Quantcast

Torsional Dampening

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

lehanover

Member
Joined
Jul 12, 2005
Messages
10
Location
Hebron Ohio in summer. Zephyrhills Florida in wint
When you start dynoing engines, it becomes obvious that a mass of sypathetic harmonics will hunt you down, and damage your equipment. The first thing I noticed was that the grease was flying out of the needle race in the universal joints at each end of the drive shaft. In 4 cylinder piston engines there are two major RPM ranges where this happens. Both are below where any race engine would be tested, so as long as the lower of those is avoided, there is no problem. Flat 4 cylinder engines, like all VWs and airplane engines, pound out the center main bearing saddle at cruise RPM. All certified airplanes have a propeller and engine combination that must not be changed. There is a minumum diameter for each propeller that must be maintained, or the blade cannot be returned to service. This is where the lost blades and broken cranks come from. The narrow RPM ranges of aircraft engines must be observed lest the crank find one of its unhappy nodes, and begin twisting and untwisting like a Slinky. Note that Lycoming has just gone through the recall of hundreds of crankshafts for massive failures, that resulted from a change of forging in one strike of the dies, rather than 2 strikes that has been the case since day one. Several lives were lost so Lycoming could experiment with production cranks made a few cents cheaper each. Note that some 6 cylinder cranks have counterweights mounted on pins with plastic sleeves between the two? And yes, sometimes those weights come off. Not a confidence builder in my book.

All of this has to do with harmonics generated by combinations of weights and coupling stiffness in the drive lines. So when you dyno any engine, you are doing so with a set of sympathetic devices included. The attach plate connecting the engine to the drive shaft. The drive shaft. the U joints, the dyno impeller and feedback from the vane count inside the dyno absorbtion unit. In any case where you linger at an RPM where this harmonic or one of the family of harmonics agrees with one of the engines problem harmonics, your equipment will decompose quickly. The flywheel effect of the prop is part of this function as well. A metal prop might add or subtract from the problem by moving the problem up or down the RPM range.
A wood prop might eliminate the problem or move the problem well outside the normal RPM range.

Whatever the reason, the combination of Tracy's adaptor plate weight and the Shore of the pucks he uses, seems to do the job nicely. His rotary powered RV has won the Sun 100 and is still fast with the Renesis engine. He also is completing a RV-8 with a three rotor.

The rotary is happy up to 6,500 RPM and as slow as 5,000 RPM. So the 2.78:1 gearbox seems the better choice. The engines will run 200 degrees lean of peak EGT and deliver close to piston engine economy.

Lynn E. Hanover
Rotary racing engines
since 1980.
 

Midniteoyl

Well-Known Member
Joined
Sep 3, 2003
Messages
2,406
Location
Indiana
YA, prolly not, but I'm sure that if you have he info they could build on for it. ATI also builds custom dampeners.
 

Starman

Well-Known Member
Joined
May 16, 2009
Messages
2,011
Location
High in the Andes Mountains
Another method is to change component stiffness to drive frequencies away. There is not much hope for trial and error here.
I think the best solution is to use a torsion bar for a drive shaft. It would have a natural frequency of only one or two cycles per second, way below any possible problem frequencies. Easier to implement on a pusher than a tractor.
 

Topaz

Super Moderator
Staff member
Log Member
Joined
Jul 29, 2005
Messages
14,208
Location
Orange County, California
Has anyone tried this?
There have been quite a few helicopters that drive the tail rotor this way, amply supported along the length to prevent 'whipping'. The Porsche 924 used a similar technique for its main driveshaft, too.

I've never seen it done for higher-power applications. I can imagine that making sure the higher-order harmonics in such a setup are also outside the engine's power range must be... challenging.
 

PTAirco

Well-Known Member
Joined
Sep 20, 2003
Messages
3,562
Location
Corona CA
I think the best solution is to use a torsion bar for a drive shaft. It would have a natural frequency of only one or two cycles per second, way below any possible problem frequencies. Easier to implement on a pusher than a tractor.
I have seen it on a Tailwind with a Type 4 VW - it drove the shaft from the flywheel end via a belt and carried it over the top of the engine to the front where it had another bearing supporting the hub . The shaft appeared to be about an inch in dia.
 

