I am building a tractor, side by side, engine behind pilot gyroplane.

[dino]

Stemme S10

 

[Sabine]

Besides the already mentioned thread on torsional behaviour, you might find this one useful;

Torsional Vibration and Resonance - Basic Theory and Issues

The whole topic of PSRU's and drive shafts and torsional vibration has been talked around and I keep seeing the same difficulty in the thinking of the involved folks. Since we on HBA tend to be an intelligent bunch, I got to thinking about how much I had to work to get my arms around the whole...

www.homebuiltairplanes.com

Next, I can tell you that I helped some folks with a Jabiru four cylinder engine, a 44" long shaft, and prop, all in direct drive, tame their vibration. An elastomeric coupling at the engine helped with the shaft alignment issue and allowed the engine to deflect and vibrate on its mounts without overloading the shaft. This also allowed the system to operate without any lash in the system. The shaft was the primary torsional compliance having an adequately low spring rate that, when combined with a substantial increase in engine flywheel inertia, drove resonance down near engine idle rpm, and tested adequately. I modeled their design and adjusted inertia using SolidWorks and was also able to find all other nearby modes were significantly above operating input vibrations. The system tested well, and we verified no hazardous vibrations using Rotec Munich measurement system and laser tachometers pointed at targets applied to engine and both ends of the prop shaft.

Your system with a high revving outboard head will require a prop reduction system somewhere with considerable attention to torsional vibration management, particularly toward driving :

• If you run the shaft at engine speed and make it a lashless system, you may need a substantial flywheel at the engine end, allow the shaft to be small diameter (engine torque) and relatively torsionally flexible, then the gearbox at the prop hub can operate in a vibrationally clean area. This is the Allison/Airacobra system, and it worked. This system must have the shaft designed for a specific spring rate to put resonance above cranking speed and below idle speed;
• The alternative is to run the gearbox at the engine end. UGH. First you will need a significant flywheel on the engine, then a soft element such as a clutch disc center spring set or a rubber giubo or other compliant element to prevent the gearbox seeing all of the engine vibe. Then you need a U-joint of some sort to the shaft, then the shaft has to be sturdy enough to carry prop level torque and speed.

Both systems require a bunch of design and analysis work to be even close to being right when you build the first copy. The first type will be lighter and simpler, and thus easier to to make work. The second type will be more difficult to make work as you must isolate torsional vibration of the engine in a small package, then the speed reducer, then a beefier shaft, then the prop. Oh, besides Hertog's book or other ME vibration text, you should look up critical speed calcs for shafts. Whirl mode must be prevented, with several ways of preventing it, including increased shaft diameter (and concurrent increased torsional stiffness), breaking the shaft into more pieces with U-joints and supports (like are done with many trucks), and a system of support bearings as is seen in helo tail rotor shafts.

You did not pick an easy one...

Billski

Goodness, thank-you Billski, sadly I am having to go the option 2 route with this plane as I will have the reduction drive immediately after the engine- wow you do make a man's brain engage and thanks for that as I guess this is what the forum is for. My brain says redrive, clutch (centrifugal/cone/motorcycle????) damper driveshaft.

 

[Sabine]

Stemme S10

thanks so much - problem is not knowing what their clutch and dampers are and if its a off shelf item which would be preferable

 

[Sabine]

What I said above even looked confusing to me.

Just pointing out that you need good rudder authority in a tractor gyro and I would always use a H tail to get the rudders clear of the fuselage for cleaner airflow.

I also said I "could" put triple rudders if I wanted to.

Jan Kritzinger (aeronautical engineer) calculated that we should have sufficient rudder authority with the single rudder- hopefully this will prove correct- try whenever possible to keep to William Bushnell's theory of "simplicate and add lightness . It keeps in sync with my brain!!!

 

[Blackhawk]

This shaft drive setup aslo works well; by keeping the engine and drive shaft seperated except for the belt.

 

[bmcj]

I was going to suggest the Belgium-built D-Motor, but GeraldC beat me to it. They are liquid cooled, so cooling should be easy. They are direct drive and have a six cylinder version that rates 121hp at only 85kg weight (including coolant).

good luck on choosing the long driveshaft... that science is mostly Voodoo to my feeble mind. Billski (wsimpson) is really our resident expert on redrives, vibration and resonance, so you can be confident trusting whatever he says in regard to that.

This next question is for Billski... could the touchless magnetic reduction units be used to eliminate engine impulse effects from the driveshaft?

