Can someone enlighten me please (Vibration absorbers)

[Phoinix]

What is this yellow thingy above the main rotor head? What is it's function? How does it work?

http://makete.net/galerija/albums/album36/untitled.jpg

[BlenderPilot]

Those are vibration dampers, and they are a good way to differentiate between the Mi-8 and Mi-17, amongst other things like the tail rotor on the opposite side, or the longer hump one of them has due to the pneumatic start versus the lead battery start on the older models.

How they work? I just know they are heavy and damp out vibration, same as in Bell 412, or Sikorsky's, Bo105's, someone else will know how to better explain this.

[SilsoeSid]

My guess is that it that this is a Mi-17-1/2 export variant with its mast-mounted vibration absorber.

http://www.b-domke.de/AviationImages/Rotorhead/Images/3828.jpg


SS

[Phoinix]

I did know about the differences you gave, but i didn't know that those dampers are also one of them. This damper is from a Mi-14PL, naval version, tail rotor on the left.

I only saw this damper on Mi-8MTV and this Mi-14.

[SilsoeSid]

A few more examples.

http://www.military.cz/russia/air/helicopters/Mi_8/images/mi17md_3.jpg

http://www.military.cz/russia/air/helicopters/Mi_8/images/mi8s.jpg

http://www.china-military.org/units/acft/cfte/photos/helos/CFTE-Mi-17_017.jpg

[Phoinix]

First is 17, other two 8, right? 17 has partical separator FWD of intake, right?

[Graviman]

If it works anything like the torsional dampers fitted to engines, then they work like this:

The mass of the dampers acts like a flywheel (i looked hard, but they all seemed to be connected together). The flywheel is connected to the top of the rotor shaft by a torsional spring. The frequency of the flywheel/spring system is chosen to be the same as the main frequency of concern in the rotor drive system. This will be a torsional mode, and could be free power turbine mass twisting drive shaft against rotor mass, or a particular blade mode that the designers were trying to avoid.

When the drivetrain resonance rears its head, the damper becomes excited. In this way the modal energy is transferred into the damper. Normally the torsional spring is an elastomeric rubber, with a high level of hysteresis. The practical upshot is that the drivetrain seldom gets into a resonance, which prolongs the time-life of the entire drivetrain (worst case for drivetrain resonance is low cycle fatigue failure).


Hmmm, Powertrain dynamics - theres an interesting area...

Mart

[NickLappos]

Graviman has it almost right. Those are in-plane vibration absorbers, similar to the Bifilar used on Sikorsky helicopters. The free weights are inside the yellow covers in the photo, and are not connected to each other. The covers keep ice and dirt away from the weights. The weights are held in place with bearings that let them swing about the arms, so that they are pendular in motion.
They are normally tuned to absorb the 5 per rev (5R) vibrations that the rotor sends to the fuselage.
The weights are held in the CF field by the rotation, so they act as though they are on springs, and oscillate about their bearings. They resonate at the 5 per rev frequency so they can efficiently absorb the vibrations that the blades give the head.

Here is a close-up:
http://www.b-domke.de/AviationImages/Rotorhead/3828.html



Here is a shot of a Sea Hawk head, where the weights are exposed and the bearings can be seen:
http://www.b-domke.de/AviationImages/Rotorhead/1642.html

[noooby]

Phoinix

Looking at four Mil8-MTV's right now here in Tajikistan. One with particle sep, 3 without. Interstingly, none of them have the Vibration Absorber fitted. Might only be MTV-1 that has it fitted, and -17.

noooby

[Graviman]

Thanks for putting me right, Nick. Learn something new every day. :D

Is this a problem that is particular to these 5-bladed articulated, or are 1P modes something that bugs every new design? I imagine that the other way of solving it is hydromounting the gearbox, to isolate the fuselage from all rotor dynamic vibration.

Out of interest have you found that the rotor dynamics guys get it right on a new design, or would this normally be an area reserved for flight testing (rotation order frequency plots, etc)? My reading of both Prouty and Newman (Westland) indicated to me that while the cause of 1P,2P,3P, etc was generally understood, the magnitude of these was normally only determined once the first machine was flying (although i imagine some whirl tower dynamics would also be used).


(Ahhh - If i had a vast cash reserve i would while away the hours thinking about physics, designing engines and flying helicopters :rolleyes: )

Mart

[NickLappos]

Graviman,

The natural vibration that we see at N per revolution (where N is the number of blades) is a product of the fact that each blade is delivering a chunk of the lift. The more blades, the more (and smaller) chunks and the less N per Rev that the rotor delivers to the aircraft, so a 5 bladed rotor has less natural vibration than one with fewer blades.

