Bell Boeing V-22 Project

tamuct01

Well-known member
"Background"

I embarked a few weeks back on what I think is a fairly ambitious project. I’ve long been a fan of the Bell Boeing V-22 Osprey. For those not familiar, the V-22 is an “American multi-mission, tiltrotor military aircraft with both vertical takeoff and landing (VTOL), and short takeoff and landing (STOL) capabilities. It is designed to combine the functionality of a conventional helicopter with the long-range, high-speed cruise performance of a turboprop aircraft.” (Wikipedia: https://en.wikipedia.org/wiki/Bell_Boeing_V-22_Osprey)



I’ve studied a few commercially produced V-22 radio control (RC) models, but they all seem to come up short. In the cases I studied serious design concessions were made to control costs and the model did not properly reproduce the real-world functionality of the V-22. Here are a couple of my observations:

E-flite V-22 Osprey VTOL BNF Basic 487mm
(https://www.horizonhobby.com/v-22-osprey-vtol-bnf-basic-efl9650)



Horizon Hobby, the makers of E-flite, are incredibly popular with RC enthusiasts. They have an extensive line of RC models for land, water, and air operations. I’m not sure of their market cap, but they are probably the largest RC manufacturer in the world.

On the surface, this model looks pretty good. It has a foam body with working tilting proprotors and the price is fairly reasonable. ($230 at the time of this writing for the BNF basic). However, closer examination reveals some significant flaws:

  1. The main rotors are FAR too small. On the real Osprey, the rotor blades are so large that landing in airplane mode would cause the rotor blades to strike the ground. It appears that this model is using some high-KV motors requiring the use of smaller props. They may have also decided that for beginners having small props that would not strike the ground when landing as an airplane was an acceptable trade-off
  2. The main rotors are fixed-pitch blades and not capable of cyclic or collective pitch inputs. For details on how helicopter inputs work, there is a really good article on Wikipedia (https://en.wikipedia.org/wiki/Swashplate_(aeronautics)
  3. There is a 3rd propeller in the center-aft of the model. In order to provide stabilization in hover mode, E-Flite has added a small 3rd propeller in the rear to provide pitch control to the model. On the real V-22, this control is provided through the cyclic pitch of the proprotors.

Rotormast V-22 Osprey
(https://www.rotormast.com/)



Rotormast is a small niche producer of scale V-22 kits. This model is much larger than the E-flite version and features working collective and cyclic pitch inputs. Each proprotor control utilizes a helicopter-style swashplate. The mixing of control inputs is done via a proprietary flight control package.

This is a much more realistic model, but still has some limitations:


  1. Price. Kits range from $1400 to $1750 depending on the desired scale look. This is out of my price range at this time.
  2. The controls on this model are done entirely with the collective and cyclic pitch of the proprotors. There are no traditional flight controls (aileron, elevator, rudder) in forward flight. As the model transitions from hover to forward flight the flight controller changes how control inputs are translated to the swashplate controls.

A while back an RC club member gave me a box of parts for the Blade mCPX BL (https://www.horizonhobby.com/mcp-x-bl-bnf-with-as3x®-technology-blh3980)



This helicopter is now discontinued, but parts are still available. In the box of parts were enough components to build 2 complete models with many parts to spare. I figure that with these parts I should be able to construct an RC model of the V-22 at a very low cost.

(https://aggiehobbies.blogspot.com/2019/05/v-22-osprey-project-background.html)
 
Last edited:

tamuct01

Well-known member
"Flight Control"

Now that I have a conceptual idea of how to build a V-22 using leftover mCPX BL parts, how do I go about controlling such an aircraft? I own a Taranis X9D+, and it can control up to 32 channels using dual receivers, so that won’t be a problem. I’m sure that I can create a mixing profile in the transmitter to handle just about anything, but I’m also sure that just like any multi-rotor “drone," this thing will need some form of automatic flight stabilization.





