# Torsional Dampening

### Help Support HomeBuiltAirplanes.com:

#### rv6ejguy

##### Well-Known Member
Bill, most of my work was with relatively small displacement engines and the designs dated back to the '70s in many cases. I was curious as to the spring rates used in clutches and tested a few of them just as I described. In no case was the spring rate at full deflection greater than the stock torque rating of these engines. Common practice on Japanese OEM clutches then was to have 3 double springs and 3 rubber dampers on each disc. These all worked fine but were typically combined with very heavy flywheels in those days. Obviously things to today are somewhat different with DM flywheels of which I'm not a great fan of. While they might solve an issue down low, they make hard shifts unpleasant in many cars I've driven including my 3.8 Genesis which rubber bands terribly in hard shifts in the lower gears. My general feeling is that modern cars are smoother throughout the range, probably in large part to NVH work, but few if any new cars I've driven have really nice clutch action as older cars had. Engineering is full of compromises so while fixing one issue, sometimes another one is less well addressed.

In racing, the sprung center discs and heavy flywheels typically were the first things to go and even in street use, I never saw any TV down below the idle rpm range.

I just don't like to see people saying that installing a clutch disc between engine and drive will somehow magically make TV a non-concern like so many using Ross drives seem to do. If TV was present, it would likely bottom these and destruct anyway. The same applies for rubber dampers as we've seen them go away rapidly on the Marcotte drives when operated in the resonance ranges for very long on certain engine/ prop combos. Clearly you need to make sure resonance is outside your normal operational ranges if you expect things to last.

#### wsimpso1

##### Super Moderator
Staff member
Log Member
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.
When you start with a road car then raise the compression ratio, improve the intake and exhaust breathing, and end up with peak torque 20-50% higher than it was and max rpm 20-50% higher than it was, lots of parts will no longer be right for the car. If no one makes a clutch disc for the race engine and road springs won't live at those torques and speeds, you gotta do something. Remember that when you pumped up the engine's output, you also needed a clutch cover with more force to hold the clutch, and frequently the transmissions and drive shafts and rear axles needed upgrades too.

Remember a couple other things about race engines - they are usually run at high rpm. Vibration travel is alpha/omega^2, where alpha is firing acceleration, and omega is firing frequency. Units are rad/s/s and rad/s. At low rpm, the travel is large, can be several degrees, but as rpm goes up, the travel gets almost smaller, so your systems can more easily tolerate the stiff disc. Last part is that manual clutches in stiff systems slip a little on every firing pulse, so the pulses are limited to clutch capacity, which can help with torque control, but is inefficient and tends to not live the kind of hours we would like in an airplane.

Billski

#### larr

##### Well-Known Member
Off the top, I want to say I really appreciate these technical threads - they make me think. They also give me a headache, but there you go.
I was wondering if there was a particular engine layout that reduces TV.
For instance, a single row 7 cylinder radial would have a 1-3-5-7 firing order on the first revolution followed by 2-4-6 on the next so there's going to be a cyclic imbalance in power production. Also there is only one node on the crankshaft.
If a two row radial is configured so that it runs 4/3 one revolution and 3/4 on the next is that ideal since the crankshaft is always equally loaded or is it a problem that it is going to be loaded mostly at the front on one revolution and then mostly at the back on the next?
And so on for in-line engines. Is a V-12 a better solution because there are six impulses per revolution or is the longer crankshaft a problem? Is a flat four a better solution because the crankshaft length is reduced over an inline four?

#### rv6ejguy

##### Well-Known Member
When you start with a road car then raise the compression ratio, improve the intake and exhaust breathing, and end up with peak torque 20-50% higher than it was and max rpm 20-50% higher than it was, lots of parts will no longer be right for the car. If no one makes a clutch disc for the race engine and road springs won't live at those torques and speeds, you gotta do something. Remember that when you pumped up the engine's output, you also needed a clutch cover with more force to hold the clutch, and frequently the transmissions and drive shafts and rear axles needed upgrades too.

Remember a couple other things about race engines - they are usually run at high rpm. Vibration travel is alpha/omega^2, where alpha is firing acceleration, and omega is firing frequency. Units are rad/s/s and rad/s. At low rpm, the travel is large, can be several degrees, but as rpm goes up, the travel gets almost smaller, so your systems can more easily tolerate the stiff disc. Last part is that manual clutches in stiff systems slip a little on every firing pulse, so the pulses are limited to clutch capacity, which can help with torque control, but is inefficient and tends to not live the kind of hours we would like in an airplane.