Lastly, a lot of people are suggesting the P39/P63 as an example of how to design a driveshaft. Didn’t they have issues on those planes too? It seems that a more modern design could be done better than a 70+ year old design... unless you also need to fire a cannon through the driveshaft.

 

[henryk]

This shaft drive setup aslo works well

=bigger momentum after reductor=heavy shaft !

I am building a tractor, side by side, engine behind pilot gyroplane.

Stemme S10

www.homebuiltairplanes.com

=S10, better !

 

[wsimpso1]

Goodness, thank-you Billski, sadly I am having to go the option 2 route with this plane as I will have the reduction drive immediately after the engine ... My brain says redrive, clutch (centrifugal/cone/motorcycle????) damper driveshaft.

Your scheme - the gear box can not be allowed to see anything more than a minor piece of the major vibe orders from the engine or it will need to be an order of magnitude beefier to stand the transmitted vibe. You will need a very substantial isolator between engine and gearbox. Once you have taken off 95% of the vibe, the rest of the system should be straightforward. The big thing to do is drive the natural frequency of the engine/flywheel inertia then spring, then gearbox inertia more than half an octave below min operation firing frequency but more than a half octave above firing frequency at cranking speed. You will have your hands full with that.

Billski

 

[Sabine]

Your scheme - the gear box can not be allowed to see anything more than a minor piece of the major vibe orders from the engine or it will need to be an order of magnitude beefier to stand the transmitted vibe. You will need a very substantial isolator between engine and gearbox. Once you have taken off 95% of the vibe, the rest of the system should be straightforward. The big thing to do is drive the natural frequency of the engine/flywheel inertia then spring, then gearbox inertia more than half an octave below min operation firing frequency but more than a half octave above firing frequency at cranking speed. You will have your hands full with that.

Billski

Thanks Billski, right now what I am thinking is a cogged tooth belt from drive pulley to driven in a 2:1 ratio and possibly putting a motorcycle clutch with built in Cush dampers in the driven pulley, so I can clutch the pulley to start and then slowly (soft start) the prop and prop-shaft. Your thoughts would be appreciated and it is obvious your knowledge is far,far greater than my simple brain

 

[dino]

You might want to use trapezoidal belts instead and clutch them by tightening similar to Robinson helos

 

[rotax618]

A 2:1 ratio is a bad idea, the power pulses will be transmitted by the same belt teeth on every revolution, successful PSRUs always have an uneven ratio.

 

[cheapracer]

A 2:1 ratio is a bad idea, the power pulses will be transmitted by the same belt teeth on every revolution, successful PSRUs always have an uneven ratio.

You of course meant to say every 'X' amount of revolutions, depending on the length of the belt, and only if the belt has an even number of teeth.

There's likely a billion engines out there running around with 2:1 camshaft belt drive ratio, most, but not all, have odd number of teeth.

 

[rotax618]

There is no comparison of the TV induced by a camshaft compared to a propellor. As I said all of the successful PSRUs, either toothed belt or gear use an uneven ratio, Rotax is the best example.

 

[wsimpso1]

A variety of belt drives have been used.

Most builders of geared engines, geared turbines, even steam turbines, as well as automotive tranny folks and turbomachinery engineers have found that whole number reductions are bad juju. Best in geared or cogged belt drives is where both sets of teeth are prime numbers. If you go with a V belt, pick pitch diameters such that ratio does the same thing. This prevents having a beat that is both annoying and destructive.

The V belt takes care of your clutching and is widely used on industrial equipment. You simply pull off the idler to disengage. Works great. Design the scheme so belt replacement is easy. LOL. In the over 100 hp arena, V-belt will likely need multiple belts of matched sizes.

The cog belt design is pretty straightforward and is used. If you are set on engaging the prop after engine start, it will require a separate clutch. If a motorcycle clutch of suitable torque capacity is available terrific. Aren't most motorcycle clutches wet clutches? That means an oil filled housing with seals. Lots of stuff to work out.

Where is the "cush" element? If it is downstream of the clutch, the clutch is still seeing engine vibe orders, and the cush is designed to protect the downstream parts from engine vibe order while the clutch is designed to take engine vibe. If the clutch is up stream, I bet that it is designed to vibrationally isolate the clutch and gearbox from engine vibe orders. Here is the rub. The stiffness of the cush section is usually designed as part of a system - it is a spring between two inertia (rotary masses) - and is designed to make the resulting natural frequency of the system both safely below idle vibe and safely above cranking vibe. When you stick this device in between your engine and your reduction drive, you doubtless will have higher inertia on both ends resulting in substantially lower natural frequencies. I worry that you may fall in resonance at cranking speed or just above, which can give difficulty getting the engine to accelerate to idle speed. Yeah, it can hang right there at resonance with all of the engine energy going into increasing the vibration amplitude. Besides making it tough to get to idle speed and above, this tends to break stuff...

