# Torsional Vibration on a long shaft?

### Help Support HomeBuiltAirplanes.com:

#### RSD

##### Well-Known Member
In the case of a ducted fan engine setup that I am working on I have the room/need to put a shaft between the output of the engine and the fan. Would putting two flexible couplings on the shaft help with the torsional vibration and other rotary problems?

HBA Supporter
Log Member

#### wsimpso1

##### Super Moderator
Staff member
Log Member
Have you read this yet? Posts 1 and 16 seem to be required reading for your topic.

Elastomeric couplings help bunches in that they allow the engine, shaft, and prop bearing to each do their individual movements while having a lashless connection between them. We found that their torsional compliance was small (spring rates in torsion were high) and so contributed to isolation only modestly, and may preclude a "stiff" system, but so will the shaft...

In a ducted fan system, you have a bunch of choices to make. Is it direct drive or will the fan be driven by a PRSU? If a PSRU, how will it be connected? Each has good and bad points. Let's go through a few options:
• If direct drive, the ducted fan will likely be the big inertia, with the engine being another large inertia. With the shaft in between, you have a two mass vibe system with a single spring (shaft and couplings). This is relatively straightforward to analyze. Big issues are: First order resonance must be "low enough" and shaft whirl mode must be precluded within the operating mode.
• Normal theory is that you put 1st resonant order of the engine and fan oscillating opposite each other more than one-half octave below min firing speed. This done by modeling the engine and prop/fan inertia and the the torsional spring rate of the shaft and couplings, then doing the Eigen calcs. The shaft is your soft element. Your adjustable features are torsional spring rate and engine side inertia. Your givens are the entire system must be strong enough. If the shaft fails, it does not just leave you engine out, the broken shaft will flail about, with risk of impact to humans and the airframe;
• That being said, I worked on a system that put resonance slightly above idle firing frequency and it was fine. This worked because at idle level speeds while running a prop or fan, torque is small (torque that a prop/fan can express is a function of rpm squared), so firing pulses (four cylinder, so 2 per rev) looked like and coincided with the next order (4 per rev) from piston accels, so we were still isolating relative to the primary forcing function at idle power. You will need good FEA Dynamics and Eigen analysis tools to design in this corner;
• Whirl mode must be checked for and prevented. This is also called shaft critical speed, and it must be above max operating speed. Calcs are straightforward, solutions are too, but run against shaft sizing for carrying torque to the fan. If a soft enough shaft for isolation is present, a one-piece shaft may put flail mode inside the operating range. Alternatives are breaking the shaft into multiple pieces (mid shaft flex joints and support bearings) and single shaft with multiple support bearings (Bell 47 tail rotor drive shafts are visible);
PSRU at engine and isolated - If you bolt the PSRU to the engine and put a soft element between them, the soft element must put 1st vibe order below min operating speed. Possible exceptions discussed above may come into play. The down side is that the shaft will now be at torque of the engine times the torque ratio of the PSRU, needing a beefier shaft to carry that torque. The above caveats about isolation and shaft issues apply;

PSRU at engine and not isolated - If you bolt the PSRU to the engine with no isolator between them, the engine and PSRU must then be a stiff system, with 1st vibe order resonance 1-1/4 octaves or more above max firing frequency. Torsional variation made in the engine is felt by the PSRU and so requires strength commensurate with the maximum seen. http://www.epi-eng.com/piston_engine_technology/torsional_excitation_from_piston_engines.htm is an excellent presentation on the topic of engine torque variation. The above caveats above still apply.

PSRU at prop - The shaft between engine and PSRU may be adequately soft for vibe isolation or you may still need a soft element at the engine end. This is all to isolate elements from engine torsional vibe. All other caveats above also apply.

Then there are the issues of the fan and its torsional vibration inputs to the system... I suggest that this is territory for talented engineers skilled in airframes and torsional vibration management. I further suggest than all options be analyzed and taken through preliminary estimates of weights, sizes, and vibe management before picking a path.

Billski

Last edited:

#### Hot Wings

##### Grumpy Cynic
HBA Supporter
Log Member
Have you read this yet? Posts 1 and 17 seem to be required reading for your topic.
Billski
Ok, what happened to post #17?! Was it so required that someone has it on permanent loan?

When I look at the first page of that thread now I see the post numbers skipping from 16 to 18. Odd..........

#### RSD

##### Well-Known Member
Ok, what happened to post #17?! Was it so required that someone has it on permanent loan?

When I look at the first page of that thread now I see the post numbers skipping from 16 to 18. Odd..........
I can see post #17 - still having troubles?
Have you read this yet? Posts 1 and 17 seem to be required reading for your topic.

