QuadPlane Pitches Up using v3.8.0 beta 5

Lol yeah! :slight_smile:

I just checked the first video again and itā€™s really hard to see the twisting there.
I donā€™t have a Ranger, but I so have a Skywalker 1880ā€¦and that thing was as twisty as a curly fry!
Itā€™s been stuck on top of a cupboard for the last 5 years.

As mentioned in one of the posts to Greg, quad arms structurally just donā€™t belong on the wings at all. Just because most other QPs are built like that doesnā€™t mean itā€™s right. The only ā€œperceivedā€ advantage of having the arms parallel to the fuselage is that one would think it reduces drag. But itā€™s not as simple as that because thereā€™s a bunch of other factors at play.

It really depends on why you want a QP. If you want to hover around and look at things close range itā€™s better to use a quad. If you want to go further, faster and for longer then a plane is better. IMHO a QP should only give the aircraft VTOL capability, but for most of itā€™s flight it should use itā€™s really efficient wings, not itā€™s inefficient motors. You canā€™t have an efficient hover on a QP because thereā€™s just too much ā€œplane weightā€, but you can reduce the quad weight substantially while actually improving the plane performance from stock after a QP conversion. Birds have wings and not props for a reason!

So if the above goal of ā€œonly VTOLā€ and then fly with the wings is the goal, all the design parameters change. The biggest differences are 1) you can nearly neglect hover efficiency and go for the smallest possible quad propulsion setup that gives you enough lift and control 2) you donā€™t have to configure your forward motor to produce static thrust for takeoff, and can completely optimise the forward motor and prop for best forward cruise. If done like this, in our experience a QP conversion results in a platform VTOL, carry more batteries/payload and fly further and faster than a stock plane can.

If you like you can take a look at what we use on our ā€œMozzieā€ QP the beta docs are here.
Weā€™re going to be releasing the docs publicly once we finally finish the flying section.

No problems. Iā€™m actually pretty glad it wasnā€™t a AP bugā€¦I donā€™t like those at all, because they tend to be platform independent and I might catch it too! :slight_smile:

Although I can see the reason to have a quick way to convert it back to a normal non-QP, I think youā€™ll only get the most out of it if you configure it to always work as a QP. What are your mission goals with these platforms, or is it just for fun and try it out? Iā€™m not sure if reinforcing the wings will completely solve the issue, so depending on what you want to achieve Iā€™d maybe look at another airframe. Structurally that Ranger just doesnā€™t look up to the QP challenge in my opinion, and adding structural elements to compensate might end up being counter productive.

Jeff,

I like the Mozzieā€¦thanks for sharing! I can see on some of the designs like the Mozzie and Darrellā€™s X8 QuadPlane wing that by eliminating some of the normal plane capabilities you can alleviate some complexities. My Ranger EX can still take-off using a normal wing but can also become a QuadPlane. This creates additional trade-offs.

On the medical transport subject, my standard gimbal mount from the FireFLY6 is mounted twice on my Ranger EX since it can carry more weight. Here are some images from my OpenGrab EPM dump of a FirstAid kit.

Nice! I like the delivery method.

One of the guys in our group also has a Firefly but we are yet to fly it.
Really need to give it a go, it has a bunch of nice running gear on it.

Are you trying to create a medical delivery platform? What range/payload is required?

Jeff,

I donā€™t have any requirements other than hobby fun and interest. I do have a second FireFLY6 that I will likely convert to Plane tiltrotor during the winter.

What is the base model for the Mozzie? Is it hobby based or proprietary?

Edit: I see now that it is a XUAV Mini Talon.

Here are a couple videos of the same Ranger EX QuadPlane conversion as mine but using the PX4 stack. His was copied from mine using the same boom length and offset but mounted closer to the fuselage. Iā€™ll admit that his donā€™t wobble as bad as mine but it seems to work fine. It will be interesting to see the result after my strut installation. Perhaps I need to tweak my gain PIDs as well.

Hi Greg.