Billski
Very true. I've got 35 years modifying cars and building turbocharged performance and road racing engines professionally, some engines developing over 5 times the stock power output with almost nothing changed in the engine except pistons, valve springs and cams in these engines. At these levels though, the stock clutch materials and transmissions last only minutes and we had to go with feramic friction material and solid center discs plus much larger transmissions so the bearing balls didn't turn square and black in short order or the shafts deflect and grenade. On the other side of less radically modded engines, lots work just fine with all stock driveline stuff, depends a lot on factory strength margins. Some brands are really marginal in design, others amazing in what abuse they can take with good reliability.

With regards to TV in aircraft PSRUs, almost all problems we've seen on a wide variety of engines, redrives and prop combinations have been below the 1500 rpm range somewhere. Not to say it can't happen higher up as we've seen a few of these too but most which survive flying for a few hundred or more hours either have no significant TV issues or they occur down in an RPM range where the engine does not spend much time and there is enough damping or TV is low enough amplitude that is does not break anything. Three of us with Subaru fours and sixes who all had TV down around the 1250 rpm range all went to higher inertia flywheels a second time around and have seen (in one case measured) a huge reduction of hammering down here and much improved bushing life. My math model confirmed that this was the easiest/ cheapest route to take compared to designing a new damper system. The standard bushings supplied with the Marcotte M300 gearbox are way too soft to help much- only 25 ft./lbs./ degree- I suspect essentially TLAR (that looks about right) engineering in this case (no disrespect intended to Guy Marcotte).

Last edited:

#### Autodidact

##### Well-Known Member
I've always heard that the spring clutch centers won't solve the problem either, and I have always believed it is most likely true, but that's because their properties have been an unknown to me, not having done the testing with a vice and torque wrench like Ross did to determine what torque they bottom out at.

But Billski said a a few things in post #59 that make me wonder, now; 1) he seemed to suggest that automotive engineers usually try to drive resonance below operating range for the engine/transmission combination as a system, 2) that the spring dampers on manual transmissions operate w/o bottoming out, throughout the operating range (the ones he worked with/is familiar with), and 3) that the inertia of a propeller is generally greater that the inertia of an automotive transmission, so that a propeller/PSRU would also have a generally greater inertia than an automotive transmission. What that implies to me, is that if a spring disc drives resonance below operating range for a car transmission, then there is some likelihood that it should also do the same for a different transmission that happens to have a greater inertia, such as a propeller/PSRU. If what BIllski said about the spring disks he's familiar with is true, such as not bottoming out, and having a torque capacity @ 125% of engine peak torque, then I can't see why it wouldn't work as long as there wasn't shaft misalignment and that sort of installation issues.

I've looked at lots of spring disks and they appear to have many different styles of designs, some with very limited angular travel and some with somewhat greater travel, some with only four springs and some with more. Some have different size springs on the same disc, apparently to provide a rising rate? If so it may be possible that some clutch centers are only designed to work in the low range and bottom out at higher torque values after the clutch has been fully engaged, my point being that there are possibly many different styles/flavors of clutch centers even though they all look very similar.

Billski, if a person did do the calculations to determine the correct spring rate and torque capacity for a spring disc TV attenuator, would an automotive clutch center that met those specs do the job?

Bret

#### Himat

##### Well-Known Member

Billski
I do as others have said find the thread highly informative and educational. What “thread drift” there might have been, have as I see it has not moved the thread outside the scoop of torsional vibrations.

It might be said that the thread have not drifted, it only gyrates somewhat around.

#### Billrsv4

##### Well-Known Member
I've always heard that the spring clutch centers won't solve the problem either, and I have always believed it is most likely true, but that's because their properties have been an unknown to me, not having done the testing with a vice and torque wrench like Ross did to determine what torque they bottom out at.

But Billski said a a few things in post #59 that make me wonder, now; 1) he seemed to suggest that automotive engineers usually try to drive resonance below operating range for the engine/transmission combination as a system, 2) that the spring dampers on manual transmissions operate w/o bottoming out, throughout the operating range (the ones he worked with/is familiar with), and 3) that the inertia of a propeller is generally greater that the inertia of an automotive transmission, so that a propeller/PSRU would also have a generally greater inertia than an automotive transmission. What that implies to me, is that if a spring disc drives resonance below operating range for a car transmission, then there is some likelihood that it should also do the same for a different transmission that happens to have a greater inertia, such as a propeller/PSRU. If what BIllski said about the spring disks he's familiar with is true, such as not bottoming out, and having a torque capacity @ 125% of engine peak torque, then I can't see why it wouldn't work as long as there wasn't shaft misalignment and that sort of installation issues.

