Just read the sticky on TV. It is still not clear to me if there is a relationship between the rpm of the shaft and the overall force of the vibration. In a straight-4 engine would the torsional vibration be worse at higher rpms than the same type engine at lower rpms at the same power. In other words, take two straight-4 engines of the same displacement, one engine turns at 3000 rpm at 100HP, and the other at 6000 rpm at 100HP. All else being equal would the higher rpm engine produce more vibration, or less, or equal?
Torsional Vibration Question
- Thread starter Nims11
- Start date
Help Support HomeBuiltAirplanes.com:
rv6ejguy
Well-Known Member
The concerning points of TV are generally where system resonance occurs and peak torque values might be 10 to 50 times the mean torque values just above and below the point of resonance. This can easily break things. The highest TV amplitude can occur at any rpm.
Manufacturers 'tune' and or dampen the resonance to suit the application, a car engine will be smooth as silk at idle because it spends most of it's time there and no one wants their dash vibrating and the car shaking as they wait for a red light. A Lycoming is balanced for 2700 rpm because that's where it spends most of it's time, etc, etc.
So there is no answer to your question other than any period of resonance can be dealt with, but it takes effort and sometimes added weight in flywheel (often in the case of a diesel) and or balance shafts or extra cylinders.
Autodidact
Well-Known Member
Very good demonstration of forced vibration resonance with an elastic system's natural frequency. Just imagine a torsion bar spring and flywheel with cylinders igniting, instead of coil springs and mass with a solenoid providing the forced vibrations to the system. The natural frequency of the system can be changed by changing the spring rate, or the mass of the weight, or both:
Last edited by a moderator:
autoreply
Well-Known Member
And for a more extensive explanation:
https://www.youtube.com/playlist?list=PLrwuNGSwGLHebv2R8bMnMhpxVOE0gtugQ
Covers both TV and flutter basics.
https://www.youtube.com/playlist?list=PLrwuNGSwGLHebv2R8bMnMhpxVOE0gtugQ
Covers both TV and flutter basics.
Good find! I'll be forwarding this to some ASTM members that aren't engineer types to explain why a proposed change in engine certification is a bad idea. Someone needs to set something like this up at Oshkosh for the crowd to play with. Take it one step further and turn down the amplitude of the exciting force to show that it is largely irrelevant.
Last edited:
Good information. I understand better what torsional vibration is, but I’m still trying to get my head around all that causes it in a piston engine.
In a straight-4 engine every power stroke at the piston, especially farthest away from the output shaft, will create a flex perpendicular to the shaft, or cause a rocking tendency, correct? With a boxer engine the force of one power stroke on the shaft will be countered by a force 180 degrees, opposite side of the shaft and therefore counter balance the forces on the shaft, correct? Does the fact the boxer engine has a shorter shaft than the equivalent cylindered straight-4 have a significant effect? If so then it follows that a straight-3 engine would be less prone to TV than a straight-4?
Also, I’ve noticed that the boxer aircraft engines, like the Lycomings, have a much larger bore to stroke, the large bore give it the torque needed at lower rpm, though that could have also been achieved with a larger stroke. With a shorter stroke the power pulse on the shaft will be more evenly distributed over the half turn of the crank than a larger stroke of the same torque, which seems intuitively to be a way to decrease TV, is this true?
In a straight-4 engine every power stroke at the piston, especially farthest away from the output shaft, will create a flex perpendicular to the shaft, or cause a rocking tendency, correct? With a boxer engine the force of one power stroke on the shaft will be countered by a force 180 degrees, opposite side of the shaft and therefore counter balance the forces on the shaft, correct? Does the fact the boxer engine has a shorter shaft than the equivalent cylindered straight-4 have a significant effect? If so then it follows that a straight-3 engine would be less prone to TV than a straight-4?
Also, I’ve noticed that the boxer aircraft engines, like the Lycomings, have a much larger bore to stroke, the large bore give it the torque needed at lower rpm, though that could have also been achieved with a larger stroke. With a shorter stroke the power pulse on the shaft will be more evenly distributed over the half turn of the crank than a larger stroke of the same torque, which seems intuitively to be a way to decrease TV, is this true?
