Part one of this project covered the building and testing of a two rotor aircraft
that used 3D printed helicopter rotor heads for control.
These rotor heads were simple by design with each blade being freely hinged at the hub
and with some precise motor movements they were able to provide sufficient control
to keep the aircraft in the air.
However it didn't fly first time due to some vibrations with the frame
but this was fixed with a stronger wing spar and redesigned motor mounts.
I then bought these slightly larger blades which are a bit longer in length
and they're also completely unbranded in the hope that these will provide more thrust
because once we add the fuselage to this thing it's going to weigh quite a bit more.
But this brought back the vibrations and I think the cause of these vibrations
were more serious than I first realised.
From this slow-mo it seems the small vibrations caused by the blades pivoting
causes the frame to flex slightly which is then fed back into the rotor head
causing crazy unwanted blade pitch movements.
These extreme blade angles then cause unwanted load on the motor
which then amplifies the vibration to the point where the whole rotor head rapidly loses lift.
Bear in mind these vibrations are bending this 20mm carbon fibre tube
like it's an uncooked spaghetti and these tubes are very strong.
I could go back to the old blades but if such a small change in the blades
causes such a drastic change in performance
then I hate to think what would happen once I get the fuselage on this thing.
For this reason I started designing an actual helicopter rotor head
with a swashplate and linkages.
But whilst developing this I received an email from the original designer
of this freely hinged rotor head system.
If you haven't seen my older video on the swashplateless helicopter
the original idea was invented by a guy called Jimmy
as part of a research project at the University of Pennsylvania
and his motor controller was developed by his friend Matt.
Matt has since developed this motor controller further as part of his business Vertic
and he's actually designed his own freely hinged rotor head system
that uses a single central hinge instead of one hinge per blade.
The clear advantage to this is obviously it's simpler having less hinges in the rotor head
but more importantly to my specific use case is that all of the blades are fixed together.
Let me show you why this is important.
Imagine this is the rotor head and all three blades are evenly spaced apart by 120 degrees.
The center of mass should be in the center of the rotor head
meaning it's perfectly balanced
and when the motor wants to create a control output
it either accelerates or decelerates to cause the blades to lead or lag.
In an ideal world the rotor head should stay balanced
as all the blades will swing by equal amounts
but if there is a slight variation in hinge friction
or one blade experiences more drag than the other
they might lead or lag by different amounts
which would shift the center of mass away from the axis of rotation
causing an imbalance and therefore a vibration.
So having all these blades fixed together will hopefully remove the vibrations I'm getting
and Matt also told me that he successfully tested a three-bladed version of this rotor head
with a single central hinge.
So with this information I have to build my own.
This rotor head has a central hinge at 45 degree angle
which means when the motor accelerates or decelerates
the blade in line with the hinge will change in angle of attack
but the other two blades will actually move up or down.
This looks a little odd when rotating as one blade maintains the same plane of rotation
but the other two seem to follow a different plane of rotation.
Once on the aircraft it was clear that the previous vibrations were gone
which was very promising.
However it doesn't seem to have much control
so the aircraft could only hover for a short period before pitching over and crashing.
But then Matt showed me another rotor head he's been working on
which has the same central 45 degree hinge as well as another horizontal hinge
which means the rotor head can tilt in any direction almost like a universal joint
but because of the 45 degree hinge it can still be controlled by the motor speed
to vector the thrust of the blades.
I found this fascinating that adding a couple of hinges to the rotor head
can cause this thrust vectoring effect by simply adjusting the motor speed
so of course I have to build my own.
This is very similar to the previous design
but features the extra horizontal hinge through the center
and what's interesting is this pivot direction works fine due to the horizontal hinge
but because of the 45 degree hinge to get it to pivot left and right
the blades have to twist side to side
and it's this exact movement in reverse that will be used to control the thrust vectoring.
When mounted on the motor it only requires a small amount of motor pulsing
to get the rotor to pivot which is really cool to watch.
Also because of the reduced vibrations of this rotor head
I removed the bulky motor mounts and went back to the 14mm wingspan to save weight.
