Video

260g dry, 370g with batteries.

@@wowthatflies Nice. That is about the minimum weight I got to, but I sacrificed weight on other components to add battery capacity. Probably just 5-10 grams. We end up at the same result. Have you solved the transition to forward flight issue?

I'm still in the coding and tuning phase. After this video I rewrote the mixer (and liked the result) and ran autotune on roll, which took two battery sets to converge and had some minor strange effect at the end anyway. Once I am satisfied with the tuning, which will take maybe two to three more sessions, I will proceed with some FPV tests, which will include, at some point, an equivalent of forward flight. Currently I don't think SN1 is ready for it.

@@wowthatflies Thanks for the info. I presume this all is related to your shrouded propeller? Is that really a crucial element of your design that you cannot do without?

I would argue that the lack of spare time contributes more to the slow progress than the duct :) You see, videos are typically uploaded a month after the event happened, and this is not because it takes me a month to do them...

It flew, and there is a video of it on the channel. But because of my move to a different country, I had to leave it, and now it likely attracts visitors in one of local drone shops. Its successor, CX7, is currently active, and there are quite many videos about it on this channel as well, including an FPV video that did not lead to a crash...

I can easily imagine attaching a 3D-printed head to the top of this. With some lights added, it can make a nice Halloween trick. The question of proper clothing is less trivial though... For the monospinner, I liked quite much the extension of this concept into a "large propeller" by Nicolas Rehm. Currently I'm not quite interested in vehicles with reduced degrees of control, but who knows: the more you build, the more you want to build :)

My other active vehicle, CX7, is even more like the Hiller Flying Platform, and it even has some similarities in aerodynamic deficiencies (which I hope to overcome, which the original platform could not do).

This one has two test programs in plans. The main one is about testing different arrangements of flaps (more specifically, related changes to control algorithms). The one that will likely follow is about different flaps themselves, like how much worse a flat flap is compared to an aerodynamic one, or how grid fins behave in comparison with single-blade flaps. This is all somewhat conservative because a hard crash would effectively result in rebuilding the entire vehicle, which I would like to postpone as much as possible. What follows after these test programs, however, can be much more risky and exciting...

@@wowthatflies can you make a more flexible frame to absorb impacts and keep the stiff flaps? Or would there be added issues with stability?

In the flight logs I already see some resonances related to either flaps or the frame. A flexible frame would only get it worse. A better solution is to separate structural strength and shock absorption by adding some soft landing endpoints to an otherwise rigid frame, like I did on two other vehicles. This one just has too many legs :) but quite likely I will do something like this here as well.

Its own wet weight, in the flying configuration with big flaps and extended legs, is 840g. The second section of the video featured an added weight of 540g, which still lifted off easily. The motor test data says on 6S the maximum thrust of this motor/propeller pair is around 4.4kg, so assuming cosine losses for 45 degrees and a bigger wet weight of 900g, theoretically the payload can be up to 2kg. The current frame design was not made for large payloads though (and even not for hard crashes), it's mostly a test vehicle for modifications of control algorithms.

@@wowthatflies if it's 3D printed means better try LW-PLA ya

I would rather say PETG-CF, since unlike planes it cannot handle loads by the whole volume - as there is no particularly big volume. Except for flaps: flaps should be ideally made of LW-PLA indeed (or fiberglass for hardcore makers). Ideally, the top assembly that supports motors and directs loads to legs should either be made of carbon or be designed in a more rigid way (in particular, to be non-flat). And for the sake of letting in more air it should rather reduce volume than strength, so my heuristics say LW-PLA is not exactly suitable for the frame.

You would need to consider what thrust can a cooler fan generate, then think what mass budget you have, where you have to fit a frame, batteries, electronics, and maybe some additional components like servos. I think there are videos of quadcopters built with such cooler fans, which fly, kind of. Depending on what exactly fans are used, controlling their speed with necessary precision can be a challenge. Singlecopters don't seem to require as sharp and precise control as quads do, but you still want to be quite exact with throttle.

> I know quite a bit about aerodynamics and physics Let's maybe try to solve the CX7 not-exactly-Coanda problem then? (The last CX7 video on this channel is dedicated to it) I've talked to some fellow physics people about that machine. They all thought that all the problems came from insufficient control authority of flaps. The CoG is indeed not very high, and the flaps are primitive, but the stand test showed that there is something they missed, something we don't understand where it comes from. I'm saying Coanda, but the duct on CX7 is just a thin wall, literally a plastic bottle with some 4mm printed structural rings. There is no dedicated surface which increases lift but is subject to asymmetrical forces when flying in the wind - and I feel there is something that we have all overlooked. I observe this all also on SN1, which has a much higher CoG and an even smaller top surface. Could you maybe have a look?