Bret

Bret,
The thing to remember about Billski's post is the idea is to drive the range below or above NORMAL RPM operating range. If you have a higher RPM additive TV mode it is likely to bottom those springs instantly. We need to be sure that we don't get into that. It can be caused if you have a high RPM misfire, or wear causes a compression change in one of the three chambers, or 1 coil fails. It only needs to be something that causes a consistent periodic change. When they were testing the modifications made to the BD-5, (The changes to make it a super soft "broomstick" system), They found that simple engine compression could cause TV when turning over the engine with the starter! They had a compression release to allow the starter to start spinning the prop, then they would close the compression release and start the engine. The TV almost caused the engine to come off the engine mounts. They solved it with a 1 way phased sprag clutch. Which was only ok because it was below engine RPM. If it was in the normal RPM range the clutch would have worn out within a few hours.

T.O. Bill

#### Autodidact

##### Well-Known Member
I'm pretty sure I understand Billski's post, I just wanted to make sure I was on the same page and to see what he said about the clutch center's potential as a TV attenuator; no reason I can see that it won't work as long as it is "selected" correctly and not indiscriminately picked out and bolted up to the flywheel. It would make things simpler to accomplish (cheaper and more readily available), but one piece of the puzzle that is still missing is how you determine the torque capacity and spring rate of all of the clutch centers that are available on the market without buying a bunch of them and testing them all. Ie, how do you get the detailed specs that are not normally provided to the aftermarket customer?

Consider points 1, 2, and 3 in my post again and think about them in the context of the fact that either less stiffness, more inertia, or a combination of less stiffness and more inertia, will always lower the resonant (natural) frequency of the system. If, as Billski seemed to say, a propeller has greater inertia than the rotating elements of a manual car transmission, then a clutch center should create an even lower natural frequency for a propeller/PSRU on the same engine - if that propeller/PSRU does indeed have greater inertia than the car transmission, with the danger being, as you pointed out with the BD-5 account, that it might get low enough to be a problem at engine cranking speed. But, if the specs of torque capacity and spring rate can be found without a lot of trouble, it may be possible to "select", in an engineering sense, the correct clutch center to use for a specific engine/propeller/PSRU application.

As far as the BD-5, if there was resonance during starting caused by engine compression, then the drive shaft was soft enough to put resonant frequency at or very near cranking speed and it should have been possible to change the stiffness of the shaft or even lighten or increase the flywheel's weight if it could be done without upsetting the engines idling characteristics or the aircraft's CG location. The extra weight of the sprag clutch could have been added to the periphery of the flywheel to increase the inertia and drive the resonance to below cranking speed without the extra complexity, IMO. But somewhere in between cranking speed and engine idle is probably best because as the battery runs down, cranking speed would also decrease, so you would always want both motors - the starter motor and the engine itself - to only move away from resonance. So increasing the stiffness of the drive shaft might have been best.

#### wsimpso1

##### Super Moderator
Staff member
Log Member
In racing, the sprung center discs and heavy flywheels typically were the first things to go and even in street use, I never saw any TV down below the idle rpm range.
Makes sense when tuning a production car into a racer or a performance street machine: Stock clutch discs usually have too little reserve capacity for a engine tuned for more horsepower, and will be destroyed if run at much higher torques. Your point about shift character probably has something to do with using solid discs too. Remember also that if you put a fat cam, light flywheel/pressure plate, and higher compression ratio in an engine, you probably don't care about idle noise.

I just don't like to see people saying that installing a clutch disc between engine and drive will somehow magically make TV a non-concern like so many using Ross drives seem to do. If TV was present, it would likely bottom these and destruct anyway. The same applies for rubber dampers as we've seen them go away rapidly on the Marcotte drives when operated in the resonance ranges for very long on certain engine/ prop combos. Clearly you need to make sure resonance is outside your normal operational ranges if you expect things to last.
I share the concern that some folks behave like "this stuff is easy, you just do such-and-such". There is no substitute for doing this stuff well. If they put a resonant mode inside the operating range, vibration will amplify and broken parts are likely. I am going to get picky about the term Torsional Vibration. Ross calls this TV. Hey, it happens, the engine makes a bunch of it, and the rest of your system responds to it. The issue is if it is small or large. If we are taking an engine vibration and amplifying it in our system, we will break metal parts, cook rubber parts, etc. If instead, we can adequately manage vibration by keeping our engine vibrations away from the resonant modes, the vibrations downstream will be small, we won't notice them, and flying will be better.

