Thread Tools
Sep 23, 2020, 11:00 AM
Registered User
Thread OP
So that's how fast a year can feel. And what a year it has been. September 2019 feels like a different lifetime.

This year I was planning on reducing the time I spend with this hobby, not because of lack of motivation but because other areas of my life need attention. Wanted more hours out in the world instead of in the workshop. It's a great winter hobby in the dark/wet/windy months but summer was supposed to be about travel and new experiences. Well I certainly got new experiences, we all have, just not the ones we probably wanted.

Anyway, if I am going to be constrained to this new local working-from-home life I can at least make the best use of the hours. So that brings me back to this old project.

Pictured is the new version. Base dimensions of the hovercraft section are 70cm wide by 80cm long. Outboard wings make the total span at the rear 1.4m which is getting unwieldy.

Lift is 2836 950kv motors with 10x3.8" counter-rotating props. Thrust is 2830 1000kv with 8x4.5" props. 3S 2200mah '40C' battery. There's no such thing as a 40C battery, only a battery with 40C printed on it.

Anyway this design incorporates everything I learned from the previous one. Lots of refinements which I will detail in a following post.

I have tried this out briefly and the hover performance over grass is far better than the previous one. Got a good air gap under it now and it certainly hovers over grass instead of barely skimming and dragging. Tried flying and it was very twitchy and borderline unstable, that should have been the sign to take it back to the bench and get the CG forward, I tried to fly it anyway and could handle it in a straight line but an attempted turn stalled it and dropped it from 5m or so up straight down onto the nose. Remarkably the damage was quite minimal and easily fixed.

Previous version needed a considerable forward speed to fly properly rather than hang off the prop thrust, the brief flight of this one shows less speed and thrust required which is expected from the larger span.

It's ready to fly again but the place to do it is troublesome. There are a few parks nearby with the minimum of space and that's if they're quiet enough. I now have access to a proper club field nearby but the lockdown rules may undo that. All of us at the club are gonna be if our activity is shut down while other objectively riskier ones are still allowed.

We'll see. I'll keep this updated
Sign up now
to remove ads between posts
Oct 05, 2020, 05:24 PM
Registered User
Thread OP
Got my first proper test done.

RC flying hovercraft - new version (3 min 22 sec)

It's certainly more sedate and easy than the previous version due to lighter 'wing' loading. Thrust is a little lacking on the 8" props running here, have since gone up to 9" and hope to test it again soon.

Hover lift is good but I think there's room for improvement. It actually can run on the much thicker grass here though the front digs in a bit. There's a bit of conflict between the ideal CG for pure hover and the ideal CG for flight. 5 years ago (!) at the start of this thread I talked about how the flyable CG of the bixel planform coincides with the center point of the hovercraft section. Testing shows that's not right on.

Looking at the hovercraft section alone, the base plate is 80cm long, with a 10mm peripheral jet and 10mm thick perimeter foam the outside total length is 84cm. CG for hover alone is logically the middle or 50% chord as you might call it. On my single motor craft only meant for hovering I'll move it to 45% chord for better stability at forward speed. Don't need to worry about that on this model as the outboard wings take care of pitch stability.

Now, looking at the whole planform, the leading edge extends 8cm or so forward of the hovercraft and the elevons go 15cm back from the rear of the hovercraft. So the 50% chord on the hovercraft alone isn't 50% chord of the whole thing. It's not far off but neither was my CG on the previous test that crashed so it pays to be careful.

I need to experiment more but 45% chord on whole thing is about the limit for flight. That's further forward that 50% of the hovercraft alone. So a flyable CG means the nose of the craft digs in under hover. If I wasn't sick of the sight of 2D CAD I'd draw this up properly. Basically you need to accept a nose-heavy hovercraft for it to be stable out of ground effect.

Looking at it another way, the center of lift of this whole thing is too far forward for the ideal hovercraft CG. How to get the AC back? More span or more chord behind it. The elevons going 15cm back is an attempt at this. If you made the elevons go 50cm back you can see this thing as a really low aspect ratio flat airfoil. Then the AC can be ~25% of *that* chord which would be behind the best hovercraft CG. So there's some trailing edge extension length that gets it right. Whether that's practical I don't know.

So, I either need to extend the trailing edge and elevons back further, or I add span to the outboard wings, or maybe both together. Perhaps an extra 5cm span on the outboard wings is all it needs. It's totally workable as is, a bit of thrust and up-elevator raises the nose just fine and levels the hovercraft section. I would like to see even hover without needing to do that.

Anyway I've bored myself enough with this. Attached is a quick sketch of different slot types I've tried. This thing is using the 6mm ramp approach at the far right. Any sort of protruding plasticard will get torn off so it needs to be set back from the bottom of the craft. The ramp gives a deep slot and good air jet and also allows the plasticard to stop short. It also keeps the base plate a light 6mm except at the perimeter. I could probably reduce the width even more.

I got a 3D printer about a year back and am now very comfortable with Fusion 360. It's been a revolution for this hobby, there are so many parts it's easier to print than try to fabricate by hand. The thrust motor mounts are a great example, I can design in any upthrust angle I like. The lift motors are bolt clamped around the center balsa spine of the hovercraft. Makes for very accurate placement of the motor hub, easy replacement of a motor and also free adjustment of the position along the beam for minor adjustment.

A persistent challenge with this design is keeping the top plate down under lift power. With no reinforcement it balloons out under the air pressure. The central pillar is a handle but also keeps the nearby edges of the top plate down. At the sides of the props are 3D printed brackets that go through the bottom plate and have holes for an R-clip at the top. This does a really good job of keeping the plate down as the force is all tension.

Stabilisation ports under each prop are 50% diameter in a cutout of 75% of the diameter. That allows easy testing of different areas. A few days back I tested various configurations of stab port on water on my single motor craft. Will post it in the relevant thread soon, I found that a conical diffuser from 50% prop diameter to 75% massively reduced turbulence on the water. It is my ambition to get this one on water one day so I'll keep that in mind.

Said in the video but for completeness:
  • 6mm Depron Aero construction. Less rigid than regular depron but lighter. Bit of flexibility is good as it bends a lot before snapping.
  • Lift is D2836 950kv motors and 10x3.8" props. Motors run hot under constant power but not excessively so. Better lift on 10x3.8 than 9x4.5.
  • Thrust is D2830 1000kv and 8x4.5 props (pictured) or 9x4.5 (awaiting test).
  • Servos are Hextronik MG14. They have a bit of work to do with the size of the elevons. I've built in an aerodynamic balance on the outer edge to relieve some strain.
  • ESCs are 30A multirorotor models as they're very light.
  • 3S 2200mah
  • 3A BEC for power
  • Frsky X8R and FLVSS voltage telemetry. While this will glide to a hard landing it's better to land with thrust power and full hover power. That takes some current so telemetry lets me know to keep a bit in reserve for landing.
Last edited by FiftySlicks; Oct 06, 2020 at 04:33 PM. Reason: Updated specs.

Quick Reply

Thread Tools