The long shaft to the prop will be a fairly soft element and may be designed to put system resonance safely below operation speeds. If it does, terrific, but if it does not, you will need to add a soft element at one or both couplings. With such a system, you can usually design to avoid need for clutching a propellor. Resonance is set at least an octave above cranking firing vibe and at least an octave below idle firing vibe. With the reduction unit between engine and shaft, you will need a somewhat overbuilt cog or V belt system - which is common - as engine firing pulses and other engine orders are being seen in the reduction drive.

The lightest way to build the system is to put the reduction unit at the prop end. The reduction unit is then isolated from firing order vibe and so can be smaller, and the shaft is carrying less torque and so can be smaller and lighter too.

It should be evident that such a system will need to be engineered. Simply linking together some existing parts is unlikely to produce a successful system. An engineering text on theory of vibration must be in your future. While the books usually work in terms of masses and springs moving linearly, it translates directly to rotation inertia and torsional springs.

Please tell us why you feel that you must put the reduction drive at the engine, where things are both more complicated and heavier.

Billski

 

[Sabine]

A variety of belt drives have been used.

Most builders of geared engines, geared turbines, even steam turbines, as well as automotive tranny folks and turbomachinery engineers have found that whole number reductions are bad juju. Best in geared or cogged belt drives is where both sets of teeth are prime numbers. If you go with a V belt, pick pitch diameters such that ratio does the same thing. This prevents having a beat that is both annoying and destructive.

The V belt takes care of your clutching and is widely used on industrial equipment. You simply pull off the idler to disengage. Works great. Design the scheme so belt replacement is easy. LOL. In the over 100 hp arena, V-belt will likely need multiple belts of matched sizes.

The cog belt design is pretty straightforward and is used. If you are set on engaging the prop after engine start, it will require a separate clutch. If a motorcycle clutch of suitable torque capacity is available terrific. Aren't most motorcycle clutches wet clutches? That means an oil filled housing with seals. Lots of stuff to work out.

Where is the "cush" element? If it is downstream of the clutch, the clutch is still seeing engine vibe orders, and the cush is designed to protect the downstream parts from engine vibe order while the clutch is designed to take engine vibe. If the clutch is up stream, I bet that it is designed to vibrationally isolate the clutch and gearbox from engine vibe orders. Here is the rub. The stiffness of the cush section is usually designed as part of a system - it is a spring between two inertia (rotary masses) - and is designed to make the resulting natural frequency of the system both safely below idle vibe and safely above cranking vibe. When you stick this device in between your engine and your reduction drive, you doubtless will have higher inertia on both ends resulting in substantially lower natural frequencies. I worry that you may fall in resonance at cranking speed or just above, which can give difficulty getting the engine to accelerate to idle speed. Yeah, it can hang right there at resonance with all of the engine energy going into increasing the vibration amplitude. Besides making it tough to get to idle speed and above, this tends to break stuff...

The long shaft to the prop will be a fairly soft element and may be designed to put system resonance safely below operation speeds. If it does, terrific, but if it does not, you will need to add a soft element at one or both couplings. With such a system, you can usually design to avoid need for clutching a propellor. Resonance is set at least an octave above cranking firing vibe and at least an octave below idle firing vibe. With the reduction unit between engine and shaft, you will need a somewhat overbuilt cog or V belt system - which is common - as engine firing pulses and other engine orders are being seen in the reduction drive.

The lightest way to build the system is to put the reduction unit at the prop end. The reduction unit is then isolated from firing order vibe and so can be smaller, and the shaft is carrying less torque and so can be smaller and lighter too.

It should be evident that such a system will need to be engineered. Simply linking together some existing parts is unlikely to produce a successful system. An engineering text on theory of vibration must be in your future. While the books usually work in terms of masses and springs moving linearly, it translates directly to rotation inertia and torsional springs.

Please tell us why you feel that you must put the reduction drive at the engine, where things are both more complicated and heavier.

