive gearboxes and differentials using proper high pressure gear lubricants. Separate lube also separates potential contamination of the two devices so if one fails, it won't cause the other to by circulating metal bits.

Again, TV rears its ugly head. Rotax 912 gearbox failures are common when operators don't keep the idle above 1200 rpm. Egg has now recommended that the H6s idle at 1400 or so. I don't know if this is due to bad experiences in the case of the Subaru package but I don't see any other reason for the recommendation. I did see a TV problem on these right above the starter cranking rpm which manifested itself in some odd kickbacks during some start cycles. Quite surprisingly, this was solved by advancing ignition timing on the ECU. H6s/ Gen 3 with the factory ECU apparently don't have this issue- do they David? The prop MOI may have a small influence here as well compared to the MT. Our tests were with the really light Sensenich hollow carbon fiber blades.

IMO gears don't need to be very big to reliably transmit a mere 200hp but shafts and bearings do need to be stiff and handle the loads especially with large spans between support. Shaft deflection and overloaded bearings were the primary causes of high hp transmission failures. TV is another thing to add to that list with PSRUs.

Sometimes the flying part is less interesting to me than what's driving my RV up front. It is amazing that all that stuff works up there- so far.

BTW, I too do some of my best thinking in the shower or at the gym doing weights- not sure what that might mean?

Heck yes

rv6ejguy said:

Interestingly EPI discusses plain bearings in certain places being superior to ball or roller bearings.

Click to expand...

Isn't plain bearings (babbitt bearings) used on the crankshaft and connecting rod big end bearings in a "modern" engine? However it takes pressure lubrication and careful design.

Wikipidia:
"In some applications rolling-element bearings such as ball, or roller bearings have replaced Babbitt bearings. Though such bearings can offer a lower coefficient of friction than plain bearings, their key advantage is that they can operate reliably without a continuous pressurised supply of lubricant. Ball and roller bearings can also be used in designs and configurations which are required to carry radial as well as axial thrusts. However, rolling element bearings lack beneficial damping and shock load capability provided by fluid film bearings."

In engineering school there are course's just on machine elements like bearings and gears. 20lb text and design books are written on the subject. Its not new technology (invented in 1839 by Isaac Babbitt). The science of how the lubrication film or hydrodynamics works in a bearing is interesting. How would you supply pressurized oil to the gear box? independent pump or from the engine. Either way could work. Its understandable why designers want to use splash lubrication and roller bearings.

It is possible to use a separate pump like some gearboxes have in order to use plain bearings however this is another complication IMO. You could also just tap oil pressure externally off the engine like the Ross drives used to.

I like bath lubrication for existing auto engine PSRUs which already have a crank seal anyway. Dead simple, nothing to go wrong and the lube has EP additives already. In auto transmissions no soft shell is often used. Hardened steel gear with oil slots rides on hardened steel shaft. Lasts forever. Don need no stinkin' 90 psi oil supply.

The big thing with using engine oil for the gearbox is to have easy access to oil for a hydraulic prop. If you must have one of those, this is the easier way to go.

rv6ejguy said:

Quite surprisingly, this was solved by advancing ignition timing on the ECU. H6s/ Gen 3 with the factory ECU apparently don't have this issue- do they David? )

Click to expand...

No kick backs - about 80 hours with 2.5 engine and some 225 with H6, both with Subaru ECU's.

Idle was moved up to 1400 with GEN3 out of concern of the gear lash noise or whatever it is. Curious, when the spring loaded fly wheel was installed with GEN2, idle was OK at 700-800 with no gear noise which was present with GEN1 and a solid fly wheel. Then GEN3 enters the picture and gear noise returns and does not go away until 1200-1300. Seems like the ratio and prop speed are factors along with the spring loaded fly wheel.

This stuff is really complicated. I wish we had an inflight survey of hamonic vibration. Maybe that's what MT did with Andy's airplane. Would be good to know if there are any rpm ranges we should avoid other than less than 1400. I will contact him about it.

DanH said:

John,
<<Is that why Rotax uses the Sprague clutch? >>

Friction clutch with limited rotation. Nice design, a true damper.

Click to expand...