Starman

Well-Known Member
Joined
May 16, 2009
Messages
2,011
Location
High in the Andes Mountains
I think that if you start making your torsion bar to short or too stiff you are asking for harmonic resonance problems. If you make it loose and soft there should be no torsional resonance problems. I envision a good torsion bar would be one where it's weak enough so that it will wind it up about 90 to 180 degrees at it's operating torque. If a bar is wound up 180 degrees then it doesn't matter how much torsional jiggling is going on at the busy end because the spring is just too soft to transfer it. It will not transfer because what we are talking about transferring is torsional momentum, and torsional momentum at a radius of 1/2" is insignificant over the length of the bar.

An alternate way is the natural frequency. Like an pendulum spring (like in a clock) it will have a natural frequency. You wind it up a bit and let it go and it will spring back and forth. If it does one or two cycles per second I would call it loose and flexible and should never be a problem - as far as torsional harmonics are concerned -

BUT

There's another concern, and that is the flexibility of the shaft in bending. A short stiff torsion bar won't have any bending issues.

If your bar is sloppy enough it will need center bearings, and placing them can be tricky. You wouldn't want to put a bearing at the 50% point, or 33%, or 25%, etc, not should the end ones go right at the ends. Figuring out where to put the bearings on a floppy bar could be done with trial and error if you can make it so you can slide the bearings on the bar.

SO

The ideal bar is one that is stiff enough so that it doesn't need any center bearings but loose enough so that there is no chance of torsional harmonics transferring.

However, if you use a proper torsion bar you can be sure there will be no torsional problems - ever!

Another consideration. If the torsion bar is stiff enough it could be possible to get rid of the flywheel on the engine so you can save some weight.
 

cnctube

Member
Joined
Dec 28, 2010
Messages
8
Location
Boise
Just brainstorming here.

Why not integrate a pendulum damper into the flywheel?
Custom flywheel with milled pockets to hold the pendulium weights.

Model the system mathematically, create a Campbell plot to show areas of concern.
Then see if a pendulum damper would help.

How to model a pendulum damper in the flywheel is way beyond me though.

Anyway just throwing out a crazy idea.
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Some other people are way ahead of you. LuK (a tier one automotive supplier of clutches, torque converters, and other drivetrain stuff) has been trying to interest people in the topic for several years. The Rattler (no longer produced) was on the front end of the driveshaft, but the flywheel allows it to weigh less by increasing it operating radius. In cars and trucks with automatic transmissions, the best place is after the damper springs. It can do its work with minimum mass there. Works great, and is now in production on Diesel engined vehicles by several brands. Daimler may have been first to market, but others are close behind for diesels.

Billski
 

slociviccoupe

Well-Known Member
Joined
Apr 1, 2011
Messages
282
Location
Palm Bay Fl.
was already previously mentioned but why not have the flywheel made more like a fluidamper or the ones made by ati? and even simpler some driveshafts in cars or trucks either being rwd or fwd axles have a rubber donut around its shaft to absorb some of the resonant frequencies. im interested in the torsional dampening problem also but of a different nature with a subaru flat 4 and choosing a redrive.
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Resonance is is a little tough for a lot of folks to get their brain around initially and I have not tried teaching the topic yet, so here goes.

First case - A weight on an arm that pivots at its grounded end is a pendulum. If it is on earth it will have a swing frequency related to its length, its mass and the acceleration due to gravity (32.2 feet/s/s). Disturb it, and it will swing at that basic frequency. Now if that pendulum is a swing in the park, and you give it a little push each time it goes one way, you increase its magnitude, it you put in more energy than friction takes out, you can cause the amplitude of the cycle to grow. This is a natural harmonic with amplitude that is being grown through resonant pumping.

Second case - One weight that is grounded through a coil spring that can be compressed or lengthened. Shove the weight and let go of it and it will oscillate at a frequency that is entirely a function of spring rate and the magnitude of the mass. It too can be pumped to higher and higher amplitude if an input is put in that is larger than friction taking energy out. Now here is the thing to remember about this resonance - if your input matches the resonant frequency, there is no upper bound to the growth except physical stops or breakage.

