Here's my crazy idea

Have a glow engine in the middle providing most of the thrust plus the 4 electric motors providing the stability on the 4 corners. Also the electric motros could be powered from a generator hooked to the glow engine. What do you guys think, would this work?

Hardyus.

Hardyus: I love your crazy idea. It would be the most weird DF on earth. I think it could work and the flight time would be longer. But I think the main problem would be vibration; it afects the gyro readings rather badly. I installed my gyros onto a dampened platform but I am afraid that wouldn't be enough if you put a glow engine.

wouldnt it rotate?

you would have to find a way to counter it, but yes, on it's own it would rotate.

The whole point of counter rotating pairs of motors is that they counteract eachothers torque and keep the craft from rotating. If youre gonna do a glow engine, why not just have 2 counter rotating props in the center. It becomes more of a helicopter then, without a tail rotor of course. You could do 2 things i guess at this point, have small thruster motors like you said to control pitch and roll, but more likely you'd just use servos at the main rotor just like a heli. Basically this accomplishes a flying platform using counterrotating rotors, one on top of the other. I'm sure you've seen those models around, or even those post WW2 prototypes that dont have tail rotors.

One Large prop (or 2 in this case) is surely going to be way more efficient than 4 small ones, and more stable too. One thought I had is if youre gonna do a main rotor and skip the outside ones... why not put the motors at the end of big prop blades. Because the power is coming from the blade itself and not the body of the heli, theres no torque, no need for a tail rotor. It doesnt take much force to turn a blade. Pretty interesting idea really

are you talking abotu two different glow motors or are you talking about two counter rotating propellers on one crankshaft? if you are talking about having them on a single crankshaft, you will not be able to use a two stroke motor because you need to run a second drive shaft through the crankshaft that you can attatch the second prop to, and the two shafts are tied together with a gearbox. if you are talking fourstroke, that is a lot of weight and will require re-balancing of the shaft and an out of balance prop could mbend the shafts. if you are talking about two seperate glow motors, you may find that the two are not identical and produce slightly different amounts of thrust for the same amount of throttle. i think one motor and other electrics around the sides countering it are the way to go

Space, Do you think 4 electrics together could counter the glow prop? Also how about generating enough electric power from the glow engine to push the electric motors? Would that work?

Damn, no sleeping tonight...

dunno. i suppose it is possible, i dont know any statistics so i cant say yes, no, or why

hmmm I think the most feasable is counter rotating props; but then again its no longer a DF anymore, you can ditch the 4 electric motors ;-)

How could this bea complished-
2 seperate glow motors with different speed for the same throttle- simple to correct for in electronics. Probbable most feasable idea i think. Coaxial bearings and shaft setups for counter rotation are pretty simple (haha ish). Not sure about the power to weight ratio for 2 small glow engines- whether its worth ur effort.

Oh well- keep dreaming anyways

oh for generators, check out sullivan products or something like that. They make generators for UAVs. 800mA @ 4.8v for $100. Thats not gunna power @#$@# but then if uve got 2 counter rotating props u only need real small electric props for drive, and to power the lectronics. Anyway at this stage its not gunna resemble a draganflyer anymore anyways
Sweet

Danvel,

You mentioned in an earlier post that it is hard to keep the gyros from drifting.
How about using two electrolytic tilt sensors, for roll and pitch, like these: http://www.frederickscom.com/product..._overview.html ?
The tilt sensor could provide a gravitational reference to recalibrate the gyros every few seconds. I've seen these sensors in digital construction levels and they are very accurate.
I don't know what vibration and acceleration does to them, but they say they make specific ones for dynamic environments...
Anyway, let me know what you think about this idea.

i have heard that bad vibrations, either too much or at a certain resinance will cause bad signals via the peizoelectric crystal itself. unless you are going to be pulling mega gs, i would reccomend using one layer of the gyro mounting tape or mirror tape for electrics and two layers for gas. two layers lowers the frequency at which it absorbs vibration the best, and the vibration from reciprocating engines is at a much lower "tone" and much much harder than it's electric counterparts. however, if you have too much mounting stuff, the gyro may flop around just enough to cause bad signals as well, so be ware. there are a few commercially produced gyros out there with almost no drift at all, such as the futaba heli gyros. i have used these in the past and i was extremely impressed. owever, calibrating the gyros every few seconds would be a great idea providing that the vibrations do not alter any re-calibration signals because if it does it will throw the gyro way off and it will re calibrate under the false reading coming from a vibration pulse affecting the whater thingies you mentioned for gravitational referance.

Hi All,

This vibration problem is a bit of a mystery, when you consider that the original DF has no physical damping of it's 3 Tokin gyros. I wonder how they've solved this problem ?

Cheers ADI

Thats exacly what kalman filtering is used for. Minimizing noise and drift.

Also the electrolytic sensors (generally) do not play nicely in a dynamic reference frame. The bubble type arragnement they use to sense movement .... moves.