Autodidact
Well-Known Member
In general, the aspects of crankshafts that you talked about will increase the stiffness of the crank; that won't decrease TV necessarily, it will just put the natural frequency at a different place in the rpm range. With the four cylinder engines: on a boxer it is pretty much the same as an inline; the 180° opposite cylinder going off is still turning the crank in the same direction, so the effect is almost the same, and there are often only three main bearings dispersed among the crank throws on a boxer compared to five on almost all modern inlines (the old model A has three) while the boxer crank is a little shorter maybe so the crank stiffness there is about the same or maybe a little less.
Torsional resonance, like that demonstrated in the video, is when the frequency of the forcing vibrations (cylinders going off) coincides with the "natural frequency" of the crank/flywheel/coupler/PSRU system as a whole. That is, when you twist a rod with a mass (or masses) on it, it will untwist and then twist up the opposite way, and then spring back to the position it was in when you first let it go (when you first twisted it up). If another cylinder goes off right when it twists back up to its original position then that force is added to the force that is already there, and every time a cylinder goes off it is added to the previous sum of forces and can theoretically reach infinity, which is why things break because nothing is infinitely strong.
Changing the natural frequency of the system by changing spring rates and/or masses is how TV is managed. If you can't drive the natural frequency below the operational range of the forcing vibrations so that it is never encountered during the normal operation of the engine, then you need to put it somewhere in the operational range where it is "out of the way" like at an rpm that you never dwell at, and you also need damping in that case because damping reduces the amplitude of the resonant vibrations allowing you to pass through resonance safely and/or to operate closer to resonance without damaging the engine/PSRU system.
Torsional resonance, like that demonstrated in the video, is when the frequency of the forcing vibrations (cylinders going off) coincides with the "natural frequency" of the crank/flywheel/coupler/PSRU system as a whole. That is, when you twist a rod with a mass (or masses) on it, it will untwist and then twist up the opposite way, and then spring back to the position it was in when you first let it go (when you first twisted it up). If another cylinder goes off right when it twists back up to its original position then that force is added to the force that is already there, and every time a cylinder goes off it is added to the previous sum of forces and can theoretically reach infinity, which is why things break because nothing is infinitely strong.
Changing the natural frequency of the system by changing spring rates and/or masses is how TV is managed. If you can't drive the natural frequency below the operational range of the forcing vibrations so that it is never encountered during the normal operation of the engine, then you need to put it somewhere in the operational range where it is "out of the way" like at an rpm that you never dwell at, and you also need damping in that case because damping reduces the amplitude of the resonant vibrations allowing you to pass through resonance safely and/or to operate closer to resonance without damaging the engine/PSRU system.
pictsidhe
Well-Known Member
Once, in an electronics class, we were doing an experiment on resonance. Being a hobbyist, I already knew all about it. While my classmates were happy to be getting 10V from a 0.5V signal generator, I had rummaged and found the right bits to give a sky high Q factor, my first and best effort was over 1kV...
That's a pretty extreme electrical example helped by the biggest air cored inductor I've ever come across, but mechanical systems frequently have a high Q if there is no deliberate damping.
pictsidhe
Well-Known Member
Forget the power pulses, they don't affect the resonant frequency, they just excite things.
Take a piece of coathangar wire, bend the ends 90 degrees so you can grip them.
Twist, that's a torsional spring. Go up to crankshaft size: IT STILL TWISTS, it's just a helluva lot stiffer.
Put a weight on each end. you have two masses connected by a spring. excite those at the right frequency, they resonate in a torsional version of the video a few posts up. In that video, there are two springs and one mass. The end result is still a large amplitude but torsional instead of linear when the particular resonant frequency of an undamped system is reached. The damping in the video is probably mostly air resistance. if a piece of card, increasing air resistance, had been attached to the weight, the amplitude would have been much lower.
The crank only appears to have one big mass (the prop) on one end. But the crank is not infinitely light, the con rods and other stuff attached to it all weigh something.