But how does it perform in the air?
Not great.
After lots of tuning and tweaking of the gyro stabilization it wasn't getting much better
but then I had a thought, what if I turn the gyro stabilization off?
My theory was that the gyroscopic stability of the rotor
would tend to stay in position as the aircraft would rotate
so is this fighting the electronic gyro stabilization?
The fact this thing flies with no stabilization on the pitch axis is really interesting
and it's something I never considered would be possible.
The only issue is it relies completely on the position of the rotor blades to keep it stable
so if a gust of wind causes the blades to adjust their angle the whole aircraft will follow.
Ultimately I have to make a decision between
are we building a strange two rotor test vehicle
or are we trying to build a V22 Osprey scale model?
And from the title of this video I think you know the answer.
I really wanted to hold on to this freely hinged rotor head system
but in reality it's not feasible for this application.
The final rotor head was the closest to success
but because it's essentially a thrust vectoring design
this can be achieved with a servo-actuated motor pivot
and a servo-actuated motor pivot has the advantage of
being able to be locked in position during forward flight.
I'm sure I could spend a few more weeks or months trying to tune this rotor head
to perform as well as Matt from Vertica has done with his
but I also feel like that's following the sunk cost fallacy
so I removed the hinged rotor heads and have gone with a simple servo-actuated motor pivot
which because these will only tilt about 20 degrees in either direction
I can use a small servo with a 3 to 1 gear ratio.
Then there will eventually be a large servo in the centre of the frame
to rotate the whole wing spar
so this isn't just far simpler electronically
but it's also lighter as it just requires these small servos
to provide all the control for hover
removing both the thicker wing spar and one of the large tilt servos
and when the aircraft transitions to forward flight
these small servos will keep the motors in a fixed position
essentially just acting as regular aircraft propellers.
I think it's probably time we move on to the fuselage portion of this project
which I gave a little teaser to in the last video.
The plan is to 3D print the fuselage using a material called lightweight PLA
and some unique design methods to save weight.
This lightweight material foams up when it is 3D printed
and weighs 40% as much as a regular 3D print
but the issue with this material is the expansion caused by the foaming process
leaves behind some residue if the printer nozzle is lifted away from the part.
To get around this the model needs to be designed
so the printer can create the part without lifting the nozzle.
I briefly covered this in my previous VTOL project
and I created a short tutorial over on my second channel
but that was for just a simple wing shape
so how will it work for something more complicated like this fuselage?
I first started by creating a number of sketches in CAD
to follow the profile of the real Osprey
which took quite a bit of time to get right
but when combined together ends up looking somewhat like the real thing.
I then created another fuselage that is slightly smaller
which will be important in a second.
From this I can start cutting the reinforcement ribs
that will help the fuselage keep its shape.
Currently this would simulate a regular 3D print infill
which fills the whole part
but for this application it's only needed around the outer skin
so I can use a smaller fuselage to subtract all the inner ribs
leaving us with a skeleton frame.
The final step is to remove these ribs from the solid fuselage model
as this will trick the printer into printing the part
without lifting the nozzle.
If we were to model the fuselage as a shell
and add solid ribs inside
the printer will have to lift the nozzle every time it creates a new rib
but if the fuselage is a solid model
and we cut the ribs out of it
the printer will draw each layer without lifting the nozzle.
Creating a clean lightweight part.
Also this is a new version of lightweight PLA from Colorfab
that is able to withstand higher temperatures
as I've previously had 3D printed wings melt in the sun
and fortunately this is the perfect color for the Osprey.
You may notice there is a combination of large and small ribs
spanning the inside of the fuselage
and that's because I ran a lot of test prints
to figure out how I can keep these parts as light as possible
whilst also maintaining a strong outer shell.
Here's just some of the prototypes I printed whilst designing this thing
like this one which has a high density of very small ribs
which holds the skin quite well
that's also quite bendy and flexible.
Then there's this one which has a lower density of larger ribs
which is quite a bit stronger
but the ribs don't support the skin very well
so it flexes between the ribs.