Id recommend u guys actually spend 12 seconds looking into kalman filtering...................

Yes ... but why would you NOT use some form of physical dampening, in order to increase the gyros' sensitivity ? It's a bit like NOT using an elastic type basket to hold a studio microphone on it's stand, but instead, rolling off the bottom end in software after you've recorded all the low frequency bumps eminating from the floor via the mic stand.
How much assembler code are we talking about with Kalman filtering ? How complex is it ?

Cheers ADI

I would definitely try to stay away from glow or any other type of combustion engine. The only type I would even have considered would be Wankel engines, as they probably have less vibration because they don't have a piston. AFAIK it's rotary all the way?
Every glow person I've been in contact with, airplane or heli, who tried to record video or take pictures said that the main problem is vibration. As electric motors are vibration-free it seems to me to be the obvious choice. After being out of the !QUOT!loop!QUOT! for the last 2 last years it seems to be even much more so now.
Now we have access to relatively safe LiPo cells, brushless engines are much more affordable and the price of their speed controllers is also getting more affordable.
I'm considering using brushless instead of Speed480s, but as I already have the 480s I suppose I'll try with them first. I need another couple of speed controllers though..
I believe in big props, but I guess that will also cause it to be more sluggish in its responses?
We want thrust and not speed, which traditionally calls for more diameter and less pitch.
At the moment I'm searching the net for good tutorials on making your own carbon props. There is a lot out there.
I haven't decided yet if I should go for rotor blade single blades type, or for one-piece props.
Any tips/advice on this?
I guess it would be very useful to be able to control the pitch to get them balanced properly?
Does anyone know why rotor blades seems to have the same angle all the way along the blade, while props always twists a lot, so that you get a lot smaller angle at the tip? Why isn't that necessary/desirable on rotors? Wouldn't this mean that a rotor will have much more lift at its tips, and thus also be more likely to have their tips stalled?

And how about using ducted fans? I don't really think that any of the ones available on the market would be useful, because they all seem to have quite small diameter and a lot of pitch, making them useful only for high speed airplanes that need to be launched with a catapult.
According to what I've read about ducted fans a well shaped duct can increase the thrust with up to 30-40 %. That certainly would be a welcome bonus, and very useful for our DF-type flyers The problem is that such big ducts would definitely have to add quite a lot to the weight, especially since to get a good effect it needs to be a pretty precise fit, which means that it have to be stiff (carbon?)
But at the same time they would also protect the environment from the dangerous props, as well as protect the props from damage.


It's a bit surprising that there seems to be no interest in discussing the usefulness of such a device as the Co-Pilot.
It seems to be able to do the job of 2 gyros?

Hmm interesting analogy. I agree that you should do it physically where possible but damping high f noise from a glow motor running at several thousand revs is gunna be especially difficult (to do mechanically). On the other hand there are many other approaches (maths, software, etc) which;

1) Consider that high f noise (above some threshold say) could not possible be coming from the gyro (as these only updata at like couple hundred hertz maybe (im just guessing))
2) Consider low f drift and adjust accordingly

Kalman is not the only way, and its probbably not the absolute bestest way but it wrks and is (relatively) easy to understang. Go to rotomotion.com where they sell IMUs and give avay Kalman C code. As for assembly- hows ur floating point assembly skills ;-) ???

Also the main thing is the more you try and do with these Df's (in terms of autonomous flight control, hovering, landing, UAV etc) the more important the sensors. Absolutely and totally spend 90% of your time eveluating the best sensor package that gives you the data that you need. I believe that means NOT re-inventing the wheel- universitys, hobbyists, companies, have done all this stuff before; use their advice in terms of sensor choice (IR vs sonar vs gvro/accel etc). There is infinate flexibility in what you can accomplish in software with the right data. This means (if ur interested read this)

Consider the maths and physics, its essential because then youl realise that to control certain parameters; position (used for landing etc, path planning), acceletration (hover, landing), velocity (hover, landing, path planning), you must have the sensors in place to sense this.
Anyways thats the last im gunna say on the theory of it really (unless someone asks me).

good luck

Hello all!

I am a 3rd year Mechatronic engineering student at Curtin University in Western Australia. I have been reading this thread for the last week, whenever time has permitted, and have found it highly interesting.

Like a drunk at an AA meeting: "Hi, my name is Josh, and I'm addicted to quadrotor VTOLs!" AND proud of it!

A couple of weeks ago, I decided to base my final year project (start of next year) on designing a quadrotor VTOL system. So I stumbled across this thread in my curious researching endeavours.

Very impressed with everyone’s knowledge and generosity to help one and other, particularly being from all corners of the globe.

Spaceclam, your tha man! 14 years old, with the knowledge and concepts of an undergraduate engineering. Your gonna make a great engineer one day if u stick with it. You’d make a good uni lecturer too.

