Torsional Dampening

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

GESchwarz

Well-Known Member
Joined
Oct 23, 2007
Messages
1,179
Location
Ventura County, California, USofA.
There is a lot of concern about torsional vibration. The reason why springs don't do the trick is because all they do is store the energy and then release it again at their own frequency...and the springs will fail quick due to the high cycle rate. The ideal damper woud work just as shock strut does; it absorbs the impulse energy, then releases it slowly. The problem is that we don't have anything that can do that that is adaptable to a piece of rotating machinery.

Another way to dampen impulses is with shear mass...dead weight. I realize that dead weight = dirty word in aviation. A manual tranny flywheel will take a lot of fight out of the torsional vibration. A manual flywheel is awfully heavy, so how about a custom flywheel of a larger diameter to achieve a greater moment arm but of lighter weight...all the weight at the outside diameter.

Any thoughts on that?
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
To a point that is a potential solution and the proper balancing mechanism is the one that Powersport used in the development of their 200 hp engine. Their product has a hard and direct connection between the engine and the prop - no damping. The whole system was redesigned from the original Mazda configuration by developing a precise dynamic model, then sizing all the balance mechanisms and components for the service environment. But this was not simple to do and it was much more of an issue than just designing a bigger flywheel.

Furthermore, while the flywheel may be a partial solution, it's not the total answer since the power impulses will still exist so you'd still have to somehow counter the engine's energy input from the props harmonic response. Seeing the rather lengthy (and expensive) development seen by Powersport, a damping mechanism is still the ideal solution.
 

GESchwarz

Well-Known Member
Joined
Oct 23, 2007
Messages
1,179
Location
Ventura County, California, USofA.
Agreed. Great article. Just overnight and this morning I've been tinkering with an idea along similar lines..using moving weights that would absorb and release the torsional load spikes. Because it may be patentable that's all I can say for now.
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
You might want to look into the work Powersport did since from what little I seem to remember, their solution was very similar in configuration to the P&W development.
 

jumpinjan

Well-Known Member
Joined
Oct 3, 2004
Messages
313
Location
Dayton OH
This a bit like the dead-blow hammer, where a cavity is partially filled with lead shot. It effectively lowers the peak amplitude and spreads out the duration of the load spike.
This is a pendulum damper. Think of a pendulum (the engine's torsional pendulum) that has a second pendulum attached to the 1st pendulum's mass.
(In essence, that's probably what's happening. I think it's the period differences, where it can't reflect the energy back in a constructive manner. It's not a "tuned" system at all.)
Jan
 
Last edited:

Midniteoyl

Well-Known Member
Joined
Sep 3, 2003
Messages
2,406
Location
Indiana
Exactly. The Rattler is a pendulum system that works at all RPMs. The 'tuning' comes from the weight and size of the main mass, and number of secondary weights (pendulums).

Using lead shot would not work in my opinion as the shot would tend to 'spread' and not act as a single mass.
 

GESchwarz

Well-Known Member
Joined
Oct 23, 2007
Messages
1,179
Location
Ventura County, California, USofA.
The Rattler looks like it would be an effective device. Does anybody have data regarding the amplitude of torsional vibration of the Rotary 13b or the Renesis? By amplitude I mean the actual angular oscillation. The disk travel should be based on the angular oscillation.
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
Just a point of clarification here; the internal torsional issues on the Mazda engines have been taken care of. What we're discussing here is the behavior of the engine/redrive/prop assembly. As such, the primary point of interest is the behavior of the system with the candidate prop you've chosen, and of course the reduction drive (and ratio) you're going to use. Then, each new prop you (or someone else) installs will change those characteristics. For this reason you need to either focus on the particular combination of components that you yourself will want to use, or try to come up with a design that accounts for the widest variety of prop models.

It's these types of issues that account for the fact that if you want to change a prop on your certified engine, you first have to research whether that particular prop is actually approved for your particular engine model and serial number.

Regarding the power pulses and other characteristics contributing to the harmonic behavior, you'll need to go to someone who specializes in the rotary, who understands what it is you're after, and who of course is willing to give you what you need. If you purchase a redrive from an established producer, there is a good chance that they've already gone through the process so they can provide you with a list of props that are OK for your application.
 

expedition2166

Well-Known Member
Joined
Dec 1, 2007
Messages
135
Location
bainbridge ga
well heres just a thought / idea it may be wayyy off topic and have no usefullness at all but

what about a fluid dampener almost like a auto torque converter basically a hollow sealed tube with a vane internal thats independant of the outer allowing the ouside to travel and the internal vane rotate in opposition to the outer it may help in dampening the overall harmonics and possibly dampen the initial jolt caused by the combustion cycle just an idae there but i guess there would be issues possibly with gyro forces dunno just a thought ive had for a while figured id share

i know crazy idea

lol
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
The torque converter configuration, as well as several other similar mechanism have been evaluated at one time or another but the reality of their use has always been determined to be too problematic for a success full application. It is a good idea but in reality it just ends up too heavy and is not 100% efficient in the power transfer unless you figured out a way to lock it - but then you'd loose the benefit of the installation. Other issues are those of cooling the fluid, fluid storage and the "what if" case of loosing the fluid.
 