It has two "operational modes" actually. It is a spring loaded dog gear with a friction clutch. When the dog gear climbs on the ramps, the spring is compressed so it will soften all hard torque impulses. The dog gear has a free play of 30 deg or something and within that free play the friction clutch is in operation. I guess this will be approximately analogue to a spring and chock absorber of a car suspension.

Earlier non certified versions (at least 912UL) only had the dog gear, but I think the clutch is now standard on all 912/912S/914, it can also be retrofitted on older ones. I havn't given this much consideration before (none ), so I don't know if the Atec I fly have the clutch or not. Maybe it doesn't, and this is the reason for not flying with zero load on the propeller shaft, since this could probably cause the dog gear to oscillate within its free play.

Anyway, according to Rotax, the clutch/dog gear assembly including correct tension of the spring is the key element regarding reliability of the entire gear train including the crank shaft. If this is important on 100 HP, it surely must be important on 150+ HP with CS prop.

<<It has two "operational modes" actually. It is a spring loaded dog gear with a friction clutch. When the dog gear climbs on the ramps, the spring is compressed so it will soften all hard torque impulses. The dog gear has a free play of 30 deg or something and within that free play the friction clutch is in operation.>>

I looked up drawings this AM for details. I think both of us were wrong <g>

Previously I said the clutch was rotation limited, thinking as you do above that the clutch slips within the constraints of the 30 degree dog freeplay. No so. The dog on the front side of the big gear drives the clutch dog (4) which is free to rotate on (2), limited only by clutch plate friction. The clutch plates drive outer case (1), which in turn drives (2), which drives the propshaft. Clutch plate preload is provided by spring washers (8). The clutch will slip without limitation if shaft torque rises high enough, or if the spring washers lose their ability to provide the axial clamp force on the clutch plates.



The axial force holding the dogs in contact is provided by an entirely different set of much larger spring washers positioned forward of the clutch assembly (and not shown in this drawing). These are the subject of a Rotax service bulletin. It states you can monitor the condition of the springs by examining the propeller torque necessary to slide the dogs through their 30 degree freeplay range.

Engines without the clutch get a different set of dogs. If they provide less than 30 degrees of freeplay, you should be able to tell if a particular engine does or does not have a clutch merely by moving the prop and noting the degree of rotation between dog contact. Can someone confirm?

Ahhh, we're off track again...enough about a Rotax clutch.

Ross, a straight question. Are you subject to a confidentiality agreement with Jan? I understand if you are (I am honoring a few CA's myself), and I'll not bother you further about the gearbox guts.

New thought. Let's suppose you wanted to be sure operators couldn't idle their engine package at a speed low enough to clatter gears at the F1 intersection. However, as David mentioned you might need to allow a low throttle setting for initial starting. Given that the EFI/IGN is software controlled, might one consider a simple idle-up dashpot or solenoid at the throttle body butterfly, with the purpose of bumping idle speed up after, oh, say, 5 seconds of steady state engine operation?

Torsional resonance control via software.

I think you are correct (this time). I was looking at some of the drawing in the service bulletings, but they are not drawn accurately enough so one can figure out what is rotating in relation to what. They are more like sketches, and they clearly state that in a note as well

But this means that it is a "pure" clutch, a torque restrictor, and the original dog gear doesn't really do anything when the clutch is installed ?

rv6ejguy said:

The OE ignition requires around 300 rpm to fire, not 600-700 or the engine would never start.

Click to expand...

I knew if I went by memory I'd be in trouble. You are correct 300RPM min. to fire the dual Ducati ignitions.

I read in a post, in this thread, something (can't find it now) about Rotax 912 / 912S gearbox failures. I'm no expert, but I do have 400+ hours flying them. I know of NO Rotax gear box failures. News like that travels pretty fast, but I have never even heard of a gear box failure on a Rotax 912 / 912S. I'm sure there have been failures, but by the time the Rotax engine made in here to the US the bugs were pretty well worked out is my guess. Good reason why we should look at it closely, it works. If anyone knows of any please post.

Great info by all in this thread. Almost 10K views WOW!

DanH said:

New thought. Let's suppose you wanted to be sure operators couldn't idle their engine package at a speed low enough to clatter gears at the F1 intersection. However, as David mentioned you might need to allow a low throttle setting for initial starting. Given that the EFI/IGN is software controlled, might one consider a simple idle-up dashpot or solenoid at the throttle body butterfly, with the purpose of bumping idle speed up after, oh, say, 5 seconds of steady state engine operation?