Iā€™m quite passionate about designing stuff to meet goals, so sometimes I spoil the fun! :slight_smile:

That 360 video of the other Ranger QP is also quite revealing, in that it still has a lot of arm movement albeit less wing contortion. Iā€™m pretty sure that the wings/structure will fatigue over time. After watching that I really think the ranger is better off being hand thrown or bungee launched and land it with reverse thrust for STOL rather than being run as a QP if you want to keep the arms on the wing.

How much do you think your Ranger weighs without the QP components? The spec says itā€™s about 1.5kg without batteries? If its under 2.5kg or so Iā€™d recommend building the Mozzie quad arms and propulsion system that can be found here and attach them to the Ranger instead. Parts are listed here. They can still be dismounted with just two screws for normal non-QP flight, but only weigh 150g per arm and still produce 4kg of thrust which would be ample for the Ranger 2.5kg airframe. The only thing Iā€™d possibly change is increase the width of the arms by about 50mm as it looks like the Ranger is a bit wider than the Talon and you might have to bush the mounting method to fit the Ranger fuselage curvature better. Iā€™m 100% confident that it will not only fix all wing twisting but also improve quad hover control, will hover at around 40A, and will reduce the forward current draw if you optimise the forward prop for cruise.

The biggest problem with most RC aircraft is that propeller efficiency is so low. Most RC aircraft only operate in the 10-20% with around 40% being the high end. So in simple numbers this means that if you improve your prop/motor efficiency by just 1% on a 10% efficient setup, youā€™d increase your range by 10%. If you go from 10% to 20% or beyond you get twice the range just by optimising your prop/motor and without changing anything else. This is why aircraft are designed from the propeller backwards for efficiency, Most RC planes donā€™t do that because they a designed to either imitate a larger aircraft or be more aesthetically pleasing or easier to construct. Prop optimisation works especially well with QPā€™s because one can completely ignore specifying props for static thrust for takeoff and have a prop that is only optimised for the best cruise. (which is another thing that needs to be tested to find out where it is on that particular airframe at that weight)

Let us know how you go.
Regards
Sam

1 Like

Greg,
Although I am in agreement with all the feedback provided in regards to locating the boom under wings, Iā€™ve still seen many successful Ranger conversions work this way. For another quad conversion, I would definitely go with a ā€œmozzieā€ setup. But in an effort to complete a successful conversion for the ranger, I strongly believe the ā€œsweet spotā€ to its stability is bringing the booms closer to the fuselage. The farther from its main support, the more succeptible you are to higher levels of deflection and torsion. This will also allow you to avoid having to retrofit any struts or secondary framing which may lead do additional issues. Iā€™m considering doing the same on my end. Is the person in the video an ardupilot member we can reach out to? I think it would be beneficial to bring him in to this discussion. I would invite him to hear his thoughts and feedback.

Sam (who I keep calling Jeff),

I agree with and appreciate your feedback. Iā€™ll admit that the Mozzie mounting technique is a better approach to eliminate wing torsion. I missed the wing torsion issue on the Ranger EX since I had not seen it on the smaller Bix3 conversion. The Ranger EX (like a Skywalker) is a great flying FPV plane for long distances. Since parts are available from multiple vendors, it seemed like good candidate to convert to a quadplane with little effort. My goal of keeping the stock plane frame, so that it can be converted to a quadplane with just a wing change, has created additional trade-offs.

I will still try mounting some struts (or thin braces) to see if it fixes the torsion issue. These can be easily connected or disconnected while not losing the normal plane functionality. I will also glue some flat carbon rods across the wing chord to help reduce its ability to twist. You can see from my videos that the Beacon Foam Tac really creates a tough bond between the square carbon tube and the foam wing. With all the torsion and crashes, they are still bonded, so it would be a difficult task to remove them without destroying the wing.

I also plan to build a Mozzie as I now see the benefit of a dedicated quadplane design. A frame like the Skywalker X8 wing that Darrell converted is another good example. Many of us would love to have his machine shop! :slight_smile:

1 Like

Alex,

He is RC Groups member, rollys, who posts on my RCG APM thread here. He is using the PX4 stack with QGC.

Sam,

I have a few off-topic questions on the Mozzie.

  1. Will someone be selling the 3D printed parts as a kit?
  2. What is the purpose of the Companion Pi?

Thanks!

