Quantcast

Torsional Dampening

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Hi Bill,

I can't resist a little fun. You do know that there's a 200 hp Renesis flying (successfully) on an RV-4 with the planetary's input shaft bolted hard to an aluminum flywheel, right?

Charlie
Nope. Tell us more. Pictures, names, places, hours flown, actual power it is running, etc are all great.

Billski
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Looking at post #37, that is the first good photo I have seen of the pendulum absorber. Wow, a piece of art, and yeah it would be expensive to build. Looking at it appears to have rollers in tracks set up for three different orders, and I sure would like some more details if they are available. At the time, Ev Hatch was kind of closed mouth on what he did with the torsional pendulum. Wish I had pushed him back then.

There appear to be six rollers with the largest swing radius, two on the back end and four on the front end. Because there are so many, I am guessing these are tuned to firing order at 2/rev. Correct?

There are also four with a smaller swing radius, two front and two back. I am guessing that these are tuned to 2x firing order at 4/rev to take out the oscillation due to accelerating/decelerating the rotors, similar to piston order. In piston engines this one is real but pretty small. Perhaps with Wankels the large rotor mass means this one must be tamed?

There is one more on the front top, with an even smaller swing radius, and I can not even fathom what that one would be. 3x or 4x firing?

Do Wankels have serious 2x/3x/4x firing content? Maybe others? If these are serious, it would account for how Wankels can be problematic to do a good job of isolating the engine from the downstream stuff.

In the car business, we found that we become way smoother than the customer can feel or the parts can benefit from when we do a good job on absorbing firing order. While 2x firing order (pistons accelerating up and down in the bores) is real, it is small amplitude and at 2x firing frequency, its throw on the input shaft is small. Higher orders in piston engines tend to be tiny. Makes me wonder if the Wankel's had even more going on torsionally than piston engines... Billrsv4 or anyone else, what were the orders above firing aiming at?

Billski
 

Autodidact

Well-Known Member
Joined
Oct 21, 2009
Messages
4,513
Location
Oklahoma
The solution they ended up with is rock simple.
Rock simple in concept, but not so simple to manufacture. The parts, including the gears, are all specifically made for their design, and are the result of complex and detailed calculations, I would bet. In other words, they didn't use an automotive (or truck) planetary, and the simplicity was the result of a large amount of development work. I do have respect, sorry I didn't make it clearer.

If someone took a heavy duty planetary and made a reliable rigid system (hard bolted input shaft), without doing the calculations, that would be luck, and "threw something against the wall and it stuck" funny. If the decision to try the rigid connection was influenced by Powersport's experience, then I don't understand the funny part. But if it works, it works...I'd like to see some details, too. If it's real, then it's real, why not show it?
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Rock simple in concept, but not so simple to manufacture. The parts, including the gears, are all specifically made for their design, and are the result of complex and detailed calculations, I would bet. In other words, they didn't use an automotive (or truck) planetary, and the simplicity was the result of a large amount of development work. I do have respect, sorry I didn't make it clearer.

If someone took a heavy duty planetary and made a reliable rigid system (hard bolted input shaft), without doing the calculations, that would be luck, and "threw something against the wall and it stuck" funny. If the decision to try the rigid connection was influenced by Powersport's experience, then I don't understand the funny part. But if it works, it works...I'd like to see some details, too. If it's real, then it's real, why not show it?
This was not Ev Hatch's first rodeo. He quit engineering school to go race hydroplanes, and got into the race business a long time ago. His race engines were noted for being fast and lasting all season. He did engines and whole cars and hydroplanes, and his stuff won races and ran all season. He was used to making things work, including Mazda rotaries. So he came to airplanes with background.

I won't defend all of Ev Hatch's decisions, but he did come to the conclusion that automatic transmission gear sets had too much lash. In a soft system, lash downstream of the isolating element is no big deal as there is little oscillation on the isolated side of the system and the system stays loaded throughout. They had trouble with isolation hardware with the Mazda rotaries, so he explored stiff systems. In stiff systems, all of the engine firing pulses are transmitted through the whole system, and lash means the working teeth unload and then reload with impact on every firing cycle. The bigger the lash, the larger the impact load. In production car/truck planetaries, they have a bunch of lash and the gears would have to be big indeed to stand that pounding. Ev had tried a C6 planetary (in use up through the 5R110 with 650+ ft pounds of torque) and had failures. After the high torque aftermarket C6 parts, there was nothing bigger except some really big Allison stuff, and he decided he did not have a path that way. The Ford and GM boxes capable of 800-1000 ft-lb were years away yet, but now their planetaries might be useful.