Third case - Two weights with a spring between them. Disturb this system, and it will behave much as the single mass system did, except now you have two masses moving. If the input pulses occur within 25% of the resonance freq, it will amplify rather easily, and if they occur at the same frequency as the resonance, it WILL tear itself apart.

Fourth case - you have more three masses with springs between them. You will have a resonant freq for mass 1 vs mass 2, mass 2 vs mass 3, and mass 1 vs mass 3. Any one of these resonances can be excited by having an input at near that frequency. This can be extended to as many masses and springs as your system has. And any of the three can be pumped if something else is putting energy in at close to any of those resonant frequencies...

Now I mentioned coil springs and talked about this like they were linear vibrations, but they can be torsional masses (inertia) and torsional springs just as well. For an engine - psru- prop, you actually have a springy and massive crankshaft, a flywheel, shafts and gears (more springs and masses), and the prop (another big inertia). There can be a bunch of resonant frequencies in such a gadget.

Engines have many orders of vibrations and a wide range of engine speeds. For instance a flat four cylinder (Subaru/ Lycoming/ Continental/ VW) has firing at 2nd order (2 per rev, and this is the biggest one when the engine is running well), pistons going up and down at 4th order (4 per rev, second biggest under normal operation), 1st order and 1/2 order due to variation in power between cylinders (these are usually pretty small, but can be big if one or more cylinders are not making power). Then there are other orders due to accessories too. A Subie can run from 1000 to 5000 rpm, so it makes vibratory inputs at frequencies ranging from 8 to 167 Hz.

Now if any of the input vibrations is within 25% of any of the resonant vibrations, they can amplify, and if they amplify much, you can break stuff. Really.

Now how to fix this stuff? The usual way is to just (by design) drive all of the resonant modes out of range of the major vibratory inputs. This can drive very stiff and massive parts (the reduction drive on the various big recip airplane engines come to mind), deliberately very flexible parts (deliberately soft shafts, spring devices, and rubber elements all come to mind), and other solutions.

As to absorbing frequencies, this can be done. To do this, you can install a device that picks off energy when the motion is big. Let's see how. A clutch that slips or a hydraulic damper - both convert energy to heat and can be made large enough. So what is wrong with this? Well, if the vibration mode is within your flight mode, there can be a lot of energy that is converted to heat that has to be disposed of, and that can be a pretty big cooling job to keep the parts at a reasonable temperature. And the energy that is being picked off in a propulsion system is propulsive energy that you burned fuel to make and are then throwing away... Another way is to figure out what frequencies you want to pick off, and put a tuned absorber on at that frequency. This is just an inertia hung on a spring that is tuned to that frequency. We have them on automotive crankshaft pulleys. They work great on exactly one frequency. So if you have just one frequency to worry over, we have a gadget. I suppose that you could hand several, but again, this is kind of a limited tool.

Yeah, you can put in some rubber components. They pick off a few percent of every deflection and rebound that they see, convert it to heat, that then has to be dumped overboard. Some thing to remember about rubbery materials - their spring rates and damping characteristics change with temperature and they are hard to cool effectively. So, they have all of the problems of a hydraulic damper, only more so. But rubber has other issues. Rubber pieces are springs that lower the resonant frequencies of pieces near them. Now that can be good if you have a frequency that is close to the lower end of the working range, as it allows you to shove that frequency off range low. Trouble is that it can also drag a frequency that used to be off range high down into your operating range. Careful how you do this...

So, by changing inertia or spring rates of various pieces, you can move the various resonant modes. The goal is to drive them all off range.

There is another method, which is the torsional pendulum. This is an order absorber. Design one to pick off firing order, another to pick off twice firing order, another to pick off any other bothersome orders. These guys temporarily absorb energy on the acceleration side of a pulse and give it back on the deceleration side of the pulse. Neat! No lost energy, works great. In airplane engines, they have been around since the 1930's. Every Continental O-470 has four of them, but they do have a few issues. Most are kind of fragile. Careful with that prop speed lever, you have to move it slowly. Now some folks have been making them for cars, quiet at start up and sturdy too. Neat, but again, you do have FEA with dynamics packages right? And a rotational vibration measurement system to check everything out with once it is built?