I own a few multirotor aircraft and I’m familiar with how they function. Some research into the popular flight control software (Betaflight, iNav, Cleanflight, LIbrepilot, etc.) all came up short. No one has written code for off-the-shelf flight controllers to manage a dual-swashplate aircraft. I would have to build my own.

I have not delved into writing code for flight controllers, but I have built a number of Arduino projects, and some of them utilize RC inputs to control servos and motors. It occurred to me that I might be able to take the stabilized outputs from the FC and pass them through an Arduino to create a custom mix for the control surfaces.

At the same time, I remembered a gentleman that I follow on YouTube named Tom Stanton. I am always impressed by his projects, especially those concerning RC flight. A while back he developed a largely 3D printed VTOL aircraft (https://www.thingiverse.com/thing:3191118). He ran into a similar case where he wanted to phase-in the main aerodynamic controls as the craft transitioned from hover to forward flight. To solve this, he daisy-chained an Arduino microcontroller between the flight controller and the servos. This allowed him to do some custom mixing in the Arduino that was not possible in the FC board.


Thankfully, Tom posted his code to GitHub and with some of my own modifications I’m on my way! (https://github.com/TomStanton/VTOL-Bicopter-transition-mixes)

(https://aggiehobbies.blogspot.com/2019/05/v-22-osprey-project-flight-control.html)
 
Last edited:

tamuct01

Well-known member
"Two helicopters on a stick"
For my first prototype, I want to attach two of the mCPX BL helicopter frames to a central boom to test the control system in a hover. Essentially, two helicopters on a stick. Upon examination of the stock mCPX BL frame, there is not a good way to attach the plastic frame to any cross support. I would have to design my own.


I took the measurements from the stock frame and started sketching them up in Fusion360. After a few test prints to confirm the dimensional accuracy, I created a way to attach the frames to a 15mm square dowel.




I’m using a Flip32+AIO FC board (https://www.readytoflyquads.com/flip32-aio) that I had in stock. It’s a Naze clone, and already quite old as FCs go, but it should provide enough functionality for this build.

The stock mCPX BL controller has a built-in ESC for the brushless motor, but I will have to use different ESCs since I’m not using the stock control board. I happened to have 3 12-amp ESCs in my stash that should do nicely. They are not the smallest form factor, but should do fine for this prototype. For now I’ve been testing with an Arduino Mega board, but it’s likely I’ll switch to a smaller form factor Arduino for the final craft.
The linear micro servos with the mCPX BL have a JST-SH connector on them. These are very tiny and don’t fit with the standard Dupont 2.54mm servo connections. I’ve removed the JST-SH connector and soldered on my own Dupont connections.




The real V-22 has counter-rotating proprotors with the left rotor turning clockwise and the right rotor turning counter-clockwise (as viewed from above in hover mode). The mCPX BL rotor is setup for clockwise rotation, so no changes are needed in the control setup for the left rotor. The right rotor, however, has some interesting issues:
  1. The motor can be reversed simply by swapping two of the three leads. It’s using a standard brushless DC motor.
  2. The rotor blades use a symmetrical airfoil so they won’t care which way they turn.
  3. The rotor head cannot be flipped on the mCPX BL. The linkages from the swashplate to the blade grips have a 90-degree right-hand twist to them that won’t fit if the blade grips are reversed (upside down).
This leads to an interesting problem. On nearly all helicopters, the control inputs are done from the front of the blade 90 degrees in front of the direction of rotation. This is due to gyroscopic precession (http://www.copters.com/aero/gyro.html). Essentially, the input of a rotating mass (gyro) will have an effect 90 degrees later. Therefore, to pitch a helicopter forward additional lift is given to the blade as it passes the 3 o’clock position. The lift actually occurs behind (90 degrees later on a clockwise rotating rotor), causing the helicopter to pitch forward.