Done correctly, all of the major resonances are at least 1/2 octave away from all of the major forcing functions - resonance can not occur. There can be a lot to accomplishing this. If resonance is present in any significant amount, some resonant mode was within reach of some vibe from the engine or other parts. There are a bunch of ways this can happen. Anticipated Errors are:

Spring Rates (Soft Systems) not Low Enough - This is the slope of the torque vs travel curve. If rate is not low enough, 1st mode fn will not be low enough and problems will occur due to amplification during starts and low rpm operation;

Spring Capacity (Soft Systems) not High Enough - This is preload plus rate times travel. If total capacity is too low, springs bottom, rubber elements fail, the mechanism goes solid, and 1st mode jumps to higher frequency but within or near operating range and problems will occur due to operation near the new resonance mode;

Forcing Functions beyond 1st Mode - There are other vibrations coming from engines. Modes beyond firing order exist. 2x firing order is significant in pistons and apparently large in Wankels. Orders higher than firing can be problems in soft systems and order lower than firing order (1/2x firing order and 1/number of cylinders order and 2/number cylinders order can show up in stiff systems;

Higher Order Modes within Operating Range (Soft Systems)- There may be all sorts of other modes above 1st order. We may have other vibe modes that can have their fn's reduced by the spring element, bringing their mode near or into the operating range, and give problems;

Elastomeric Springs and Hysteresis Devices (Soft Systems) turn some vibration energy into heat and do not cool well. If you operate near resonant modes, these parts will warm up, cook and fail;

Wear, Spring Set, and Corrosion - Many manual transmission clutch discs have the springs exposed to environmental air and contaminants, steel sliding on steel, and exposed to engine heat. If not properly designed and/or protected, all sorts of problems can occur;

System not Stiff Enough (Stiff Systems) - Mode fn's not high enough to avoid amplification will cause problems;

There are some parts about that give me hope for adequate. AutoPSRU uses some stout coil springs cast in polyurethane. You get a rising rate spring with lots of torque capacity, some hysteresis (vibration energy turned to heat) for handling transients, and corrosion/rusting precluded. They are supposedly drag racer parts...

So, this business is not for the faint of heart. All of the makers of big piston engines wrestled with this issue. Gotta be done right...