Elastomeric couplings help bunches in that they allow the engine, shaft, and prop bearing to each do their individual movements while having a lashless connection between them. We found that their torsional compliance was small (spring rates in torsion were high) and so contributed to isolation only modestly, and may preclude a "stiff" system, but so will the shaft...

In a ducted fan system, you have a bunch of choices to make. Is it direct drive or will the fan be driven by a PRSU? If a PSRU, how will it be connected? Each has good and bad points. Let's go through a few options:
• If direct drive, the ducted fan will likely be the big inertia, with the engine being another large inertia. With the shaft in between, you have a two mass vibe system with a single spring (shaft and couplings). This is relatively straightforward to analyze. Big issues are: First order resonance must be "low enough" and shaft whirl mode must be precluded within the operating mode.
• Normal theory is that you put 1st resonant order of the engine and fan oscillating opposite each other more than one-half octave below min firing speed. This done by modeling the engine and prop/fan inertia and the the torsional spring rate of the shaft and couplings, then doing the Eigen calcs. The shaft is your soft element. Your adjustable features are torsional spring rate and engine side inertia. Your givens are the entire system must be strong enough. If the shaft fails, it does not just leave you engine out, the broken shaft will flail about, with risk of impact to humans and the airframe;
• That being said, I worked on a system that put resonance slightly above idle firing frequency and it was fine. This worked because at idle level speeds while running a prop or fan, torque is small (torque that a prop/fan can express is a function of rpm squared), so firing pulses (four cylinder, so 2 per rev) looked like and coincided with the next order (4 per rev) from piston accels, so we were still isolating relative to the primary forcing function at idle power. You will need good FEA Dynamics and Eigen analysis tools to design in this corner;
• Whirl mode must be checked for and prevented. This is also called shaft critical speed, and it must be above max operating speed. Calcs are straightforward, solutions are too, but run against shaft sizing for carrying torque to the fan. If a soft enough shaft for isolation is present, a one-piece shaft may put flail mode inside the operating range. Alternatives are breaking the shaft into multiple pieces (mid shaft flex joints and support bearings) and single shaft with multiple support bearings (Bell 47 tail rotor drive shafts are visible);
PSRU at engine and isolated - If you bolt the PSRU to the engine and put a soft element between them, the soft element must put 1st vibe order below min operating speed. Possible exceptions discussed above may come into play. The down side is that the shaft will now be at torque of the engine times the torque ratio of the PSRU, needing a beefier shaft to carry that torque. The above caveats about isolation and shaft issues apply;

PSRU at engine and not isolated - If you bolt the PSRU to the engine with no isolator between them, the engine and PSRU must then be a stiff system, with 1st vibe order resonance 1-1/4 octaves or more above max firing frequency. Torsional variation made in the engine is felt by the PSRU and so requires strength commensurate with the maximum seen. http://www.epi-eng.com/piston_engine_technology/torsional_excitation_from_piston_engines.htm is an excellent presentation on the topic of engine torque variation. The above caveats above still apply.

PSRU at prop - The shaft between engine and PSRU may be adequately soft for vibe isolation or you may still need a soft element at the engine end. This is all to isolate elements from engine torsional vibe. All other caveats above also apply.

Then there are the issues of the fan and its torsional vibration inputs to the system... I suggest that this is territory for talented engineers skilled in airframes and torsional vibration management. I further suggest than all options be analyzed and taken through preliminary estimates of weights, sizes, and vibe management before picking a path.

Billski
Cheers Billski - a lot more to this subject than I was aware of!

The first iteration of this will be a re-engine and cockpit modernisation of an existing early 80's ducted fan aircraft - it was originally designed for an Allison turboshaft, but was acquired without an engine, and am looking to replace it with a Mazda rotary of similar horsepower. The 2.17:1 PRSU is mounted to the fan, the Allison had a shaft + two flexible couplings between it and the PRSU on the fan. The second iteration will be a twin engine/twin much smaller fans version.

Last edited:

#### wsimpso1

##### Super Moderator
Staff member
Log Member
Ok, what happened to post #17?! Was it so required that someone has it on permanent loan?

When I look at the first page of that thread now I see the post numbers skipping from 16 to 18. Odd..........

I typoed in my post here. Try Posta 1 and 16 ... with apologies. Will edit.

#### Hot Wings

##### Grumpy Cynic
HBA Supporter
Log Member
I typoed in my post here. Try Posta 1 and 16 ... with apologies. Will edit.
No apologies needed! I still can't see post #17 so it must be an artifact of the move to the new forum software.