Let's tally up the issues with automotive planetaries. They are helical gears with big thrust loads in the elements. They have lash in the input spline, the output spline, between the sun and the planet gears, between the planet gears and ring gear. Then the planet carrier is usually pretty soft, and so is the ring. These things all make for destructive levels of impact loading in the gear set and can make the system too soft to keep resonant mode well above max forcing function frequencies. The last problem is that the parts coming off the shelf are all heat treated splined parts for going in the rest of a specific transmission - there are no parts that can just be "bolted" to a flywheel and the prop shaft. They will still require custom parts, either to adapt the splined gears or as integral gear/shaft parts in a PSRU. Easy for that sort of thing to get into more weight and larger length dimensions for the PSRU. Larger lengths also work against you in building a stiff system, as each part is more flexible when longer, unless you also increase its diameters, which means more weight...

Ev Hatch got tired of breaking planetaries and had eliminated the torsional pendulums as too expensive, so he decided to try a simple gear set with small lash and big gear teeth. Yeah, that did mean custom gears, but only two of them, and both are straight cut spur gears using standard gear making tools. One gear bolts to the engine eccentric shaft and the other gear is integral with the prop shaft - he had to make parts to do these things anyway, so he made them gears too. Nothing fancy and available for reasonable cost. Small lash was achieved which kept impact within range for durability, it was light, stiff, and compact.

As to complex and detailed calculations, nope, none of that. Ev talked with me about his selection process. He did standard gear calcs to select gear size and pitch, taking into account what he knew about firing pulse torque and impact. Then he called his gear shops to find out what standard cutters they had, and went up a couple sizes from what the calcs said. The tech specialist for gears at work looked at the sizing and said he was "way overbuilt but would do well with impact".

This guy who bolted a standard planetary to the flywheel of a Mazda rotary, I would love to see pictures and hear all about how he did it. Website? Reports? Mission profile engine torque and speed? Pictures? How many hours has he got on it? Has he done any snap rolls or spins with it?

Billski
 

Autodidact

Well-Known Member
Joined
Oct 21, 2009
Messages
4,513
Location
Oklahoma
Ev talked with me about his selection process. He did standard gear calcs to select gear size and pitch, taking into account what he knew about firing pulse torque and impact.
How hard is it to judge the influence of what little backlash is possible, on the natural frequency? I guess I'm asking that after the long road that led through that beautiful pendulum damper, was there an unknown that had to be accepted, or if there was an unknown, was the direction of the error known and able to be reasonably accurately accounted for?
 

Billrsv4

Well-Known Member
Joined
Sep 29, 2016
Messages
135
Location
NW Oregon
Rock simple in concept, but not so simple to manufacture. The parts, including the gears, are all specifically made for their design, and are the result of complex and detailed calculations, I would bet. In other words, they didn't use an automotive (or truck) planetary, and the simplicity was the result of a large amount of development work. I do have respect, sorry I didn't make it clearer.

If someone took a heavy duty planetary and made a reliable rigid system (hard bolted input shaft), without doing the calculations, that would be luck, and "threw something against the wall and it stuck" funny. If the decision to try the rigid connection was influenced by Powersport's experience, then I don't understand the funny part. But if it works, it works...I'd like to see some details, too. If it's real, then it's real, why not show it?
Autoaddict,
When testing their early rotary engines Powersport had the engine snap a quill shaft on a dyno designed and running 500+ HP V8s. Their planetary was a custom built low backlash unit. Remember that if you have a additive TV the addition goes in an asymptotic curve. It goes towards infinity. There isn't a tough enough gearbox that can be made that won't break under those conditions. With respect to direct bolt up shafts, all I can say is you can get lucky, but shouldn't count on it.
Bill
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
QUOTE=Autodidact;354332]How hard is it to judge the influence of what litltle backlash is possible, on the natural frequency? I guess I'm asking that after the long road that led through that beautiful pendulum damper, was there an unknown that had to be accepted, or if there was an unknown, was the direction of the error known and able to be reasonably accurately accounted for?[/QUOTE]

There are problems with the question. Let's see if I can explain the system behaviour a bit.

In any system with an engine running at the head of it, the engine has steady state torque with huge torque oscillation. Yep, crank or eccentric torque is going big positive then big negative, and it is not quite sinusoidal, but close. If you are running a stiff system (first resonant mode frequency significantly above the max firing frequency) with no lash, the whole thing will see steady state speed and torque and the big oscillation superimposed on it. If instead you are running stiff system with lash, the system will lift off the lash, maybe crash through to negative torque with impact, then go back through lash and crash again. Kinetic energy of the parts at contact must be absorbed elastically by the system parts, and the loads can be multiplied many times by impact. Looking at impact we use impact factors that are a multiplier of the steady state load.