Billski

Hi Billski, goodness I wish I was close to you and could feed off your good sense. The difficulty i would have in placing the reduction drive in the nose is quite simply space, and without tearing the plane apart completely would be virtually impossible to do. What I have is a two bearing (for want of a better word "plumber block") housing that bolts onto the face of the nose. This supports the prop on a stubshaft approximately 300mm in length the stubshaft then bolts onto the pto which has conventional knuckles on each end- the engine side knuckles have a splined shaft as per any conventional rear driven motor vehicle. the pto is filament would carbon and is 80mm in dia and the whole sits in a 100mm dia kevlar tube for safety. The only way right now that I could see to put the redrive in the nose would be if I could find a small enough planetary gear box so as to have an in line reduction. There are dry belt driven clutches- Harley have one that drives some of their large cruisers. I totally get your comments regarding the math and physics surrounding this engineering- sadly not a strength of mine and your comments remind me that this is where I need to go. I must sound like the Mad Hatter, as I try to find a light enough, powerful enough engine and reduction drive (sub 100kgstotal weight) to drive the beast. As always thanks for your time and wisdom.

 

[wsimpso1]

I can tell you from a number of people's hard experience that guessing and trial and error on this will require a separate test rig where you may tear the thing up and have to rebuild it. I personally took over vibration management on a project where they were attempting to make a direct drive shaft system work by just trying stuff without doing any analytical work, and their test rig was all scarred up from broken parts flying about and thrashing around. The airport where this project was working had exiled their test runs to a far corner of the field away from EVERYTHING else on the field. Some of their schemes lasted less than 5 minutes. I am here to tell you that I modeled their system and tamed it with one modification. They had full time folks fighting with this problem for 18 months prior to my involvement... No, this retired engineer is not for hire, but I will check your math if you want. This is not a place for trial and error.

The PTO systems you see in agricultural equipment have a big flywheel on the engine, a torsional spring set in the clutch hub, then the gearbox, then the PTO output, shaft, and then inertia of the load. This is very similar to cars and trucks, with first the engine, the spring set, the gearbox, then the output shaft, then the load. The inertia of the engine is isolated vibrationally from the inertia of the rest of the system by a purpose built device that has a spring set designed to make the resonance frequency of the system be lower than the operational firing frequencies at their ground idle.

You have already chosen the engine, prop, and shaft. That defines the minimum engine inertia, prop inertia, and shaft spring rate, and number of firing pulses per engine turn. That leaves to be defined your flywheel inertia, cog belt system inertia, spring rate of the coupling and min idle speed. This is a large design space to explore, and doing so analytically is the way to do it. Once you have the basics for your cog belt drive established, you can figure out its inertia and fix that. You will have to find or build a coupling with somewhere around the right amount of torsional spring rate and more than enough torque capacity to handle the engine output. I would start with an engine idle speed no lower than about 1600 rpm, pick the lightest flywheel that will allow that idle, and see what the available spring rates (that have torque capacities above your engine max torque) do to your resonance frequency. If it is too high, you can add inertia to the flywheel to bring it down or find lower spring rate couplings. You may also raise your minimum idle speed.

The best way to do this will involve doing all of this work analytically. Yes, you or someone you trust will have to get friendly with vibration theory to do so. It is not that hard, many thousands of engineering students do it annually in their sophomore or junior year of college every year. Doing it any other way is to basically heading for random trials and error.

Billski

 

[wsimpso1]

Airframes are generally either designed to allow what ever comes along for engines to be hung or are designed around engine systems that are already defined. Doing it the other way around makes for difficult times and can make for fatal flaws.

A cog belt drive or gear drive at the prop end would normally have the bearing set to support the prop built into it, so the argument that you already have a prop bearing to work around is not really much of an issue. Packaging the system may then be the hard part. I propose the altering your airframe is the least of your issues... The powertrain has to work reliably to ever fly your bird.

One path for a coaxial drive would be a double cog belt drive, where you do part of the reduction going off line then the rest of the reduction to bring it back in line. Yes, two belt reductions, but the prop bearings can be integral parts of the system, and it is a coaxial arrangement.

Another path for coaxial arrangement will involve getting friendly with automatic transmission planetary gear sets and selecting one that has suitable speed ratio, torque capacity, and rotational speeds, then building a case around it for containing lube and reacting torque to the airframe. Again, prop bearings would be integral to such a gearbox design.

I suggest these options only down stream of thorough isolation of engine vibration. Isolation is achieved by either a soft enough shaft or a combination of a shaft and a suitably springy coupling. Tuning of the natural frequency and thus the amount of isolation achieved is by adjusting spring rates and engine flywheel inertia. Analytical approaches are definitely preferred to experimental ones.

I understand that becoming capable of the analytical work on this topic is daunting. I feel that it is way more preferable to an expensive and potentially dangerous trial and error series that has little likelihood of producing a successful powertrain.