@ RSD = Just read the whole thread. Billski's posts are worth the time.

#### pictsidhe

##### Well-Known Member
Is post 17 by someone on your ignore list?

#### wsimpso1

##### Super Moderator
Staff member
Log Member
All this fuss over post 17 is thread drift...

#### RSD

##### Well-Known Member
All this fuss over post 17 is thread drift...
I was thinking that too...

Let's get back on topic folks as obviously I need to make this beastie fly again and it needs more horsepower than most aircraft discussed around here.

#### wsimpso1

##### Super Moderator
Staff member
Log Member
The first iteration of this will be a re-engine and cockpit modernisation of an existing ducted fan aircraft - it was originally designed for an Allison turboshaft, but was acquired without an engine, and am looking to replace it with a Mazda rotary of similar horsepower. The 2.17:1 PRSU is mounted to the fan, the Allison had a shaft + two flexible couplings between it and the PRSU on the fan. The second iteration will be a twin engine/twin much smaller fans version.
OK, that narrows it down. Was it successfully run and flown with the Allison?

Turbine engines do not have firing pulses per se. They do have blade passing frequencies, and usually the number of blades in each wheel (rotors and stators) are prime numbers, and different ones between adjacacent wheels, all to drive blade passing frequencies as high as possible. The result is that the turbine only has really high frequency content.

Usually turbines are geared way down to drive a prop or insert into a helo gearbox. These are usually well built with precision gears and bearings to keep the tooth passing amplitudes small and prime numbers tend to be used on the tooth counts. Not that your 2.17 ratio is most likely a 37:17 tooth gear set, two prime numbers.

In the end, the output of these engines is pretty darned smooth. Piston engines have firing order and 2x firing order as their primary. With a long shaft and PSRU at that end, you will most likely need to isolate all of that by driving the 1st torsional order way low. Tuning tools will be adjusting the spring rate of the total driveline and inertia at the engine side. The prop or fan tends to have the largest inertia of the system. You will be able to get a fair amount of springiness from the shaft and elastomeric couplings, but you may need to add more isolation at the engine flywheel.

Fans running in shrouds will have varying angle of attack as they go around, so you will have at minimum as many torsional pulses sent back up the system as you have blades on the fan. If there are straightening vanes, they too will produce inputs. Then if the fan is in the wake of the wing or other control surfaces, the they will add in additional torsional inputs at the fan and resonance from this train of pulses must also be checked for. My biggest check would be no fan pulse trains line up on 2x or 4x engine revs.

Your choice of a Mazda rotary though may get interesting. They are known to be more trouble to adequately isolate from powertrains than piston engines. The same fundamentals apply as for a four banger, but the Wankel can get interesting because it is more severe.

The existing shaft and PSRU should be reviewed carefully for several issues:
First is that it is likely to see more torsional vibration input. Max momentary torque on the shaft and PSRU may be higher;
Second is a check of flail mode in shafting;
Third is a survey of torsional resonance is in order as the original designers were not running with significant firing pulses and their reflections, but you are.

Billski

#### RSD

##### Well-Known Member
OK, that narrows it down. Was it successfully run and flown with the Allison?

Turbine engines do not have firing pulses per se. They do have blade passing frequencies, and usually the number of blades in each wheel (rotors and stators) are prime numbers, and different ones between adjacacent wheels, all to drive blade passing frequencies as high as possible. The result is that the turbine only has really high frequency content.

Usually turbines are geared way down to drive a prop or insert into a helo gearbox. These are usually well built with precision gears and bearings to keep the tooth passing amplitudes small and prime numbers tend to be used on the tooth counts. Not that your 2.17 ratio is most likely a 37:17 tooth gear set, two prime numbers.

In the end, the output of these engines is pretty darned smooth. Piston engines have firing order and 2x firing order as their primary. With a long shaft and PSRU at that end, you will most likely need to isolate all of that by driving the 1st torsional order way low. Tuning tools will be adjusting the spring rate of the total driveline and inertia at the engine side. The prop or fan tends to have the largest inertia of the system. You will be able to get a fair amount of springiness from the shaft and elastomeric couplings, but you may need to add more isolation at the engine flywheel.

Fans running in shrouds will have varying angle of attack as they go around, so you will have at minimum as many torsional pulses sent back up the system as you have blades on the fan. If there are straightening vanes, they too will produce inputs. Then if the fan is in the wake of the wing or other control surfaces, the they will add in additional torsional inputs at the fan and resonance from this train of pulses must also be checked for. My biggest check would be no fan pulse trains line up on 2x or 4x engine revs.