Find a broom, rotate along the axis of the shaft, it can be twisted quite easily. That's low rotational inertia, the torsional version of mass. Now hold it at 90 degrees to your arm and twist your wrist again, way harder to change direction! It's got much more mass at a large diameter and a higher rotational inertia, although it's still the same broom.
Crank stiffness depends on length. It's not hard to work out that a 1/2 length crank is twice as stiff as full length crank. Bigger diameter increases stiffness, that actually varies with the 4th power of diameter.
Straight shaft are fairly easy, but calculating the stiffness of something shaped like a crankshaft is the kind of headache I prefer to solve with actual experiment...
HTH
So, do manufacturers know where resonance will occur in their engines at a particular rpm? Is that data that is available?
Bottom line, if I want to use an inline 4 in direct drive with a prop, what is the concern I need to have about resonance? Can it be easily negated with a flywheel dampener, and if so how much would such a flywheel weigh for a 100HP engine.
I think I'm confusing TV with another phenomenon, maybe torque pulses, or power pulses, which has relevance to why aircraft engines in direct drive are almost always boxer engines.
the big companies probably do
Not to you and me that I've ever seen
Billski might know better
The engine or anything attached to it may fail very prematurely and catastrophically at any time
maybe, maybe not. you will need to build, ground test and test fly.
The general approaches to deal with TV is to avoid or damp TVs.
To avoid, you can try to design a torsionally soft/stiff system so the resonance frequencies are below/above the operating range of the system.
Or have no-fly rpm ranges to avoid resonance frequencies.
Or attempt to damp them with the one of the many available dampers for boats, industrial engines, cars, trucks.
This will likely involve lots of ground vibration studies and test flying. There are lots of discussions here and elsewhere.
http://www.homebuiltairplanes.com/forums/general-auto-conversion-discussion/14215-torsional-vibration-resonance-basic-theory-issues.html
http://www.vibrationdata.com/tutorials/torsional_vibration.pdf
rv6ejguy
Well-Known Member
OEM aero engine manufacturers have certainly done extensive instrumented TV studies and engineered the scary points out by the time the engine is in production. You don't skip this step or you might be in for a world of trouble down the road.
If you are converting an auto engine and using a PSRU, a light flywheel usually compounds TV issues as many have found out. Most successful installations are using flywheels of at least 15 lbs. plus some sort of damper like Centaflex or Lovejoy or a friction type (Raven) or a fluid filled "dual mass" type. A higher flywheel MOI should always reduce TV amplitude but there are obvious limits on weight in an aircraft installation.
The cylinder configuration has little to do with TV downstream. A four cylinder inline has the same excitation frequencies as a four cylinder opposed layout.
Some individuals like Dan Horton have done TV studies on their conversions, both mathematical and by measurement with strain gauges. Dan helped me do a math study on mine and it was very enlightening.
pictsidhe
Well-Known Member
what matters to TV frequency is RPM, shaft springiness and the various masses that the springy shaft connects. Bhp makes no difference, engine configuration changes the shaft dimensions but also affect the amplitude of the stimulation. If all the pistons and rods are accelerating at the same time, such as a single or boxer twin, the torsion is highest. Experiment is by far the best way to check for a TV issue. There's no one size fits all bandaid to fix a TV issue, it's one of those things that really does need experiment. TV used to be measured by the monitoring the angular acceleration of the shaft ends. That would be easiest for a homebuilder to improvise. I believe more modern method is to measure the actual twist of the shaft, while running. that's somewhat harder as it needs high resolution and to be done many times per rev.
I keep thinking there would be a market for a homebuilder friendly TV measurement system that didn't cost $Anarmanaleg like the modern stuff does. I've got an idea about that that I don't really want to broadcast yet...