Steady state load gives no multiplication, impact factor of 1, max load is 1.0*static load.
If the load goes to zero, but the parts still touch, then the full load is released, load factor is 2, max load is 2.0*static load.
If the load does not come off completely, the load factor is between 1 and 2.
If the load is removed, and a gap opened in the load path, then load factor will exceed 2.0 and might be much higher.
So, any lash, if it is being used, means peak load is more than twice the steady state load.

Now what happens when the lash is closed is impact. Hit something with a hammer and you will excite ALL of the natural frequencies in the structure. This is what the design of bells and percussion musical instruments are about. To absorb this impact, the elastic system has to store all of the kinetic energy of the moving parts in the elastic deformation of the parts. If it is a stiff system, the loads to do this can be huge, and your stiff system has to be strong enough to carry those huge loads.

When the system is loaded up, it has the resonant modes and frequencies of the whole system as if there is no lash. When the system has lash opened up, each of the parts will have their own resonant frequencies, but they may vibrate in isolation of each other with low loads internally. When the parts come back into contact under load, they will have impact loads and are then back to having their system modes. Watching a system we have individual parts ringing and the whole system vibrating like it does as a unit. Interesting thing happens with a lot of lash - the total system resonant modes may be difficult to excite and amplify, but the impact loads will tear apart all but the sturdiest of systems..

From talking with Ev (and I am sure Steve Beckham can fill in a LOT of details), after they saw what happened with a lot of lash, they decided to see if they could take off the vibrations from the engine. He did the academic work to figure out what vibe orders were big in a Wankel, what the relative magnitudes of the vibe orders were, and figured out what he needed to hang on the eccentric shaft. This stuff was all worked out prior to WWII with big round engines, and he did a bunch of reading and thinking. After he figured out how expensive his order absorber was, he backed away from it and looked for another another way.

The stiff system with very little lash appealed to him. I was trying convince him that a soft system with a well designed elastic element would work - that we did it all of the time in cars and trucks, and then lash became irrelevant. By this time, Ev had already gone after stiff systems. I did point out that gears would have to be large, lash would have to be very small, and he would need to put some crown in the gear teeth. He remarked at the time, that he was planning on lapping gears together anyway, and lapping in some crown would be no big deal.

Unknowns accepted? There are always unknowns. Sometimes you just do not know all of the things that you are up against. Other times, you have knowledge of system stuff, but you are guessing as to actual loads and that your system is sturdy enough. With a stiff system, they had several unknowns, but I did not know if they were dealing with them, accepting them or were ignorant of them. I doubt they knew just how high the torque oscillations would run. I doubt that they knew if the small lash they could build would keep impact loading within the capability of their parts. I know that they underestimated the loads in the small block Ford bottom end bearings they supported the prop shaft with because those plain bearings failed right at the end of the Sun 60 one year. I am sure there are others. That is why I will decide which, if any, I will buy based upon test flight hours and maneuvering history and customer experience... I will not be an alpha customer, and will not be the first beta either.

Billski
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
927
Location
Jackson
Well, first, I suppose I should specify my credentials. BS. In economics. In 1973. From a liberal arts school. And I'm too cheap to stay in Holiday Inns.

The Renesis I mentioned is flying on an RV-4, the owner's test mule. The drive is a RWS 6-planet 2.85-1; otherwise stock. Not sure about number of hours since conversion to 'direct drive' (note the quotes), but if the weather holds, he'll be flying it up here to Slobovia Outernational next weekend (~2.5-3 hr flight). I'll ask him if he minds being identified, & how many hours are on the mod.

In response to the stuff about lash, resonance frequency above max excitation frequency, etc, here's what my economics degree (and some pretty smart engineers, including the designer of a successful redrive) tell me. Remember, this is the economics major trying to replay what he's been told by engineers.

First, torsional resonance can be avoided by keeping the system's natural frequency either above normal operational excitation freqs (Powersport style), or below (typical planetary systems, belt drives, etc, using 'soft' couplers). For 'soft' systems, the dampers don't 'damp', except at resonance; they are too stiff to give at normal impulse levels. That's why I don't understand the talk of gear hammering in planetaries during normal operation. The factory a/c engines with planetaries only have hammering issues when the prop is allowed to drive the engine. Power input would seem to be only tangentially related to resonance problems. Example: A number of years ago, there was a canard pusher flying a 13B with a Mazda manual transmission locked in 2nd gear. He had years of successful operation, until one day he cranked it up and it was only firing on one rotor. Even though it was idling (poorly, of course), it stripped every tooth on the gear set. (Resonance.) The mention of the original Powersport failing multiple planetaries is news to me (not that I ever knew a lot); all I've ever heard about was the destruction of a coupling shaft on a dyno. That more or less concludes my economics-major 2nd hand knowledge of torsional resonance.