Generally, large transports need more vibe solutions not because the rotor puts out more vibration, but because the bigger fuselages with larger spans across and along the cabin create more susceptability to an adverse reaction to the naturally imposed vibration. In other words, bigger things shake more easily at the typical frequencies (think of a diving board, the longer it is the very much lower the natural frequency is, and the more likely it will react to a disturbing vibration.)

1 per rev is not helped or cured by the bifilar, only N per Rev.

Here is what I posted previously when a question about Black Hawk bifilars came up:

I helped develop the bifilars on two of our models, and understand them a bit:

They are simple swinging masses that are resonent near the N/rev frequencies (4 per revolution in the Black Hawk and S-76)- they love to hum along at the same beat frequency as the blade passages. They are actually tuned to N-1 and N+1 because they are in the rotating system, where the blade passage frequency is made up of the sum of those two frequencies.

The bifilar is much lighter than other vibration absorbers, and right at the source. Also, there is no spring on a bifilar, they achieve their frequency by CF, so they automatically tune to the rpm they are at, allowing large rpm changes without large vibration changes.

Bifilars work in-plane, the most effective direction. Generally, if only one is needed, the one bifilar is tuned to 3 per rev (N-1) which is the bigger, more objectionable component. Nobody removes bifilars by themselves, because they are part of the airworthiness of the machine, not an option. The vibrations at many airframe stations, not just the people, depend on them, so just tinkering with them might make some nice piece of equipment very sad without showing up on the normal maintenance gear. We want our engines to be very very happy, don't we?

Bifilars were invented for piston engines decades ago, and are not rocket science.

The orientation of the arm over the space between the hub arms is specifically designed to allow ease of maintenance and to prevent interference with blade flapping. There is no issue with the exact orientation, the weights all act to absorb the in-plane vibrations.

The only vibration they act on is N per rev. NO other frequency is helped by the bifilar. If one weight is badly slung, and does not swing properly, some odd frequencies can be flet, because the other weights conspire to upset the vibration balance. Typically, a 3 per rev is felt if the bifilar is badly maintained.

[outhouse]

Igor was a clever chap and the Bifilars are a cool way of correcting for the change of disk c of g with cyclic changes to disc path plane on the S61 and S76. Bell has used a similar method with the pendulum dampers on the 412. All works well as long as the lateral movement on the bearings is kept to a minimum. Hint to 412 drivers look out for brunelling(?) on the needle bearings during preflight inspections.
Outhouse

[PhilJ]

now for my helpful addition to this thread:

The spelling is brinelling, named after the swedish guy who came up with a hardness test.

[outhouse]

Thanks for the addition to the thread PhilJ, I was unsure regarding spelling thus I included the? I will add to my dictionary and hope to perform better regarding spelling bees.
As a friendly extension to the discussion regarding the subject would anyone like to suggest a formula that enables the calculation to achieve maximum effect? Hint rotor speed X.
To 412 drivers, what would be the effect of lateral movement on the damper weights and or any brinelling on the needle bearings?

[Phoinix]

Thanks to everyone. Damn, i thought i knew the basics. Great reading.

[NickLappos]

outhouse,

The vibrations that the blades deliver to the head have nothing to do with CG effects due to flapping. The dynamics of how stuff vibrates is complex, but understandable in a sense.

The cg effects are tiny, it is the aerodynamic force and the response of the blade as a structure that tell the story. Every diving board has a natural frequency, as does each tile on a xylophone. The way these vibrations arise is due to the tuning of the blade and head response to the fuselage response.

[Dave_Jackson]

MODELING THE BIFILAR PENDULUM (http://www.brazd.ru/books/b0005/book/58af110.pdf)

[Graviman]

Blimey! It took the combined talents of Prouty, Newman, Lappos, and Jackson to shove it into my brain - but i got it...

In forward flight, as the blade sweeps out it's path, each position of each blade is subject to varying angles of attack and sweep from the ingested air. The fast revolution of the rotor means that these forces likely excite all of the blade fundamental resonances. Additional vibration sources exist in the form of the varying profile (and to a lesser extent induced) drag through azimuth, and any manouvreing input (particularly for an articulated hub). This inevitably leads to the hub suffering input forces with fundamental harmonic of Np. Higher harmonics also exist, but are less in magnitude (and may be trimmed out with Higher Harmonic Control).