Supposedly, there are folks who have done successful setups. For me to be convinced that they have an excellent working setup, I would need either terrific FEA plus validation with rotary vibration measurement equipment or a lot of history with the engine and prop in question. The builders of your PSRU did model the parts and system, perform FEA with an Eigen solver, right? How about a few engines with your propeller that have gone 2000 flight hours? Maybe someone has put a torsional vibration measurement system on the power team and run the whole operating range including induced engine misses? Hmm. When I talk with these guys, the best I have gotten is mention of how many are running and how a couple have a bunch of hours... That's nice.

I personally have fixed an airplane system by adjusting inertias. It never did fly due to other issues (that pesky issue of design, tooling and flight test costing more money than anticipated). It did run 180 hours on a stand with zero vibe issues, while some of the previous schemes ran for minutes before tearing themselves to pieces...

Billski
 
Last edited:

Billrsv4

Well-Known Member
Joined
Sep 29, 2016
Messages
135
Location
NW Oregon
Billski,
I also want to thank you for your fine write ups. I am late to this party but wanted to update some of the info mentioned about Powersport. Everett did design a pendroulus damper for their superlite engine. It worked great. The reason they didn't produce the superlite for sale was cost pure and simple. The later model single reduction psru was designed to a stiff model. This places all the orders of TV above the operating range of the rotary. This uses straight cut gears. The reason is the lash can be controlled and adjusted. Planetary gears required too much dampening. Note that all the big radials with reduction used planetarys with straight cut gears. These guys were well aware of helicals. The straight cut gears could be controlled with center distance. Also less power is wasted in end thrust. Dave Leonard ran Tracy's transmission derived planetary at high power and cooked the thrust bearings. While helical gears could be used but they bring another set of problems with them.
Bill
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Thanks for the info on Powersport. I knew that Steve and Ev had built a pendulum damper, and that they vetoed it because it was costly. The new production ones for automotive use are running about $40 delta in a $100 torque converter with production tooling and machines. With Powersport volumes, that would have been quite a chunk of change.

I continue to look forward to Powersport.

Billski
 

Billrsv4

Well-Known Member
Joined
Sep 29, 2016
Messages
135
Location
NW Oregon
Just in case everybody thinks that it is easy to design a dampener for a rotary and planetary, I wanted to pass this along. dampened eshaft.jpg This is why they decided it would be too expensive to produce the planetary reduced version of the 200 HP engine.
Bill
 

rv7charlie

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

Autodidact

Well-Known Member
Joined
Oct 21, 2009
Messages
4,513
Location
Oklahoma
Something is not adding up here. Lots of time and money expended to find a solution for TVR on wankels - supposedly because there have been problems and a solution was needed - but, turns out all you need to do is bolt the input shaft straight to the crank and it works fine so why are you guys going to so much trouble? If it is so simple, is everyone else just clueless?
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,135
Location
Saline Michigan
Going to so much trouble? Have some respect here. Smart people broke everything that was conventional and already available in the 1990's. Ev Hatch (RIP my friend) started with 215 HP and the Ross planetary type. They broke them in short order, designed their own and broke ever more beefy versions, destroyed rubber dampers, etc. Alan Tolle did a bunch more dead-stick landings than you would tolerate. They were running a real 215 hp, and flying at that power in order to test hard. They were also going after a big weight reduction, which probably gave them some of their problems. When they looked at the other setups, they concluded that the other guys were running a little at 160 hp, and a lot of their time at 125-150 hp. Sure the engines were capable of more, but they were prop'd for cruise at less, and really could not get to the high power.

The solution they ended up with is rock simple. No torsional absorbers or rubber dampers at all. Just an external spur gear bolted to the eccentric and supported on the other end, a big internal gear integral with the prop shaft and a couple bearings of suitable size, a case, and lube.

In between the planetary at the beginning and the external-internal spur gear set at the end, Ev Hatch built the torsional pendulum absorber shown in post 37. It worked, allowed a very lightweight setup, but was going to cost them out of the market. That was why he developed the external-internal spur gear set. Simple, buildable, robust, reliable, and it made for an affordable powerplant. The folks who bought the design and tools from Ev Hatch's estate did get a couple products out before their parent company went bust, and I never heard anything bad said about the PSRU. Those guys also had funny ideas about engine management resulting high fuel use, but that is a whole 'nother story.

I look forward to Bill and Steve getting to the market.

Billski
 
Top