Regardless, there are numerous RWS planetary type drives flying on higher HP installations, including turbo 13B's, a supercharged Renesis on an RV-10, and a P-ported 20B (3 rotor) on a Lancair ES. I'm aware of several different failure modes that are documented with the RWS drive, but I'm not aware of any failures of the planetary itself, if it's been installed properly (alignment issues) and kept fed with oil. The Lancair, in particular, has partially failed (caught before total failure) some of the hard parts of the damper plate, but not the dampers themselves, and no problems with the planetary. It's been operated at roughly the same cruise speeds as a/c engined ES's (with likely higher cooling drag), so HP levels have to be in the same league as the intended 6 cyl a/c engine.

Sorry for the rambling answer, and I'll ask if it's ok to share more info on the direct coupled planetary.

Charlie
 
Last edited:

Autodidact

Well-Known Member
Joined
Oct 21, 2009
Messages
4,513
Location
Oklahoma
Thanks Bill and Billski. Awesome post Billski, and thanks as always for going to the trouble to explain these things to me as well as any other interested folk.
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Well, first, I suppose I should specify my credentials. BS. In economics. In 1973. From a liberal arts school. And I'm too cheap to stay in Holiday Inns.

The Renesis I mentioned is flying on an RV-4, the owner's test mule. The drive is a RWS 6 pinion 2.85-1; otherwise stock. Not sure about number of hours since conversion to 'direct drive' (note the quotes), but if the weather holds, he'll be flying it up here to Slobovia Outernational next weekend (~2.5-3 hr flight). I'll ask him if he minds being identified, & how many hours are on the mod.

In response to the stuff about lash, resonance frequency above max excitation frequency, etc, here's what my economics degree (and some pretty smart engineers, including the designer of a successful redrive) tell me. Remember, this is the economics major trying to replay what he's been told by engineers.

First, torsional resonance can be avoided by keeping the system's natural frequency either above normal operational excitation freqs (Powersport style), or below (typical planetary systems, belt drives, etc, using 'soft' couplers). For 'soft' systems, the dampers don't 'damp', except at resonance; they are too stiff to give at normal impulse levels. That's why I don't understand the talk of gear hammering in planetaries during normal operation. The factory a/c engines with planetaries only have hammering issues when the prop is allowed to drive the engine. Power input would seem to be only tangentially related to resonance problems. Example: A number of years ago, there was a canard pusher flying a 13B with a Mazda manual transmission locked in 2nd gear. He had years of successful operation, until one day he cranked it up and it was only firing on one rotor. Even though it was idling (poorly, of course), it stripped every tooth on the gear set. (Resonance.) The mention of the original Powersport failing multiple planetaries is news to me (not that I ever knew a lot); all I've ever heard about was the destruction of a coupling shaft on a dyno. That more or less concludes my economics-major 2nd hand knowledge of torsional resonance.

Regardless, there are numerous RWS planetary type drives flying on higher HP installations, including turbo 13B's, a supercharged Renesis on an RV-10, and a P-ported 20B (3 rotor) on a Lancair ES. I'm aware of several different failure modes that are documented with the RWS drive, but I'm not aware of any failures of the planetary itself, if it's been installed properly (alignment issues) and kept fed with oil. The Lancair, in particular, has partially failed (caught before total failure) some of the hard parts of the damper plate, but not the dampers themselves, and no problems with the planetary. It's been operated at roughly the same cruise speeds as a/c engined ES's (with likely higher cooling drag), so HP levels have to be in the same league as the intended 6 cyl a/c engine.

Sorry for the rambling answer, and I'll ask if it's ok to share more info on the direct coupled planetary.

Charlie
Charlie,

I do want to know more about these guys with a supposedly stiff system.

Let's talk about soft vs stiff systems. If we have a system that has one natural frequency, and we put in an oscillating input on one side of the spring that is at least twice as fast as the natural frequency it will isolate, that is only a fraction of the input vibration will show up on the other side of the spring. The bigger the ratio of f/fn, the better it isolates. No matter how strong or stiff the system feels by hand, this system has low stiffness compared to its mass, and this is a "soft" system, based entirely upon the fact that we have f/fn >> 1. If we have no springy elements built in, its stiffness is high compared to its mass, and f/fn is << 1, it is stiff system - basically, the whole thing accelerates together. Now if you approach f/fn = 1, the thing wants to amplify. Output vibration will be greater than input vibration. f/fn = 1 and there is no upper bound.

Yeah, PSRU's from RWS are out there, running fine. They are supplied as soft systems, that is they have spring elements between the flywheel and the planetary, and the spring rate is low enough that the 1st resonant mode of the system is substantially below idle. Yeah, I know, it has rubber elements. They are still springs...