The bifilars are basically just pendulums with clever mechanisms for a compact geometry (using centripetal accel). They are no good if tuned to Np since the pendulums on (say) the advancing and retreating sides of the rotor would "wobble" in phase relative to rotor rotation, but out of phase with repect to airframe, thus cancelling any fore-aft force. By tuning them all to either Np+1 or Np-1 then you ensure that bifilars on (say) the advancing and retreating sides "wobble" out of phase wrt to rotor rotation, but in phase wrt to airframe fore-aft direction. This means that both go forwards and backwards together at just the right frequency to cancel the rotors Np fore-aft wobble - and likewise for sideways wobble.

In laymans terms:

System behave like a dod of metal on the end of a spring, tuned to Np... :}

Mart

[Edit: To rephrase first paragraph in accordance with Nick Lappos' subsequent posts]

[NickLappos]

Graviman,

Your second paragraph about the weights rotating in a centrifugal field as if by spring force is right on!

The first paragraph however is still hanging onto some idea that the N per revolution is due to a blade movement/mass balance issue, which is not correct. It is funny how mechanical engineers see the mechanism, but not the air! The blades go from upwind at Mach .9 to downwind at near stall. They see flipping and shoving by tons of air per minute, and are constantly ringing at their natural frequencies as the loading on them shifts with the air's forces. They attempt to average out 1 x the aircraft's weight while doing these gymnastics.

This creates a natural N/Rev vibration, (N+1 and N-1 in the rotating field) that shakes the aircraft.

Why is it not blade motion? Two proofs: 1) The max flapping occurs highest in many aircraft in level flight at 40 knots, but peak N/rev is usually (always) at max speed.
2) Rigid rotors with the least flapping motion have the highest N/rev!

[Graviman]

Nick,

Good point well made. The use of compliance to explain the force input was not the best approach.

Mech heads just see the world in terms of things they can affect. Air is that viscousy stuff that you need to heat to make engines whirr, and throw at the ground to make aircraft whizz...

Mart

[Edit: Shpelleeng]

[Dave_Jackson]

There will be symmetry of lift between the advancing side and the retreating side during forward flight, therefor the induced drag on the advancing side and the retreating sides must be very similar.

However, the profile drag will be very different between the advancing and the retreating sides. The rotor drag force (drag from the forward velocity) and rotating shear force (drag from the rotation) will be summed on the advancing side and in opposition on the retreating side.

This difference in profile drag between the advancing and retreating sides will increase as the forward velocity of the craft increases.

At higher forward velocities, retreating tip stall and advancing tip compression must further increase the rotating shear force. God knows what this does to the basket of assorted/P vibrations.

One thing that I have difficult understanding is why an increase in the rigidity of a rotor will cause an increase in the rotor's in-plane vibration. Perhaps, in a 3-blade rotor it might, but in a 'theoretical' 4-blade rotor with no provision for led/lag the total profile drag of the complete rotor should be quite constant throughout the full rotation.

Could it simply be a case of the vibration increasing with forward velocity, and faster craft tend to have greater rotor rigidity?
__________________________

The Bifilar vibration damper is a means of isolating some of the problems of the rotor from the rest of the craft. The slowed rotor concept (http://www.unicopter.com/1090.html) appears to be a means of reducing the actual cause of the problem.

Dave

[NickLappos]

Dave,
You are looking at it with the right set of glasses, we call the N/Rev disturbences "root shears" to mark their equivilent vertical and in-plane vibratory shear force acting on the head. These have strong content that varies with the number of blades, so that going from 2 to 3 or 3 to 4 blades makes a non-linear improvement on the size of the root shears. The root shear magnitude drops markedly due to the number of blades, and its frequency rises at the same time, making the net effect a reduction in root shear effect that is not quite N squared (so that 4 blades has something like 15% of the vibration of 2 blades.)

The root shears rise due to the rigidity of the blade attachment (since the blade can't flap or hunt to relieve some of this shear mechanically.) Thus, your hope that an awsomely rigid blade would be immune to N/rev vibration is actually backwards. Most of the time, we try to tune the blade to be compliant (softer) in a given frequency where it can help relieve these vibrations or absorb them.

The response of the fuselage is important to vibration supression. The low N/rev frequency of a 2 bladed rotor is a problem all by itself, because as the frequency drops, more things on the aircraft "like" that frequency, and try to resonate at it. This makes the ride quality much worse, of course. Also, as fuselages get bigger, their natural frequencies drop, so they need more vibration treatment.