Fixes for TV involve either shifting the problem frequency somewhere that it isn't a problem or reducing it's amplitude, adding a flywheel will drop the resonant frequency. whether or not that is a good thing depends on whether the TV issue is below or above your operating rpm range. Manuals cars usually have a problem when used at too low rpm, that horrible shaking when you give it throttle at too few rpm. Planes don't disconnect the prop at idle, so the TV frequency needs to be below idle. You then assume that the engine spends extremely little time going through the problem frequency on starting and stopping. To reduce a TV amplitude, it gets harder. Tuning a damper to a particular frequency can work (a mass hung on lossy spring), but it also tends to warm up. that can be a problem. those ones are generally using a rubber spring, they usually harden over time and go out of tune. There is another class of absorbers that react to a particular torsional harmonic and aren't lossy. They're a bit trickier to design.
N
NEMAN
[QUOTE = pictsidhe; 266762]., Что важно для телевизионной частоты является RPM, вал упругость и различные массы, пружинит вал соединяет [/ QUOTE]
Сoaxial propellers as influences frequencies?
Сoaxial propellers as influences frequencies?
The relationship that matters is if any of the engine's torsional pulse frequencies lines up with any of the engine/psru/prop resonant frequencies. If they coincide, the vibration will grow and is likely to break things. The size of the pulses matter, but not as much as the issue of amplifying the vibration.
Billski
Automotive and other base engines have all of their internal resonances out of range high. In general, the primary frequency of the crank shaft/cams/accessories is above usually above twice primary firing rate. This is because most automotive engines will have significant vibration content at 1/2 order, 1st order, 1/2 of firing order, firing order, and twice firing order. Firing order is by far the largest, and twice firing order is generally the next largest. Once you have designed the primary freq above twice firing order, the others are too small to interact.
Once you get into powertrains, where there is something beyond the flywheel, it gets more interesting. In direct drive aircraft engines (Lycos, Contis, Franklins, etc) they still make the crank stiff enough to still keep the resonance of the system (Crank and prop) above twice firing order. But in engines with PSRU's, that high level of stiffness is tough to get. So you go the same way the car/truck/ag/construction powertrain guys do. Put a spring into the system that brings the primary resonance well below firing order, usually below half firing order at idle rpm or even lower. The device that has the spring will usually have some friction device in it too. The design of such systems is where I lived for the last decade of my working life. Inertia on the flywheel can help, but what it is doing is reducing the resonant frequency too, and everyone in both businesses hates doing it, as it adds weight, and WEIGHT IS THE ENEMY.
Billski
When we talk torsional resonance, we actually ignore the block. We are talking about the gas pressure pushing down on the piston and accelerating the crank, then slowing down until the next firing pulse (firing order), the imbalanced torques from accelerating the pistons up and down in the cylinders (twice firing order), and others. We are not even considering the various rocking orders in torsional resonance - they are not torsional... The prop is a big inertia, and the crank/flywheel is a big inertia, and the gears and shafts and any isolation devices are a soft spring between the big inertia giving a low enough resonant frequency.
Now if you want to talk about engine mount reaction modes and frequencies, these come about from the same firings with gas pressure applied to the heads and other inertia accelerations effects, and the mounts are designed using similar theory. The mount stiffnesses are chosen to put the primary freq's below firing frequencies at idle, and they go through the resonance at start up and during shut down. This is relatively well known and relatively easily solved compared to PSRU design, and that is not the Torsional Vibration we usually discuss in PSRU design.
Billski
Nope...
You seem to be confusing engine block rocking couples with torsional resonance. The block accelerates about because the firing pressure is applied to one combustion chamber at a time and is almost never balanced - the firing pulses are usually spread out. same for the reactions from the crank, cams, etc from accelerations to the pistons, con rods, and valve gear. With PSRU's, we get more, as the firing pulses in the crank will be reacted to the case from the gear set.
A three cylinder is generally worse for torsionals - firing order is at lower frequencies and this requires isolation to an even lower frequency, which means lower spring rates and more travel to isolate firing pulses from downstream components, both with the internal torsionals and external block vibration. On top of that, three's have big rocking couples in both pitch and yaw axes to deal with in block isolation.
The stroke is largely a matter of what max rpm you intend to run the engine, as there are pratical maximum piston speeds (in ft/min or m/s) that engines run at with long lives. If you want an engine to make 6000 rpm, it will require much shorter stroke than one that only has to make 2700 rpm.
Billski