The electronics guys here have already helped me out a tone, as I will also be using an ATMEL AVR microcontroller. I look forward to conversing with you all though next year as my project gets underway.

The principle behind my thesis is to design and build a VTOL which can also “fly� (winged) horizontally. So to begin with, I will have to master the hovering and manoeuvrability of a quadrotor system in hover mode. Because this vehicle will be unmanned with visual feedback, it will not require a horizontal passenger cavity, thus eliminating rotor tilt systems, like previous VTOLs capable of horizontal flight.

I would love to get into more detail, however, I have a tone of reports due and exams coming up, which I must concentrate on. As soon as my exams are finished in late November, I will begin the design phase, and become a regular on these forums.

Look forward to bouncing ideas of you guys!

Until then, happy hovering!

Josh

you know, you could put a wing on servos, (try digis because you can slow them down) put one motor on each tip, and simply use what would be ailerons to control yaw because they would be in the propwash. you could just use one vtail to controlthe motor speeds. however, 4 rotores would be more dificult. what kind of design are you thinking about that would allow 4 rotors to fly in propetual forward motion like a regular fixed wing aircraft? two on each wing?

Not a bad idea Spaceclam, but I'm actually trying to avoid as many mechanical parts as possible. The idea is, it will be a light-weight tail-sitter setup (i.e. a quad rotor plan that takes off vertically, thus can also hover). When coming time to design (maybe build, pending sponsorship) the final VTOL, I plan to incorporate ducted electric rotors, most likely in a squished x-formation, with wing(s) parrallel to the horizon of the x. I'm not ruling out a + formation, however I have a gut feeling about greater stability with the x in horizontal flight. Will all be revealed in the maths! Also, because I am using a microcontroller, and making my own feedback, the x system isn't too hard to implement in code.

Possible layout looking down while hovering (or its flying towards you [X(]):

(X)\ /(X)
====================
(X)/ \(X)


or maybe:

=======(X)==(X)=======
|| ||
===(X)==(X)===

Won't know until I start the aerodynamic modelling.

Could even go the + layout like this:

(X)
=======(X)==(X)=======
(X)

By avoiding mechanical rudders, it significantly decreases weight. I must admit though, the control algorithm, the aerodynamics, and the transition between modes are going to be a challenge, but that’s what makes it fun!

To start with, I am going to make a Dragonflyer, no ducts and no wings, to get familiar with the gyroscopic control.

Anyways, I'm going to try avoid spending too much time her until exams are finished… so much work… so little time!

Catcha guys!

however, when you alter the speed of the motors on a certain axis, that axis has a tencancy to become somewhat unstable. having small control surfaces in the wings would avoid all that. you could even hook them up via pull-pull to a single micro servo.

Dammit, can't keep away!!! haha

Didn't realise multiple spaces are ignored... try again.

Possible layouts looking down while hovering (or flying towards you):


Good point about the stability, but hopefully my electronics and programming can rectify it. Otherwise I will have to have variable control surfaces.

Cheers, l8a!

Hardyus: I looked into tilt sensors but they are useless here because they can't tell the diference beetween gravity and acceler ation. If you draw the forces in a simple diagram you will realize that the tilt sensor will always think that the DF is horizontal because; whenever the gravity is tilted, it creates a sideways acceleration that counteracts the tilt exactly!. Tell me if you want me to draw the diagram, i dont' have the time right now.

ADI: I agree with you, it is better to do some physical damping first if possible. However, I'm sure that the original DF uses software or/and hardware filtering (as they use analog gyros a simple analog R-C filter could work). However, I never said it's impossible to fly without the dampers; but the movements are somewhat more "jerky".

Another thought: For those of you that use off the shelf motor drivers; they normally limit the acceleration of the motor to preserve the gearbox. This is in fact acting as a low pass filter because very rapid transients are not allowed in the motors.

[sm=tired.gif][:@]

Danvel Wrote:
> tilt sensor will always think that the DF is horizontal because; whenever the gravity is tilted, it creates a sideways acceleration that counteracts the tilt exactly!. >

This is precisely what happens. I've tested the Analogue devices ADXL202 accelerometer chip.
They're great for absolute tilt angle sensing when they're stationary though.
For anyone who doesn't quite understand how these work: If you imagine a small weight anchored at each end with a metal spring and then picture what happens to the weight when you tilt it while stationary, or accelerate it forwards or backwards.

__\\\\\\\___|Weight|___\\\\\\\__

<<------Direction of Travel------->>

Hey Danvel ...... got some cool pushpull rail to rail output voltage comparator chips for driving FET gates. Check out National's LMC7211 SOT 23-5 package. They sink fully to ground and source to top +ve rail.

Cheers ADI

Does anyone have some suggestion on how to attach the gws motor mount to the carbon fiber rods? Space, i see you're using the same one, how did you mount that thing?