The challenge of gyrocopter design.

[mnowell129]

Having studied for some time and designed and test flown my gyrocopters for the past few months I came to a tiny amount of understanding about gyrocopter design. This is the beginning of a thread intended for future designers that I intend to discuss the challenges of gyrocopter design that I have encountered. It may generate a discussion among experienced designers as well and that would be welcome. According to David Ramsey this is part 1 of "the book of mickey". I thought I would start with some basics. A rotor in autorotative forward flight with coning is diagrammed. The rotor in side view and top view is shown. Clock wise (from the top) rotation. The "V"s are local air velocities. Vf is the forward speed but it is drawn backwards since this is what the blade "sees". Note that the rotational air velocity vector is drawn backwards to the rotation because that is what the blade "sees".
In side view is the rotor with coning with Vf resolved into the parts down the blade and perpindicular to the blade. The part down the blade does nothing to change the lift of that blade. The part perpindicular to the blade has the effect of increasing the local angle of attack and local velocity of that blade. note that the forward blade has a bigger increase than the backward blade due to coning. If the blades were straight the increase would be the same on both.
On the top view the left blade sees the Vr+Vf, the right blades sees Vr-Vf, creating the asymmetric lift problem.
There are two ways to resolve the asymmetric lift problem (ALP), feathering or flapping. Feathering and flapping are BOTH forms of cyclic pitch (more later).
Either one or both can be used to resolve ALP due to forward flight differences.
Here lies the first great gyrocopter challenge. Flapping cannot be used to resolve the ALP due to coning. If flapping were used to resolve the coning induced lift asymmetry, the forward blade would flap up on the right and the back blade would flap down on the left. The rotor disk would then be tilted to the left by some amount. The rotor always produces a force perpindicular to the plane formed by the rotor tips, called the tip path plane (TPP). So if the TPP is tilted left a left roll is created that must be trimmed out to fly level. This has to be done with feathering. This is accomplished by either shaft tilt feathering or swashplate type feathering. The effect being the same. The blade in front is pitched down as it crosses the front, the one in back is pitched up, this keeps the rotor level. (In my models I make the head rigid with very little coning, thus very little roll trim is needed and it changes very little with airspeed.)
So here is the first great gyrocopter challenge. If you build in flapping hinges to alleviate the forward motion ALP you allow more coning angle to develop. The higher coning angle requires roll trim to stop the coning induced ALP. Further, the faster you go, the more roll trim is required. In fact the forward flight speed becomes limited by running out of roll trim! This is also true of helicopters! On the other hand going faster speeds the head up, causing it to flatten out, reducing coning. The question becomes does the rotor speedup that reduces coning and therefore reduces the roll trim needed end up being more than the increased effect a higher forward speed on coning creates that increases the roll trim needed!!!! Obviously some models will favor one vs the other and is related to the rotor speed, flight speed and flapping stiffness....
First frustrating outcome : Fixed rotor head gyros with coning and shaft tilt to correct for coning are only usually trimmed at one airspeed. If this is takeoff speed then cruise is wrong, if this is cruise speed then takeoff speed is wrong. This is frustrating indeed. What does this lead to? Hand launches right into cruise speed.

The great compromise is just enough flapping to relieve forward motion ALP, but no so much as to allow so much coning that you run out of roll trim....
First conclusion : coning is bad, flapping hinges aggrevate coning.
......
First design choice : Roll cyclic either by shaft tilting or swashplate is probably a good idea, even if you don't use pitch cyclic....so you can maintain roll trim at multiple airspeeds.

[mnowell129]

section 2...
This is a diagram of flapping and feathering. The goal is to reduce the asymmetric lift due to forward flight. We are looking at the left side of a clockwise rotating rotor, moving forward to the left. The blade in the middle is the left hand advancing blade that recieves the sum of the forward velocity and the rotating velocity. This gives it more lift than the retreating blade. We must reduce this lift (and conversely increase the lift of the retreating blade) to achieve balance.
Reducing lift can be done two ways 1) slow down 2) reduce angle of attack. Obviously slowing down the blade at one spot in the circle is not possible. So we must reduce the angle of the attack of the advancing blade and increase the angle of attack of the retreating blade. Since this increase/decrease is zero at the back, max increase at the left side, zero again at the front and max decrease at the right, the pitch is varying around one blade cycle. Hence the term cyclic pitch. Asymmetric lift is removed by cyclic pitch. No one has invented another practical way to do this.
There are two ways to achieve cyclic pitch to remove the asymmetric lift, flapping or feathering. Both achieve the same result of reducing the local angle of attack of the advancing blade.
There are three sets of rotors in the diagram. The top is the rotor with no flapping or feathering. The advancing blade sees the same angle that the retreating blade sees, the same as the blade in any position. But we know the advancing blades sees a higher airspeed and must be pitch reduced to remove the difference in lift.
In the second rotor we allow the rotor to flap along the dotted line. The airflow that the blade sees is now relative to the flapped up dotted line. In other words the blade is climbing as it goes from the back to the front. As the blade passes through the left hand position its angle is no longer +a but is now -b. I have exaggerated the angles to show them better but the effect is the same. The blade, being allowed to flap up, sees a less positive angle of attack and thus does not attempt to climb even further in front, creating an increasing nose up condition. The retreating blade is descending into the airflow down the dotted line. It sees an increased angle of attack and thus more lift, preventing it from trying to flap further down adding to the nose up pitch that would occur with forward motion.

In the third rotor we limit the flap but we mechanically feather the blade in a cyclic manner. Feather means to tilt the blade on an axis down the length of the blade. We can do mechanical feathering by tilting the shaft ("direct control") which we will see later is in fact mechanical feathering or so called cyclic pitch, or by a feathering spider mechanism with a hollow shaft, or by a swashplate mechanism. No matter what mechanical means is used it still results in mechanical cyclic feathering being applied to the blade. Whatever the mechanism used the result is the same. The advancing blade is mechanically feathered so that it sees the same -b angle of attack that would have been produced with flapping.

(Aside: Delta three hinging is used to reduce the amount of flapping needed. I will deal with this as a seperate topic. Delta three actually performs automatic mechanical feathering as well as flapping.)

Important result :
Flapping and feathering ("cyclic pitch") have the same effect on the advancing/retreating blade in balancing asymmetric lift differences due to forward flight.
Important result:
Cyclic Pitch/Feathering provides asymmetric lift difference balancing without flapping thus has a more nose down rotor plane. This gives more tail clearance. This also results in less drag, because as you will recall the force vector is perpindicular to the tip path plane. The flapping alleviated solution is tilted further back resulting in more aft rotor thrust, and therefore more drag.

Important design consideration:
Feathering/cyclic pitch allows a more compact model ( less tail clearance needed), reduces the coning thereby reducing the roll trim needed for forward flight and requires less power to fly. As a non insignificant side benefit mechanical feathering can be used to directly control the rotor, not just trim out asymmetric lift due to forward motion and coning.
To the negative, feathering is more mechanically complex to build.
On the whole, the advantages of mechanical feathering usually outweigh the disadvantages.

The next question is what are the ways to perform mechanical feathering?
There are 4 common ones, shaft tilt, pitch spider, rotor servo flap and swashplate. They all are used to mechanically feather blades in a cyclic manner.

This brings up the issues of controllabilty and stability...
I guess that will be the next ramble...

If you are getting any benefit from my ramblings, please post something to that effect. This will be much more fun for me if I know that at least one person is benefitting from me sharing what I have learned....

[David A Ramsey]

Just got back from a forced vacation to Hawaii. My wife had to yank me from my hallowed basement, kicking and screaming. In the hotel bedroom night table I found that the Gideons and the Hindus were there. Humm. I went right down to the hotel desk and shouted; Where's the Book of Mickey??????

Informative and detailed. Thanks.

David

[mnowell129]

Just got back from a forced vacation to Hawaii. My wife had to yank me from my hallowed basement, kicking and screaming. In the hotel bedroom night table I found that the Gideons and the Hindus were there. Humm. I went right down to the hotel desk and shouted; Where's the Book of Mickey??????

Informative and detailed. Thanks.

David
we gotta meet sometime..
Are you coming down for the autogyro fly in Feb 24,25th?
I'll keep going. ...

[David A Ramsey]

Would be an honor to meet you. I'd like to make the Gyro Fun Fly, but unable to this year.

Prior to your input, much of my autogyro imformation was from "autogyro.com". As helpful as that was it didn't cover recognizing the myriad of problems, solutions, cause and effect. Please go on.

[mnowell129]

Would be an honor to meet you. I'd like to make the Gyro Fun Fly, but unable to this year.