With the right rotor of my prototype turning counter-clockwise and the blade inputs coming from the rear of the blade (90 degrees behind), the controls are a bit weird! For starters, I knew that the collective inputs would be reversed (up is now down), and I thought that all other inputs would similarly be reversed from normal, but they aren’t. Because I’m controlling the blades from the rear and the blade is rotating in the opposite direction, cyclic inputs remain the same as the left rotor.

IMG_3038.jpg IMG_3039.jpg IMG_3040.jpg

(https://aggiehobbies.blogspot.com/2019/05/v-22-osprey-project-prototype-1.html)
 
Last edited:

tamuct01

Well-known member
"First Flight Pre-Checks"

I had to take a few weeks away from this project while we moved to a new house, but I'm back at it and nearing first flight of the prototype. I've built a little power distribution board to connect the power for servos, Arduino, and flight controller. All of the electronic components I'm using are stuff I had lying around and are definitely not what I'll use on the final aircraft. For now, I'm using:
  • FLIP32 + AIO flight controller running iNAV and NAZE firmware
  • Arduino Mega
  • 12 Amp ESC
The all-up weight at this time is 418 grams. For comparison, the original Blade MCPx BL (without battery) weighed in at 48 grams. I have serious doubts this will actually fly at all in the current configuration.



Before any first flight, I need to verify the throws on all the cyclic servos so they match positive and negative travel. The flight controller should take care of small differences, but it's best to start off as neutral as possible.

I'm already looking into replacing many of the electrical components. I'll leave the FC alone for now, but I'll replace the discrete ESCs with a 4-in-1 model that can mount under the FC. I'll replace the Arduino Mega with an Arduino Micro. Only certain Arduinos have the number of external interrupt pins necessary to read the outputs from the FC. I have some Uno and Nano boards, but these only have 2 interrupt pins. For reference, the interrupt pin compatibility for Arduino is here: https://www.arduino.cc/reference/en/language/functions/external-interrupts/attachinterrupt/.

For now, the battery I'm using is a 2S 2200mAh. The original MCPx BL used a 2S 300mAh battery, so I'm way up on capacity as well as weight. I'll investigate connecting some of the smaller batteries in parallel or just purchasing a smaller battery in the 1000mAh range.

That's all for now!

(https://aggiehobbies.blogspot.com/2019/07/v-22-osprey-project-first-flight-pre.html)
 
Last edited:

tamuct01

Well-known member
"Prototype 1 Ending"


I'm calling an end to the first prototype build for this project. A few things changed between the first attempted flights. I gave the thing a serious diet by:

  • replacing the wooden boom with carbon fiber and redesigned/reprinted the motor mounts
  • replaced the separate ESCs with a 4-in-1 drone ESC module
  • replaced the Frsky X4R with a FrSky XSR
  • replaced the 2S 2000mAh battery with 2x 1S 600mAh batteries in series
  • replaced the Arduino Mega with an Arduino Micro
  • generally lightened everything up where I could
The new all-up weight is 217g, but it's still too heavy and the tolerances in these micro helis aren't good enough. I tried a few hanging control tests, and there's so much slop in the control linkages that it ping-ponged all over and never stabilized.

For Phase 2, I have a pair of KDS 450 helicopters that I'd like to use for the next prototype. They run on 3S 2200mAh batteries and are certainly more capable of lifting a larger airframe. They are also more capable of doing damage, so safety is going to be paramount. I think the control system is sound using the Arduino to mix the flight controller outputs. Better tolerances should lead to less slop and more authoritative control. Initially, I'd like to modify as little of the helicopter airframes as possible in hopes of returning them to separate helicopters if this prototype also fails. The plans, for now, are to remove the tail rotor assemblies and figure out a way to connect them together with a fixed boom similar to my 1st prototype. I've already checked, and it is possible to reverse the motor and the one-way bearing to have the two rotors counter-rotate.




Stay tuned for more!

