How would you design a cyclocopter?
- Thread starter NickRehm
- Start date
2jujube7
Well-known member
Great! Good to hear that the airframe flies, can't wait to see it.I've maidened the airframe with a regular prop just to make sure it flies, but it was crazy windy. So still waiting on a nice day with sun + less wind + I'm off work. At this rate, maybe a flight by next year?
NickRehm
Member
Had a great day for flying today (minus extremely cold temps...) so I took the cycloplane out for a flight--Not enough power for takeoff.
I'm going back to revisit my blade twisting issue to see if I am actually loosing a lot of performance from that. I've started by covering the blades in tape, and am printing out an additional 3DP piece to brace the control linkage on the blade that was also bending a little bit.
Another thing I'm concerned about is blade angle of attack at high forward speed. I'm estimating they are rotating at about 3000 rpm with the 1900kv motor and 3:1 gear reduction:
(1900kv * 12v) / 3 = 7600 rpm, *guessing* closer to 3000rpm under load = 314 rad/s
So the blade velocity with 0.075 rotor radius is about:
314rad/s * 0.075m = 23 m/s
On the ground I guestimate it was rolling along at ~8 m/s, so the wind angle relative to the blade at 0 aoa is about:
sin(8/23) = ~20 degrees
Subtract this from the blade aoa at zero forward airspeed (the geometric angle I built it of about ~40 degrees), the actual angle of attack becomes closer to 20 degrees ~= half the thrust I saw for the static case. Normally high forward speed isn't really a problem for props because they spin so much faster; that relative wind angle that was subtracted out goes down and the true blade aoa doesn't go down as much.
Basically I'm worried that the blades aren't going fast enough to produce high enough velocity thrust for the airspeed my plane needs to fly. Like putting a slow-fly prop on a jet--it might produce the same static thrust as an EDF, but the thrust velocity isn't high enough for the airspeed the jet needs to stay airborne. I had a similar issue a while back with a high-wing-loading flying wing that I put a slow-fly prop/motor on. It had above 1:1 thrust:weight ratio, but refused to fly at all.
So there's a few options to explore:
And I thought this project would be easy
I'm going back to revisit my blade twisting issue to see if I am actually loosing a lot of performance from that. I've started by covering the blades in tape, and am printing out an additional 3DP piece to brace the control linkage on the blade that was also bending a little bit.
Another thing I'm concerned about is blade angle of attack at high forward speed. I'm estimating they are rotating at about 3000 rpm with the 1900kv motor and 3:1 gear reduction:
(1900kv * 12v) / 3 = 7600 rpm, *guessing* closer to 3000rpm under load = 314 rad/s
So the blade velocity with 0.075 rotor radius is about:
314rad/s * 0.075m = 23 m/s
On the ground I guestimate it was rolling along at ~8 m/s, so the wind angle relative to the blade at 0 aoa is about:
sin(8/23) = ~20 degrees
Subtract this from the blade aoa at zero forward airspeed (the geometric angle I built it of about ~40 degrees), the actual angle of attack becomes closer to 20 degrees ~= half the thrust I saw for the static case. Normally high forward speed isn't really a problem for props because they spin so much faster; that relative wind angle that was subtracted out goes down and the true blade aoa doesn't go down as much.
Basically I'm worried that the blades aren't going fast enough to produce high enough velocity thrust for the airspeed my plane needs to fly. Like putting a slow-fly prop on a jet--it might produce the same static thrust as an EDF, but the thrust velocity isn't high enough for the airspeed the jet needs to stay airborne. I had a similar issue a while back with a high-wing-loading flying wing that I put a slow-fly prop/motor on. It had above 1:1 thrust:weight ratio, but refused to fly at all.
So there's a few options to explore:
- Increase rotor radius to bring blade velocity up -- this might start overheating the motor, so a larger gear ratio would be needed
- Larger gear ratio anyways to bring the brushless motor back into a more efficient location on the rpm-power plot. Right now, I'm definitely on the right side of the curve due to my heating issues
- Increase geometric blade pitch to ~60 degrees so in forward flight the effective aoa is closer to 40. This will probably screw up any static thrust/efficiency measurements, but who cares
And I thought this project would be easy
Flightspeed
Convicted Necroposter
Y’all nerdy peoples rc airplane hobby is completely different! 🤣 if you ever get one working good I now know somebody who has a 3d printer. Would love to have my own cyclopter!Had a great day for flying today (minus extremely cold temps...) so I took the cycloplane out for a flight--Not enough power for takeoff.