All rotors produce N/rev, it is not a sign of a poor rotor. Pilots have this mythological belief that if they feel an N/rev, there must be something unhappy in the rotor, or it needs redesigning. Actually, all rotors give off about the same vibration (based on the number of blades, the hinge offset and the rpm) but what you feel is a measure of how well the manufacturer has learned to hide it or absorb it, or how the place where your seat is bolted is responding to it. It is possible for a complex airframe to have variability of 5 to 1 depending on where the seat is attached.

[outhouse]

Many thanks NickLappos, I had a very simplistic explanation when doing a S61 conversion many years ago. I presume that the dampers on the 412 have the same function as they seem to follow the theory. Thanks to all and very interesting.

[Graviman]

"...3-blade rotor it might, but in a 'theoretical' 4-blade rotor..."

Dave, i think you need to find a good aerodynamic efficiency justification for 4 blades per rotor. Removing vortex wake interaction between both rotors will by itself greatly reduce vibration. [Slightly off topic]

----

Nick,

"Most of the time, we try to tune the blade to be compliant (softer) in a given frequency where it can help relieve these vibrations or absorb them."

Has any work been done on blade material damping? I also imagine that the gearbox mounting is carefully designed to maximise lateral compliance (good isolation in X and Y), while minimising torsional compliance (good manouvreing). Are hydromounts the norm (hydraulically damped elastomers - usually Lord or Simrit), and have active mounts been considered?

"The response of the fuselage is important to vibration supression."

I've done a lot of work on modal response and input point force spectrum sensitivity. Powertrain mounting systems are normally the best means of isolation, although you can tune structures to some extent. Since i'm understanding that Np ties up with one of the main airframe bending modes, i imagine structural tuning to be limited.

Mart

[Edit: typing while watching Disco do it's dance with the ISS]

[Dave_Jackson]

Nick,

I understand your concern about high rigidity, and the efforts that are applied in attempting to tune components about a selected RRPM. I hope that the following response will clarify my 'brute strength' approach.


Graviman,
Dave, i think you need to find a good aerodynamic efficiency justification for 4 blades per rotor. For a quick justification, how about Sikorsky's coaxials? The S-69 (XH-59) ABC (http://www.unicopter.com/0891.html) had three blades per rotor. It did not quite achieve the anticipated maximum forward velocity due to excessive vibration. Note that the X2 coaxial (http://www.unicopter.com/1465.html) shows four blades per rotor.


For a longwinded justification; I did a crude and error prone comparison of a 3-blade rotor (http://www.unicopter.com/0871.html) and a 4-blade rotor (http://www.unicopter.com/1218.html) for the twin main rotor UniCopter, a couple of years ago. It was done to evaluate the lateral dissymitry of lift, but I suspect that its conclusions can be applied to drag.

The addition of a 4th blade provides a reduction in the drag of the individual blade, as mentioned by Nick. In addition, it appears that the moment about the craft's X-axis, from the advancing side of an 4-blade 'theoretical' absolutely rigid rotor is a constant, irrespective of what azimuth(s) the (one or two) advancing blades(s) are at.

What I surmise from this is that the individual blades will be subjected to varying in-plane shear as they rotate through the 360º. However, if the rotor is 'absolutely' rigid, the total rotor will not experience any cyclical moment about the Z-axis or cyclical forces along the X and Y-axii.

This makes me think that the closer the rotor can come to 'absolute' rigidity the less the vibration (from this source) should be. In other words; the very high frequencies will be well above what could be detrimental to the structure, components and occupants of the craft.


Dave

[Graviman]

Been thinking about this a while, Dave. :uhoh:

It depends very much on the lift distribution of the rotor. A single rotor could be argued to maintain constant lift throughout the azimuth, so there is "theoretically" no source of vibration regardless of blade number. This was why i had to use the "N bladed rotor is being slammed backward N times per revolution" explanation to explain cyclic drag variation. I'm hoping Nick will expand on the root shear = ~1/(N^2) relationship.

The main vibration source to my mind is during roll or pitch. Since i favour feathered retreating (i know we are at a difference of opinion on that one :rolleyes: ), then this cyclic input could also be seen as the hover condition for ABC intermeshing. Since i gather S-69 ABC had at least reduced retreating lift, if not fully feathered, then perhaps i should conclude that your calculations about 4-blade over 3-blade rotors is right.