If the spring rate in a "soft" system chosen at idle power gave a 1st resonance order at 1/2 of the normal min firing frequency of the engine, and you have one chamber go completely flat, you have a one per rev firing rate and your idle speed is now coincident with the resonant mode. Resonance can grow and tear things up. This is why with a 2 chamber Wankel, you should probably drive for 1st mode resonance substantially lower than 1/2 of min firing frequency. In a piston four banger, losing one cylinder reduces you to two per one rev and one per the next rev, and it really needs a somewhat lower 1st resonance too, but not as much as the rotary.

When a gear set with lash in it operates anywhere near resonance, those amplifying vibrations result in the system parts cycling into big positive and then big negative torques on every vibration swing (firing pulse), going through the lash and crashing dynamically back and forth. If the stresses in the gears exceed their strengths, they break. Tooth breakage is a common failure mode in overloaded gears with tooth face surfaces, splines, bearings, cases, and shafts also being jeopardy.

If the system has lash but is a soft system, the stuff downstream of the soft element (the spring that ensure fn is low) sees only small oscillations compared to the main torque, and the lash never opens, at least while power is on. The GO-480's and the like are reputed to not appreciate being run at small throttle while going downhill fast. What was supposedly happening was the firing pulses were small but still in there, and the mean torque was close to zero, so they oscillated back and forth across the lash, battering the gear teeth on each other and causing pitch line fretting of the gear teeth and other damage over time. I will suggest that their gearboxes were a tad undersize for their mission if this is really true. Might have relevance to RWS issues you mention too.

Now let's think about stiff systems. All parts are stiff and strong compared to their mass, 1st resonance mode is above operating range. This is what we are used to in airplanes. Prop is solidly bolted to the crank flange. Everything vibrates together. Prop sees all of the firing pulses from the engine, including the decels between firing pulses. If you have any lash in the system, the parts will unload during the decel between firing pulses and then close the lash and reload during the firing pulse. If the lash is big enough and the parts have small enough reserve strength, the impact will damage things. If the lash is small enough, and the parts have big enough reserve strength, the kinetic energy obtained during the accel across the open lash will not damage things, and it will run fine.

Ev Hatch went to trouble to get very small lash and beefy gear teeth and substantial associated components. I can tell you (18 years doing transmission engineering for Ford, and knowing that C6 planetary) that even the aftermarket planetaries intended for racing have quite a bit of lash. I do not expect that they will survive long at 200 plus horsepower on a Mazda rotary. My basis is a smart guy that had done all kinds of engine and powertrain builds for all kinds of stuff was breaking these same gearsets with 200 plus HP Mazda rotaries. Ev broke them on his waterbrake dyno, and when he had them running OK on the dyno, Alan Tolle took them aloft in his RV3 and broke them in flight. Alan had multiple dead stick landings when planetaries broke. These were painful lessons and drove them to have gears made that had no lash, and lapped them to finish size. The also had failures of soft systems too, and we never got into those issues, but in retrospect, I wish that I had pushed Ev on that. The info would have been interesting.

Making a 3 rotor run well with a soft system should be OK if it is strong enough. Yeah, spring rate may need to go up, and engine inertia will go up the same amount, but the really big inertia is the prop and it won't go up 50%, so f/fn will get higher and isolate better with the three rotor. Yeah, it will have to stand the bigger torque, but the vibe problem should be easier.

So, if someone else has found a way to make the C6 planetary live in a stiff system, I want to know about it.

Billski
 
Last edited:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,866
Location
Calgary, Alberta, Canada
I read quite a bit over the years on Tracy Crook's development of his gearboxes with regard to TV and gear lash. He took a step outside the norm and INCREASED lash to avoid a problem he saw was there. It worked at the 160hp level as evidenced of his 1500 or so flight hours on his own aircraft as well as thousands more on customers aircraft. At higher HP levels, there is way less time on them to prove the design but a still many hundreds of hours.

Charlie, soft elements damp or absorb some of the torque impulse at all levels, they are much softer than you might imagine- as low at 30 ft./lbs./ degree of deflection. Rubber has a rising rate with deflection.

There are lots of ways to skin this cat and in the end, only a ton of flight hours will truly validate the design, not matter if a TV analysis was done or not but a TV study will give you a ton more confidence that your design is likely to be good before you start ground or flight testing.
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
927
Location
Jackson
Billski,

I was just having some fun with Bill, after he posted the image of PS' pendulum damper. Intent was not to claim the 'direct' coupled planetary is a stiff system; it isn't (see Ross' post above). Just that it might not be that tough to make a planetary work with a rotary.