Billski

#### weasel

I request that Ross Farnham, Billski, and Dan Horton form a joint venture to develop a reduction drive specifically for a LS series V-8 conversion that will reliably support a Hartzell CS prop across the normal range of operation and produce it to the market for 5000.00 along with complete TV study analysis. :ban: While we are at it Ross can go ahead and work out the SDS fuel control and ignition system and we will have a package. From there we will start the design teams on drawing airplanes specifically around this power system :gig: #### wsimpso1 ##### Super Moderator Staff member Log Member I've always heard that the spring clutch centers won't solve the problem either Well, they seem to work pretty well in a very complicated vibe environment of car/truck powertrains. You can not just randomly pick parts from O'Rielly's. It requires either some engineering or some testing with a bunch of suitable rotational vibration measurement equipment. Trial and error is a risky way to do this... But Billski said a a few things in post #59 that make me wonder, now; 1) he seemed to suggest that automotive engineers usually try to drive resonance below operating range for the engine/transmission combination as a system, All engineers wrestling with vibration will consider isolation by installing suitable "soft" elements in suitable places. Isolation is the best. When f/fn > 2, very little incident vibration is transmitted. Drive the particular resonant mode far enough below the operating range, and you hardly know the vibration existed. In car and truck, this is done by everybody. The big mode is Engine (and flywheel/cover or flexplate/TC) vibrating at firing order. The engine inertia is vibrating opposite the transmission inertia. We all (everyone in the world) put in a torsional spring device carefully designed to pull 1st mode low and cover the enire torque range of the engine. 2) that the spring dampers on manual transmissions operate w/o bottoming out, throughout the operating range (the ones he worked with/is familiar with) I did not actually say that. In anything resembling normal driving, you won't hit the stops. At the Big Three, this was common practice long before I arrived in 1990... There are purposeless and abusive driving techniques that can drive the damper to its stops. Starting off in a high gear with lots of throttle and slipping the clutch can bring the engine rpm down below idle and into resonance. Side stepping the clutch pedal with the engine revved up can do it. Power shifting with fast clutch work may do it. One other way is to drag the engine down below idle with a hill or with brakes and floor the gas pedal. Most modern cars will not open the throttle under these circumstances and may just cut spark and fuel. My Cruze shuts of the engine if you let it get to about 500 rpm during launch... , and 3) that the inertia of a propeller is generally greater that the inertia of an automotive transmission, so that a propeller/PSRU would also have a generally greater inertia than an automotive transmission. What that implies to me, is that if a spring disc drives resonance below operating range for a car transmission, then there is some likelihood that it should also do the same for a different transmission that happens to have a greater inertia, such as a propeller/PSRU. If what BIllski said about the spring disks he's familiar with is true, such as not bottoming out, and having a torque capacity @ 125% of engine peak torque, then I can't see why it wouldn't work as long as there wasn't shaft misalignment and that sort of installation issues. I think that manual or automatic trans clutch dampers could work. Resonant frequency is proportional to the square root of (k/m). With bigger m, and thus smaller k/m, the resonance will be lower . Trouble is there are lots of tails here. I have another response on this thread talking about all of the ways steel springs could get into broken parts. I've looked at lots of spring disks and they appear to have many different styles of designs, some with very limited angular travel and some with somewhat greater travel, some with only four springs and some with more. Some have different size springs on the same disc, apparently to provide a rising rate? If so it may be possible that some clutch centers are only designed to work in the low range and bottom out at higher torque values after the clutch has been fully engaged, my point being that there are possibly many different styles/flavors of clutch centers even though they all look very similar. Some have as many as three and four stages. Gives higher rates when input torques are higher. Sort of rising rate. All are designed to cover the whole torque range of the engine with margins of 25% or more. Using existing dampers requires engineering work or big surprise await. Billski, if a person did do the calculations to determine the correct spring rate and torque capacity for a spring disc TV attenuator, would an automotive clutch center that met those specs do the job? I said above that they could do the job, but the analyze/design/build/test cycle would be necessary and could be daunting. In the auto industry, we rarely reused a damper assy in another engine/trans combination. Something would be adjusted. Different springs, different friction package, smaller packaging space, more travel, requirements for transmitted vibration marched on to lower numbers in support of smoother/quieter vehicles. I do not doubt that you would end up needing to change specifics to make it work for you. I suspect that the best way to do a soft system PSRU is elastomeric springs. Low initial spring rates, rising spring rates as torque comes up, some damping for start/stop. They can be designed, tested, and developed to give resonance out of range, good starting/shutdown behavior, and long life. You can buy urethanes in a variety of durometer, play with ID, OD, thickness and number of elements until you get the characteristics you need. RWS and Marcotte and Autoflight all use them. I said it before. As much as I know about gearboxes and vibration management, I am not leaping forward with my cash to design, build, and market a PSRU. I might be talked into doing some engineering on one... Billski #### wsimpso1 ##### Super Moderator Staff member Log Member Bret, The thing to remember about Billski's post is the idea is to drive the range below or above NORMAL RPM operating range. If you have a higher RPM additive TV mode it is likely to bottom those springs instantly. We need to be sure that we don't get into that. It can be caused if you have a high RPM misfire, or wear causes a compression change in one of the three chambers, or 1 coil fails. It only needs to be something that causes a consistent periodic change. When they were testing the modifications made to the BD-5, (The changes to make it a super soft "broomstick" system), They found that simple engine compression could cause TV when turning over the engine with the starter! They had a compression release to allow the starter to start spinning the prop, then they would close the compression release and start the engine. The TV almost caused the engine to come off the engine mounts. They solved it with a 1 way phased sprag clutch. Which was only ok because it was below engine RPM. If it was in the normal RPM range the clutch would have worn out within a few hours. T.O. Bill The bombproof simple way to do it is to go stiff. Even that has tails. For a rotary, you probably have to drive fn safely above max 2x firing frequency because 2x firing has got to be big with those big rotors. Also, all of the parts have to add up to some pretty high overall stiffness, gears have to be beefy, and lash has to be kept really small. Hmm, this sounds familiar. Grin. Billski #### wsimpso1 ##### Super Moderator Staff member Log Member I'm pretty sure I understand Billski's post, I just wanted to make sure I was on the same page and to see what he said about the clutch center's potential as a TV attenuator; Let's try "Isolator", because that is what we are talking about. Also, Isolators do not need to lose a lot of energy, most is returned to the system later in each vibration cycle. Consider points 1, 2, and 3 in my post again and think about them in the context of the fact that either less stiffness, more inertia, or a combination of less stiffness and more inertia, will always lower the resonant (natural) frequency of the system. If, as Billski seemed to say, a propeller has greater inertia than the rotating elements of a manual car transmission, then a clutch center should create an even lower natural frequency for a propeller/PSRU on the same engine - if that propeller/PSRU does indeed have greater inertia than the car transmission, with the danger being, as you pointed out with the BD-5 account, that it might get low enough to be a problem at engine cranking speed. But, if the specs of torque capacity and spring rate can be found without a lot of trouble, it may be possible to "select", in an engineering sense, the correct clutch center to use for a specific engine/propeller/PSRU application. But somewhere in between cranking speed and engine idle is probably best because as the battery runs down, cranking speed would also decrease, so you would always want both motors - the starter motor and the engine itself - to only move away from resonance. So increasing the stiffness of the drive shaft might have been best. Getting too low a frequency is a potential problem. To fully spec this thing, you would want to put fn (1st stage) safely above cranking/initial firing rpm and safely below idle speed. The idea being to avoid resonance while the starter is turning the engine, and just barely getting firing to happen, and to avoid resonance at idle. Now that means you have to go through resonance while the engine accelerates from initial firing rpm to idle speed. Most engines will do that quickly, and only expose the system to a few firing pulses while within a half octave of fn, so amplitudes do not get a chance to build. Having a 2nd or 3rd stage spring or rising rate elastomeric elements will help in that if a little amplification occurs, the system spends some of each spring cycle in a different fn, and that tends to stop amplification from getting any bigger, that is, until torque and firing f approaches fn for the second stage... As to increasing stiffness, well, once you drive down fn1 with a soft spring, you have to see if you dragged any of the other fn's into the operating range. Might have to worry your way through either making stiffness low enough to draw down any other fairly low fn's... Or you could consider a stiff system. The fun never ends. Billski #### wsimpso1 ##### Super Moderator Staff member Log Member I request that Ross Farnham, Billski, and Dan Horton form a joint venture to develop a reduction drive specifically for a LS series V-8 conversion that will reliably support a Hartzell CS prop across the normal range of operation and produce it to the market for5000.00 along with complete TV study analysis. :ban: While we are at it Ross can go ahead and work out the SDS fuel control and ignition system and we will have a package. From there we will start the design teams on drawing airplanes specifically around this power system :gig:
I will not be a major investor. I might be talked into some serious FEA and Eigen analysis. I will watch what is out there and buy what suits based upon success in test and flying...