I'm going back to revisit my blade twisting issue to see if I am actually loosing a lot of performance from that. I've started by covering the blades in tape, and am printing out an additional 3DP piece to brace the control linkage on the blade that was also bending a little bit.
Another thing I'm concerned about is blade angle of attack at high forward speed. I'm estimating they are rotating at about 3000 rpm with the 1900kv motor and 3:1 gear reduction:
(1900kv * 12v) / 3 = 7600 rpm, *guessing* closer to 3000rpm under load = 314 rad/s
So the blade velocity with 0.075 rotor radius is about:
314rad/s * 0.075m = 23 m/s
On the ground I guestimate it was rolling along at ~8 m/s, so the wind angle relative to the blade at 0 aoa is about:
sin(8/23) = ~20 degrees
Subtract this from the blade aoa at zero forward airspeed (the geometric angle I built it of about ~40 degrees), the actual angle of attack becomes closer to 20 degrees ~= half the thrust I saw for the static case. Normally high forward speed isn't really a problem for props because they spin so much faster; that relative wind angle that was subtracted out goes down and the true blade aoa doesn't go down as much.
Basically I'm worried that the blades aren't going fast enough to produce high enough velocity thrust for the airspeed my plane needs to fly. Like putting a slow-fly prop on a jet--it might produce the same static thrust as an EDF, but the thrust velocity isn't high enough for the airspeed the jet needs to stay airborne. I had a similar issue a while back with a high-wing-loading flying wing that I put a slow-fly prop/motor on. It had above 1:1 thrust:weight ratio, but refused to fly at all.
So there's a few options to explore:
- Increase rotor radius to bring blade velocity up -- this might start overheating the motor, so a larger gear ratio would be needed
- Larger gear ratio anyways to bring the brushless motor back into a more efficient location on the rpm-power plot. Right now, I'm definitely on the right side of the curve due to my heating issues
- Increase geometric blade pitch to ~60 degrees so in forward flight the effective aoa is closer to 40. This will probably screw up any static thrust/efficiency measurements, but who cares
And I thought this project would be easy
2jujube7
Well-known member
Haha I had the same thoughts when I was building my cycloplane. Interesting effects that you're experiencing though, I hope you can get around it.And I thought this project would be easy
(Just a note: the positive camber thrust results regularly fluctuated +/- 10g while on the scale. I just took what was in the middle of that range, which happened to be a nice even number. Also, there was no power or thrust increase when I went from 5/6 to full throttle, which is probably due to an calibrated ESC? Therefore I didn't include it.)
I just did some efficiency tests with an asymmetric airfoil. It turns out that the positive camber (as seen at the top of the rotation, which somewhat negates virtual camber) is much more efficient than negative camber (which increases virtual camber). This is the opposite of what another cyclocopter researcher that I was talking to said would happen, so it's an interesting result. I'm planning on building some symmetric airfoils today to test those too, but I think that they should be pretty similar to the positive camber just based on previous thrust/throttle results, and maybe a little higher as I got a full 300g at full throttle.
Attachments
NickRehm
Member
Those are some pretty definitive results...almost 2x power loading. Awesome find!! It may be because 'negating' the virtual camber eases forces on the blades that try to twist them. The blade twisting leads to a lot of efficiency loss, and especially more friction in the control linkages.
NickRehm
Member
Had a quick flight this evening in a nearby parking lot after reducing weight by about 100g across the whole plane. Just enough to get it off the ground and plop it back down again under the street lights. Then while taxiing back, I guess I hit a rock or something and snapped two of the endplate arms on one of the rotors So I've got some new parts in the printer, but now I'm confident this thing can actually fly, so I just need to find a time during the day to take it out for a real flight and catch some video...
I also just got some LW-PLA so I'm gonna have a go at printing some solid blades to save me some time making the blades and give more consistency. I also hope they'll be a bit stiffer in twisting so I can improve my thrust a bit more. Anyone have any tips/settings for and Ender 3?