Prior to your input, much of my autogyro imformation was from "autogyro.com". As helpful as that was it didn't cover recognizing the myriad of problems, solutions, cause and effect. Please go on.
Will do.
I want to give credit where it is due. The material organized and presented on autogyro.com is very helpful, a tremendous effort and very unselfish in its presentation. The amount of work represented in design and experimentation is huge and not to be taken lightly. So my discussion here is not meant to diminish that work in any way. The difference is that much of the information presented there is very practical and the result of real experimentation. It is difficult however to use that form of information to make an educated guess as to what is right or wrong with something, it is more of report of what works rather than a predictor of what to do even though there are some guidelines on the site. My goal is to approach the topic from a theoretical perspective with the goal of establishing some way to steer a new design so that it will be successful. I believe if one understands the various forces and problems involved, then one is better able to make design choices and is more aware of the consequences of the design choices they are making. I believe that this understanding was vital to my success and may benefit others. As an example, I knew I could make a PT gyro/Gyro Schtick/ECDC type autogyro fly. But knowing that the direct control head creates very high servo loads ahead of time guided me into another form of control for my design that I wanted to use pico servos on. There are only two other real choices 1) A hollow shaft with pitch spider or 2) a swashplate and resulting stuff. My plan for development was to learn to fly a gyro first and continue to develop once I had a working platform and flight skill. THe hollow shaft spider method is something I am now working on that I will add as the single unknown to my already proven G3PO platform. I chose #2 because I knew it would work and get me in the air and because the parts were cheap, available and required no manufacturing on my part. In the process I sorted out general layout, landing gear setup, ground handling, tail requirements, power requirements, CG sensitivity, tractor vs pusher arrangements, and I learned to fly gyrocopters. From an experimental viewpoint it has been highly successful and I am now much better equipped to do the spider control version...

I'm hoping that I can explain the reasons that lead me to my conclusions and my design with the hope that a theoretical understanding of the problems will let other gyro designers make good choices in what they are building rather than just guessing....
mickey

[mnowell129]

Cyclic, version one.
Having determined that cyclic is probably a good thing to use in the autogyro we go about implementing it. The first obvious and simple choice is to tilt the shaft on which the rotor is rotating. The picture shows a side view of a four bladed rotor in the level flight condition in red. In green is the rotor if we were to gimbal the shaft and try to tilt the shaft backwards. What would happen would be that the advancing blade, the one on the left where we can see the airfoil, gets an increase in local angle of attack (LAOA) proportional to the shaft tilt. The blade on the other side decreases in LAOA, the same. The two blades fore and aft get tilted up or down but don't change in pitch. If we were to turn the rotor around slowly we would see that the front and back blades as they turn would increase/decrease in angle as they approached the side positions and then return to the nominal settings back at the front, when measured against the original steady state rotor plane.
This is very important, the forces are too high to instantly change the rotor shaft like this, so the change in pitch at the sides requires very high forces to overcome the centrifugal forces in the fore/aft blades, essentially causing them to bend to allow the shaft tilt to take place. In reality this is basically impossible with stiff rotor blades turning at any kind of rpm. Further even if you could force the pitch into the side blades and force the fore/aft blades up and down because the rotor is fairly rigid it would precess at something less than 90 degrees. A gyroscope precesses 90 degrees out of phase only when it is not constrained. If you add stiffness to the gimbal the precession angle gets smaller. For rigid rotors this angle may be as small as 65 degrees. What this would do is when you put in an up elevator command you would get rotor precession 65 degrees later which would give you 2/3 up command and 1/3 right roll command. Helicopters have to deal with this problem too and the solution they use is to turn the upper part of the swashplate to match the non 90 precession angle.
In any event this solution generates way too high of control forces to be of value. The next thing to try is the bottom diagram. The blades are hinged allowing them to pivot in the flap direction (flapping hinges). Now when the shaft is tilted up the left blade can get its increase in pitch and the fore aft blades don't have to be lifted to accomplish this. However, then hinge is usually at some distance from the shaft, called the flapping offset. As you may be able to guess from the diagram, because of the flapping offset as you tilt the shaft, the fore and aft blades are pulled in a tiny amount, and are no longer lined up with each other. They try to pull the hub back flat due to centrifugal force. This is bad because this force tries to flatten the hub out against the control input so the servos have to overcome this force. This is why so called "direct control" heads require heavy duty servos, they are carrying the centrifigal load of the blades during control inputs. Note that after a few revolutions the cyclic pitch applied to the left and right blades as they come around has forced them to precess up and down at the front and back and the whole rotor system is now realigned with the shaft which is now pointed in a different direction. Because the thrust is perpindicular to the tip path plane the thrust now doesn't line up with the aircraft CG anymore and goes in front of it. The coupling of the rotor thrust ahead of the CG now causes a pitching motion of the aircraft. Just what we wanted in the first place.
There is another hidden cost. Because the rotor is now aligned with the shaft the rotor is not applying any side load to the shaft. The pitch is strictly due to the thrust of the rotor and the CG not lining up. The result is that direct control cyclic heads have very little control power, so the roll and pitch rates are not very authoritative. In some respects the direct control head controls more like a hang glider than anything else. You could probably just shift the CG around under a fixed rotor and have the same control power as a direct control head if you could shift the cg an equivalent amount to the rotor tilt.
A very light gyro could potentially be steered by hanging the battery on a gimbal and shoving the battery around. There was a person in europe who built a weight shift autogyro hang glider employing just the method.
Notice that we introduced flapping hinges, this allows coning. Coning requires roll trim that varies with airspeed.
Here is the first great gyrocopter design challenge.
With a direct control head, flapping hinges are required to reduce the servo loads. The softer the flapping hinges and the shorter the flapping offset the less servo load is created. However the softer the flapping hinges and shorter the flapping offset the more coning is allowed and less control power is created. So at a certain flexibility of flapping hinge you run out of control power that is needed to counter the coning induced roll. This is why some gyrocopter designs will not turn to the retreating blade. The control power is already all used up countering the coning induced roll trim!
Now what does delta 3 add to this. I won't go into it now but the result of delta three is that it reduces the overall flapping magnitude. So the flap up in the front of the rotor is reduced, lessening the amount of coning induced roll trim that is needed. However you still don't have very much control power with this design because all the forces against the shaft are in the opposite direction of the way you are trying to steer. So the initial reaction to the control input is to fight the control then the rotor precesses and the CG difference pitch or roll sets in and the control begins to take effect.
This is why the direct control head fell out of favor. Only one kind of direct control head remained, that being the two bladed bensen/wallis type. This head solves some of these problems but creates some nasty other ones, one of which is that it doesn't scale up or down very well from the man carrying size.
The solution to getting more control power without the coning penalty and the high servo loads lies in control mechanisms that fix the blades to a non tilting shaft and adjust the cyclic pitch using other methods....
These other methods involve letting each blade feather by itself on a "feathering hinge" rather than a flapping hinge.
more to come...

[mnowell129]

Two "spider" controlled cyclic pitch mechanisms.
Verbage to go with the pictures later.
mickey

[mnowell129]

Questions that remain that I intend to address...
What are the characteristics of the two bladed teetering head (bensen/wallis) ?
Why does a scale benson head not work right?
Why is the flybar used on your model?
What does collective do and does it help the gyrocopter?
What is following rate and why is it important?
What's the big deal about CG correcting and/or tip weights on
the blades?