(https://aggiehobbies.blogspot.com/2020/03/prototype-1-ending.html)
 
Last edited:

tamuct01

Well-known member
"Prototype 2 Control Tests"

To start on the second prototype of the V-22 Osprey, I took the pair of KDS 450 helicopters and stripped them of everything but the main shaft housings and landing skids. Each rotor disc is 28" (711.2mm) in diameter. If I compare that to the full-scale rotor disk of the V-22 of 11.61m, I get a scale factor of 16.87:1. In Fusion360, I scaled a 3-view drawing of the V-22 to the full size of the model and measured the distance between the rotor shafts and came up with a scale value of 33.2". I modeled a connecting block in Fusion360 that utilizes the original bolts and pattern of the KDS helicopter to attach to each side. The connecting boom is a piece of 1/2" x 5/8" C-channel aluminum that is 1/16" wall thickness.

I moved the control electronics over from the first prototype but redid the Arduino mixing code. I found some other PWM reading code that implements Pin Change Interrupts instead of the dedicated external interrupt pins which are very limited on most Arduino models. Using this code, I'm able to almost immediately detect any change on the input pins coming from the SPRacing F3 flight controller and mix them to control the 6 cyclic servos. The reduction in the delay is noticeable, and I was able to switch from the Arduino micro board to an Arduino nano and shave a few grams of weight. (https://create.arduino.cc/projecthu...-rc-receiver-input-and-apply-fail-safe-6b90eb)

The first suspended control tests went pretty well, but it became apparent that many flight controllers don't implement a heading hold gyro in their code. Despite enabling the HEADING_HOLD mode in iNAV, the copter would begin to rotate one way or another on its own. The flybarless "eBar" control module from the original helicopter does implement this, so my next try will be to pass only the rudder output from the flight controller through the KDS eBar box to keep the copter heading hold locked. If this succeeds, I will likely swap out the flight controller for a FrSky S8R stabilized receiver and pass the yaw control through the eBar. This will potentially simplify the control stack.

The eBar did try to keep the heading on track, but the response was too slow and rather violent. I was concerned it would twist the aluminum bar as it tried to correct for yaw. I tried just using the FrSky S8R stabilized receiver and immediately had a couple of successful, and more importantly, stable flights as shown below:


The next steps are to draw up some foamboard shaps to make it look more like an airplane. After additional flights with a static hover config, I will start developing the tilt mechanism. I have a few thoughts on this, but nothing definite for now.

(https://aggiehobbies.blogspot.com/2020/03/v-22-osprey-project-prototype-2-control.html)
 
Last edited:

nOtyRmOm

Member
This may not be a ton of help but ya i agree i had a v-22 thing. It was more like a bi-copter, not to good. People have had them where they are on servo gears and they work pretty well most of the time.
 

kilroy07

Legendary member
Pretty ambitious project, that I will be following closely!! (y)

Really hope you are successful so I can build one too.
 

tamuct01

Well-known member
After a successful first flight with the newer, larger design, I set out building a profile fuselage for it and made a couple of flights with it. It did ok, but the yaw motion tended to twist the aluminum channel that I used for a spar more than I would like. Lots of lessons were learned, so I started to design the tilt mechanism in CAD first and then start ordering parts. I have a few on order but with everything shut down, Amazon Prime is on hold. Here's a teaser of what's to come:

Screen Shot 2020-03-25 at 10.17.00 PM.png
 

tamuct01

Well-known member
"A Few Steps Forward and Major Design Flaw"

With the Coronavirus shutting down everything since my last post, I've made significant progress on the V-22 project. After the last post, I created a foamboard profile of the Osprey and was able to hover the craft around the yard successfully.