Billski

 

[henryk]

This shaft drive setup aslo works well; by keeping the engine and drive shaft seperated except for the belt.

View attachment 95727

=or DIFFERENTIAL gear (circa 3: -/+ 1) , natural aerodempfer, +30 % thrust force...

(70 HP, 230 kG )

Dym - zwolnione tempo.mp4

Shared with Dropbox

www.dropbox.com

IMG_6622.mp4

Shared with Dropbox

www.dropbox.com

 

[wsimpso1]

This shaft drive setup aslo works well; by keeping the engine and drive shaft seperated except for the belt.

Directly driven belt drives work OK on industrial engines where the engine output shaft is designed to carry a pulley load. The Subaru E-81 and EJ series engines are car engines and the end of the crank shaft is NOT designed to stand pulley loads. Placing a pulley on it with belt loads will usually result in failing either the crankshaft or the last main bearing on the crank or both. Some engines might run longer than others, but they will not live as many hours as we expect in our homebuilts.

Usually an automotive conversion PSRU runs pulleys (or gears) supported on both sides by bearings and requires some sort of coupling between the engine and PSRU that accommodates the small positional and angular errors you will have in aligning the crankshaft and PSRU input shaft. The coupling prevents the bearing fight that otherwise occurs from these errors, which will usually destroy the crank or crank bearings or bearings in the PSRU.

The other thing that goes on with the pictured system is that the crank pulley sees every bit of torsional accel and decel from firing pulses and 2x firing pulses, and so the belt drive system has to be much beefier than if it were isolated. Order of magnitude on this - Firing pulses are about eight (8!) times median torque of an four cylinder engine. So, you can run big enough pulleys and cogbelts, but they have to be a lot heavier, and WEIGHT IS THE ENEMY.

As long as you are putting a coupling in between the engine and PSRU, you might as well put in a designed amount of torsional compliance to reduce natural frequency of the engine - PSRU to significantly below idle firing frequency. Properly done - selection of flywheel inertia and torsional spring rate - the firing ripple going through the belt drive is modest. By the time the prop sees the ripple, isolated further by the shaft's compliance, it is almost non-existent.

Billski

 

[Sabine]

I can tell you from a number of people's hard experience that guessing and trial and error on this will require a separate test rig where you may tear the thing up and have to rebuild it. I personally took over vibration management on a project where they were attempting to make a direct drive shaft system work by just trying stuff without doing any analytical work, and their test rig was all scarred up from broken parts flying about and thrashing around. The airport where this project was working had exiled their test runs to a far corner of the field away from EVERYTHING else on the field. Some of their schemes lasted less than 5 minutes. I am here to tell you that I modeled their system and tamed it with one modification. They had full time folks fighting with this problem for 18 months prior to my involvement... No, this retired engineer is not for hire, but I will check your math if you want. This is not a place for trial and error.

The PTO systems you see in agricultural equipment have a big flywheel on the engine, a torsional spring set in the clutch hub, then the gearbox, then the PTO output, shaft, and then inertia of the load. This is very similar to cars and trucks, with first the engine, the spring set, the gearbox, then the output shaft, then the load. The inertia of the engine is isolated vibrationally from the inertia of the rest of the system by a purpose built device that has a spring set designed to make the resonance frequency of the system be lower than the operational firing frequencies at their ground idle.

You have already chosen the engine, prop, and shaft. That defines the minimum engine inertia, prop inertia, and shaft spring rate, and number of firing pulses per engine turn. That leaves to be defined your flywheel inertia, cog belt system inertia, spring rate of the coupling and min idle speed. This is a large design space to explore, and doing so analytically is the way to do it. Once you have the basics for your cog belt drive established, you can figure out its inertia and fix that. You will have to find or build a coupling with somewhere around the right amount of torsional spring rate and more than enough torque capacity to handle the engine output. I would start with an engine idle speed no lower than about 1600 rpm, pick the lightest flywheel that will allow that idle, and see what the available spring rates (that have torque capacities above your engine max torque) do to your resonance frequency. If it is too high, you can add inertia to the flywheel to bring it down or find lower spring rate couplings. You may also raise your minimum idle speed.

The best way to do this will involve doing all of this work analytically. Yes, you or someone your trust will have to get friendly with vibration theory to do so. It is not that hard, many thousands of engineering students do it annually in their sophomore or junior year of college every year. Doing it any other way is to basically head for random trials and error.

Billski

Thanks so very much- yes I do have a test rig and will use it and follow your advise. Greatly appreciated