If you have questions yourself, please add to the list.
mickey

[mnowell129]

Text that goes with the spider cyclic diagram.
This is the next step in cyclic control beyond tilting the shaft. You still tilt the shaft but you add a hollow hub mounted on bearings around the shaft. You now fix the blades to the hub which is connected rigidly through a big bearing to the fuselage. Now the flapping forces of each blade are carried straight to the fuselage and not carried by the control servos or human. The tilting shaft now tilts in a ball link in the center of the hollow hub. A "spider" turns with the rotor and has linkages attached to blade holders. The blade holders have feathering bearings so the spider can move the pitch of the blades positive and negative. If you look at the left side of the diagram you will see that the shaft tilts and all the blades follow in pitch all around the circle just as if you had tilted the shaft in direct control. Because the linkages are attached at 90 degrees with ball links, only shaft motion perpindicular to a blade causes any change in pitch. Thus if you tilt the shaft and follow a blade around you will see that it receives cyclic pitch just as it would with a direct control head.
This form of control puts the same cyclic pitch inputs in that are present with shaft tilting with two distinct advantages. 1) The flapping forces are not fed into the servos 2) The control power is much higher because the flapping forces are fed directly into the fuselage, so you end up with roll or pitch forces applied directly to the body.
I read someone state on autogyro.com ( I believe Jim Baxter) that with direct control the rotor moves first and then the body. This is in fact true and is indicative of the low control power. The misalignment of the rotor thrust and CG must take place to create motion. With the spider controlled cyclic you get the same rotor thrust tilt but is rigidly attached to the body so the whole unit rolls together, indicative of the much higher control power present.
THe right side of the diagram is an idea of how to build a spider control head without lots of bearings and linkages. The spider is now a thick slotted piece. THe blades now tilt on plain hinges (radially oriented). The little block and CF rod fit into the spider. Now tilting the spider has the same on/off axis effect (or lack of effect) on each blade. This might work if the hinges were good enough to not bind under load and the slot in the spider were smooth enough to also not bind under load.
The complication with this type of control system is the big hollow hub and bearing required. This is a practical, not aerodynamic problem.
The practical solution is to move the spider to the bottom and put IT on the big bearing and let the rotor spin on its normal sized shaft.
What you can probably guess is that this turns into the swashplate and the nominal arrangement for helicopters.
Maybe the next installment will show how a bottom spider works and how this easily turns into a swashplate.
The general result however is this:
1) Non controlled rotors need flapping hinges to correct for forward flight asymmetry.
2) Non controlled rotors need roll trim to correct for coning induced roll.
3) Non controlled rotors can be trimmed in roll for one flight speed.
4) Non controlled rotors can be trimmed with a wing or winglets (the original Cierva's did this)
5) Cyclic pitch MUST be used to control both asymmetric lift and coning induced roll trim.
6) Cyclic pitch can be peformed by shaft tilting, however the control forces are high and the control power is low.
7) Cyclic pitch can be performed with feathering bearings and a cyclic spider. THe control forces are light, the control power is high. The hollow hub is hard to employ.
8) A swashplate provides the same control power and light controls as the cyclic spider with less mechanical difficulty in construction, and results in the same effective cyclic pitch as shaft tilt.

This chain of results is why helicopters have swashplates.
This is why my gyrocopter model uses a cheap commercially available swashplate, etc. (no manufacturing required on my part).

WHat a designer should know is that for a fixed head you are going to have to solve the asymmetric lift problem and the resulting coning induced roll trim problem. This can be tedious and without ailerons, likely only to work at a narrow range of flight speeds.
If you put roll control in with flapping hinges you can expect high servo loads, and low control power but you can still correct the coning induced roll trim problem.
If you do spider cyclic you can get pitch and/or roll control, get by with no flapping and very little coning and need very little roll correction due to coning and will work at all flight speeds.
If you use a swashplate you get all the benefits of the spider cyclic, it is easier to construct and components are commercially available at low cost.

(www.micro-flight.com is guy who makes very tiny helicopters (< 12"), he has very small swashplates, linkages, etc. if you are into the tiny indoor problem, For anything bigger GWS dragonfly, Century hummingbird or bigger helicopter parts are available. www.like90.com has the trex450 parts, including blade holders, swashplates, etc for dirt cheap, these would be appropriate for 1.5 - 2 pound sized models.).

The bottom line is that gyrocopters are controlled by cyclic pitch, whether by flap or by feather, its just which one you chose and what effects you get from your choice.

The next question, "what about that flybar or those tip weights?" is very important. THe answer will explain why your swashplate or spider controlled (or two bladed direct controlled ,etc. ) gyro rolls over on takeoff faster that you can say spit...(and why you can't make a scale bensen! without electronics) And what to do about it.

TTFN
hope somebody is reading this chatter with some benefit...
mickey

[FlyByMike]

<snipped>
TTFN
hope somebody is reading this chatter with some benefit...
mickeymickey,
I am. I thought autogyro'ing was going to be simple, boy was I wrong :eek:
Keep it up, you're article(s) is very enlightening.
-Mike

[mnowell129]

mickey,
I am. I thought autogyro'ing was going to be simple, boy was I wrong :eek:
Keep it up, you're article(s) is very enlightening.
-Mike
Thanks, I will.
The advice to the gyrocopter beginner "build a proven design" is
very good advice indeed....

[David A Ramsey]

Still reading. Are ya gonna get to "ccpm"?

[mnowell129]

Still reading. Are ya gonna get to "ccpm"?
I can and will if you like.

[mnowell129]

When your spider is on its back...
Anyhow. When you take the cyclic spider in the last installment and move it to the bottom of the rotor, this is what you get. The blade holder with feathering bearings is the same. The hub is now small and just big enough to bolt the blade holder to. The shaft is a normal non-hollow shaft and goes down into bearings in the fuse. Or the hub can have bearings but this creates some other issues. For simplicity we assume the shaft goes down into bearings below.
We take a shore piece of hollow shaft, the ball joint and a much less strong large diameter bearing and our spider arms and put it together as shown. We attach a little tiller arm off to the side and now we can tilt the hollow shaft cyclic pitch spider any direction. The spider arm now has the same on/off axis behavior creating cyclic pitch in the blade holder as before.
The advantage is that the bearing only has to carry the control load, not the entire rotor head load. So the large diameter bearing required is a very light duty one instead of one or more heavy duty ones for the whole head. The rotor head load is carried by small insided diameter bearings on the shaft, down below, in the fuse, where it can be strong, not cantilievered way up high (there is a trend here).
Anyway it is now just a minor change to get to a real swashplate on the right. Swing the tiller arms out to the side, make one for pitch, one for roll and now you can connect one each to a servo and get pitch and roll. The upper part of the swashplate turns with the shaft (this is the reason to not put bearings in the head, but put them below on the shaft, otherwise you'd have to put a bearing inside the ball join on the upper part of the swashplate).
The upper part of the swashplate has links to the blade holders and this transfers the cyclic to the blades. The blades get the same cyclic pitch with the swashplate tilt as they would have gotten with shaft tilt. We still have the advantages of the spider mechanism, light controls, high control power, etc. And it is easier to build because we don't need hollow shafts. We just need one big light duty bearing with a ball joint in the middle that connects an upper and lower plate with links for the blades holders and servos respectively. The disgustingly easy part about this is that a complete GWS dragonfly swashplate assembly is a whopping five dollars and 37 cents ($5.37). I don't know what your time is worth but 5 bucks for a swashplate to get easy cyclic pitch, which I MUST have to have to achieve trim and control seems like a no brainer, but that's just my opinion.
Further, the entire rotor head and swashplate parts is just $40. I also bought and entire helicopter wreck with all the parts I need for $35.
Ergo, my two gyros have helicopter heads on them with gyrocopter style blades, no coning, good control power and no trim change with speed.
This is how I got to where I got.
If you go look at closeups (rcuniverse, they take bigger image files you can zoom in on) of the G3PO model and the BeGi pusher you'll see exactly what was developed here ( plus that naggy flybar, but that's coming).

While we're at it. Imagine what happens if you made that swashplate
rise or fall instead of tilt. What would happen would be that the pitch of all the blades would go up or down together at the same time, errr...umm... collectively! The process of changing the pitch of all the blades at the same time is called collective pitch. Heli's do it to increase the up down lift control without having to speed the motor up and down (the motor up/down control only on some heli's being called "fixed pitch" for obvious reasons, even though they still have swashplates and cyclic pitch).
Sometimes collective pitch is done by running a rod through the center of the shaft or along side the ball joint and connecting to little levers on the blade holder. When the rod slide up and down the collective pitch motion occurs in all blades at the same time.
If however we just slide the swashplate up and down we accomplish Cyclic and Collective Pitch Mixed (CCPM) into the same little levers. We can do this by using electronics and three servos (ECCPM) ( they all go up and down for collective, opposite each other for pitch and roll) or mechanically (Mechanical CCPM) by mounting the pitch and roll servos on a rocker arm and using a third servo to move the whole rocker arm. Either way when the swashplate rises and falls for collective and tilts for cyclic this is CCPM. There are many ways of doing collective pitch, but if you have a model in your hand and can make the servos move its pretty easy to see what control is doing what. They become the stuff of dreams for mechanical designers, all those littlle linkages.
Enough for one session.
I guess we have to deal with that nagging flybar issue soon....
mickey

[mnowell129]

Q:Where do the household insect pests go to relax?
B: The flybar of course.

What about that old flybar.
Here's the deal if you take a rotorcraft, roll left until the rotor is tilted left 15 degrees, then apply full right and time how long it takes the rotor go from 15 degrees left to 15 degrees right you get a time called the following rate.