We all learn something, all the time. For SC1 to fly well, I had to set the yaw error filter to zero. Looking back, I should have done that for my bicopters too, two years ago. And since you were actually flying the SingleCopter version, you would have benefitted from this too...

As I also mentioned in the Ardupilot discussion forum, I have a large queue of things to do as of now. Fixing the Bicopter part of Arducopter was on my table two years ago, but I had to move to a different country, and it is currently less feasible for me. If/when I break SN1, and if most of the meat is still intact, I will probably build a smaller bicopter to finally nail that task. In this case, the next step may be to convert it to a tailsitter plane, but which exactly shape it would take, is still quite unclear to me, with a lot of decisions to make.

I'm glad you liked it! This is definitely not the first singlecopter ever, and not even the first on CZcams. Maybe the most popular design is like this: czcams.com/video/LGTTuYx5jpo/video.html, also built by this guy: czcams.com/video/YdvwP_g6c2M/video.html and many other makers. My build is a little bit special because it uses single-blade flaps instead of "grid fins", and because it can be converted to three blades in a matter of five minutes. It appears to be somewhat more stable in yaw, but that's probably due to a different firmware with a different control algorithm. For the video format, I usually release a video per day of testing, but this time testing was quite tight in time, and I thought it may be more natural to make a single video for this.

@@wowthatflies thanks for sharing those videos! Now I've got a new rabbit hole to go down! 😅

It requires bigger and better control surfaces though. For pitch and roll, the force arm is from the CoG to the lift application point of a flap, and is easily controllable by where you put your battery. For yaw, it is the distance from the central axis to the same point, and cannot be changed much without starting from scratch, so only increasing flaps (to increase the force) is the way to deal with it. Once I am done with the main test program for this machine, it is likely I will perform tests on how different flap designs work. Likely, by putting the reference design on roll servos, the tested design on pitch servos, and doing some flights and log analysis.

@@wowthatflies What system would you think is more efficient, single rotor which deflects air to counteract the torque or a coaxial? I watched that video you suggested on the conada effect of ducted flying platforms. Made perfect sense. 👍I see this model isn't ducted so won't be affected by the coanda effect.

SC1 does not seem to have this Coanda problem indeed. Generally, a single rotor, at least on a smaller scale, should be more efficient - per unit of power - than coaxial rotors, because the efficiency of the lower rotor is smaller as it has to work with turbulent air. But other design considerations turn that into a fight of engineering trade-offs: like, how many additional structures you need to get a single rotor work, and how much they add to the overall weight, and so on. Per unit of mass, the single-rotor SC1 hovers at roughly 185 watts/kg, while the coaxial CX7 hovers at roughly 308 watts/kg. (These are random single-point measurements, so take this with a grain of salt anyway). The ratio is 5/3 in the favor of SC1. But SC1 uses a 10" prop, whereas CX7 uses 5" props, and the huge motor of SC1 definitely has more efficiency than two much smaller motors of CX7. So the comparison is obviously not very clear. It seems that the choice of motors and propellers may have a comparable or even bigger impact on efficiency than single vs coax. Losses on control surfaces should be smaller in coaxial designs, as they don't have to be constantly at a noticeable angle to the flow. Bigger surfaces will need smaller angles, but they will introduce more drag. A trade-off theater indeed.

Well well well, who do we have here

The remake of SN1, to be uploaded soon, is even better. Your Javelin project is also quite a cool thing, I've subbed too.

The OSD in the FPV view (with black fields on the sides) is the standard OSD drawn by HDZero. The left panel displays the info about the HDZero goggles, the right panel is about the vehicle. The overlay plots for pitch, roll and yaw, as well as the stick widgets, are produced during video editing by my custom code. This will be made public on GitHub soon-ish, but I doubt it will find any use in the current form, unless you are a true nerd.

Yeah, a way to be a little bit greener by not letting these be cracked by car tyres.

The duct, functioning as a tube mitigating propeller vortices alone, does not have Coanda effect. The upper brim, which is often added to ducts, is subject to Coanda effect, which is used to increase lift at the same motor power, but at the same time this creates problems when flying sideways. CX7 does not have this upper brim. However, it has parts that are close to where it would be, such as tops of the batteries, the flight controller, and the 4mm wide structural ring. I suspect these may function as areas where the difference in air speed due to side wind can add unwanted rotation. The resulting force is not as big as what would follow from a "proper" duct brim, but still enough to saturate flaps at 45 degrees of pitch forward. At least this is my current theory. It may be that lifting the CoG can increase control authority of flaps enough to forget about that. Or it may be necessary to design the shape of that brim so that all the air speeds are less similar at all pitches. I still don't know for sure.