So I've got some new parts in the printer, but now I'm confident this thing can actually fly, so I just need to find a time during the day to take it out for a real flight and catch some video...I also just got some LW-PLA so I'm gonna have a go at printing some solid blades to save me some time making the blades and give more consistency. I also hope they'll be a bit stiffer in twisting so I can improve my thrust a bit more. Anyone have any tips/settings for and Ender 3?
Last edited:
tamuct01
Well-known member
I'm using the eSun PLA-LW on my Prusa Mk2.5S. I started with the Colorfabb testing process printing squares, etc., and came to the conclusion that 50% extrusion and 240C were appropriate. I then found this article from 3dLabPrint (https://3dlabprint.com/faq/prusaslicer/) and they have profiles already built for Prusaslicer with very similar settings. I'm sure you can port them over to Cura/Ender for your needs.
I'm using this on some wing ribs and formers for my V-22 and it seems pretty strong and lightweight. For instance, I had previously carved the leading edge from balsa, but this go around I printed the leading edge in 3 sections and glued it together. I think it's accelerated the build process with comparable weight. A single perimeter is very flexible and could be easily damaged. Layer adhesion looks strong, though. A small amount of infill can really add strength without adding much weight.
My V-22 v2 wing weighed about 167g with just the carbon spar, balsa, and 3d printed (PETG) ribs (no electronics). This new version with carbon spar, PLA-Pro, PLA-LW, and balsa sheeting is about 156g. We'll see how it all holds up when China post delivers the other carbon tubes I need to complete the other wing half for v3.
I'm using this on some wing ribs and formers for my V-22 and it seems pretty strong and lightweight. For instance, I had previously carved the leading edge from balsa, but this go around I printed the leading edge in 3 sections and glued it together. I think it's accelerated the build process with comparable weight. A single perimeter is very flexible and could be easily damaged. Layer adhesion looks strong, though. A small amount of infill can really add strength without adding much weight.
My V-22 v2 wing weighed about 167g with just the carbon spar, balsa, and 3d printed (PETG) ribs (no electronics). This new version with carbon spar, PLA-Pro, PLA-LW, and balsa sheeting is about 156g. We'll see how it all holds up when China post delivers the other carbon tubes I need to complete the other wing half for v3.
2jujube7
Well-known member
I would not recommend trying 3D printed airfoils. Early on I tried to get them to work with regular PLA, but a 40mm chord 140mm length came out at 14g even when I did a single wall and minimal ~8% infill for rigidity (Even LW-PLA airfoils would still end up being a significant percentage of this weight). In comparison, my 50mm chord 150mm length XPS foam airfoil is just about 1.1g. Even if you can design it stiff enough to withstand these forces, the added ~13g per airfoil is a killer when you have to build 4 of them. That plus the stronger end plates puts the added weight well over 50g.
I'm using a hot wire cutter on ABS templates that are pinned on to the xps foam. Now that I've made around 20 airfoils or so, I've gotten my technique to the point where I don't even need to sand them. I'd recommend trying this.
EDIT: Lol nvm
Last edited:
NickRehm
Member
50% and 240C worked great on my ender 3! Also very excited to see v3 next summer at FF
New blades come in at about 3.5 grams each, a little more than my handmade foam ones, but they're producing more thrust because they're not as twisty
Attachments
2jujube7
Well-known member
Wow that's good to hear. I guess they're much smaller than what I was doing.
That sure cuts down on production time.
2jujube7
Well-known member
We have flight footage! + bonus cyclos exploding from a short kiss with the ground
Will probably have that out tomorrow
There better be slo mo
2jujube7
Well-known member
Well, I did some more tests and it turns out that that a bunch of the different airfoils and pitches are pretty similar.
(This is with "new" style linkage, see paragraph below for explanation)
But I've been wondering why I haven't been getting the ~300g thrust that I got a month or two ago, so I switched my vector linkage back to the old version (the kind that Nick is currently using) and ran the efficiency test again. The thrust is around 50% more than the "new" style linkage that I've been using, with a negligible increase in power used. The funny thing is that the "new" style linkage just transfers the attachment of the vector linkages on the offset point bearings to separate bearings, which should only reduce friction with no other side effects. I have no idea why it reduces thrust, but it does so I'm switching back to the old style.