Following rate is the measure of how quickly the rotor follows control input. There is a formula for this time, but since we've been formula free so far we'll skip and and just go to the result. The result is that
the following rate increases with rotor span and rotor speed and decreases with rotor mass and flapping hinge stiffness. All this makes sense to most people except the rotor span. It would seem that a shorter rotor would follow input faster since the centrifigal forces would be lower, those being the ones trying to hold the rotor in place. This is not true because the centrifigual forces get larger proportional to the radius of the rotor but the lift forces get larger proportional to the square of the radius. Thus the lift grows stronger than the radius in effecting the following rate.
The bottom line is that if you have a low mass, loose flapping hinge, high rpm rotor the following rate will be very high.
Here is the magic criteria. For a human to fly it the following rate of the rotor must be between 1/4 and 2 seconds. If the rate is greater than 2 seconds the command is so late that the pilot gets into pilot induced oscillation (PIO) trying to chase the control. If the rate is less than 1/4 second the rate is so high that the pilot can't react fast enough and overcontrols and crashes ( is this starting to sound familiar?).
If you turn the crank on the forumula (email me if you want the reference and forumula, i'm not going to put it here.) for model autogyros at model rotor speeds, with model rotor rates you get following rates like 20 millseconds (1/50! of a second). The issue is that model rotors have to turn fast because they are not as efficient as full sized rotors. At that higher rpm the following rate goes ballistic. So what happens with a control input is that you get an extremely high response, faster than human reaction time. Because most model gyros have a fuse but the fuse tends to damp and slow down the pitch response, but without a wing there is no stopping the roll response and upside down you go, often in the blink of an eye.
Now we reach the great challenge. Following rate and control power are not the same. Following rate is how quickly the rotor follows the control input, control power is how much authority you have when the rotor acts on the body. Here's the $64 answer to many modelers dilemmas.
The direct control head with flapping hinges has low flapping stiffness. At model sized RPM's the following rate is very high. The DC head has low control power. THis is a recipe for disaster because you have a rotor that responds way too quickly to the stick but then has very little control power to do anything about what happened. So here's the scenario, taxi or takeoff. The model is a tiny bit out of trim (they always are) at launch. Because the ground or your hand is holding trim nothing happens. Right at launch the tiny bit out out of trim roll causes a super high model type roll response, way beyond human response time. Typically because the model has flapping hinges that as soon it is launched and comes under its on weight the rotor cones. Because you tossed it forward the coning induced roll kicks in and the model that was perfectly trimmed in your hand now has gobs of coning induced roll. The pilot, now way behind puts in the opposite control but because of the low control power you hit the ground before the opposite control can take effect and then you crash. The general solution has been to limit the roll and pitch input to small angles and try to guess at the amount of trim to counter coning induced roll. This helps once the model is trimmed but actually generates less control power when flying. So you can achieve a delicate stable balance without much control, but you can fly. This is why these things are so hard. You have to hit the nail squarely on the head with trim, you have only milliseconds to respond to the out of trim condition and then the control (and trim) authority is tiny compared to the problem.

Now if we think about a full sized bensen. Because of its size, rpm, rotor mass, etc. The following rate is in the magic 1/4 to 2 second range. If a bensen were any bigger the following rate would get slower and impossible to control by being too sluggish. If it were any smaller (say ... model sized?) the following rate gets so high that a human can't control it. As an aside to a future topic the two bladed teetering head design is used because it has lower control forces (i'll explain that later). But in a teetering head the flapping offset is 0!. Remember back, the larger the flapping offset the more the rotor resists the control, this reduces the following rate. A teetering rotor has ZERO resistance to flapping and thus has the highest possible following rate for a rotor of its size.
Thus the bensen is able to fly with its two bladed direct control head only because it is exactly the right mixture of rotor mass, rpm, flapping stiffness and a human being.
At model size the roll and pitch rates are so high that a small bensen can roll at something like 3 revolutions per second! This is just way too fast for a human to control and thats why they crash in model form.
The original bell heliopters had the same problem. The solution there was to mount a non aerodynamic gyroscope on the rotor to hold it in a stable position. That weighted bar on early bell helicopters like early hueys and bell 47's is actually a big gyroscope, linked to the cyclic that controls the rotor so the following rates are in line with human response. In later bell helicopters the gyroscope became smaller and electromechanical and went into the fuselage. In rotorway and robinson helicopters what looks like a teetering head is not. They have dampers in the head or more than one bearing to deal with this issue.
The outcome is that with two (roll and pitch axis) electronic rate gyros (not an autopilot which is something else all together) like used on the tails of model heli's you could bring the following rate down on a scale bensen to a flyable level. This is probably the only way to make an exact scale model of that design fly in model form.
Note that the following rate of three and four bladed heads is always slower than a teetering head because multibladed heads always have some flapping offset just due to mechanical design. This is probably why you see lots of multiblade gyrocopter designs but virtually no two bladers except those with wings or large vertical pylons providing the necessary roll damping (except mine which uses a flybar gyro stabilizer like a bell helicopter).
Here's the bottom line. For model gyro's you have to slow the following rate of the rotor if you want to make it controllable by a human. You have a couple choices : 1) Bell stabilizer bar (this morphs into the flybar, more later) 2) Multi-bladed head with large flapping offset (high control forces become the issue) 3) Electronic rate gyros 4) Very short blades (tough to get enough lift) 5) Add tip weights to the blades (very good solution)

My thoughts are that if you absolutely insist on doing a direct control head you should make the flapping offset (low servo forces) as small as possible, put tip weights in to keep the following rate down and increase the throw to get some more trim authority and limit the flapping to keep the coning induced roll down.

The solution for my model was a bell/hiller flybar stabilizer with swashplate type cyclic controls. This has a slow following rate and very high control power. It is very easy to fly compared to high following rate and low control power. Because of the slow following rate, no wing or horizontal tail is needed to damp the rotor system, it is already damped by itself.

Note that you can put a two bladed teetering head with a flybar on a direct control tilting shaft and have this work. I actually built one of these to prove to myself it would work but I only tested in my bike tunnel. It will have a slow following rate and be very flyable but still have the low control power intrinsic to a teetering head. I have pictures of it if anyone is interested.
If you are satisfied with low control power like an indoor model a shaft tilt ("DC") cyclic controlled two bladed teetering head with flybar might be an option.
Enough for now.
Questions, comments encouraged.

[mnowell129]

The vicious circle.
What I failed to mention in the last post was that in an effort to tame high following rates, most modelers limit the pitch and roll controls. Because this reduces the amount of pitch and roll control, there isn't enough to correct for asymmetric lift due to forward flight. So they make the blades flap more to allow for flap back. This increases the coning. The coning induced roll now is a higher amount than before and there isn't enough trim. Because the flapping resistance is lower the following rate increases. So the modeler reduces the pitch and roll input some more. This reduces the pitch available to correct for asymmetric lift due to forward flight. So the blades are made to flap more..... well you get the idea.
The original problem is that the following rate is too high. The most likely solution to this problem is to start adding tip weights. The more tip weight, the lower the following rate, allowing more control input. This allows less flapping because you now have more control power to correct asymmetric forward flight lift. With more flapping stiffness you get slower following rate and can thus have more control input... well you get the idea. I think the general solution should be to add tip weights to get the following rates down and start reducing the flapping looseness and increasing the controls, not the other way around.
If you keep adding tip weight you will eventually reach a point where the model becomes very controllable and you have enough pitch and roll control to manuver (sic) effectively.
Tip weights do other good things too, more later.
Me, i'm sticking with the flybar and swashplate.
mickey

[mnowell129]

We're in the home stretch.
Blade balancing.
If you take a rotor blade and hang it by the bolt hole as shown, most blades will hang with the leading edge swept forward. This condition known as "leading" happens because the blade CG is behind the line drawn through the bolt hole down the length of the blade. When spinning the centrifugal force pulls the blade to the leading position. The problem is that the blade is unstable in this position. Like putting your hand out the car window with forward sweep, its hard to hold it in place. Because the blades are always flexible and the linkages always have a little slop, a leading blade will always try to wander off and put in control you didn't ask, sometimes in dramatic and forceful ways. You see as the blade starts to bend up the destabilizing force increases making the blade twist up even more adding more and more unwanted control.
The solution is to correct the CG of the blades so that the CG is along the line in line with the bolt hole and the feathering axis. Then the blade doesn't try to pitch up or down by itself. Note that you can overdo it and get the blade nose heavy. In this case the blade is too stable and will resist the controls, feeding control forces back into the servos. The correct decision is just enough weight to make the blades neutral or slightly lag.
Now back to following rate. When you add weight to the rotor tips to slow down the following rate you should do so along the CG line so as not to upset the lead lag balance. The two weights are for entirely different purposes. One corrects the CG of the blade making it stable. The other slows down the overall following rate of the rotor. The general solution is to weight the blades correctly for the CG. Seal that slot up. Then make another recepticle for the following rate weight, somewhere along the CG line. Then add weight as necessary until the following rate is acceptable (note that it may be a lot). Note that weight towards the tip is more effective and slowing the following rate than at the root.
Happy flying.
I guess we touch on flybars next and this thread is done....