@@wowthatflies Thaks for the explanaition.

It's coastal middle Wales. I guess the forests here and there have much in common.

@@wowthatflies Same trees, same plants.

Sometimes I try to explain things, in written form, to the best of my then-current knowledge. Occasionally the explanations are bullshit. I try to avoid telling plain lies, because I pretend to have critical thinking, but as I have no formal education in control systems and fluid dynamics, sometimes bloops happen. In this case, tests themselves (and synchronized logs where they exist) make more sense as pure experimental observations.

@@wowthatflies I have read and studied a great deal, but most of my work is done by guestimation. Fluid dynamics are luckily not so important on this scale and it is easy to just tinker and adjust.

@@wowthatflies Thats right, design it, build it, send it, repeat.

It is a rather straightforward modification to Ardupilot's "libraries/AP_Motors/AP_MotorsCoax.cpp". Currently it exists only locally, because this is not something production-ready, and I am not totally happy with how it works. However, I am going to generalize it and make a pull request to Ardupilot, in which case everyone can test it. Hopefully this will happen this summer, maybe even in July.

On this craft, there are two motors and two vanes, not four. From it, you cannot remove anything without losing the ability to control. (I'm not talking here about non-rigid propellers, such as helicopters or recent swashplateless designs such as HALO) You can build something with one motor and four vanes. In fact, three vanes are enough, but Ardupilot does not support that yet. I have already constructed a reconfigurable craft for improving this part of Ardupilot, which I probably manage to complete this summer. (You can easily configure this in Betaflight and maybe INAV, but they do not currently support proper output control for vanes, so you can tune it to hover nicely, but everything else will be not)

@@wowthatflies You can counteract the motor torque with 2 vanes, you don't need a coaxial motor. Betaflight supports everything you need as a fixed propeller craft with one motor and two vanes is just an airplane without a tail and wings, just ailerons. What you are trying to do is what 3d airplane flyers do when they stand the plane on its tail in the air. You just need a much lower center of gravity so the software does not have to work so hard to keep it balanced in higher angles of attack. Try the center of gravity of a helicopter. I have just finished designing a craft on the same principle for sale and expect to get racing speeds in forward flight (Just writing the build manual now). The prototype was easy to fly. I hope this helps.

For classic VTOL craft control, you need four degrees of freedom (equivalent to pitch, roll, yaw and thrust). Every time you have less degrees of freedom, you have to sacrifice something. On a fixed wing, you basically have high speed and don't care about yaw stabilization, you have thrust, pitch and roll, exactly in the way you described. If you want to reduce the number of degrees of freedom on something which can hover, it seems you need to sacrifice a lot. Following what you say, I would imagine something with an extremely low CoG, which can throttle, yaw and pitch, but cannot roll (e.g. to move left, you would have to first turn left then pitch forward). This seems like a natural extension of a fixed wing, with all pros and cons, and maybe this is exactly what you have. Personally, I would not like to fly this, but I see that it can have some use.

By the way, something like this sounds like a nice-to-have failsafe mode for single-axis copters: in the case some vanes fail during the flight, it may descend and try to keep attitude using remaining degrees of freedom and yawing when necessary. Seems quite complicated and might be tricky when yaw is also by vanes, but could be tested.

@@wowthatflies Yes, it cannot roll, but neither can a helicopter. This can be an advantage as it is inherently stable in hover. The problem I have seen with designs that have a high center of gravity is that they have problems transitioning from hover to forward flight. To hard of a stick input flips it over, because once you get over a certain tilt you lose lift. Do you need to roll? This kind of craft is perfect for cameras as you get 360 degree field of veiw with just a tilt servo, as you can rotate the entire craft. So you have a choice between putting it on top, where it has less use, or putting it at the bottom which automatically lowers your center of gravity. I am not sure what flight characteristics you are going for. Is there anything you can compare it to?

I will soon publish drawings of all the recent stuff on GitHub, so stay tuned.