. .
.
With basic hovering out of the way, I started fine-tuning the CAD drawings for the tilt mechanism. Other folks have used multiple servos to control the tilt of VTOL motors. Tom Stanton used a gearing mechanism with a servo at each wing-tip to rotate the props on his fixed-pitch VTOL. The folks at Rotormast used a custom servo at each wingtip to operate a rack and pinion system. The servo was modified to use an external potentiometer. There is a gear on the rotation shaft that actuates the potentiometer through the 90-degree travel while the servo's rack and pinion system can gain additional torque through several rotations. I didn't have sufficiently strong servos on hand to handle the torque required to rotate the heavy rotors. I did have a standard size digital high torque servo on hand from my Sbach 342 build, so I designed a system that would utilize a single servo.


The design consisted of the servo at the wing root driving a translation horn through the 90-degree travel. The translation horn was connected via a dual-linkage system to the servo. The main shaft handled the transition to the wing through a U-joint. The Osprey has about 4.4 degrees of dihedral and 6 degrees forward sweep to the wings. Closer to the wingtip, the small 4mm shaft was connected to a 12mm aluminum tube via a 3d printed coupler with 2mm bolts pinned through the aluminum, plastic, and steel shafts. At the rotor end, a 3d printed mount connected the aluminum translation tube to the rotor body. Each of the shaft points was supported via bearings in the wing ribs.

All seemed to be going good and I built the right wing with a combination of the mechanics, 3d printed ribs, carbon fiber, and balsa components. The 3d printed center wing block connected the wing to the driving servo.

.

During the initial tests of the servo, though, it was immediately apparent that this would not work for a flying model. There was significant slop in the torque tube setup, specifically in the U-joint and translation horn. This slop would undoubtedly cause uncontrollable oscillations in flight. I needed to go back to the drawing board and change to a dual-servo configuration.

Without any appropriate servos on hand, I studied various sources for a while longer to select the right servo and design the mechanism that could get optimal torque in the smallest package. Frustration in this process makes that $2000 price tag from Rotormast more attractive! Finally, I settled on a pair of Hitec D85MG servos from ServoCity. These are high-torque metal gear digital miniature servos. They're slightly bigger than the cheap 9-gram servos I use for most of my foamies, but these really pack a punch!

I redesigned the wingtip mount and recessed the servo to maximize the clearance between the translation gears and the main rotor gears. I added a tension screw to maintain full engagement with the gears and keep everything tight and free of backlash. I also designed this piece so that hopefully I can gently cut the old rib free and glue this one in with minimal damage to the hard work I put into the previous wing.


With the servos and gears on hand, I performed a few tests just holding the mount and the movement feels much more secure than the previous design. I'll graft this onto the wing and see how well it works as part of the greater model.

(https://aggiehobbies.blogspot.com/2020/05/a-few-steps-forward-and-major-design.html)
 

JasonK

Participation Award Recipient
are you using full on heli variable pitch on the props for this?

I was working out a design in my head/paper that is similar, but was just planning on using thrust vectoring. Care to talk about your choice on rotor control?
 

tamuct01

Well-known member
are you using full on heli variable pitch on the props for this?

I was working out a design in my head/paper that is similar, but was just planning on using thrust vectoring. Care to talk about your choice on rotor control?

I'm using two 450-sized heli rotor guts for the power plant. These are stock from a pair of KDS Innova 450s that I got cheap at a swap meet. Each rotor is stock with 3 cyclic servos arranged in a Y config like most RC helis. Other RC models of the V-22 use fixed-pitch rotors (Eflite) or only have pitch cyclic in each rotor (Rotormast). When I started the project, I spoke at length with a college buddy who flew the V-22 for the Marines and I picked her brain on how they actually work. The full-scale bird does have full pitch and roll cyclic on each rotor.

I'm using an Arduino to take the RC receiver pitch, roll, and yaw outputs (Stabilized by the S8R receiver) and map them to the 3 cyclic servos in each rotor. The mapping is as follows in hover mode: Pitch input causes standard pitch response in the rotors equally--the two forward servos in each rotor move opposite the rear servo. Yaw input causes differential cyclic pitch response in the rotors--one pitches forward while the other pitches back to induce the yaw motion. Roll control is a little special. For now, roll input causes a differential collective response in each rotor--all 3 servos in each rotor work together to induce the roll. In the real craft, there is also a slight roll input to each rotor to cause the craft to translate left or right (slide). I have a placeholder in the mix right now for this, but I haven't implemented it yet. One option would be to tie the stabilized roll output to the collective pitch of each rotor for stability and level hover and pass a second, unstabilized roll input to the mix to serve as the "slide" input. That way the craft remains straight and level but allows the pilot to strafe around left and right.