[mnowell129]

The flybar.
Lets start with a two bladed head with a feathering bearing. To make it easy the rotor head is one piece on the feathering bearings. So both blades are free to feather together. You then happily hook up your swashplate and go fly. What you soon find out is that the following rate is so high that you can't control it. Helicopter designers had this same problem. There are several twists and turns to the story but we'll get to the bottom line.
With a flybar you add a small teetering rotor with very small blades at 90 degrees to the rotor head. You now hook up the swashplate cyclic controls to the flybar so that it gets cyclic pitch. So imagine the main rotor was gone and you have a very small rotor that you fly with cyclic pitch. It is as teetering design so in theory the following rate should be very high. Except the rotor is very short (slows following rate). The rotor is tiny, just little paddles (slows following rate more). The result is that the following rate of the flybar itself is very managable. In fact you can add weights in the form of wheel collars to the flybar shaft and get the following rate as slow as you want. You can make the paddles bigger or the flybar longer to increase the following rate ( aside : i made the flybar longer on my gyro copter since the original flybar length was for the RPM of a small heli, at the slower RPM of the gyro the following rate was too stable for me so I made the flybar longer to get the following rate faster to suit me.).
In any event you can tune the flybar to fly at any following rate that you are happy with. Model heli sites list all kinds of different flybar paddles for tuning for 3d flying, etc.
Now you let the teetering motion of the flybar provide cyclic pitch to the main rotor. What happens is that the main rotor is slaved to the flybar and now has a following rate exactly the same as the flybar. So it doesn't matter what the natural following rate of the main rotor would be, the following rate is whatever the flybar rate is.
This is the basic reason that virtually all model helicopters have a flybar and full sized helicopters don't. The small diameter, high RPM models have following rates too high for a human. The flybar slows the following rate down to a human flyable value. Note that there have been a few non flybar models and this is accomplished with large amounts of tip weight and they are still "twitchy" to fly.
You'll have to play with it in your mind a bit but the flybar is 90 degrees off because of precession. To roll right (CW rotation)for example, you put right cyclic in. The flybar gets plus pitch at the back of the model. 90 degrees later the flybar precesses up , now on the left side of the model. THe main rotor is now to back of the model and the precessed up flybar tilts the rotor such that the aft blade gets plus pitch. 90 degrees later the aft blade has precessed up on the left hand side, and there's your right roll.
This system was invented for hiller helicopters and adapted to models. Note that often on model helicopters the swashplate will apply cyclic directly to the blades rather than the flybar. This is called bell type control after bell helicopters. Because the blades get hiller type input from the flybar tilting and direct swashplate input this is called bell-hiller mixing. Virtually all model helicopters from the trex450 and shogun up have bell hiller mixing. Most models smaller than that don't, and most small electrics don't have collective pitch either. These linkages are fairly complex but provide great performance. The fixed pitch, hiller only head from the hummingbird is good enough to provide good control for an autogyro without the extra complexity of collective and bell-hiller mixing. Anyhow that's a digression.
The bottom line is that the flybar makes a two bladed head flyable in model size. The only other way this has been done with a model is very large tip weights (or electronic rate gyros on pitch and roll). This has an obvious disadvantage to the autogyro where spin up is always an issue. With a heli running constant rotor speed the mass is not important but with the autogyro it is.
So there's the whole story. To make a two bladed autogyro with no wing and no rotor pylon, etc. you pretty much have three choices 1) Large tip weights 2) Electronic rate gyros on pitch and roll or 3) The flybar.
Since the flybar/swashplate/head parts were cheap and I already knew how they worked I went that direction.
Well that's the end of my story. I hope someone will come back with a question or a comment because I'd like to be having a conversation and not giving a speech.....
mickey

[JochenK]

This is excellent, Mickey. At last somebody is telling me all the things I should have known before I got interested in autogyros.
In your second posting you mentioned the delta three hinge and said, you'd deal with this later. Can you keep up the educational process and elaborate on this?
Thanks a lot, Jochen

[mnowell129]

This is excellent, Mickey. At last somebody is telling me all the things I should have known before I got interested in autogyros.
In your second posting you mentioned the delta three hinge and said, you'd deal with this later. Can you keep up the educational process and elaborate on this?
Thanks a lot, Jochen
Yes, thanks for reminding me.
Here's a diagram of typical direct control head. Top view, CW rotation. When the blades flap on the hinge they stay at the same relative angle. If we introduce a hinge angle (thats a greek delta, all the rotor angles have fairly standard engineering terms starting with alpha, beta, etc. delta just happened to be the one assigned here, possibly because it creates a little "delta" shape in the hinge area, i really don't know for sure) then the blades change angle as they flap. Notice in the end view that when the blade flaps up it also reduces in pitch and when it flaps down it increases in pitch. Note that you could make positive delta three and the blade would pitch up when it flaps up, but this is de-stabilizing, not stabilizing.
Now how this affects the rotor is subtle. The side view of the rotor shows the normal non flapping case as reference. Note that the blade will now flap, but as it is flapping up the pitch is being reduced. Thus the blade flap is now introducing mechanical feathering, ala cyclic pitch. There is a subtlety however. As the blade passes through the left hand position it has (due to coning induced flap) less pitch than the nonimal case. This reduces the flap up in the front, but does not completely eliminate it. This is not the major effect with delta 3. The flap in front is. Remember the flap in the front and the coning combine to produce a coning induced roll. The delta three really helps with this because as the blade is flapped at its highest in the front it also is getting the maximum negative pitch due to the delta three hinging. Thus this blade now sees much less angle of attack than the plain hinged case. As the blade is out front and getting less lift than the plain hinge case, when the blade is 90 degrees around on the right side it precess up less. Thus the delta three induced feathering automatically reduces the coning/flapping induced roll trim and the subsequent control input to correct for it. As an added benefit when the blades flap due to gusts the delta three action automatically applies the right corrective action to the blade.
Note that delta 3 doesn't reduce the control forces or increase the control power, it simply reduces the amount of flapping that takes place, thereby reducing the coning/flapping induced trim problems. It may appear to give you more control since the amount of corrective trim is less providing apparently more control in the trim direction.

I guess while we are at it we might as well mention that coning is not the same as dihedral and is not a stabilizing factor. Remember that the flap up in the front produces a roll, not a pitch. When the gyro copter sideslips or gets a cross wind gust, if the rotor is coning the blade that is in the direction of the slip or the gust gets some positive pitch. Instead of lifting that side of the model like dihedral does, this extra blade lift takes action 90 degrees later. If the gust was on the advancing blade then the model reaction is nose up. If the gust was on the retreating blade then the model reaction is nose down. Neither one of these responses serves to lift the low side of the model, in fact it requires corrective action on the part of the pilot to lift or lower the nose. This is part of the reason that all rotor craft are inherently less stable than fixed wing.
Note however that the flap up of the rotor with forward flight asymmetric lift DOES provide dihedral like stability. As the model slips the lift assymmetry shifts around to the side with the gust or slip. If to the right for example, the blade in front is now the advancing blade, sees more airflow, flaps up 90 degrees later, and applies the proper left roll corrective action. The result is that a flat rotor with no coning is more roll stable than a rotor with coning.
The real truth is that coning has no positive value. It is a by product of the blades lifting and the body pulling down in the middle. A topic we haven't touched on is coriolis. This is what makes objects spin faster when the mass is more concentrated, like an ice skater pulling in their arms and legs to speed up a spin. When a blade cones/flaps up its effective radius gets smaller (think about the fact that if it flapped 90 degrees the radius would get tiny). As the radius gets smaller the angular momentum has to be conserved and so when the blade flaps up that blade has to speed up to maintain energy. What this means is that the blades speed up and slow down a little as they go around the circle flapping up and down. In real helicopters this is a serious issue and thus real helicopters have lead/lag hinges to allow for this speedup/slowdown around the circle. Cierva was stubborn about lead/lag hinges initially and only after destroying rotor hubs due to coriolis induced lead/lag fatigue failures did he reluctantly add lead/lag hinges.
In models this is why the blade attachment bolt is always a tiny bit loose (This is why I use a plastic stop nut and a single blade bolt. The stop nut lets you adjust just enough tension to hold the blades in place for spin up, but not so much that a little lead/lag can't take place. There was a method behind my apparent madness.) so as to take up the lead/lag motion. If you rigidly fix your blades with two bolts and they are allowed to flap you will always have a vibration present due to coriolis induced lead/lag not being able to be absorbed by your blade attachments.
This adds to the problems with coning.
Just to reiterate, coning is not stabilizing, its just the opposite. Coning is just a problem that has to be dealt with.
mickey

