RinksCustoms: Yes, thats the idea. Now with the LED strips around the edge to flash various patterns and some nice colors and a good video camera, I suspect a lot more UFOs will be sighted LOL. A PropUFO (TM).
BTW 2lbs sounds a bit heavy - i would think we could get away quite a bit lighter than this. Instead of paint or gelcoat, the colored heat film used on rc plane wings should work - lighter and able to use multiple colors.
If only they weren't so dangerous, laser leds could be used with good effects :shakehead: Just how difficult is it to project a small hologram underneath - a miniature green man LOL
PFO or PPFO
It would probably weigh alot less than 2lb empty. Though not shown in the pictures, i would fibrglass in balsa, thin model aircraft ply, or single layered (pre cured) fiberglass sheets in between the ducts to act as PCB mounting. Also i was considering using Nomex "Honeycore" as a reinforcement within the inner half of the structure. Also, two thinner coats of gelcoat could be used on the outside with single layer glass & selective carbon fiber or carbon veil reinf of the two halves of the shells.
the laser led comment -i knew you were heading for a 50's B movie design with this! LOL
@ Heater - while the UAV will be relatively light with respect to surface area, you can compensate for most turbulence by way of weight distrubution, ie; put the heavy battery packs and ESC's in the outter compartments near or at the circumfrence. This gives the effect of increased stability in turbulence but has the drawback of increasing roll rate response time and if not accounteded for in software can shorten flight times by making the motors work harder trying to keep up with the gyros outputs. For most tasks some drift is ok, for video survelance, your craft can wobble some in the air turbulence as long as it compensates for the change in attitude quickly enough and using a somewhat averaged gyro output to stabilize the shell while using near raw gyro rates to stabilize the camera. What i'm trying to say is to avg out gyro/G sensor rates to control the ESC's in a way that is proportional to the saucers mass relative to the response rate to that particular gyro direction. In contrast, the camera having much less mass than the saucer can be stabilized with a much shorter avg or raw gyro data. Digital servos respond best to this and a high quality gyro like the digital ones avail for model heli's are often paired with certain digital servos to give best drift vs rate correction perf.
With respect to power savings, instead of constantly varrying the speed of the motors and thus rotating mass (which bleeds energy from the system and generates heat under braking -yes you can setup most aircraft ESC's for braking), why not instead use constant speed props? ESC's are controlled via a single PWM signal from the prop and is only varied according to loading of the motors and/or flight mode (hover vs aerobatics), the control comes from varying the pitch of the props instead of varying the speed of the rotating mass of the rotor mass + prop mass. There's a hidden benifit from this way of control, kinetic energy stored in the rotating mass of the prop and rotor which can be translated to english as really fast transient response of the craft as far as control or sudden movements are concerned. It would be pretty cool to see the saucer config'd with battery packs in the center for faster roll rates and to see this thing do somersaults in mid air! like a coin spinning on its side on a table but spinning horizontally in the air instead!
Constant speed prop setup may cost a bit more but i think its a better choice for control and power savings over just relying on the ESC's. Your thoughts?
Hmmm... So it's worse than I was casually thinking.
The body shell:
1) Greatly increases surface area and hence susceptibility to turbulence.
2) Increases the crafts moment of inertia with all the consequences you described.
3) Increases mass and hence decreases flight time and rate of climb.
If any of that outweighs the sheer coolness of the finished machine is another matter:)
Distributing the weight away from the centre just does not seem right to me. The resulting increase in moment of inertial may help dampen the effect of turbulence on the shell but it also has a negative effect on roll rate control and energy consumption as you noted.
I'm not at all up to date with modern model engineering so are there such things as variable pitch prop systems?
Is there any mileage in using some kind of directional vanes under the propellers (perhaps entirely in the ducts) that redirect thrust for stabilisation rather than motor speed or prop pitch control?
Yes there exists constant speed prop setups for models, it requires the use of a hollow motor shaft, or in the case of this saucer can be implemented from the top side since there is/can be structure above the tip of the props. Trust vectoring implementations tend to be a bit slow responding and kind of complex mechanically, nFINS / duct x nDUCTS x nControl axes for fins = total fins, hinges, linkages needed for control. If you had 2 fins per axis (X,Y) each duct would need 4 fins and two servos then, eight hinge points would need to be installed for the 4 pins and two linkages to link eack two fins with its respective servo. Thrust vectoring via a rotatable tube having a 45 degree swivel in it is less complicated than the fin idea but response time would probably be pretty slow compared to a constant speed setup. And after all that, you still need to vary thrust individually. In essence the best opperation for this quad copter is to mimick the opperation that of a heli. By using variable pitch props you have effectively 4 helis on the same chassis, each prop acting like the main rotor baldes of a heli and by varying the pitch, thus thrust on each duct you get your attitude, altitude, and direction movements. The simplest design will generally be the best with respect to reliability and robustness, any efforts to implement a more complex design will usually yield better efficiency and power but at the cost of reducing reliability and robustness.