Your choice of a Mazda rotary though may get interesting. They are known to be more trouble to adequately isolate from powertrains than piston engines. The same fundamentals apply as for a four banger, but the Wankel can get interesting because it is more severe.

The existing shaft and PSRU should be reviewed carefully for several issues:
First is that it is likely to see more torsional vibration input. Max momentary torque on the shaft and PSRU may be higher;
Second is a check of flail mode in shafting;
Third is a survey of torsional resonance is in order as the original designers were not running with significant firing pulses and their reflections, but you are.

Billski
Was successfully flown with the Allison, I believe that the aircraft model has about 20,000 hours flown amongst the 20 or so produced.

Lots of things to consider by the look of it.

#### pictsidhe

##### Well-Known Member
It will take some significant engineering effort to drive a fan via a shaft. If it is possible to relocate the fan so it can be directly driven, that is the easy way. A shaft won't have a simple design way. The design may well look simple once done, but it has to be just right.

#### RSD

##### Well-Known Member
It will take some significant engineering effort to drive a fan via a shaft. If it is possible to relocate the fan so it can be directly driven, that is the easy way. A shaft won't have a simple design way. The design may well look simple once done, but it has to be just right.
It may be possible to position the engine so that it is directly driven. What are your thoughts on this Billski?

#### Vigilant1

##### Well-Known Member
Just to check the box--I assume you've learned all you can from Perry Mick's Wankel-with-a-ducted-fan experience.

#### RSD

##### Well-Known Member
Just to check the box--I assume you've learned all you can from Perry Mick's Wankel-with-a-ducted-fan experience.
Umm... no?

#### Vigilant1

##### Well-Known Member
Umm... no?
Perry Mick used a modified Mazda 13B to turn a ducted fan to power his Long-Eze . More here, and in many other places. He experimented with several different fan configurations. In the end, he went back to a propeller for better performance. At any rate, there's no point in re-learning what he's already learned. In addition to his web site linked above, check Contact! magazine, do some searches of the site here at HBA, and at the other Mazda aviation sites. The case was pretty well documented and might have information of use to you (fan configurations, etc). I'm sure you already know that, unless your cruise speeds are quite low, a ducted fan is not at efficient as a prop. But, it sounds like the configuration of the airplane makes a prop impractical/undesirable.

#### RSD

##### Well-Known Member
Perry Mick used a modified Mazda 13B to turn a ducted fan to power his Long-Eze . More here, and in many other places. He experimented with several different fan configurations. In the end, he went back to a propeller for better performance. At any rate, there's no point in re-learning what he's already learned. In addition to his web site linked above, check Contact! magazine, do some searches of the site here at HBA, and at the other Mazda aviation sites. The case was pretty well documented and might have information of use to you (fan configurations, etc). I'm sure you already know that, unless your cruise speeds are quite low, a ducted fan is not at efficient as a prop. But, it sounds like the configuration of the airplane makes a prop impractical/undesirable.
OK cheers for that. Some interesting reading for me to do there. I'm fortunate in my case that all of the hard work for the duct/fan design is done, I've acquired an airframe with all that installed and working, but without the very expensive 650 hp Allison turbine. Not having a spare $300,000 or so available I'm looking to drop a 650 hp rotary in there with the aim of making it go like this again - #### Vigilant1 ##### Well-Known Member Lifetime Supporter OK cheers for that. Some interesting reading for me to do there. I'm fortunate in my case that all of the hard work for the duct/fan design is done, I've acquired an airframe with all that installed and working, but without the very expensive 650 hp Allison turbine. Not having a spare$300,000 or so available I'm looking to drop a 650 hp rotary in there with the aim of making it go like this again.
That's quite a project. As a professionally engineered unit, there should be good information on the inertial characteristics of the fan assembly and any gearing, which may be of use to you.
They are exotic looking airplanes. One of our frequent posters on this site still holds them up as good performers.
I'm sure you've looked into the weight and balance issues, I would think they are significant. A 650HP Wankel will be lighter than a piston alternative, but heavier than a turbine. You'll get some of that weight back due to the lower fuel requirements, hopefully it will be in the right place.
You'd probably prefer to maintain the looks of the plane as is. However, if the engineering challenges of mating a Wankel to the existing driveline become too much, you might at least consider going with an open pusher prop with a PSRU, if you can make it work from a packaging perspective. The drag from that shroud is likely considerable, I'd imagine service parts for the proprietary fan unit will not be easy to get, and a conventional PSRU and prop may provide increased cruise thrust in comparison.