[mnowell129]

I hope after this "tome" I have been able to explain why I made the design choices I did.
Swashplate cyclic with flybar head with no coning meant :
1) high control power
2) small servos could be used
3) better roll stability with less coning
4) No flapping means predictable tail clearance
5) Stable rotor (flybar) means no need for wings or tails
6) Good gusty wind performance

This let me build the stick with a rotor and rudder on it.
The basic rotor system on my model is very stable and
reproducible. It turns out that the pusher model BeGi was
almost falling down easy to do after I had the geometry worked
out on G3PO. Once I got the pitch trim right for the thrust line
I guessed at on the pusher motor, it basically flew right out of my
hand. I've only crashed it once because I flew in a 30 mph wind and
tried to dive it to get back to the field and stopped the rotor.

I've shared my experience because I can tell those of you who are struggling just don't have enough information to know what to do to fix your own design and there is very little written describing the theoretical problems with model gyrocopters.
I don't expect everyone to rush out and build swashplate controlled flybar head gyrocopters because that doesn't appeal to everyone. Everyone has their own interests and there are those who like the purity of the Cierva and Bensen designs. I've just tried to point out what makes this difficult to do in model form. I think a scale bensen is possible but I believe it will be done with lots of tip weights and rate gyros on the roll and pitch axis. With the gyros hidden in the body this would be very scale. However the probability of success of a scale bensen with pure aerodynamic control is very low. The three bladed direct control head will always have poor control power and high servo loads, but with proper blade weighting can be made quite flyable when the following rate is brought in line. The swashplate/flybar model will always be more mechanically complex but will have better control and lighter control forces. But life is always a compromise right?
As for me the flybar doesn't bother me because it is on my helicopters and I fully understand why its there. Two blades take less time to make as well. And since DC,spider and swashplate are all cyclic pitch, I chose the easiest one to build (I just bought it!). My goal was to get an operational aircraft, not a scale model. I achieved that goal. I hope, that armed with a little theoretical understanding of the model gyro, that others can make improvements to their models. Hopefully you can see now that following rate is serious problem that must be dealt with somehow to make a flyable model. Its up to you how you go about getting the following rate under control for your model.
more questions welcome
mickey

[mnowell129]

I guess one more topic that needs explaining is the pivot offset in a direct control head.....

[David A Ramsey]

Great discription on rotor blade balancing; Took me back 18 years ago.

Your discriptions have shead much light on aspects of the autogyro rotor that I have mostly only observed; Like having pictures without captions.

"PIVOT OFFSET?"

[mnowell129]

Great discription on rotor blade balancing; Took me back 18 years ago.

Your discriptions have shead much light on aspects of the autogyro rotor that I have mostly only observed; Like having pictures without captions.

"PIVOT OFFSET?"
Thanks.
Yes, if you look at the DC head designs, the rotor shaft is behind the elevator pivot point by some offset. This is also true of the bensen style heads.
When I can get back here I'll explain why that is.
Credit goes to Emilio Cabezes for bringing the offset pivot to models with his ECDC model.
mickey

[mnowell129]

Pivot offset is a design solution where the rotor axle of a direct control head is behind the pitch (elevator) pivot point. Here's why. In the first diagram is a rotor flying along. The blades are trying to flap but because the flapping hinge is not perfectly free ( it never can be truly free) the blade flaps a little (exaggerated in the drawing to see it). Now the rotor thrust is forward of the axle and causing a nose up pitching moment. A little down cyclic will fix this problem except...
In diagram two we've applied the down cyclic and the rotor is now realigned with the down cyclic and a new more nose down rotor tip plane. Unfortunately now the thrust that is perpindicular to the tip path plane is not lined up with the elevator pitch point (1) anymore. There is now significant thrust ahead of the pitch point trying to pull the rotor back against the elevator servo. So the solution is to move the pivot point forward so that in the trim condition the rotor thrust goes through the pivot point and does not generate any force back into the elevator servo.
Note that this isn't perfect and only is exactly right for one speed and trim condition but it goes a long way towards relieving the load on the elevator servo. Note that the roll function doesn't suffer this problem because the asymmetric lift problem doesn't happen on that axis. But a similar problem occurs with coning induced roll trim, but clearly to a lesser degree.
Note that you would not have this problem if the elevator pivot point were in line with the line drawn through the rotor tips because there would never be a misalignment. The conclusion is that the longer the DC rotor axle the worse this problem is, so as designer the elevator pivot point needs to be a close as possible to the center of the head. To make this point it is easy to visualize that a very long shaft would produce very large elevator servo forces.
This condition further aggrevates the high control forces just due to the flapping offset fed back to the elevator servo, reinforcing the need for a very beefy elevator servo.
I observe that there are more than a few gyro models with just roll control on the head and elevator out back. This is a compromise that works mostly. I say mostly because cyclic pitch control on elevator works at very low airspeed ( essentially 0 ) but an elevator looses effectiveness (the effectiveness goes to 0 at 0 airspeed) as you slow down. So the elevator and cyclic rolled control gyro will lose elevator control at some point when slowing down, the cyclic one will not. If you are willing to always fly above the elevator effectiveness airspeed then this might be reasonable compromise. In my models I often do motor off dead stick landings for kicks and the cyclic control remains effective all the way to touchdown in a near vertical descent. The elevator controlled model will require some forward airspeed in this situation otherwise you reach a point where the elevator quits working and you have no way to adjust the pitch of the model and therefore you never get control back without restarting the engine. This makes an engine out situation in an elevator controlled model somewhat risky in that if you get it too slow, you lose the elevator and never get it back.
all for now.
mickey

[mnowell129]

Evidence that the bensen will fly if it is the right size.
A 10 foot, 10 hp version that is said to fly.
At this size the following rate would likely be manageable.

http://www.rcuniverse.com/forum/fb.asp?m=2588410

[Rus]

hope somebody is reading this chatter with some benefit...
mickey

I'm reading - learning - enjoying!!

[Don Sims]

Excellent thread!!

[mnowell129]

I'm reading - learning - enjoying!!

Thanks. If you have questions please post them.
I do well with a little prompt...

[mnowell129]

Excellent thread!!
Thanks.
Post questions if you have em.

[TMorita]

Still reading. Are ya gonna get to "ccpm"?

Sorry to interrupt this fine tutorial, but this thread may help:

http://www.rcgroups.com/forums/showthread.php?t=323392

Toshi

[TMorita]

The flybar.
...
This is the basic reason that virtually all model helicopters have a flybar and full sized helicopters don't. The small diameter, high RPM models have following rates too high for a human. The flybar slows the following rate down to a human flyable value. Note that there have been a few non flybar models and this is accomplished with large amounts of tip weight and they are still "twitchy" to fly.


This correlates with the stuff I've read from people who have done flybarless conversions. The twitchiness is manageable by dialing down the swashplate mixing of the aileron and elevator, though.

People say a flybarless head is really good for doing stationary flips and rolls, but not very good for forward flight. Most people who have experimented with flybarless heads on helicopters have gone back to flybared heads.

Toshi

[mnowell129]

This correlates with the stuff I've read from people who have done flybarless conversions. The twitchiness is manageable by dialing down the swashplate mixing of the aileron and elevator, though.

People say a flybarless head is really good for doing stationary flips and rolls, but not very good for forward flight. Most people who have experimented with flybarless heads on helicopters have gone back to flybared heads.

Toshi
I can unfortunately admit to being old enough to have seen people flying Hubert Bitner's Horizon with no flybar. It had what seems to be a hockey puck sized piece of lead in the tips and was still twitchy to fly. My observation is the same as yours, ever year or so someone will try a flybarless conversion and making the same realization about touchiness and will go back to the flybar.
Pertaining to gyrocopters I read many accounts that included a phrase like "rolled over immediately upon takeoff" and I started recognizing the symptoms and figured the solution had to be a flybar head.
It's a nice dream to make that simple looking flybarless two bladed head but the darn things are just too much work to fly.
thanks for the other link.
mickey

[TMorita]

BTW, how much does the G3PO weigh?