In short, K.I.S.S. works best and just using the ESC's for thrust control will work well, i think if a constant speed prop could be implemented it would be worth the slight decrease in reliability and robustness since the basic design in the pictures would be very reliable and robust initially. Reliability decreases proportionaly as moving parts count increase IMHO. Constant speed props is a good, safe tradeoff between reliability, liability, and the pros (increased eff, control).
It is a shame the reaction time (motors included in this) wouldn't be fast enough to use all motors rotating in the same direction causing the saucer to rotate while flying. That would be so cool!
So why can't two opposite motors run at high thrust and the other two opposite motors at lower thrust. The first two provide the bulk of the lift the other two just maintain stability. Now there is a net rotational reaction and the ship starts to rotate.
As I understand it, the front/rear axis has the motors turn one way (say cw) and the side axis has the motors turn the opposite (say ccw). This is to prevent the copter from spinning. Now to move forward, the front motor slows, and the rear motor speed up by an equivalent amount, causing the copter to dip in the front and hence move forward.
AFAIK, if the copter spins then the requirements of the motors will vary way too fast to actually control them.
Some have tilted the motor mounts the cancel the rotational force and so use all motors rotating the same way.
Maybe it's doable but it's just too complex - if so, then hey, we have seen the impossible here before. Who said you couldn't emulate a Z80 fast enough with a prop??? Or USB???
Ah there is my problem, I can't tell which is the front:)
It seemed natural to me that opposite motors spin the same way and adjacent motors spin the opposite way, as you say. Thus canceling out rotational reaction.
I said:
So why can't two opposite motors run at high thrust and the other two opposite motors at lower thrust. The first two provide the bulk of the lift the other two just maintain stability. Now there is a net rotational reaction and the ship starts to rotate.
Ignoring UFO spinning effects, isn't that what you have to do anyway to change heading, that is if you want your "front" motor at the front as you move?
heater: They usually mark the front with a color so you know where the front is. The software also know this. Obviously, any axis could be the front. In fact, I thought about ensuring the copter remained in the same magnetic direction (or true) and dipping the appropriate wing for movement. Then a friend said OK, now what about when the craft gets behind you... then everything is reversed again.
Now, there are also some who fly "X" instead of "+".
However, if the craft is rotating fast, I am not sure of the effects of speeding and slowing the motors as the craft rotates. Need to get of the ground and back safely first, then master the flying. Maybe then we can think about alternatives. I think flashing the leds to simulate spinning will be easier :smilewinkgrin:
PS I never did refine the art of landing my RC planes, but I did learn how to build them to save destruction from crashing :idea:
If you have a 3axis compass and combine with acc/gyro you can track the rotatation and so the "front". When calculating the motor speeds you combine the roll/pitch/yaw from the sensors with the RC inputs and then translate for the motor positions. My original code had fixed code to do this for +, my latest code uses a table per motor, so the code for + or x is the same - you sum the roll/pitch/yaw multiplied by a table value and add the vertical speed (throttle). The table value is different per motor and per roll/pitch/yaw i.e. a quad has 12 table values. This code is run every time round the main loop at over 50/sec, if you want the quad to rotate then you are adding a fixed yaw constant to the yaw calculations.
This is the code I use to calculate the motor speeds both for + and x, it will handle upto 8 motors.
mptr := MotorPtr
mixptr := motortable
m1 := long[SpeedPtr+FOFFSET] ' pitch
m2 := long[SpeedPtr+SOFFSET] ' roll
m3 := long[SpeedPtr+TOFFSET] ' throttle
m4 := long[SpeedPtr+YOFFSET] ' yaw
'Mix using a configurable table
repeat i from 0 to NoMotors-1
long[mptr] := m3 + (m1*~byte[mixptr])~>7 + (m2*~byte[mixptr+1])~>7 + (m4*~byte[mixptr+2])~>7
mptr += 4
mixptr += 3
So Tim, I gather you think it is possible to allow the copter to rotate a little and still be able to control the craft. I guess with 50 updates/sec we could only rotate at say 5 times/sec.
What rate are you reading the sensors (gyro, accel, compass)?
I wonder how much rotation there would be with all motors rotating in the same direction without angling the motors to counteract the spin?