Good luck with your design!
 

JasonK

Participation Award Recipient
thanks. and yes, I absolutely expected the real thing to have full heli type controls on the rotors, I am just trying to get as light as possible, which means as few moving parts/etc as possible. I saw the arduino in there, nice job. and I agree, having all that control in your modules, will definitely give a much better hover experience.

Do you know what your CWL is going to be (or a guess at it)? I am pondering that issue myself because of the relatively small wing sizes on all the examples I have seen (at least the real ones). So far the only thing I have flown are some quads in angle mode (off the shelf things) and a Tiny Trainer with the 3 channel wing on it (I just built the 4 channel wing and am planing on getting it out this weekend.)

Don't suppose you have a video of the hover working? would be interested in seeing it.
 

F106DeltaDart

Elite member
Very cool and excellent work! I’ve been considering buying the Rotormast machine for some time now, and even considered building a V-280. I’m a test Engineer at Bell, and would love to add a tilt rotor to my fleet. Will be watching closely!
 

L Edge

Master member
Welcome to one of the few areas that are left that is totally different.

After reading your thread, I am building up the idea to pull mine from the boneyard and make an other attempt at it. I tried to duplicate Tom's project last year and ended up with a terrible vibration with one arm trying to hover it. I did everything from changing the servo, motor, prop and spinner and finally put it away for another day.

You got me inspired once again, so I have to wipe the dust off it and re-evaluate it to see if I can solve the problem and forge ahead. This has been the one project that has kicked my butt, time and time again. Tried two other approaches and failed, that is why at least Tom's should work for me. Thanks!!!!
 

tamuct01

Well-known member
I haven't calculated the wing loading. I've been going of 3-views in Fusion360 to get the scale-ish shape correct. The prototype right now is a symmetric airfoil, but I expect that it will not have great forward flight properties. This is accurate on the real aircraft as well. The surface area of the wing is small for an airplane and would have a high wing loading. The full-scale craft will tilt the rotors up as they slow down essentially trading reduced lift from the wings to the rotor lift. Essentially, the wings are there to keep the rotors apart and provide a little bit of lift in forward flight.

Most of the VTOL craft in the wild use brushless motors and fixed-pitch props for simplicity. I'm quite concerned about the mass of the rotor assemblies hanging way out there on the ends of a ~36" wing. This won't be designed from the get-go to carry much cargo. The fuselage will be DTFB and its only real purpose will be to hold batteries and look nice.

I only have a video of the rotors mounted vertically in post #6 before I added DTFB around it and built a quick profile fuselage. The performance seemed about the same before and after the DTFB additions. My early attempts were using micro helis, but those just don't have the power needed to lift much of anything. That's why I switched to the 450 sized helis.
 

JasonK

Participation Award Recipient
So if I understand correctly, the real thing is doing more of the "big enough engine + correct CG => anything flies" thing that FT videos often say? Perhapse I need to be ok with a really bad WCL.
 

tamuct01

Well-known member
A couple of quick photos to document the progress:

The right wing is nearly completed and sheeted with 0.025" paper-backed maple veneer. I'm not sure if I'm going to cover this with Monokote or similar or seal the veneer and paint it.

IMG_3531.jpeg

I started work on the left wing and have most of the basic structure completed. The 12mm aluminum rod used for the translation tube is slightly oversized for the 12mm bearings, so I need to chuck the rod in my drill and sand/polish it down to the correct OD to fit.

IMG_3530.jpeg