I'm rather obsessive about watts per kg to fly/hover, and was wondering what was typical for autogyros to fly.

Toshi

[TMorita]

...
My observation is the same as yours, ever year or so someone will try a flybarless conversion and making the same realization about touchiness and will go back to the flybar.
...
mickey

This year we have started early on the flybarless conversion idea.
Here's a thread which started up YESTERDAY about making a flybarless head... :)

http://www.rcgroups.com/forums/showthread.php?t=328933

Toshi

[mnowell129]

BTW, how much does the G3PO weigh?

I'm rather obsessive about watts per kg to fly/hover, and was wondering what was typical for autogyros to fly.

Toshi
My thumbnail is about 100 watts/pound for good performance.

[mnowell129]

Not to pick on the guy flying, but this video is very good and showing a gyro that has a high following rate and poor control power. Watch the video on the top right. You can see that in rolls it very quickly from the gust but then appears to not have enough control power to correct in either pitch or roll....

http://www.auav.net/autogyro/videos.html

A graphic illustration of what happens when you reduce control throws in an attempt to solve the following rate problem.

[David A Ramsey]

Enjoyed all the videos. On the top right video, seems he could have used a healthy input of left rudder. If he could have kept the bank angle flatter he might have been ok.

David

[mnowell129]

Enjoyed all the videos. On the top right video, seems he could have used a healthy input of left rudder. If he could have kept the bank angle flatter he might have been ok.

David

Yea, it needed something more. Just kinda reinforces that not having enough control is not the right direction.

[flicka5]

Not to pick on the guy flying, but this video is very good and showing a gyro that has a high following rate and poor control power. Watch the video on the top right. You can see that in rolls it very quickly from the gust but then appears to not have enough control power to correct in either pitch or roll....

http://www.auav.net/autogyro/videos.html

A graphic illustration of what happens when you reduce control throws in an attempt to solve the following rate problem.

The garbage code as seen in my title #$%&@(**&^(& is what I get on my screen when I try to download any of the videos from your URL above. What software am I missing to play these videos? Also, congrats on the superb gyro designing as you kind of put those old designers at Spring Hill to task including my friend R. Ogren. Yep, it is too bad that deBolt could not have seen your miracle fly!

[mnowell129]

The garbage code as seen in my title #$%&@(**&^(& is what I get on my screen when I try to download any of the videos from your URL above. What software am I missing to play these videos? Also, congrats on the superb gyro designing as you kind of put those old designers at Spring Hill to task including my friend R. Ogren. Yep, it is too bad that deBolt could not have seen your miracle fly!

Thanks for the vote of confidence. Miracle is a strong word, but I thank you.

I think most gyro designers up to
this point have been working real hard to make a direct control head
gyro and continue to insist that only a direct control head gyro is
a real gyro. My definition of gyrocopter/autogyro/autogiro or gyroplane (FAA definition) is a rotor craft pulled along by a motor with an unpowered rotor for lift. ( I might even show with a ducted fan version one day..) Everything else is just a design choice detail. The direct control head rotorcraft went of fashion the instant the swashplate was invented because the direct control head has lots of problems that will never be overcome. The only surviving direct control head rotorcraft are one/two person bensen/wallis type aircraft that as I've explained before on this forum only work due to the unique intersection of their size and aerodynamics and the use of a two bladed head which is unstable in model form. A bigger one doesn't work nor does a smaller one. It is abundantly clear to me both practically and theoretically that the direct control head gyros will never perform as well as one similar to mine with a swashplate, etc. You can make a nice direct control head model fly well (i.e. PT Gyro, etc.) and it will be a good sport/beginner model, but it will be extremely difficult to ever make it perform like G3PO. My goal was a functional, operational aircraft, not a scale replica. I feel fairly satisfied that I accomplished that goal. I'm sure that there are some grumblers out there that insist I've cheated somehow or commited some heresy, but frankly I don't care, I have a great flying gyrocopter.

I try to not get on my soapbox too much and stick to the engineering facts but I felt like I had to state this once for the record.

That link is from Phil Ploof's website and came from this post on
RCU, happily I'll refer you to him...
http://www.rcuniverse.com/forum/fb.asp?m=2707766

mickey

[grant31781]

Mickey,

You have really cleared up many of the questions/ headaches i had in my gyro design in 30 mins of reading! thanks! I built a spad gyro with coroplast blades i know not the most effiecent blades but it did takeoff every time and then crash rolling to the advancing blade. I made a flapping head with hinges with no up stops and no feathering so i guess the coning induced roll casued it. It also would not fly with out a tilting head to overcome the right roll.

I see your not a real fan of the DC head gyro. I just ordered a PT gyro and this is a dc head with a delta 3 hinging method i think. Do you think this is a good 1st gyro?

I want to build another one with the swashplate flybar setup like yours except in a 25 to 60 size version. When i was experimenting with my gyro a few months ago i though hard about the swasplate idea. i though yes a swashplate would fix these problems beacuse it works fine on a heli. I just didnt have a clue what to by for it.

A few questions about the swashplate setup. Since there is no flaping involved, How are the rotor blades and the flybar setup? Is there a tad of negitive trim in the advancing blade and positve in the retreating blade?
I would guess it would have to have this to over come the lift dissymetry.

Does the flybar attach to the rotors to chage their pitch or is it strictly the flybar that is controlled?

Sorry but i still dont quite understand how the swashplate works.

Thanks,
Grant

[mnowell129]

Thanks for the post.
My DSL line is down and your answer may take a while.
I'll get back to you in a few days when my connection comes back.
Meanwhile I posted to another thread about the swashplate,
that may help for now.
mickey

[mnowell129]

Mickey,

I see your not a real fan of the DC head gyro. I just ordered a PT gyro and this is a dc head with a delta 3 hinging method i think. Do you think this is a good 1st gyro?


I'm not a fan of the DC head because of the stresses it puts on the servos. I don't like having to have the flapping hinges either as this creates all the coning related issues. It seems quite difficult to balance all the interactions to get one to fly properly.
However. I've flown a pt gyro and it seemed to fly just fine. I think what the PT gyro offers is all the details worked out so that you have a good chance of success if you follow their instructions.
I think the absolute beginner needs to start with a two rotor model gyrace/tango/spin doctor/etc. just to get a feel for the rotor speeding up and slowing down behavior and the sluggishness compared to a plane. My flying buddy, a pretty good pilot, built a spin doctor and had a quite a few
encounters with the ground before getting the "feel" of flying a gyrocopter.
I think the PT is great next step or a good place to start for the advanced pilot or a pilot who has helicopter experience.
And just to be fair to myself there is a guy in Arizona flying a G3PO with no rotary wing experience at all.
One aspect of rotary winged aircraft is that it demands a much more attentive pilot. Not everyone is prepared to expend that much energy flying.
mickey

[grant31781]

With a direct control head, flapping hinges are required to reduce the servo loads. The softer the flapping hinges and the shorter the flapping offset the less servo load is created. However the softer the flapping hinges and shorter the flapping offset the more coning is allowed and less control power is created.

Mickey could you explain what flapping offsets are?

[mnowell129]

Mickey could you explain what flapping offsets are?
The distance from the flapping hinge line to the center of the hub.
For the typical three bladed DC head its from the center of the hub to
the point where the blade actually pivots.

[grant31781]

Pivot offset is a design solution where the rotor axle of a direct control head is behind the pitch (elevator) pivot point.

Does it matter if the elevator servo is on the front or back side of the mast?

[mnowell129]

Does it matter if the elevator servo is on the front or back side of the mast?
I'm just answering based on thought not experience. I've seen both done.
It makes more sense to me to put the servo in back, this way it puts the
pushrod in tension rather than compression. Seems less likely to buckle
the pushrod that way.
Maybe someone with DC head experience will comment as well.
mickey

[bobandris]

Mickey,

Since this thread did such an excellent job of explaining the importance of following rate, I have a related question. Does your approach also apply to the dual-rotor (side-by-side) gyrocopters? We have several flying at our club field, and once you get the rotors setup they are a real joy to fly, using throttle, rudder, and elevator. We have some very small and some of the Slow-Stik size, and they don't seem to have a scaling problem.

Bob

[mnowell129]

Mickey,

Since this thread did such an excellent job of explaining the importance of following rate, I have a related question. Does your approach also apply to the dual-rotor (side-by-side) gyrocopters? We have several flying at our club field, and once you get the rotors setup they are a real joy to fly, using throttle, rudder, and elevator. We have some very small and some of the Slow-Stik size, and they don't seem to have a scaling problem.