I currently read the rc data from hub, the sensors and feed to imu hub variables, imu output hub variables, the roll/pitch/yaw/altitude pids and the roll/pitch/yaw/altitude to motor calculations on 1 cog at ~60/sec with the pids being the slowest part (~10ms per loop).
I wouldn't have all the motors spinning in 1 direction. The motors are spining at a few 1000s rpm even allowing for the difference in angular inertia of the props/motors to the quad thats still a fast spin. very rough approx - motors + props x 4 ~100gr, quad is ~1000gr, so if props spin at 10K rpm, the quad would spin at ~1000rpm.
That's probably more than Cluso wanted for his UFO effect:)
Now, excuse my ignorance here, having defined the "front" of the copter and painted it up to make it obvious when flying what is the normal way of navigating the thing?
That is, say it is heading north and you want to then head east. Does one turn the machine so that the front points east or does one just "slide" sideways in an easterly direction.
I've always assumed the former, which implies the thing can be turned easily already.
heater: yes, the craft rotates (i.e. turns) to the east. So, basically you fly it like a plane from that point of view.
tim: 1000rpm =16.7 rps so that could perhaps be controlled. Now, if the saucer had it's own propeller or rotor shape maybe we could take advantage of this too. The mind boggles. Probably the gyro/accel/compass would not be able to keep up with the changes.
tim: 1000rpm =16.7 rps ... Probably the gyro/accel/compass would not be able to keep up with the changes.
compass would probably be useless, as would probably the gyros, as well as anything payload wise like a camera. . . On the plus side centrifugal force would really stabilize the saucer in high winds!
[/quote]
I better just get the basics working first.[/QUOTE]
Hi all!
Back from a loooooong time waiting on the ordered parts, but finally they arrived yesterdag!
While I was trying to figure out how to solder everything together, I came across a problem: The Gyroscope ITG 3200 manual says that 2 "pull up" resistors must be connected to Vdd and to the I2C channels.
What they don't tell you is what value the resistors should have, so I'm stuck...
Can any of you guys tell me how you did it? (I saw you use the same gyro :smilewinkgrin:)
I read these resistors change the data rate some how, can anyone explain that?
say me welcome back! :-)
Sorry for leaving you for such a long time. I was involved in some projects (through my company) and was traveling a lot....so I had no chance to follow the forum.
Now that I spent several minutes to read back I hope that I can catch up with you guys and to contribute as much as I can.
Meanwhile my ordered parts from HK arrived after 2 or 3 months (true that I placed order for some out of stock items too)
Now I'm on holiday for a week but I will post you some pictures about my parts hereafter.
@Cluso, if I remember right you asked me to let you know about the quality of the carbon-fibre roads ...this will follow too.
That's probably more than Cluso wanted for his UFO effect:)
Now, excuse my ignorance here, having defined the "front" of the copter and painted it up to make it obvious when flying what is the normal way of navigating the thing?
That is, say it is heading north and you want to then head east. Does one turn the machine so that the front points east or does one just "slide" sideways in an easterly direction.
I've always assumed the former, which implies the thing can be turned easily already.
Now that I have some time I thought I should build a frame for my quad prototype. :idea: The idea was to build something rigid without using my carbon fibre material ordered form HK (I will keep it and use it for a later version...) I'm sure there will be crashes and replacement is needed for repair, so I don't want to waste too much money or time by ordering and buying carbon fibre material.
I found some cheap but really good quality AL materials in a local store. This frame is really cheap and can be built in just a few hours (I did it in 5hr). My total material cost was about 5 USD.
This is a 400mm sized frame calculated for 9-10'' props. The frame is made out 10.5 x 8mm U and 15x1.5mm I profile. It weights 245 grams (without motors).
Attached are my blender plans and a blurred picture of the ready frame (I will post new ones as I'm progressing)
Looking at DIY Drones it seems they made a very good job with there Arducopter project .
Hover1 said they sell kits for frames or full blown quads.
Have you a project using this frames and motors with a Propeller in place of the Arduino ?
It will be better and much easier for Propeller users but I dont have the skills to make such adaptation.
I had I look at separate pieces of projects from Cluso99 , JasonDorie and other : very nice work but it seems that nobody tried to put this pieces together to make a full coherent project like the Arducopter.
Any hope to see a more complete "Propellercopter" project for beginners in a more compact document (the first post of this tread was the beginning of such work , after that there are a lot of very good but separate ideas )?