However....

If blade pitch/lift is linearly varied from max at 90' azimuth to min (ie feathered) at 270' azimuth then, considering each individual rotor during full azimuth rotation:

3-bladed rotor has total lift varying from 1 1/3 to 1 2/3 blade equivalents, and moment about x-axis of constant 2/3 blade equivalent at 90'. So max blade equivalent variation 20%

4-bladed rotor has constant lift of 2 blade equivalents, and moment about x-axis varying from 0.707 to 1 blade equivalent at 90'. So max variation 29%

From that perspective there is an arguement for 3 blade, but i'm expecting Nick to jump in and tell me i'm talking nonsense... :D

Mart

[NickLappos]

Graviman,
The issue is still that you tend to see the blade as some static source of lift (as in the phrase "maintain constant lift throughout the azimuth") when in fact that blade is whipping about, getting its butt kicked by the relative wind (which changes its azimuth relative to the chord line at about 5 to 10 times per second) and also by the rotor controls, which are constantly changing its feathering in the hopes that the net lift averages out to 1/N of the aircraft's need.

At the hover, there is no N/rev because all the blades are experiencing almost no azimuthal variations, and the flapping across the revolution is approximately the same. As you build forward speed, the blade begins to see this varying velocity vector, varying as to speed and angle from the chord line.

At Vmax, the blade advancing abeam at precisely 90 degrees sees the "wind" as Vrotation plus Vairspeed. When that passes over the nose, it sees the "wind" as the vector sum of Vrotation and Vairspeed (which are 90 degrees apart). Each blade chord segment going out from the root sees the lateral angle of the "wind" differently. At the hub, the "wind" is along the blade, and no lift is produced. At 1/3 span, the "wind" is about 45 degrees to the chord line, at the tip it is about 20 degrees. Do the math for the confusion of the lift distribution, and the net wild spanwise swings of the center of net lift as that blade travels around. Then remember that this variability occurs at 5 to 10 times per second. In effect, the blade is bucked and kicked by its chore, which rings every natural frequency it has. Look up that classic blade movie to see the wild ride the blade takes.

Again, only a stolid ME with a feeling for static solutions could see this as a nice, solvable problem of statics!

If there were one blade, the pounding of the root shear is eye watering, with two blades it is less than half the variability, by 7 or 8 blades, the root shears approximate a steady flow of lift.

I have no idea how you showed the proof that you did, where 3 is the optimal number of blades, but I'd suggest two things :

1) Get a good book on helo engineering (Prouty, and Stepnewski + Keys are two that come to mind)

2) If you have a mortgage, don't quit your day job to run off designing a helo!! ;)

[Graviman]

Nick,

"Again, only a stolid ME with a feeling for static solutions could see this as a nice, solvable problem of statics!"

Not at all - i was just trying to find a way to do some rough sums for the post. I was assuming a blade made from "unobtanium" (spanwise at least), which adjusted pitch and twist in an attempt to maintain lift distribution required through azimuth (ignoring reverse velocity region). The calcs were just to show how, with the equivalent of extreme roll input , total lift and rolling moment will still vary. Prouty had a neat way of integrating a "rigid" blade lift around azimuth to predict rotor performance, but i thought that too much for one post!

From your (as usual, extremely knowledgable) posts i understand that the main cause of the hub vibration input is blade eigenmodes, excited by azimuth rotation air velocity vector variation. This was why i was curious about whether blade material damping had been considered, to get even closer to the source. If blade material technology improves (which is DJs stance), then a theoretically spanwise rigid blade can be approached.

I was really trying to highlight that even with the "perfect" blade design, there would still be sources of vibration (including of course drag variation around azimuth). This was the reason for my question regarding gearbox elastomeric isolation mounting.

It may very well be that feathered retreating is far less practical than ABC. It may be that 4-blade coaxial is the best solution (Sikorsky seem to think so). I'm just trying understand the contraints of the problem. :D

It may equally be that i should just stick to precision hammer adjustment... :}

----

Dave,

"One thing that I have difficult understanding is why an increase in the rigidity of a rotor will cause an increase in the rotor's in-plane vibration."

My understanding of this dynamic system is that stiffer blades of similar mass will have higher frequency eigenmodes. This is more likely to lead to the blade resonance which result in hub input forces. Bifilars effectively attempt to "rigidify" the root mechanical impedance, but a much better solution might be to damp the blade flexural modes directly...

Mart