Torsional resonance control via software.

Click to expand...

Such a device is installed in the auto intake but was removed for aircraft use, I don't know why. Some of the engines surge at low speed, even in flight. This is caused by the absense of that idle device. A cure for it is the 1400 idle setting. It does not do it at that setting. It is annoying in flight on final at idle power. The engine won't quit, but it sounds like it may. I had it with the 2.5, but not the H6.

prop or engine?

David-aviator said:

Such a device is installed in the auto intake but was removed for aircraft use, I don't know why. Some of the engines surge at low speed, even in flight. This is caused by the absense of that idle device. A cure for it is the 1400 idle setting. It does not do it at that setting. It is annoying in flight on final at idle power. The engine won't quit, but it sounds like it may. I had it with the 2.5, but not the H6.

Click to expand...

I may have missed this before, but is that 1400 idle set on prop or engine rpm?

Gotta be engine - 1400 is way fast for idle power on the prop

<<But this means that it is a "pure" clutch, a torque restrictor, and the original dog gear doesn't really do anything when the clutch is installed ?>>

It is indeed a torque restrictor. Above a set level, it slips. I'm sure the Rotax engineers already had a pretty good idea what shaft torque would break the box or bend the crank nose, so they probably set spring pressure for slip somewhere between that level and, oh, 3 times max mean torque.

The dogs work just like the dogs in the non-clutch gearbox. Apparently the driven dogs on the clutch are a slightly different shape as compared to the non-clutched driven dogs. That is what I would like to confirm, merely as a means of determining if a clutch is installed in any particular box. Anybody know if a non-clutched box has 30 degrees of prop rotation before it turns the crank? I don't think it does.

airguy said:

Gotta be engine - 1400 is way fast for idle power on the prop

Click to expand...

It is engine rpm. The prop is turning at 700.

The RD-1B/C PSRU (Just a point of reference)

All the rotary guys may know this by heart but it might be interesting to dissect the RD-1 drive.

Real World Solutions has articles and detailed pics of his successful drive on his site. It shows how it was developed and the philosophy behind it. The features are: stock Mazda flywheel/flex plate, rubber isolator plate, planetary gear set (I believe from Ford transmission) and pressurized bearing lubrication and spray gear lubrication. It utilizes roller bearing (fwd), plain bearing (aft) and even a flat tapered needle thrust bearing.

Here is a three part article discribing philosophy & design development of the RD-1, including mistakes (which is interesting):
http://www.rotaryaviation.com/PSRU Zen Part 1.htm
http://www.rotaryaviation.com/PSRU Zen Part 2.html
http://www.rotaryaviation.com/PSRU Zen Part 3.html

Now at the risk of making rotary guys heads expand with bragging fodder , may be, just may be, rotary engines are better suited or easier on a PSRU due to their "smoother" output? Also from the pics below, you can see the engine is more compact, thus it allows a longer PSRU. The compact in-line nature of the RD-1B drive looks desirable or elegant in simplicity. The Wankel having the pressurized oil port is also pretty handy. Clearly Tracy made this drive to fit the rotary and learned from others. The off the shelf planetary gear set (Ford Transmission) is a plus. Gear design seems easy but really needs experts with technical backgrounds. Of course having a engineering degree I may be prejudice, but I'm no gear man. Gear designing is a specialty with subtle "tribal knowledge". It's not super human but with computers today, a dude who knows only enough to put numbers in the computer gets, POOF, a CAD drawing and pretty CNC part. It may look nice but an off the self proven Ford gear set is better than a custom unknown one.

Make no mistake, there are real harmonics going on, destructive ones. If not properly designed (and tested) even a Wankel's PSRU can fail, I'm sure, even if a "Mazda goes Hmmmm".

Pics: Output Shaft; Parts disassembled; Plantary Gear Reduction assembled; RD-1 installed in plane



The open flywheel / damper was done for production reasons. I always thought it looked unfinished and crude, but it does give a view of what's going on in there. A solid bell housing casting would look better, but is not cost effective for low production runs as Tracy explains. Clearly he's gone through the Dash 1 through A, B and C. That is good. The longer it stays in service the better. I gather TBO/Major Inspection are in the hundreds of hours? That is probably a must with any PSRU, frequent inspections. Also limits on metal props and acro should put in perspective. PSRU's are not for everyone.