Hey Greg
I donā€™t really have any plans in selling the 3D printed parts, but I could if you want me too. Not sure what the freight to USA is but I can find out. Alternatively thereā€™s the option of printing them at a local maker group or store, or an online producer like 3dhubs or shapeways etc. If you donā€™t need the CC or avionics enclosure/camera gimbal etc, you only really need to print the quad arm 3D parts and can use the timber parts that come with the mini talon instead to build a QP. (In fact thatā€™s an idea Iā€™ll add to the instructions as well!)

The CC Pi is for imaging using the picamera and to georefence images to find Joe in the OBC/MEC. It also runs 4G modem and/or wifi, which you can stream as well if you want. weā€™re currently working on a updated CC image using fnoops Maverick.

The idea is to provide a reference platform for everyone that is capable of completing the Medical Express Challenge.

Ok, thanks for the information, Sam. I will likely build a modified Mozzie which uses some of the parts recommendations. It looks like a good deal of work went into the design and documentation. It is much appreciated! Iā€™ll focus on getting the quad arm parts but using as many stock timber parts as possible.

Nice documentation of the Mozzie Sam, thankĀ“s. IĀ“m just curious about how to choose the most efficient motors for QP?. (IĀ“m in that task now) Mozzie looks smaller for 2kgs than I thought for my near 3 kgs take off weight; I choose 3536 810kv type motors, 950kv info looks to have less trust to hover at 50-60%, perhaps I in a mistake how to choose, any advice?
Sorry the oftopic too.

Hi Greg, as always, Iā€™m watching your threads as I am particularly interested in the firmware developments. I have not flown my Ranger for over 7 months as I am just too scared of losing it with all the pitch up issues, etc. I have read about. I had plenty of good flights on 3.6 but changed to 3.7.1 as I wanted the additional features, but now think I will wind back to 3.6

I also wanted to let you know that there is a VERY easy fix for the wing twist problem. Firstly, carbon/glass cloth on the wing, or even extra spars, will not help with the twist. All you need to do is use some fishing trace cable to tie the diagonal motors, ie. front left motor to the right rear and front right to left rear. The cables pass under the fuselage and itā€™s the depth of the fuselage that gives all the strength in this setup. It was so easy on mine, as I just swapped out one motor mount screw on each motor with a slightly longer screw, which gave me a tie point. Absolutely no twist whatsoever control authority (esp yaw) improved significantly.

If you look carefully at the shots of my aircraft on the ground in the beginning of this video, you can just make out the tie to the front right motor: https://vimeo.com/174204873

I will get some detail photos when I am back from my holiday.

Cheers,
Jacques

2 Likes

Hi Jacques,

Thanks for the fishing trace cable solution! It may be easier than adding spars so I will look into it.

Enjoy your holiday! :slight_smile:

Greg no worries let us know how you go with the build.

Jacques
Your Ranger looks much better in the video, do you have the diagonal cables fixed on the fuselage as well?

@jacques_eloff looks like you may have saved the day! Great idea on the fix. I like that itā€™s an easy and simple solution! Iā€™ve seen your videos a few times and have seen your quadplane fly successfully. Great job. Iā€™m a step behind Greg in a quadranger build so this feedback is extremely helpful. Thanks again and have a great holiday!

On my Skywalker Iā€™m using much less than 1.5 prop diameter and I still have adequate roll control. Iā€™m using 9" props, and my booms are 10" from the fuselage, or 12" from the center of the wing.
@JeffBloggs has the right idea. Use small props and high KV motors. This results in shorter booms, more rigidity, less weight, and less QProp drag. My build is only designed to be in VTOL modes for under 2 mins. Everything is optimized for plane efficiency. On my SW1880 build I used carbon plates all the way to the fuselage. There was no torsion issues but it was unnecessarily heavy. The 1880 was a terrible QP wing as it has a nasty tendency to tip stall which required too much airspeed and burned too much power in cruise. In my SW1900 built I replaced the carbon plates with diagonal strapping tape, and brought the booms closer to the fuse. There is a bit of torsion, but Iā€™ve had no trouble with controlability and I feel the trade off was worth the 150g savings.
Fully loaded with payload my SW1900QP is 3.8kg.

1 Like

What motors are you using Iskess?