So this is the final design where it's a combination of the two
so there's a high density of thin ribs to maintain the strength of the skin
and then a low density of larger ribs to create the general strength of the fuselage.
So after about four days of printing
I now have all the parts ready to glue together
which I used regular CA glue
or sometimes referred to as super glue
which was a bit of a pain as it dries quite fast.
The fuselage is designed so it's glued together in three pieces
the nose, a midsection and the tail
which will eventually be held together with magnets.
I then need to remove the wings bar and motors from the frame
to install it in the fuselage
which is done by sliding the carbon fibre skids
into these rails printed into the design.
Once installed I need to cut an opening in the print
where the wing will be mounted
because this hole couldn't be designed into the model
as it's required to print the part using vase mode.
Now the wires can be plugged back in
and the wings bar and motors can be remounted to the frame
with access at the front and rear to tighten the required bolts.
I can then slide the tail section onto the skids to check how it fits
as well as attaching the nose
and it's really starting to look more like the full scale model.
All it needs is a few magnets to hold the parts in position.
I then printed the horizontal stabiliser and elevator
using the same lightweight material and design methods.
To create the hinge for the elevator
I chose to cut small strips of rubber sheet
that can be glued into this pre-printed slot.
I was tempted to 3D print an actual hinge pivot for the control surface
but this rubber hinge is far easier to manufacture
and is still strong yet flexible.
Then the elevator servo can be glued to the underside
which will be hidden inside of the fuselage once mounted
but first the rudders must be glued onto the tips of the horizontal stabiliser
and as you can probably guess these were also 3D printed
and the whole tail assembly can now be glued onto the fuselage.
Now it's time to cover up the motor mounts with some scale engine nacelles.
These things are huge just like the real thing
which is great for hiding all the messy wiring
and they are fixed in position with these glued on brackets
as well as some magnets to attach the two halves together
leaving just the top of the motor and propeller sticking out.
Also because the motor pivots independently to the engine nacelle
when hovering the thrust vectoring won't be tilting the nacelles back and forth
hopefully keeping the aircraft looking scale.
So before I build the wing it's probably a good time to test
if this thing will still hover with the added fuselage.
The propellers did vibrate a little to start as the blades were folded slightly for transport
but they soon straightened out and the aircraft lifted off the ground easily.
To be honest I'm surprised at how well this thing flies
it's actually more stable with the fuselage attached
which is probably because the added weight and inertia dampens the movements.
I haven't done any tuning since adding the fuselage so it could be a little better
but it's definitely a good start
and I'm also glad I moved away from the older rotor heads as this system is far simpler
and when hovering it's almost impossible to tell that the rotors are tilting
due to the stationary engine nacelles.
The next step of this build is to add a tilt mechanism and wing to the Osprey
but first let me show you some interesting mechanisms you can build yourself from KiwiCo.
KiwiCo makes hands-on projects for kids that are designed to be much more than just a toy
as they help teach science and engineering to spark curiosity and creativity.
I've always been a fan of making or doing things to learn
and these kits are an excellent way to get going
like this kit where you can build your own lock which requires an actual key to open.
Everyone knows how to unlock a door but do you know how they work inside?
The design and quality of these kits are just amazing
and they're not only educational but also good fun
and with everything you need being supplied within the box
there's no need to buy anything extra which I think makes for an excellent Christmas gift
and trust me completing a project that actually works is the best feeling
and you can experience this too by going to the KiwiCo link in the description
to get your first month with KiwiCo for free so definitely go and check them out.
Before building the wing we need a method of tilting the motors for transition
as they are currently fixed in the vertical position
so I removed the fixed wing spar mount and replaced it with a 3D printed mount
that will hold the tilt mechanism.
I chose to use a servo operated gear system for this tilt
as it needs a lot of torque to tilt both motors simultaneously
so these bevel gears give the servo a 3 to 1 ratio
where the servo rotates 270 degrees to tilt the wing spar from horizontal to vertical.
The servo I used is a cheap one I found on Amazon that's designed for radio controlled cars
and apparently can lift 20 kilograms at a one centimeter radius
so about 2 newton meters of torque
but with the gear ratio this is tripled to about 6 newton meters.