expedition2166

Well-Known Member
Joined
Dec 1, 2007
Messages
135
Location
bainbridge ga
ok i admit you all got my wheels turning i agree with you orion on conventional fabrication tech. the idea would more then likely be a bit heavier then desired but as we all know great new goodies are coming out every day just a non professional idea here as a continuation to the malay
using a carbon fiber turbine internal with magnets internal and the outer just a milled alum, or al/mag completly sealed and a mild electro mag coil assmbly to add a mild ( i highly empashize mild ) load on the internal turbine / vane assy. in theroy just theroy yould never need to add oil as far as cooling goes have the outer shell finned and let natural convection do the rest the only issue i see and im sure orion would agree is the load it presents on the engine but i think weight ,cooling , and oil loss are things that could be overcome if somone was really interested in doing it
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
6,314
Location
Saline Michigan
I know a few things about this stuff as about a quarter of my professional work is in powertrain NVH for an automotive company.

The No Short Days article takes us through the torsional pendulum in a couple of iterations. And they exist in many aircraft engines. O-470's for instance. The Rattler is also a torsional pendulum, marketed for the high performance automotive crowd. Torsional pendulums are generically an order tuned vibration dampener. The force that returns a pendulum towards its rest point is the g's it is under. Well, as the rpms increase, the g's go up, and, as it turns out, a pendulum, once tuned to match the firing order (3 per rev on a six cylinder engine) at one speed, it also matches firing order at all other speeds. The down side is that it only absorbs the exact order it is tuned for, so if you have multiple vibrations to pick off, you will need multiple tuned pendulums. The Rattler actually has rollers of two diameters, all in the same size bores, so it appears to have some tuned for firing order (4 per rev in a V8) and others tuned for 8 per rev, which is the second biggest vibration in an engine. You do have to have an even firing interval engine for this to work cleanly. When a firing pulse comes along, the pendulum swings one way absorbing energy and as it subsides, the pendulum swings back giving the energy back. They take out zero power and if they are adequately sized, they pluck off a huge fraction of the vibration - 98% of the tuned order is doable. And they date from the 1930's, so while you might succeed in coming up with a new way to do one (a design patent), someone interested should have no trouble finding a way around it for their own design. When Everett Hatch was developing the Powersport engine, he did investigate these, but decided to try the high stiffness zero play geared system that did ultimately work. Talking with Everett about his PSRU's, he recognized that it would be the ultimate way to solve things, but he also recognized that getting the tuning right might be very hard to do. I currently have a $70k piece of hardware in my locked cabinet at work that would have worked great for confirmation and tuning of a TP, but Everett was past this topic by the time he and I became acquainted, and he had left us by the time got the knowledge to have helped him with the topic.

Now the viscous dampener is also out there in many forms, including in the automotive aftermarket. The advantage to the viscous dampener is that it picks off energy from vibrations over a wide range of frequencies and does not have to be tuned. The disadvantage is that it turns the vibrational energy it takes out of the engine into heat, thus diverting your hard earned engine power. It also only gets some of each vibe that it sees.

More inertia in the flywheel? Sure, you can reduce the amplitude of the firing pulses with more inertia. And you can get the inertia with less weight by using a larger radius, but there are practical limits. I = sum(dm*r^2). The biggest problem with adding inertia to the engine as a tool for fixing things is that the pulse is still there, and now all of the vibrational modes present in the system will occur at a lower frequency. If your problem is one mode is just a little too high, you might move it off scale or to a less troublesome place this way. And you might move another mode that was out of your operating range on the high side into your operating range.

Another method is to change component stiffness to drive frequencies away. There is not much hope for trial and error here.

The really big problem with all of this stuff is that it requires really good analysis of your resonance modes so that you know what is out there trying to break your airplane, and then you need a high tech rotational vibratory measuring and data analysis system to tune it and then confirm that some deadly mode is not lurking out there to blow your gadget off the airplane when over rough terrain...

As much as I know about this, I will not invent my own PSRU.

Billski
 

jumpinjan

Well-Known Member
Joined
Oct 3, 2004
Messages
313
Location
Dayton OH
Bill,
Could you mention a little about your vibratory instrumentation? Is the sensor small enough & have the capability to where it could be attached to various shafts & gears inside the system? I'm just trying to get a feel about it's capability. What's the sensor consist of?
Jan
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
6,314
Location
Saline Michigan
The method is used by a couple companies. The outfit that supplies us is Rotec out of Munich. They supply a magneto-resistive sensor, but we primarily just use a mag pickup. This reads any ferrous gear or wheel with interruptions. That signal is converted to a TTL square wave by a signal processor, and that is used by the box. An alternate method is to print out and attach a target with a bar code looking target and read it with a laser tachometer. Yeah, print it on sticky-back labels. This is particularly useful with aluminium parts or small low inertia stuff where a toothed wheel would change the inertia too much. The laser tach can fit in some mighty small places.