Tracy was fairly inexperienced when he started with his design but seems to have hit a sweet spot after several iterations. I attribute that to: Using off the shelf parts, analysing existing drives (Ross), copying ideas that worked, testing and tear down inspections, willingness to make changes (even painful ones) and making the right design choices, either by talent or luck or may be both. I say luck because he drove a BMW and decided to use elastomer dampers which he carved out shock strut bushings. All kind of on the fly stuff. However going back to my theory, may be the nature of a Wankel is just less critical or harsh on a PSRU? So either he's a natural born designer, got lucky or the Wankel is easier to put a PSRU on? May be a little of all of the above.

The only thing I doubt is he is selling the RD-1 design for other engines? Hmmmm now that does not seem kosher. Every engine needs its own unique PSRU in my opinion. Also in my opinion, every gear box driver should have an OIL TEMP and CHIP detector in their box or oil return line, with idiot light warning. Helicopter guys have temp/chip warning on their gearboxes. When the light comes on they do an immediate emergency landing while praying and sweating.

David-aviator said:

It is engine rpm. The prop is turning at 700.

Click to expand...

Been wondering about that myself, a couple of posts that seem to be in conflict with each other way back at the start of this thread, only made sense if one person was talking prop RPM, and the other engine RPM.

For the sake of all of us following this wonderful exchange, please indicate weather you guys are talking prop or engine when you use RPM numbers.

Thanks.

As was stated earlier, it sure is nice to see this kind of exchange without the personal **** drifting in.

George-
Absolutely correct comments regarding harmonic resonance. As interesting as RWS stuff is, you might also check out the website at http://www.rotaryeng.net/ A lot more technical history and information regarding rotary engines: more theoretical than Crook's decidedly practical approach.

Regarding 912 gearbox failures. Poor choice of words on my part. In my dealing with Rotech (major Canadian Rotax distributor and overhaul facility), they said they had only seen a couple "fail" but dozens completely trashed inside when they get them and making bad noises. So, inflight failures, almost never. These trashed gearboxes showed the classic signs of low rpm operation, contrary to Rotax's recommendations. Apparently some pilots coming from the direct drive world have trouble coming to grips with an engine idle of 1400 and takeoff rpm of 5800.

Rotary engines and PSRUs are still subject to TV concerns. Tracy Crook has applied more engineering than many of the other drive makers and has therefore had better success than many others. No surprise there to me.

There can be endless combinations of gear types, bearings and TV dampers that will work successfully. Proper analysis and testing will doubtless save money and time in the end over eyeball engineering and crossed fingers.

Geico266 said:

I knew if I went by memory I'd be in trouble. You are correct 300RPM min. to fire the dual Ducati ignitions.

I read in a post, in this thread, something (can't find it now) about Rotax 912 / 912S gearbox failures. I'm no expert, but I do have 400+ hours flying them. I know of NO Rotax gear box failures. News like that travels pretty fast, but I have never even heard of a gear box failure on a Rotax 912 / 912S. I'm sure there have been failures, but by the time the Rotax engine made in here to the US the bugs were pretty well worked out is my guess. Good reason why we should look at it closely, it works. If anyone knows of any please post.

Great info by all in this thread. Almost 10K views WOW!

Click to expand...

There was one here that desintegrated in the air some half a year ago. The pilot was cruising along normally, then suddenly with no warning whatsoever he heard a big bang, lots of oil on the canopy and no power. The main gear wheel in the gear box broke and one of the parts left like a projectile through the gear box wall then through the cowling. This was a 912 ULS. You can read it here (if you can read Norwegian I just put it here as a ref), it's on the bottom of the page. http://www.nak.no/mikro/html/flytrygg/Rapporter/rapporter2007.html
The gear box has been sent to Rotax for examination, but so far no news of what might have been the cause.

This was on a Flight Design CTSW, which now also is marketed in the US under the LSA rules. Most of the 912s are used in Europe under the UL rules, like the CTSW. This means the owner are free to do anything he wants with the engine, sometimes (or maybe often??) against the manufacturers recommendations. I have never heard of a certified 912 having had any problems. If this is due to low number of certified engines, or if the certified engines are maintained much more in accordance with the manufacturers recommendations, I don't know. In general though most 912 makes it to 1500 TBO, and they routinely go 2 and 3 times past that when used as trainers.