For the aerofoil I chose to use a NACA 6412
because its profile is large enough to fit the wing spar
and it's capable of producing a high coefficient of lift before stalling
as this aircraft is getting pretty heavy so we need as much lift as possible.
Once 3D printed I can mount the aileron servo inside and glue all the wing sections together
which I still haven't improved my gluing skills but I'm hoping to clean some of it off later.
To fix the wing to the fuselage I didn't want to use glue
as I might need to access the tilt mechanism and other electronics
so I made this mid section that simply rests on top of the fuselage
then when I slide the wing sections onto the spar it holds the mid section in place
and the angle of the wing is supported by the top of the fuselage pressing against it.
I then remounted the engine nacelles and checked if the tilt mechanism was still working.
Finally we can add the ailerons which work the same way as the elevator
by using rubber strips as the hinge.
Now it's time to see if it still hovers.
It definitely requires a lot more throttle to hover with the wing attached
as a large portion of the airflow from the propellers is being blocked by the wing
and this is why the ailerons are in their max downward position
to reduce the horizontal wing area.
Oh and one small tip, don't fix an aircraft fuselage together with magnets.
Luckily I thought this might be an issue so printed these clips to clamp the parts together.
So weighing in at a little over 2.2 kilograms
with a one meter wingspan and more panel gaps than a Tesla
this thing is ready for a test flight.
Right so today is the day to see whether this thing flies or dies.
Nothing left to do but fly this thing.
I'm going to take off very quickly and transition to half forward fly very quickly
just so that we get some air under the wings.
Okay okay that's not half bad.
It's cruising at a much lower throttle than when it was hovering.
I didn't want to spend too much time hovering because the battery doesn't last very long.
Okay okay this is actually this is flying at much lower throttle than a hover which is nice.
All right coming in for another pass.
That looks amazing it's so smooth.
Do one more turn and then we'll bring it into land.
Wow wow wow I can't believe this is flying.
I'm gonna put full flaps down now and then full transition mode.
It's now in hover mode.
Okay no that's not good that's not good.
Okay right so basically what happened is when I flick the back to hover mode
because the motors shift backwards a lot the it suddenly becomes extremely tail heavy.
When it's like that see how far the motors are forwards.
Actually wait half transition mode so the motors are there
then when I flick into hover mode it suddenly becomes a lot more tail heavy because it
shifts the center of mass quite far back.
Right I've replaced the battery ready for test flight number two.
Let's try and get some forward flight action here.
Transition to half forward flight.
Flaps up to halfway.
Get lots of altitude.
I'm gonna transition now.
Okay it is not like that.
Whoa okay that's probably extremely hard to see on the GoPro
but I was like full throttle and it wasn't climbing at all.
Okay we're really high.
Get ready to transition in three two one.
It is flying.
No it doesn't like that.
Yeah um it does not like that at all.
Right let's bring it in for a landing.
Let's try not get this all tail heavy again.
I'm gonna try and slow it down enough like this.
Full flaps down and then last second.
Nope nope.
Oh okay yeah this is um.
That transition back to a hover is absolutely awful
and I'm not sure how to get around it.
It just it just pitches up instantly as the center of mass shifts back.
I gave it a few more test flights as I had a few spare batteries
but it only seemed to fly well in the half transition mode.
I think this is because it has enough airspeed under the wing to reduce the load on the motors
but still using the motors to produce lift allows it to fly at a manageable speed.
The center of mass shift from the motors tilting is still a major issue
for both transitioning back to a hover and also forward flight
as when the motors are in the horizontal position the aircraft is extremely nose heavy.
A simple fix for this could be to add a counterweight to the motor pods
to balance the weight of the motors but this thing is already heavy enough
so I think a better solution would be to design a weight shifting system
that slides the battery back and forth depending on the motor position.
I would also like to add some more scale features to the design
like windows and maybe undercarriage but for now I think I'm going to keep this as a model
for my workshop and enjoy flying my previously built high performance VTOL.