You start by telling the software how many teeth and what ratios you have on each of the targets, and then run speed sweeps. The gadget reads each tooth passage and records it, then has software to correct for regular tooth errors, missing teeth, etc, then can make many different plots of what moving by itself and moving relative to the other stuff you pointed sensors at.

On a RWD engine and automatic transmission, we usually point one sensor at the starter ring gear, another at a part that turns at turbine speed or a ratio from turbine speed, another at a 60 tooth wheel attached to the output shaft, and another at either the diff pinion or a 60 tooth wheel on the diff flange. Then if we have something specific to look at, we will find something to point a sensor at that reflects what we are looking for internally. Automatic transmissions are loaded with gears, splines, clutch packs with formed splines, etc. Likewise PSRU's have gears and shafts with splines. You should be able to drill and tap a 10mm hole someplace for each shaft. On FWD, we pickup the engine, turbine, and output shafts. Other folks work on the front end accessories, and put toothed wheels on the crank pulley and each of the accessories.

In all of these cases, we can watch the excitation (engine firing order, 2 per rev on u-joints, compressor pumping order, etc) and watch the way that the other parts vibrate. Eventually, you find the things that are moving together and figure out what can be done to fix the problems.

It is work to check out what you already have. If you are designing from scratch, the software to analyze and design systems is out there at impressive costs too. Or you can just write your own stuff from first principles.

Billski
 

lehanover

Member
Joined
Jul 12, 2005
Messages
10
Location
Hebron Ohio in summer. Zephyrhills Florida in wint
The rotary has little torque, as the output is geared up 1:3 from rotor speed, so a reduction unit between 2:1 and 3:1 is used. Where a piston engine actually sees reversals in torque, and requires extensive damping, the rotary has only positive going torque pulses, and has less complex damping requirements. The poor torque requires that the engine be run in the 5,000 to 6,000 RPM range for a reasonable horse power output. The troublesome RPM range is close to idle speed, where accelerations can damage straightcut gears in reduction boxes.

Non straight cut gears have much less a problem such as plentary sets. More planet gears produce fewer problems. Tracy Crooks damper system involves only 4 plastic pucks to absorb any unhappy accerlerations, and works flawlessly. The planetary set is from a Ford diesel truck
transmission with 6 planet gears, and is quite robust.

Lynn E. Hanover
 

orion

Well-Known Member
Joined
Mar 2, 2003
Messages
5,800
Location
Western Washington
One of the biggest and most problematic assumptions in this discussion is the notion of "If I don't feel it, it isn't there". The fact is that the rotary as applied to aircraft does have very serious torsional feedback issues and only a glance at the various developments is needed to confirm the complexity of the problem. And in virtually all cases idle is not the problem, although the harmonic of the more conventional aircraft engines certainly seems to suggest that that is the case. But at idle the torque generated is very low so even if the engine/prop combination does develop a feedback cycle, the torque multiplier generally stays below the failure threshold.

But even in the Lycs and Continentals the problems that arise usually do so at higher frequencies. It's one of the reasons that quite a number of installations have very narrow green arcs, surrounded by yellow or red ones. A C-182 I got checked out in some time back had an O-470 whose green arc extended only from about 2,200 rpm to about 2,450. On take-off you could stay at full rpm for only about 30 seconds before having to throttle back - barely enough time to get off the ground and start the climb. The plane's tiny green arc was surrounded by large yellow arcs, which means that extended operation in those rpm ranges the engine/prop combination could develop a destructive feedback so prolonged operation there was prohibited.

Two of the primary aviation developments whose products centered about the Mazda 13B both had failures in testing, both occurring at about 4,500 rpm. One of these was the Powersport program and the other was conducted by Hayes Rotary Engineering (and I was involved with the latter). Powersport got a very direct lesson in this when the testing destroyed their dyno (which was rated for torque levels nearly ten times that of the rotary). Failure analysis showed a feedback harmonic just over 4,000 rpm that caused a torque spike sufficient enough to shear the dyno's shaft from its flywheel.

Hayes had a similar experience however there we installed a sprint-car gearbox between the engine and the dyno components. In our case the torque spike (at about 4,550 rpm) pressed the gearbox input bearing right out the side of the case.

And the smaller units (engines based on the Norton and Suzuki products) have had similar issues, destroying several gearboxes and/or couplings in the process.

In short, this is not a simple task and assuming that there is no problem just because you don't feel a vibration at idle, will most likely lead to a relatively expensive and possibly painful lesson.
 
2
Group Builder
Top