Harny farny smoogly moogly

SvingenB said:

This was a 912 ULS. You can read it here (if you can read Norwegian I just put it here as a ref), it's on the bottom of the page. http://www.nak.no/mikro/html/flytrygg/Rapporter/rapporter2007.html
The gear box has been sent to Rotax for examination, but so far no news of what might have been the cause.

Click to expand...

I can't read Norwegian and Google translator can either. Can you translate?

Look there are gear boxes that have not failed and there are gear boxes that will fail, sooner or later. This idea of a mechanical device never, ever failing is a dream, yet to be discovered. That's the prime argument against the gear box PSRU's, if its there, it can fail. The corollary: Direct drive engines don't have a gear box to fail, so failure is eliminated. That is with out debate.

Now you can argue Lyc cranks fail. True, but see above, all mechanical things made by man are subject to failure. The crank however is "solid state". It's a moving part but has no internal moving parts itself (unless it has damper weights). Prop is bolted to the crank, done. The simpler you make it the better for reliability. Lyc and Continental figured out if they make the displacement large enough they don't need higher engine RPM, while still being powerful and efficient. Of course nothing is 100% efficient and compromises are always made.

This does not take away from the fact gear boxes can be made to be safe and reliable. The above is just philosophical; if its not there to fail, it can't fail. Many a Continental, Lyc, Pratt, Wright where produced with integral gear boxes. They worked fine, but you had to be careful back driving the prop. Most radials have a gear reduction. They all work fine. On the other hand they where integral, pressure lubricated and had tight tolerances, which adds cost and weight. They used magnesium where they could**. During overhaul its another thing to do and expense. Even finding someone to work on geared engines can be a challenge.

**BMW makes a in-line 6 cyl magnesium/aluminum composite engine; the core is aluminum (with steel sleeves) and the outer engine case is of magnesium, fused together. It saves 24% weight. Magnesium has pros and cons but they put the materials to best effect. The way the make and fuse/bond/marry the alum and mag parts is interesting. Some Cessna twins have magnesium rims. If they ever catch fire, you RUN. Magnesium burns like the sun (and almost impossible to extinguish).

gmcjetpilot said:

I can't read Norwegian and Google translator can either. Can you translate?

Click to expand...

I did. The rest in there is just more elaborate things about what happened during the emergency landing etc + comments from the pilot and the technical comettee. There are no conclusions (nor any speculations about the cause of the cathastrophic-like failure) since no feedback from Rotax has been received yet. If it's of interest I can post it in this forum when it is received.

Philosophies can be tricky. Another philosophy say that no system is better than its weakest link. The weakest link is the engine itself, and a good designed gear box will improve the reliability of the engine because the gear box unloads and filters out the largest and most destructive forces on the crank shaft like bending, axial oscillations, propeller vibrations, gyroscopic forces etc leaving only torque. This is not unsimilar to the suspension of a car. A good suspension is very complicated with lots of parts that takes very hard loads, but I don't think anyone will agree to the notion that the car will become more reliable if the suspension was removed. A car with the wheels bolted directly on to the chassis would desintegrate very fast.

More complicated technology does not automatically equal less reliability. Comparing the Rotax 912/914 and Jabiru 2200/3300 clearly shows that the most complicated of the two also is the most reliable, by huge margins. I don't believe that reason for the reliability of Lycomings is simplicity, I believe that the truth is that the Lycomings are extremely well designed engines.

In all my reading on the popular V12 WW2 engines, the engine itself was far more likely to fail than the reduction gear. The key is to design a reliable box which most of these were. The reduction gear is a very simple device compared to a 48 valve, 12 cylinder, ohc, supercharged engine. The same thing applies today. We just need some science and engineering applied along with some long term testing and the reduction gear is likely to exceed the reliability of any engine bolted to it.

If they could do it 65 years ago, we can certainly do it today. How many manual gearboxes have I have to repair in the last 20 years on my street cars (some with triple the stock hp)? Zero. I used to beat the **** out of them too.

I agree with Dan, buyers should lobby PSRU makers to perform more testing on their drives. Seems to me, this would help sales a fair bit if the manufacturers took this step themselves.