@@wowthatflies next month my project submission is there 🥲

Well, just in case you will be faster with your project than me publishing everything - which is quite likely - here are the basics: - The general design is called "coax-copter", at least in terms of Ardupilot, and you can check the general idea here: (ardupilot.org/copter/docs/singlecopter-and-coaxcopter.html). - The firmware is Ardupilot, I use the master branch with a small correction (which, I hope, I can expand and get accepted upstream, maybe to one of 4.6 pre-releases). Something like this is possible to configure with other firmware vendors (though likely with caveats of not being able to control vanes the way it should be done), or, if you are really curious, you can write your own controller, although it only seems easy. - The flight controller is Matek H743-WLITE, but one can definitely use a cheaper controller, provided it can control two motors and two servos. Not all AIO controllers have pads for servos, and not all of them can run Ardupilot (not F411, not 722). - If your controller is not AIO, you need to also get two ESCs to drive your motors. The ones I use here are marketed as "Hee Wing T1 VTOL FX-25A Upgraded ESC BLS", these were quite a random choice actually. It's good if you can get DShot telemetry from them, to be able to use that for filters: if they are running Blheli_s firmware, consider updating to Bluejay; if they are running Blheli_32, you may be lucky to have a fresh enough version that already does that, but if you need to upgrade that, Blheli_32 is gone, and you have to consider an alternative firmware such as AM32. - Motors are Ummagawd Aerolite 2004 motors at 2400 kV, propellers are HQProp 5x3 two-blade propellers with a matching shaft. Servos are EMAX ES9052MD. - Batteries are 4S, and this craft flies on battery cells extracted from disposable vapes. It appears that with the shape of CX7, using different batteries will be somewhat more complicated. These cells still provide enough current to fly non-aggressively. - The remote control receiver is ELRS (an old Happymodel EP1), the video system here is HDZero. All of this is of course more like your personal choice, but ELRS is lightweight enough and good enough to be considered as a good default choice nowadays.

I can be found on Discord as "maximbuzdalov" and on the ArduPilot forum as "MaxBuzz".

CX7 runs on two Ummagawd Aerolite 2004 Motors at 2400 kV, and two 5x3 two-blade propellers, all with a 4S battery assembly. This combination was borrowed from some toothpick 5" quad configuration, and overall makes some sense, especially if you aim at an energy-efficient build able to fly on vape batteries. I tend to validate my builds with eCalc, though it is somewhat conservative, does not support single-axis machines nicely, and for smaller builds it always complains about high-pitch propellers (and probably cannot handle them well). Additionally, coaxial builds lose some efficiency due to being coaxial, and ducted builds ideally require duct-optimized propellers with wide tips, but it seems as if lightweight 5" propellers optimized for ducts do not exist on the market. I have a vehicle in development which runs on 3.5" duct-optimized propellers - it is kind of optimized for thrust rather than efficiency, and its hover efficiency is way smaller than of CX7, but, overall, only time will tell.

And yet another funny contraption, which is yet to take its final shape, uses a single 10-inch propeller, and appears to be 2 to 2.5 times more efficient in hover (per unit of mass) than CX7 (upd: actually more like 1.7). So if you want to go crazy for efficiency, choose large propellers (and smaller target RPM, but not necessarily a smaller kV motor).

@@wowthatflies thanks for the great info! Would you say then, that a singlecopter with a bigger prop/lower kv motor would be more efficient than a coaxcopter (2 motor) even though you need 4 servos vs 2 to control?

It's a big engineering question, and a trade-off theater. A bigger prop may result in more structural overheads, hence more weight and less flying time, even if efficiency-per-kg is greater. Some overheads may be also incurred by flap drag, which is higher in singlecopters than in coaxcopters at least in hover. So this is all complicated.

@@wowthatflies totally undersetandable,but would u say per unit of mass a singlecopter with 4 servos is more efficient than a coaxcopter with 2 motors + 2 servos?

It is actually quite noisy. I tend to avoid testing it at night...

The stuff that happened to that craft is precisely what I am investigating now. They used weight shift in first prototypes, but the particular variation of the Coanda effect from the duct's upper envelope is that when you are at an angle to have some horizontal speed, that speed adds a force that turns the craft back. This was strong enough to end the project even when they added flaps similar to what my models have. By the way, this is explained pretty well in this old video by RCModelReviews: czcams.com/video/0stl1U9evzU/video.html. CX7 does not have this upper envelope, though what I observed during the entire course of its life, and what I am trying to quantify in some recent tests (not yet published), is very similar to that. A yet untested idea is that something similar in shape to CX4's tunnel instead can alleviate or at least reduce this effect. This will be tested in some form in the next month or so.

@@wowthatflies Thank you very much for the information and I will watch Rc model reviews video too. 👍

@@wowthatflies Brilliant video explaining how strong the coanda effect is. I've been wondering about the downside to this machine for ages, so great to have an answer! The delackner aero cycle wouldn't have suffered the same effects with rotors that weren't ducted but rotor flap back would have had some effect. I suspect there wouldn't be any rotor flapback with the flying platform but there would be with the aerocycle, same for translational lift. Not sure if you have seen my channel but I built a full scale coaxial helicopter, the flying platform certainly seems a fun project even if I could only travel at very slow forward speeds. Look forward to monitoring your progress, models are the way to test. Do you think mass above the rotor has any stabilising effect ? If you did the same thing but pilot weight shift below the rotor would that be more or less stable ?

Your projects are nuts! And I think that technically there should not be much speed limitations, unless you resort to very small pitch adjustments. The mass above the rotor... Well, I had a theory once that the center of gravity should ideally exactly match the center of thrust. But I no longer think this way. The effect from that mass was twofold: 1) it increased the weight by 10%, which increased the hover thrust; 2) it lifted the center of gravity a little, which increased the moment arm of the flaps; and both of these contribute multiplicatively to the control authority. Whatever the levelling effect comes from in the case of CX7, even if it increased after this change, it was to a lesser extent, so overall the authority improved.

@@wowthatflies makes sense when the CG is higher the control authority is better, like a longer tail on an aeroplane. 👍 I saw this video only recently. czcams.com/video/aD6Om6LoERM/video.htmlsi=0n5i__V6lq9sGdCK Don't know what happened about it. Interesting the guy thought that coaxial's are gyroscopes. Counter Rotation cancels gyroscopic effects 🙂

I can be reached on Discord as maximbuzdalov, or on the ArduPilot discussion forum as MaxBuzz.

A more proper aerodynamic study is actually the next phase...

Sick as in seasick? :)

This craft uses a pair of ESCs that are marketed as "Hee Wing T1 VTOL FX-25A Upgraded ESC BLS". Flashed with Bluejay, of course. Batteries are 12 350mAh cells from disposable vapes, connected as 4s3p. This is obviously neither mandatory nor perfect, but you will probably want to use at least two batteries (2 x 2s) to perform weight balancing. Mounting a battery on the top is also possible, but you may want to avoid interfering with airflow, which is not easy.

Yes, I liked it, but there is still a room for improvement - which I am investigating now...

There are two vanes/flaps arranged in a "+" shape. The one that comes front-to-back controls roll: when it tilts bottom-to-the-right, the pressure from the airflow causes the drone to roll right. In the same way the vane that comes left-to-right controls pitch.

@@wowthatflies can I have yr email or any contact no? Bcs I currently do the same project like yr one, but using Arduino n mpu6050 only. I m not sure the mechanism that I codes it correct or not.

@@YAPKAIHONG_AEE I think people nowadays use Discord for this kind of chatting, and I can be found there as "maximbuzdalov".

@@wowthatflies alright. Already sent. Very Appreciate.

Still quite loud for a UFO, I think :) But this one can be quite creepy too, if met outside in the darkness...

I am not sure I am the right person to do it. There are lots of tutorials across the internet about quads and planes, which cover basic stuff, from components to soldering to piloting. I only have time to explain the difference, sketchily, between these and the "less common" vehicles I am building, and to log my own failures and successes, so that people doing similar projects can learn from that. And sometimes I do really dumb things :) I will eventually open-source the designs. However, I would rather like to first be sure that I am not publishing things that fundamentally do not work. With CX7, I think I have one more series of tests about one design change - which will start after three more videos - before I put the drawings online. My resources are, and have always been, quite limited, so I cannot be a "Joshua Bardwell" or "Joe Barnard" of coax copters. Please forgive me for that :-(

@@wowthatflies Thanks for reply. I love your CX7 design. It seems safe and innovative. I truly understand your point 100% that you are running all these experiments in limited time and resources. I am truly thankful to you for sharing your experiences along the development stages. Someday, when you are all done with tests and ready to put the results to public, I wish I'd be the first person to mimic your experiment.

Few years ago I used to autotune every vehicle strictly in this way, using an FPV system. (Check some CX5 footage earlier on this channel for few examples). However, in recent years I actually found it somewhat more productive to run that in LOS, especially in confined spaces. What is more, it's easier to crash a vehicle that misbehaves, if you fly it in FPV. It's sort of easier to get a vehicle back in FPV, but one would not get a good tune if it's too windy for a craft anyway. But, you know what, at the time of me writing it, an FPV system has already been installed and successfully tested, so... :)

Yes, it's nice to know that a craft does not get broken immediately after a tiny gust of wind :)

At this moment in time, it had not yet been added :) Spoiler alert...

Do you have any video or other resources about it? And what exactly does Arduino do in your setup?

@@wowthatflies I have no any video , I ordered project related components on online commercial website. After delivery and complete of project upload a video on yt (sorry for english)

I'm looking forward to that!