Billski

#### Billrsv4

##### Well-Known Member
Let's try "Isolator", because that is what we are talking about. Also, Isolators do not need to lose a lot of energy, most is returned to the system later in each vibration cycle.

Getting too low a frequency is a potential problem. To fully spec this thing, you would want to put fn (1st stage) safely above cranking/initial firing rpm and safely below idle speed. The idea being to avoid resonance while the starter is turning the engine, and just barely getting firing to happen, and to avoid resonance at idle. Now that means you have to go through resonance while the engine accelerates from initial firing rpm to idle speed. Most engines will do that quickly, and only expose the system to a few firing pulses while within a half octave of fn, so amplitudes do not get a chance to build. Having a 2nd or 3rd stage spring or rising rate elastomeric elements will help in that if a little amplification occurs, the system spends some of each spring cycle in a different fn, and that tends to stop amplification from getting any bigger, that is, until torque and firing f approaches fn for the second stage...

As to increasing stiffness, well, once you drive down fn1 with a soft spring, you have to see if you dragged any of the other fn's into the operating range. Might have to worry your way through either making stiffness low enough to draw down any other fairly low fn's... Or you could consider a stiff system. The fun never ends.

Billski
Billski,
Just so everyone knows, the 3 rotor, (20B), does have some minor second order issues. PS's testing revealed it. The damper is very small, but needed. Stiff system.
T.O. Bill

#### larr

##### Well-Known Member
I sort of remember a post by Orion about TV where he stated that the problem was that the prop presented an elastic load. So that makes each size and kind of prop a different case when it comes to analysing TV.
Also, to sort of condense the 'No Short Days' story, Pratt and Whitney had to consider the engine, reduction drive and prop as a system that had to be tuned.
That would seem to make it unlikely that there could ever be an off-the-shelf PSRU.