(This test is with the "old" style linkage)
I suppose that I should probably go through and retake all of the data about different pitches and airfoils, but use the "old" linkage in order to get more accurate measurements. One more thing that I noticed, courtesy of Nick's data, is that my rotor is at ~20% higher than the efficiency at the same 300g thrust produced by the 5" prop, and this difference increases at lower throttle settings to around 30-35%.
(This is with "new" style linkage, see paragraph below for explanation)
But I've been wondering why I haven't been getting the ~300g thrust that I got a month or two ago, so I switched my vector linkage back to the old version (the kind that Nick is currently using) and ran the efficiency test again. The thrust is around 50% more than the "new" style linkage that I've been using, with a negligible increase in power used. The funny thing is that the "new" style linkage just transfers the attachment of the vector linkages on the offset point bearings to separate bearings, which should only reduce friction with no other side effects. I have no idea why it reduces thrust, but it does so I'm switching back to the old style.
(This test is with the "old" style linkage)
I suppose that I should probably go through and retake all of the data about different pitches and airfoils, but use the "old" linkage in order to get more accurate measurements. One more thing that I noticed, courtesy of Nick's data, is that my rotor is at ~20% higher than the efficiency at the same 300g thrust produced by the 5" prop, and this difference increases at lower throttle settings to around 30-35%.
2jujube7
Well-known member
You bet.
Below is the "new" linkage that was under preforming. In order to get rid of the little bit of plastic-on-bearing contact that results from the angles of the vector linkages slightly changing, I decided to add 4 more bearings in an attempt to reduce friction. When thinking this up, I reasoned that although the inner plate would be slightly delayed beyond the rest of the rotor while spooling up, the centripetal force would pull it outwards pretty quickly once in reached operational speeds in the range of 1000-2000rpm. Super slo mo I took at low throttle seemed to confirm this, as they appeared straight relative to the side plates. However, something else funky is going on as seen in the g/W charts above. I have no guess what it could be other than a slight delay that possibly causes ~1mm shorter linkages, but even that seems unlikely because of the high centripetal forces.
v Bearings are friction fit into the plastic and the ID is screwed into the linkages.
Below is the "old" style linkage that works much better. You can see that the bearings would rub slightly on the plastic as the rotating rotor moves the linkages slightly at high frequencies. (30hz full throttle) But I guess it works really well still so I will continue to use this. I'll probably add in some lubricant though, it's been running dry and I'm not sure if that is the best setup.
Below is the "new" linkage that was under preforming. In order to get rid of the little bit of plastic-on-bearing contact that results from the angles of the vector linkages slightly changing, I decided to add 4 more bearings in an attempt to reduce friction. When thinking this up, I reasoned that although the inner plate would be slightly delayed beyond the rest of the rotor while spooling up, the centripetal force would pull it outwards pretty quickly once in reached operational speeds in the range of 1000-2000rpm. Super slo mo I took at low throttle seemed to confirm this, as they appeared straight relative to the side plates. However, something else funky is going on as seen in the g/W charts above. I have no guess what it could be other than a slight delay that possibly causes ~1mm shorter linkages, but even that seems unlikely because of the high centripetal forces.
v Bearings are friction fit into the plastic and the ID is screwed into the linkages.
Below is the "old" style linkage that works much better. You can see that the bearings would rub slightly on the plastic as the rotating rotor moves the linkages slightly at high frequencies. (30hz full throttle) But I guess it works really well still so I will continue to use this. I'll probably add in some lubricant though, it's been running dry and I'm not sure if that is the best setup.
NickRehm
Member
Ahh yes, I've toyed with that design in my head a bit and decided against it because it adds an unconstrained degree of freedom. Glad you at least tried it and that it somewhat worked with the centrifugal force pulling the linkages out, but even the slightest lag from that center bearing plate from the main rotor will throw off the blade pitching and hurt thrust
I use this stuff for lubricating PLA/PETG: https://www.amazon.com/dp/B00C5014K8/?tag=lstir-20
My control linkages don't have bearings at the blade side, so I used the lube but beware: it gets flung off lol
I use this stuff for lubricating PLA/PETG: https://www.amazon.com/dp/B00C5014K8/?tag=lstir-20
My control linkages don't have bearings at the blade side, so I used the lube but beware: it gets flung off lol
2jujube7
Well-known member
Haha yeah I saw your picture of your 3D printed rotor a page back, and those did look kinda sketchy.