I've always had a hard time rapping my econ-trained head around Everett's difficulties, since the rotary is a 2 per rev engine like a 4 cyl Lyc, but without the negative net torque swings and with a lower peak to average ratio in the torque curve (if I can believe the internet).

2-rotor-torque.jpg4cyl-V-6cyl-torque-var (1).jpg

If these curves are conceptually correct (ignoring the apples/oranges units of measure and levels), it would seem less problematic; not more. Since so many planetaries are flying successfully, on everything from radials to flat engines to rotaries, is there a chance that he was just missing something? After all, both Lycoming & Continental have made some colossal blunders in recent years, messing up stuff they already had working.

Ross,

I wouldn't argue that some soft systems have some deflection with each impulse, but the example I've been given is the spring system in a clutch plate (IIRC, similar to what Lou Ross used in his drives). I've been told that those springs are strong enough to never deflect except (possibly) when the clutch is initially released, and when the system tries to go into resonance. This could be bad info; I'm an econ major....

I also wouldn't argue that a few thousand hours of operation is where the proof lives.

Bill,

What does the Powersport drive weigh, engine flange to prop flange?

Charlie
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,280
Location
Rocky Mountains
So, if someone else has found a way to make the C6 planetary live in a stiff system, I want to know about it.
Could it be as simple as a different oil viscosity?

In a time far far in the past I had a Fiat 128 that suddenly developed a very severe vibration and noise exiting the freeway after the clutch was depressed. Sounded and felt like a spun rod bearing. After towing the car home and rebuilding the engine* the noise reappeared - again while exiting the freeway. Long story short it turned out to be the transmission and a change from the recommended 90W gear oil to 40W motor oil completely eliminated the problem. This fix was recommended by a friend that was a factory trained Fiat mechanic. Turns out it was a common problem and a 100% reliable fix.

*No anomalies found other than slightly worn valves and the corroded head studs that made head removal difficult.
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,866
Location
Calgary, Alberta, Canada
Clutch springs are nowhere near stiff enough to absorb TV gone amok where we've seen through analysis and measurement peak torque values in the thousands of foot pound range at resonance (which is why stuff almost instantly breaks). Clamp any disc in your vice, weld an old socket to the center of the spline drive, connect your torque wrench and give it a go. I'll bet very few take even 100 foot lbs. torque to bottom them out. At low rpm (typically where most 4 cylinder type models show the worst case TV), these won't do squat. My take on clutch springs are they are there to smooth out initial engagement in normal driving. I've swapped plenty out to solid discs for racing, TV wasn't any different but they were pretty brutal on engagement.
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
927
Location
Jackson
Yeah, I worded that pretty poorly. I'd never expect them to stop actual the kind of destructive resonance being discussed, but the pitch I heard was that it had more to do with extremely low speed stuff, as in the bucking 'off the line' when the clutch is fully engaged too soon at just enough power setting to keep running.

Apologies for not being fully engaged in the discussion today; we're prepping for our annual Pumpkin Drop here at Slobovia Outernational. If you're in striking distance of central Mississippi on Nov 5, come on down, eat some Jambalaya & see if you can hit the target from 200 feet.

Charlie
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Billski,

I was just having some fun with Bill, after he posted the image of PS' pendulum damper. Intent was not to claim the 'direct' coupled planetary is a stiff system; it isn't (see Ross' post above). Just that it might not be that tough to make a planetary work with a rotary.
RWS has always been a soft system, and they work. I never said that they don't.

I've always had a hard time rapping my econ-trained head around Everett's difficulties, since the rotary is a 2 per rev engine like a 4 cyl Lyc, but without the negative net torque swings and with a lower peak to average ratio in the torque curve (if I can believe the internet).
Just because somebody plots some data and shows it does not mean it is broadly correct. Simple physics of IC engines are that they speed up during firing pulses and slow down between firing pulses. In a 2 firing pulse per rev engine (piston four cylinder and Wankel two chamber) the firing pulse lasts roughly one half revolution, and then there is one half revolution of a compression stroke, and then the cycle repeats. Rest assured that the eccentric shaft is slowing down half of the time, and speeding up half of the time. Really. The best way to tell exactly what the engine itself is doing is to strain gage the crank just forward of the flange and read it in real time.

Now if you put an engine on a dynamometer, and look at the continuous output, you won't see that so well, because the electric motor is a huge inertia attached to the engine, and they usually have a spring element in between too. Between these two, you can change the big oscillations into a gentle ripple in the torque signature. Really. Put a big enough inertia downstream of soft enough springs, and the torque signature can look very smooth. But rest assured that those great big M10.9 bolts torqued within a cm of their life holding the flex plate or flywheel to the crank are needed to stand the big firing pulses. In the Powersport, they were working at reducing weight hard, and that makes the size of the speed oscillation bigger too, which might have contributed.