#### wsimpso1

##### Super Moderator
Staff member
Log Member
I sort of remember a post by Orion about TV where he stated that the problem was that the prop presented an elastic load. So that makes each size and kind of prop a different case when it comes to analysing TV.
Also, to sort of condense the 'No Short Days' story, Pratt and Whitney had to consider the engine, reduction drive and prop as a system that had to be tuned.
That would seem to make it unlikely that there could ever be an off-the-shelf PSRU.
"No Short Days" is the classic article on how torsional vibration was figured out and managed.

Once they had an engine figured out using test clubs, they mostly just validated that the desired engine-prop combinations met Army-Navy specs, which was bit over the top, but designed to make sure that props would not fly apart and have some tolerance to field damage. Bullet and flak damage was another deal.

They had to invent methods to measure the vibrations in the various orders. These days, we can point laser tachs or mag pickups at each end of each shaft we are interested in and then record with a ROTEC or similar equipment, do the Fast Fourier to get vibe content by orders, frequencies, look at it by vibe pairs, etc.

If the inertia of tested engines and props and the inertia/spring rate of the PSRU are known, it is possible to analyze the system, know where the troublesome resonant modes are, and calculate the operating envelope of prop inertia and engine speed based on the engine inertia, number of chambers/cylinders, and torque curve. It may not be one size fits all, but each PSRU may have a fairly broad range of props and operating speeds that are acceptable.

One thing I find necessary to point out. In any system, first resonant frequency is lower than the lowest component resonant frequency. Think about it. If you had a prop blade that had an fn of 400 Hz, and everything else was much higher, well, 400 Hz would be the first resonant order. To build a stiff system, all of your parts have to be quite stiff, have very high individual resonant frequencies, and be strong. If we install a soft prop, it just became a soft system...

Billski

##### Well-Known Member
Cars (and trucks) have another card up their sleeve - soft engine mounts. This means that the engine block can effectively "rotate" about the crankshaft to some extent, which helps in the "soft spring" approach. And some modern mounts are pretty trick, fluid damped units, not the rubber blocks of old.
ps I have seen failed engine mounts in rally cars which I would swear have died due to overheating, not overload.

#### Jay Kempf

##### Curmudgeon in Training (CIT)
Cars (and trucks) have another card up their sleeve - soft engine mounts. This means that the engine block can effectively "rotate" about the crankshaft to some extent, which helps in the "soft spring" approach. And some modern mounts are pretty trick, fluid damped units, not the rubber blocks of old.
ps I have seen failed engine mounts in rally cars which I would swear have died due to overheating, not overload.
And those fluid mounts are notoriously unreliable. Great concept theoretically but much harder to assess life. My old 928 has both massive solid rubber mounts and actual engine shocks absorbers (hydraulic) between the frame cross member and the engine block. Elegant but expensive solution from the mid seventies. Unfortunately in aviation you can't react the engine against the frame, tires and ground except when taxiing. The 928 also has an engine speed drive shaft in a torque tube with the transaxle at the back and the engine/clutch at the front which has always interested me in terms of a setup for a soft drive shaft system. Damper and soft driveshaft protecting the engine from the mass moment of inertia of a heavy large prop. In the 928 the sacrifice you get is what is called driveline snatch (not my word) which is the soft driveshaft winding up and releasing during shifting. Just something to get used to in terms of timing shifts in a car. In an aircraft you wouldn't know. You would get some overrunning noises as the shaft absorbed the inertia of the prop. Wonder what the setup on the P39 was like. I am guessing just universal jointed shaft in a protective tube and some massive bearings supporting a soft shaft. There is also some ultralight with an electric motor with about a 3 foot carbon unsupported shaft sticking out of the middle with the prop on the end. Runs smooth as silk relying on prop balance alone. Out of the box but an interesting data point.

#### wsimpso1

##### Super Moderator
Staff member
Log Member
I sort of remember a post by Orion about TV where he stated that the problem was that the prop presented an elastic load. So that makes each size and kind of prop a different case when it comes to analysing TV.
Also, to sort of condense the 'No Short Days' story, Pratt and Whitney had to consider the engine, reduction drive and prop as a system that had to be tuned.
That would seem to make it unlikely that there could ever be an off-the-shelf PSRU.
I do have some comment on prop influence on the system. If you pick gear ratios badly or take a system with one gear ratio and number of blades that works, and start changing gear ratios and/or the number of blades, yeah, you might almost be starting over. The guys making big piston engines were pretty smart about that. Let's talk about why...