Bob
The side by side models are pretty simple, probably why they are so enjoyable. The rotors don't really pitch or roll because they are rigidly mounted. When the rotor tries to roll it is effectively trying to twist the whole model up or down from the wingtip. You can get a mental idea by imagining trying to roll a model by holding at the fuse versus the wingtip. Even though the rigid rotor would have a very high following rate by itself the intertia of trying to roll the whole model from the wingtip is the dominating factor. In other words the wingtip mounting provides very very high damping. So the following rate of the rotor itself is rather irrelevant. Incidently you gain some roll damping by mounting a single rotor up high. You get pitch damping by having a fuse with a horizontal stab way out back. I think this is why you see so many DC models with very high mounted rotors and horizontal tails. By way of contrast, the G3PO rotor is damped by itself, hence the mast height is only dictated by not hitting the tailboom and there is no horizontal stab.

The twins fly by having the two rotors canted inward. To turn the rudder is used to yaw the model. Since the rotor is canted inward, as it is yawed forward it sees an increase in angle of attack. This causes two things, first there is an instant lift change due to AOA, then the rotor speeds up. This provides a two stage kind of roll, one rather quick and then one that is controlled by the rotor inertia. My observation is that these models tend to "wallow" and overshoot the turns, probably due to this two part process. The control is rather sluggish for the same reason, there is no direct control over the rotor.
In pitch the rotors have a lot more control over the model. They are simply trying to twist the fuse up or down, here again the long fuse and horizontal stab provide the necessary damping.
Adding weight to the twin kind of rotor to slow down the following rate is likely a bad move, since this kind of model depends on the rotor speeding up
and slowing down to aid in the roll, the lighter the rotor mass the better.

Scaling doesn't affect them much because it doesn't matter what the following rate of the rotor is as long as the rotor mass is small compared to the total airframe weight. In essence the very small mass rotor may have an extremely high following rate and may precess instantly and apply a roll force to the model, but because the rotor is low mass, the amount of force applied is tiny compared to the relative intertia of the whole model. It doesn't matter how quickly a fly can run into a bowling ball, the bowling ball just won't move quickly in this circumstance.


Anyway. The summary answer is that the following rate would be important if you were trying to fly the rotors directly, but since you are controlling with indirect methods and the whole model is very highly damped, it doesn't come into play.
good question, thought provoking
mick

[max z]

Evidence that the bensen will fly if it is the right size.
A 10 foot, 10 hp version that is said to fly.
At this size the following rate would likely be manageable.

And this is probably it. Seen at a recent big (really big!) scale meeting near Paris.
Mickey,
I have been aware of this thread for some time, but only now have I taken the time to properly read your installments and try to digest them. I can understand most of it, but I have difficulties coming to grips with the term Precession. I know what gyroscopic precession is, a resisting force at an angle (90 degrees?) relative to the force you excert on a rotating mass. You also use it for lift forces coming into effect at a certain rotation angle after changing the blade pitch. I don't understand that, if a blade sees the air coming at an increased angle of attack, the increase in lift is immediate, and any assymmetric effects with it. What am I missing?

Regards,
Max.
(No, I did not see Popeye fly :( )

[mnowell129]

[QUOTE=max z] I don't understand that, if a blade sees the air coming at an increased angle of attack, the increase in lift is immediate, and any assymmetric effects with it. What am I missing?
QUOTE]
This point is subtle and probably the most challenging to understand.
I'm going to go sideways a second so bear with me.
Imagine a golf ball rolling in a large tube. Suppose you have an air blaster attached to a hole that as the ball rolls by, you give it a puff of air. The side force you apply to the ball is immediate. The ball continues to roll but because you gave it a side force it now travels up the side of the tube as it continues to roll down the tube. While it is traveling up the side of the tube it has no side force having rolled past your air blaster hole.
Another example. You are driving your car down the highway. You jerk the steering wheel to one side to avoid a pothole. The force of the tires is immediate and large. But it takes a great distance for the vehicle to actually move to the other lane. By the time you are in the other lane, the side force on the tires is back to nil.
Both of these examples illustrate a basic physics principle. The movement (distance moved, or displacement) from a disturbing force always comes much later (lags) than the force itself. This is because of Newtons laws of motion. When you apply a force the mass has to accelerate and cannot do so instantly.
Now to a rotor. Yes the blade sees an instant angle change and sees an instant lift change. The lag for a rotor blade turns out to be 90 degrees of rotor rotation. So the blade that sees a lift change (disturbing force) doesn't flap up till later, just like the golf ball, just like the car. The displacement comes later. In the case of rotors 90 degrees later. So 90 degrees later in the rotation the blade flaps up (or down if the lift change was negative).
Now, the next bit is not commonly understood. A rotor blade with a flapping hinge (whether a real hinge, an flexy member, or whatever) lifts the copter with centrifugal force, not blade lift. Think about it. Put you hand on a blade on your gyro when its not turning, lift. Does the gyro lift. No. The blade just tilts up. What happens is that the blades fly in a circle creating centrifugal force and the body is hung between these spinning objects. The coning angle that results is the angle through which the centrifugal force holds the body up.
So... the instant blade lift on any one blade does nothing to roll or pitch the model. What rolls or pitches the model is to make the blade move to a higher or lower angle changing the angle through which the centrifugal force is acting, causing that side of the model to be pulled up or down.
So what cyclic pitch is doing, either DC controlled or swashplate feathering controlled, is applying an instant lift change to the blade, knowing that 90 degrees later this blade will flap up or down, tilting the tip path plane of the whole rotor and changing the direction that the centrifugal force is pulling on the main shaft, causing the desired pitch or roll result.
Here's where following rate comes in.
If you replace the golf ball with a table tennis ball and apply the same force or drive a small sports car instead of a minivan, the same force being applied results in much more displacement. Newton again, same force, less mass = more movement.
In rotors if you make the rotor lighter it responds with more amplitude also. More amplitude response means that the rotor tips more, applies more force to the model and the model rolls or pitches faster. This is the following rate, how fast the model responds to control input. If this rate is too high a human can't control it. Solutions: make the blades heavier so their amplitude response is smaller, slow down the rotor system so everything happens slower, use 3 or 4 blades so centrifugal force of the blades limits the amount of cyclic that can be applied, use a rate gyro to artificially control the amplitude response (this is the flybar), use an electronic stability augmentation systems (SAS on bell helicopters?), use aerodynamic damping (wings, elevators), physical damping (tall masts, long fuselages).
Unfortunately small models have high speed rotors due to poor blade performance, and at the speeds they turn its hard to get them heavy enough to have a human flyable following rate. This is why a small two bladed gyro copter is fairly close to physically impossible without some kind of stabilizer system.
The sticky part of blade weights on small models is that most of them have three blades because of the above reasons. Weight would help the following rate, however in three blade DC type heads the blade forces are transferred directly into the servos. This tends to overload the servos. This is how you end up with the typical small model configuration of 3 bladed head with unweighted blades to keep the servo loads low, tall mast for roll damping, and fuse and horizontal stab for pitch damping.
This is why a new gyrocopter design is fairly difficult, much more so than a new airplane design. Having any one of the competing factors out of balance creates a very difficult to fly model.
howzat?
mickey

[max z]

So... the instant blade lift on any one blade does nothing to roll or pitch the model. What rolls or pitches the model is to make the blade move to a higher or lower angle changing the angle through which the centrifugal force is acting, causing that side of the model to be pulled up or down.
howzat?
mickey
Thanks, that was the bit I was missing! And I don't think you mentioned it before, or was it just me. Whatever, thanks for your elaborate explanation.

Max.

[max z]

Mickey,
Is there any way to get a sketch / drawing of your G3PO design or the BEGi? Would any of those be a good first for a raw beginner (in rotorcraft, I have flown fixed wing RC for 30 years)? I looked at your previous threads and your website, but all I found is the G3PO manual (which I downloaded) and the BoM.

Max.

[flicka5]

Mickey,
Is there any way to get a sketch / drawing of your G3PO design or the BEGi? Would any of those be a good first for a raw beginner (in rotorcraft, I have flown fixed wing RC for 30 years)? I looked at your previous threads and your website, but all I found is the G3PO manual (which I downloaded) and the BoM.

Max.

Simple as you need to go to the Brand "X" site!!!!
G3PO plans are there!http://www.rcuniverse.com/forum/m_2380064/tm.htm

[mnowell129]

Simple as you need to go to the Brand "X" site!!!!
G3PO plans are there!http://www.rcuniverse.com/forum/m_2380064/tm.htm
thanks for the help on the link.
I have kits also, everything but electronics.
mickey