While piston engine vibe is dominated by firing order with 2x firing order about a 25% of firing order (it is the pistons being accelerated in the bores), the Wankel apparently has big firing order and big 2x firing order (rotors are much heavier than pistons), plus at least one other that Ev and Steve felt a roller was justified for...

If these curves are conceptually correct
They are only right when you have the amount and arrangement of inertia and spring that the dyno had. Otherwise, they are fiction. In the auto industry, we only got reliable agreement between calculated data and the real world when we calculated instantaneous torque from gas pressure data and then did calcs down the driveline... after flywheel/clutch cover or flexplate/torque converter and damper springs, the torque ripple could be pretty small.

I wouldn't argue that some soft systems have some deflection with each impulse, but the example I've been given is the spring system in a clutch plate (IIRC, similar to what Lou Ross used in his drives). I've been told that those springs are strong enough to never deflect except (possibly) when the clutch is initially released, and when the system tries to go into resonance. This could be bad info; I'm an econ major....
It is bad info. I did this at Ford and at Chrysler FCA, retiring 18 months ago. Usually these are pretty soft and for exactly the reasons I have been talking about. Everyone was shooting to drive 1st resonant mode below 700 rpm in automatics and even lower in manual trans. That takes soft springs, small preload, and little friction. Many of the designs start moving at around 7 ft-lbs. In manual trans discs, we were trying to keep first mode significantly below idle speed. In automatics, everyone is going for lower and lower clutch engagement rpm. When I left, we were going for 1000 rpm at any throttle in V8 and V6's, so first mode was below 700 rpm, and we were trying for lower. Clutch discs for V8 manual transmission clutches ran with 15-25 ft-lb/ deg springs in the first stage, some had three spring stages to cover the entire engine range. Damper spring capacity always went to about 125% of peak engine torque or higher. In I-4 and V6 automatic trans TC clutch dampers, first stages were around 6-9 ft-lb/deg, and second stages in the 15-35 Nm/deg range, depending upon engines, and damper capacity was 120% of peak engine torque or higher. For the Cummins diesels, the dampers had much higher spring rates and capacities around 800 ft-lb (with the engine at 650 ft-lb). In every case, the compliance of the damper was designed commensurate with engine output, firing pulse behaviour, and inertias of the system.

Billski
 
Last edited:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Clutch springs are nowhere near stiff enough to absorb TV gone amok where we've seen through analysis and measurement peak torque values in the thousands of foot pound range at resonance (which is why stuff almost instantly breaks). Clamp any disc in your vice, weld an old socket to the center of the spline drive, connect your torque wrench and give it a go. I'll bet very few take even 100 foot lbs. torque to bottom them out. At low rpm (typically where most 4 cylinder type models show the worst case TV), these won't do squat. My take on clutch springs are they are there to smooth out initial engagement in normal driving. I've swapped plenty out to solid discs for racing, TV wasn't any different but they were pretty brutal on engagement.
Ross,

Some of the things you say surprise me. Maybe your data predates my time in the industry, and folks were getting away with clutch discs that bottomed below peak engine torque. That did not fit my experience at two auto makers.

I had 23 years doing transmissions starting in 1990, most of it in TC clutch and damper at Ford and then Chrysler, but occaisionally worked with manual trans and even some hybrid stuff, most of it doing vibration isolation, only retired 18 months back. Maybe some of the vehicles you have worked on had weak dampers that would bottom out at WOT, but the last time I saw one was in the 4R70 in the 1990's, and it only lacked about 20 ft-lbs on the 4.6L V8. I fixed that with more spring capacity, as the transmission made unwanted noise and vibration when the damper was bottomed.

In my history, manual clutch plate dampers are designed to give first mode below idle (manual drivers can run anywhere in the engine rpm range), while automatics have been trying to get 1st mode below 700 rpm recently to allow 1000 rpm clutch engagement. Dual mass flywheels on manual trans systems have had similar targets. They all cover peak engine torque with excess to make sure that they do not bottom out. Most have very small preload and very small friction to allow them to isolate well at low torques. Many TC and trans internals can not take full engine torque ripple, it has to be isolated, and we did at both companies. I drove the design for the 6R140 TC, and the clutch and damper capacities went above 750 ft-lb at launch, and has to have been taken higher as the big diesel keeps getting more boost. Worked with a guy on the TC for the Ram Cummins Diesel, and the damper had an 800 ft-lb capacity when the engine made 650 ft-lb.