If your propeller is absolutely face-on to the incoming air, its vibration fed into the prop hub is pretty darned close to static. But in real airplanes flown by real people with real variation in airspeed, AOA, etc, each prop blade sees a one per rev load cycle. Yeah, the prop hub sees bending and torsional moments at the number of blades on the prop per rev of the prop.

Now these inputs are just as capable of driving resonant amplification as the engine firing pulses, and you have to understand where they line up with resonance modes, how strong they are, and then figure out how to manage them...

Running a couple examples:

Two blade prop gives 2/rev at the hub, plus the blades are never identical, so you will also have a 1/rev as well. In direct drive, that is a 2/rev and 1/rev inputs into the system, with the 2/rev lining up with firing order on a piston four cylinder or two chamber Wankel, and the 1/rev lining up with the 1/2x firing order, which is sometimes significant. If you line up with the engine pulses, you will reinforce the engine inputs. If you can arrange the prop so it is about out of phase with the engine pulses, you will have a 4 and 2 and 1/rev inputs to the system... On traditional direct drive engines, which are stiff systems, the system is already stiff enough that even the 4 per rev is not a big deal. A four cylinder engine already has 4 /rev (2x firing from the pistons being accelerated up and down the bores);

Two blade prop with 2/rev at the hub, but with a 2:1 speed reduction, means that you will have a inputs at 1/engine rev and 0.5/eng rev. Add in the others and you will have 1 and 2 and 1/2 per engine rev inputs to the system. These can all line up with the engine vibe inputs and all have to be dealt with if they get too close to resonant modes;

OK, maybe whole number ratios are not a good idea. Let's look at more serious gearsets. Ideally each gear in a non-unity gear set has a prime number for its number of teeth. 7 tooth input and 19 tooth output give 2.714:1 and is in the right range for the engines we are talking about. Why prime numbers? A whole' nother topic... Anyway two blade prop gives 0.737 and 1.47/engine rev inputs and it does not line up with engine vibration inputs. Good.

For soft systems and a two chamber Wankel, we have to drive 1st resonant mode safely below 2/engine rev. Even ground ops, we will get these prop inputs, and with a 2:1, they will all coincide. Even if you get really soft and aim for 1/rev at idle, you can still feed resonance there... Even if the inputs and amplified results are small (idle speed on a the prop is small torques and small fluctuations) it will be annoying and maybe it can make trouble. What's a mother to do? Well, have a little courage, and stack the deck in your favor.

Start with a gear ratio that will not align prop inputs up with the engine inputs. Our 7:19 gear example puts the prop inputs out of alignment with the engine inputs, but also drives the lowest order that you have to get below even lower... This is where soft systems can have their difficulties.

If you only have steel springs, they can amplify even small inputs. In cars/trucks, we add in a friction device, a little brake made of Belleville springs and nylon, that can take off a little energy on each vibe cycle. They can handle modest torques at low engine speeds. As engine speed goes up, the vibe amplitudes go down (1/rpm^2), so they might be OK.

But if you have elastomeric springs, with significantly rising rates and a little more damping, they have much lower amplification because the spring rate moves around during the movement through each cycle of the vibration. Trick here is to make the elastomeric springs have low enough initial rate to cover idle resonance, then have enough rising rate that they are sturdy against the lower level but present other orders, and then they need to be durable. You can use high temp fluorocarbon rubbers and you can design to let the heat come out of them more easily, and you can play with different Durometer, all while driving for right spring rates, rising rates, and adequately low temperatures. It will be an interesting dance and you had better start with math models so you can get close to real solutions before you start the fuss with hardware. Straight trial and error might be a long game otherwise...

I am almost sure that most of the existing soft PSRU's out there only drive 1st resonance a little below idle, and count on their elastomeric elements to both prevent much amplification (through rising rate) and convert other order vibe energy to heat. Most of the time, the AutoFlight, RWS, and Marcotte systems do run beautifully. And as Ross has demonstrated, a little flywheel inertia helps a bunch with further driving these systems in the safe direction. Ross has seen these rubber elements come apart, presumably from cooking with big vibe, so it still is out there waiting for us if we get too close to resonance or too stingy on engine side inertia.

The other solution is the stiff system. The prop inputs (unless you have a big number of blades and/or smaller ratios than we talk about around here) are lower order or about the same order as inputs than the engine. If you can drive the 1st mode safely above max firing speed with a stiff system, the prop driven modes are also safely out of range. COOL!

Now this should drive all sorts of thinking and will make some folks complain of headaches. So be it.

Billski

Last edited: