Hi Everyone,
Heres a "little" project I started a little less then a year ago. It all begin with design requirements of a quadcopter with a 10 pound payload for a one hour flight time and ended up with a gasoline/variable pitch build. At the time and since then I've come across many blogs, discussions, pictures, and videos of people who have tried this kind of build and only one or two videos of anything remotely proof of concept ( a nitro and an electric single engine/motor variable pitch build that simply made it off the ground). Since as of 10 months ago I did not even know that diydrones.com existed I decided to wait until I could show up with my proof of concept. To the best of my knowledge this is the first successful gasoline powered quadcopter. I am a mechanical engineer and am working with a computer scientist and computer engineer who have their own equally ambitious plans. I have plans on selling these to farmers and other researchers some day but in the mean time I want to to lend my experience to anyone attempting a similar build and would like to see all those people who have started these builds to be able to finish.
This picture was taken on 4/17/2014 and was my first full system test. It was more then happy to to fly at the end of it's leash (which I made too short for it to hang itself) for 20 minutes before I landed because I smelt something burning. The burning turned out to be the clutch pads, which were entirely gone. Must have been too high pitch for not enough throttle.
My second test was less impressive. Lets just say that there are dozens of bolts that individually keep bad things from happening and it only takes one missing Loctite to vibrate out and let the bad things happen. I'm at least proud to say that after test two I am only waiting on two pairs of rotors in the mail.
At the moment This is what I have:
Control system- Arduino Uno and potentiometer with a spool of speaker wire in between (remember, mechanical engineer)
Engine- Zenoah G290RC
Gear Reduction- 17:57 (calculations based off having equal rotor tip speed as RC helicopters with the same engine)
Rotors- 325mm symmetric carbon/glass fiber main blades for the T-Rex450 (as cheep as I can find)
Rotor grips- Tail pitch assembly from Align 700/800 series
Servos- Protek 100SS (both for pitch and throttle)
Power Distribution- miter gears and torque tubes
Gas Storage- Two Danhakl Design tanks 32 oz each
One way bearings in each arm
Plans and ambitions:
3DR Pixhawk control hardware (my friends are scrapping the firmware) -or- RoboVero and a Gumstix (we have both hardware but are probably a month out from having a respectable control system).
Fully autonomous system
Self learning algerithiums
A foldable design that fits in a 2'x2'x4' box with 5 minutes in or out
Multiple test flights around an hour long
Auto-rotation
The ten pound payload
Individual rotor RPM sensors with Hall Effect switches
Sonar distance to ground sensor
50 miles per hour
Thoughts, comments or questions?
Comments
Took a quick look on eBay and found some of those 450 class stretch blades. These will be geometrically straight across compatible with my current build. My frame is unnecessarily big at the moment and will easily fit the 500 sized rotors too. I was planing on cutting it down but I think I'll hold off on that for a while now. I like the 425mm rotors due to their availability and price. The 8mm head would defiantly be a "more work to do" but I should be able to do it without giving up my one way bearings. The 425 mm blades I saw still had a 3mm bolt.
Does anyone know how thick the main blade grips are for the T-Rex 500 Class?
I'm mostly worried about gearing. I'm at 17:57 which is flying me in the low 3k RPM. I think I'll really need to drop my rotor RPM of 2500 if I go up to the 425mm blades. I'm using baja racing gears which go as low as 16:57. I only went up to the 17 tooth pinion because I have a weakness for helical gears and when I was looking there was a sexy helical set only available in the 17:57 ratio. I think since then they have a helical 16:57 but it's not like that will be a noticeable difference. I don't trust the regular helicopter gears because even if I could find my ratio they wont be designed for my loads. This leaves me to wander the ground pounder hobby sights or source industrial gears or probably mix and match and hope they fit together. It looks like modulus 1.5 gears are the standard for this kind of power in the hobby world. I should be able to hit just about any ratio with a hobby pinion and industrial gear but it will come with a gearbox weight penalty.
Am I right to shoot for 2500 RPM with 425mm blades or should I optimize it for efficient flying at heavy loads one way or another?
So maybe I can still claim a more wind resistant build (if only a little). I should take this someplace rather windy when I'm done and tell it to hold position to see how well it does. Thanks for the explanation of the rigid/nonrigid disk. According to MIT servo control is faster then ESC control. (http://acl.mit.edu/papers/GNC11_Cutler_uber.pdf) The article is kinda long but if you just look at the figures and read the captions it is pretty clear what is going on.
So, just an observation here. You're at 20lbs AUW, and with 4 700mm Dia rotor disks, your total disk area is 1.54 m2. An 800 gasser would typically weigh about 12-14lbs AUW with fuel, and has a rotor disk area of 2.55m2. So you're at a significant disadvantage compared to a conventional heli. This is an example of why you're on a difficult path. Once you decide to use a single gas power plant, all of the sudden the disadvantage of the quadcopter frame layout begins to come out. It's exposed because you no longer can benefit from the availability of high efficiency fixed-pitch propellers.
You should be able to help yourself a little bit if you go to "480 stretch" main blades. These are 350mm long rather than 325mm for a 450. This brings up the disk area to 1.77m2, which will help a bit. The blades otherwise fit like a normal 450.
Ideally, if your arms are long enough, you'd be better off with 425-475mm blades from a 500 size machine. But, these have different root thickness and grip bolt, so won't fit in your existing mechanics easily. You could simply install a 500 size head, but they use an 8mm shaft, so more work to do.
On the topic of stability:
There's a lot of bad guesses out there. Few people have flown both helicopters and multirotors using the exact same control system. I have. IMO, multirotors have an advantage in light wind conditions. This is because the IMU is connected directly to the frame, and the actuators are also connected directly to the frame. So the IMU has a more direct sense and control over the dynamics. Yes, you have to accelerate and slow the rotors to change the thrust for stabilization, and that does take some time. But don't forget that it takes time for a servo to move too. Then you also have to consider that a helicopter rotor disk is not normally rigidly connected to the frame. There's a least a couple of degrees of freedom there. So when a gust hits the disk, first the disk moves, then it drags the frame along. The IMU cannot sense and respond to the disk movement. It can only see the frame movement. Then it must "catch" the disk movement which has a few milliseconds of head-start.
And then it just gets worse from there, no point getting into it here.
Luckily, helicopters are aerodynamically less affected by large gusts, and have much more control power than a multirotor.
5mm... There are hundreds of 5mm bearings and hundreds of 6mm bearings to chose from. Not so much with gears, especially for my size, strength, and operating speed. Guess if I'm forced to go down this path it will be easier to have someone lathe the end of my shaft down to 5mm.
Thanks for the measurement
I just measured the main shaft of my Trex 450. It's 5mm... You could change the bearings which hold the shaft, and you'll need to find a gear that fits.
Interesting thing that they try to increase efficiency with bumps. But I think this is a long way from appearing on your quad ;).
Ok, I'll stop mentioning better stability over a traditional helicopter. And for the outward sling topic I guess it could go one of three ways. One- everything works just fine forever. Two- everything works just fine for a while before bearings and possibly other parts wear out and need scheduled replacing. Three- I took a second look at 450 main heads and they don't look that much more complicated to retrofit to one point of control. Does anyone know how big the main shaft is ware it attaches to the rotor head on the 450 class? if it's 6mm fewer things would need to be changed.
Back on the rotor form note has anyone heard about the study that reverse engineered a humpback whale fin and applied the bumps to a wind turbine? of http://www.stle.org/assets/news/document/techbeat_tlt_12-08.pdf and http://www.scientificamerican.com/podcast/episode/005e4190-9de4-32a... Apparently they increased stall pitch by %40 along with doing a variety of other good things. Wounder how long it will take the concept to catch on in rotors. Hopefully the speed difference doesn't nullify the concept. I believe that there are a few rotors out there with swept tips and other special geometry that is hopefully other then asymmetric.
Wow 30 degrees is HUGE!! A typical (model, 450-size) helicopter has 12 degrees of pitch (both positive and negative 12 deg), or maybe 14 degrees if you're really badass. I'd say go to max 12 and see if it'll hover. If not, larger blades! Make sure though that the blade grips are up to the task. Trex 700/800 tail blade grips may not be able to handle both the thrust and outward-slinging-power (what's that called in English?). They're definitely not designed for it.
About stability: what your friends say is (pardon my words) BS. Typical multirotors have fixed-pitch props. In order to control the attitude of the multicopter, these props have to spin faster/slower all the time. This takes some time. A collective pitch helicopter changes the (cyclic) pitch of the blades nearly instantly, making much faster (and therefore smaller) corrections. This is all gyro-controlled, so the speed of the correction that need to be applied is mainly dependent on the speed of the servo.
The amount of wind that is 'catched' by the vertical disc (the tail rotor) is also controlled by the variable pitch of that rotor, controlled by a (usually) ultra-high-speed tail servo (again coupled to a gyro)
Sorry, traditional, fixed-pitch multicopters are definitely less stable than collective pitch helicopters. For this variable pitch quad (or Curtis' Stingray) this is ofc a different story.
AUW is about 20 pounds including half a gallon of gas which is a kilogram of the flight weight. By my math this will let me fly one hour. I have not had the chance to run that much gas through this thing yet my but my first test proved that the felt clunks in the gas tanks were quite capable of sucking every last drop of fuel into the engine. Based on how long I felt it ran and ho much fuel I think I put in it I believe my calculations are reasonable. I am posting between my proof of concept and fully operational product so I have outside suggestions now as opposed to what I should have done weeks ago if I waited until I was fully functional. The flip side is that I really don't have much for tested performance numbers.
My statement that quads are more stable then helicopters especially in wind is based on what I've heard from others and the fact that even though the primary lift from both airframe is a horizontal disk the helicopter has a vertical tail rotor disk which would be more affected by uneven gusts. I really have no hard source if anyone can point me to something other then opinion on this topic they should point me to it. I think that any 20 pound airframe will be sufficiently stable anyways so admitting it's not the best of selling points.
I think the biggest hurdle with gas engine builds is simply the size of the engine. There is very little in terms of helicopter capable gas engines in the small scale so a builder might as well as go up a hair and get a good chunk of power to play with too. Now the builder has to transfer that power and lets just say that I'm not using the hobby grade plastic miter gears found in all the other non-belted attempts.
I think my pitch may be as high as 25 degrees in this picture (based on the position of the servo arm and I know that the geometric limit is 30 degrees), which is what I'm blaming my burned out clutch on. I'll be running higher RPM and less pitch from now on. From the research I did back when I too came to the conclusion that Stefan posted. Four blade rotor heads pack a stronger punch per footprint but two blade heads are more energy efficient, epically if you have two slightly longer blades. I believe it goes back to less disturbed air being able to provide better lift then more disturbed air. After I get everything functioning reliably I probably should try out some variations. But from the pictures I saw any 450 main head would be rather difficult to incorporate with multiple servo connection points. All things can be done but a tail grip from a bigger class helicopter with one control point is much easier to work with. I guess I should experiment with longer rotors before I cut my arms too short. Below is my collective configuration at 30 degrees angle of attack with a 6mm shaft.
Thanks for the discussion it will help me drive this design home!
Hm I checked some flight logs from a while back, and you're right! The asymmetrical blades do not appear to be more efficient (headspeed 3200 rpm on trex 450).
So how about flat bottom blades? These are used a lot in model autogyros. With symmetrical blades they just fly horrible, from what I've heard from a fellow pilot (I'm an autogyro beginner).
Also, blades which have stronger undercamber at the root vs the tip would give a more equal thrust, like a propeller does. For example, fixed pitch helicopter main blades? (as long as this huge quad is not going to fly inverted ;) ). I'm not sure about that though, because these blades will still have significant drag at 'zero' pitch when no lift is produced, so it may mess with the yaw control.
Stefan, I've done a fair amount of work with the Spin Blade asymmetrics. What I'm finding is that they are not all they're cracked up to be. The theory that they are more efficient doesn't pan out. Rather, the theory is too simple to reflect the truth of the situation.
It is true that they have a high peak Cl/Cd. But the peak is sharp, and it falls off fast on either side of optimum. And the thing about a helicopter, is only a small section of each blade is actually operating at peak Cl/Cd. Net result is that while standard blades may produce less total lift, they also produce less drag. Therefore, the actual efficiency is more or less the same.
Also, the Spinblade Asym's have a really nasty flat ridge along the leading edge that destroys the laminar flow right from the start. That doesn't help the case at all.
This is cool! I have 4 years of helicopter experience, and limited tricopter experience.
The thing with multirotors being so popular, is the simple mechanics compared to that of a traditional helicopter (with swash plates, tail drive gear, etc). With modern gyros, multirotors are easy to control.
Making it gas powered and collective pitch, does not make things mechanically simple anymore. But it's friggin' cool! I'll give you that :).
325mm blades for 450-size helicopters are designed to run in the 3000-3500 rpm range. However, if you want to fly longer/make it more efficient, you will want a lower headspeed. But not so low that extreme pitch is needed to fly... so you'll have to do some research here. A 450 size helicopter weighs about 800-900 grams, so that's what these blades are designed to carry. Times 4, this is less than the 4.5 kg you wish to lift, so you may consider having slighly larger blades.
I read a comment about multi-bladed rotors. Generally, the less blades the more efficient the rotor system, because the blade feels the distorted airflow of the previous blade. Especially at high head speed. So yes, a single blade rotor would be best, but then there are balancing issues... So I'd stick with the 2-blade heads.
Futher, to increase efficiency, you could consider semi-symmetrical blades (SpinBlades makes these). These are more efficient than fully symmetrical ones (but also more expensive). Maybe nice further down the road to optimize your system when it is working nice and smooth. Be sure you can find left-hand AND right-hand rotating blades... For symmetrical blades there is obviously no preferred rotation direction.
-
1
-
2
-
3
-
4
of 4 Next