Now I have seen the suppliers offer dampers with little capacity and at very low prices in attempts at winning the business. They knew that their product would get more expensive when the right springs were used for each engine, but they hoped that they would be the only supplier by the time anyone figured it out. I never let them get away with such shenanigans, requiring that they quote appropriate springs for each engine, but supervisors and engineers not up to speed on the topic could be fooled during supplier selection.

I have seen aftermarket manual clutch discs where the springs were just whatever they had handy, and would bottom out at any real throttle. Criminal to do that to customers, but some aftermarket parts makers did it.


Now these damper spring are not designed to control stuff when resonance is happening. They are designed to drive the 1st resonant order to low enough rpm that the engine is always safely above resonance and is nicely isolated. The challenge becomes giving low enough rate and high enough torque capacity, usually solved with enough package space and enough money for springs.

Billski
 

Autodidact

Well-Known Member
Joined
Oct 21, 2009
Messages
4,513
Location
Oklahoma
Billski, how do the summed up drive-line inertias downstream of typical manual trans spring discs in cars compare with the inertia for a PSRU/propeller - just ballpark comparison - are they ever close at all?
 
Last edited:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Billski, how do the summed up drive-line inertias downstream of typical manual trans spring discs in cars compare with the inertia for a PSRU/propeller - just ballpark comparison - are they ever close at all?
Serious thread drift alert! What the hell... Inertia numbers are almost all way less than unity in SI units and did not stick in my head, but I can talk relative magnitudes.

When you leave the clutch going towards the wheels, there is gearing which reduces the effect of those slowed down inertias. The wheels are always turning slower than the engine. The tallest total gearing I ever saw was about 2.5:1. Inertia of an element reflected to the engine speed is I/SR^2, so the inertia of a heavy element like the wheel/ brake/ tire assemblies is divided by 2.5^2 in 9th gear, and divided by even bigger numbers in lower gears.

The wheel behaves funny in the system. The tire does not just want to rotate about its axle because the tread patch rolls along the pavement. Any torsional oscillation of the tire hub gets turned into the tire rolling forward and backward a minute amount and so the correct inertia for the wheel/brake/tire is taken from the tread patch, not the axle...

Then there are a lot of low inertia, torsionally soft elements between the engine and the wheels.

All that being said, in airplanes, engine inertia is pretty significant, the prop is a bunch more inertia than the engine, and everything else has almost negligible inertia. If the prop is bolted to the crank flange, you do not need a flywheel. If you build a soft system, you need a flywheel or the engine won't idle well... For springiness, you will have a fairly stiff system even with automotive planetaries and light shafting, with fn either in the operating range or just barely above operating range unless you deliberately put in a soft element. The soft element has to be on the order of 1/20th the stiffness of the rest of the system to have any hope of driving Fn below operating range. Stiff system tends to have a bunch of modes way high, and usually you have to deliberately stiffen up the softer pieces of the system to make fn high enough. Soft systems have most of those same high modes, but the engine inertia and prop inertia are vibrating opposite each other across the low rate spring from each other.

Cars/ and trucks? Engine is same, flywheel/clutch cover or flexplate/TC are between the engine and a prop for inertia and tightly coupled to the engine. The transmission and differential/final drive are all lower inertia than the engine but big enough that they can make noise when the modes are excited, while the driveshafts and half shafts and some transmission internal parts tend to be kind of springy. Again, the damper has to be much softer than everything else to have a chance of pulling fn down below the operating range. The tires are big inertia, but their reflected inertia is reduced by 1/SR^2 between the engine and the wheels. There are many modes from any pair of inertia, but mostly you have engine-trans mode you are trying to drive low with the damper. The other parts as you go downstream are increasingly isolated by more sets of inertia and springs. There can be troublesome modes, like coupling between engine and final drive/differential inertia in rear drive vehicles. This has a specific frequency and is only excited when firing frequency coincides with it. If it is objectionable, a tuned frequency absorber (steel tube or ring pressed over a rubber tube) is installed on the shaft at the diff end to pick off the noise. Similarly, you can get wheel/brake/tire vibrating opposite the diff inertia or the engine inertia when its resonance coincides with an engine firing frequency and you again use a tuned absorber to pick it off. Smarter to try to design it off scale, but sometimes that just doesn't work, so you adjust things to have modes in the operating range that you can fix cheaply, like these tuned absorbers.

Billski
 

TFF

Well-Known Member
Joined
Apr 28, 2010
Messages
13,631
Location
Memphis, TN
I'm learning a lot with this thread. Right or wrong what rv6ejguy is describing is what every road racer does if rules allows. I have an old Triumph Spitfire factory racing manual that tells you to pop the springs out and put in blocks to lock the action, if you cant afford a custom plate. Even if running Le Mans, a race car engine is not running a lot compared to a road car or an airplane engine before tear down. Its probably apples vs bananas.
 
Top