Cats92: I wanted to do a complete project as you suggest. Not necessarily making a kit available other than the electronics, but the recommendations of what works. Unfortunately, life gets in the way of hobbies some times and a lack of time has prevented me from achieving this so far. Hopefully I will get back on track shortly. The pcbs are mostly complete - there are a number of pcbs I am doing concurrently for different projects because they use common parts and it makes part costs much more reasonable. In fact shipping is still the major component which always costs 20-50% of my parts costs -OUCH!
Comments
PFO or PPFO
It would probably weigh alot less than 2lb empty. Though not shown in the pictures, i would fibrglass in balsa, thin model aircraft ply, or single layered (pre cured) fiberglass sheets in between the ducts to act as PCB mounting. Also i was considering using Nomex "Honeycore" as a reinforcement within the inner half of the structure. Also, two thinner coats of gelcoat could be used on the outside with single layer glass & selective carbon fiber or carbon veil reinf of the two halves of the shells.
the laser led comment -i knew you were heading for a 50's B movie design with this! LOL
@ Heater - while the UAV will be relatively light with respect to surface area, you can compensate for most turbulence by way of weight distrubution, ie; put the heavy battery packs and ESC's in the outter compartments near or at the circumfrence. This gives the effect of increased stability in turbulence but has the drawback of increasing roll rate response time and if not accounteded for in software can shorten flight times by making the motors work harder trying to keep up with the gyros outputs. For most tasks some drift is ok, for video survelance, your craft can wobble some in the air turbulence as long as it compensates for the change in attitude quickly enough and using a somewhat averaged gyro output to stabilize the shell while using near raw gyro rates to stabilize the camera. What i'm trying to say is to avg out gyro/G sensor rates to control the ESC's in a way that is proportional to the saucers mass relative to the response rate to that particular gyro direction. In contrast, the camera having much less mass than the saucer can be stabilized with a much shorter avg or raw gyro data. Digital servos respond best to this and a high quality gyro like the digital ones avail for model heli's are often paired with certain digital servos to give best drift vs rate correction perf.
With respect to power savings, instead of constantly varrying the speed of the motors and thus rotating mass (which bleeds energy from the system and generates heat under braking -yes you can setup most aircraft ESC's for braking), why not instead use constant speed props? ESC's are controlled via a single PWM signal from the prop and is only varied according to loading of the motors and/or flight mode (hover vs aerobatics), the control comes from varying the pitch of the props instead of varying the speed of the rotating mass of the rotor mass + prop mass. There's a hidden benifit from this way of control, kinetic energy stored in the rotating mass of the prop and rotor which can be translated to english as really fast transient response of the craft as far as control or sudden movements are concerned. It would be pretty cool to see the saucer config'd with battery packs in the center for faster roll rates and to see this thing do somersaults in mid air! like a coin spinning on its side on a table but spinning horizontally in the air instead!
Constant speed prop setup may cost a bit more but i think its a better choice for control and power savings over just relying on the ESC's. Your thoughts?
The body shell:
1) Greatly increases surface area and hence susceptibility to turbulence.
2) Increases the crafts moment of inertia with all the consequences you described.
3) Increases mass and hence decreases flight time and rate of climb.
If any of that outweighs the sheer coolness of the finished machine is another matter:)
Distributing the weight away from the centre just does not seem right to me. The resulting increase in moment of inertial may help dampen the effect of turbulence on the shell but it also has a negative effect on roll rate control and energy consumption as you noted.
I'm not at all up to date with modern model engineering so are there such things as variable pitch prop systems?
Is there any mileage in using some kind of directional vanes under the propellers (perhaps entirely in the ducts) that redirect thrust for stabilisation rather than motor speed or prop pitch control?
In short, K.I.S.S. works best and just using the ESC's for thrust control will work well, i think if a constant speed prop could be implemented it would be worth the slight decrease in reliability and robustness since the basic design in the pictures would be very reliable and robust initially. Reliability decreases proportionaly as moving parts count increase IMHO. Constant speed props is a good, safe tradeoff between reliability, liability, and the pros (increased eff, control).
http://www.hobby-lobby.com/evp_pro.htm 9.5" variable pitch prop & outrunner motor
But that is one neat variable pitch prop set up uyou have linked to. Much preferred.
As I understand it, the front/rear axis has the motors turn one way (say cw) and the side axis has the motors turn the opposite (say ccw). This is to prevent the copter from spinning. Now to move forward, the front motor slows, and the rear motor speed up by an equivalent amount, causing the copter to dip in the front and hence move forward.
AFAIK, if the copter spins then the requirements of the motors will vary way too fast to actually control them.
Some have tilted the motor mounts the cancel the rotational force and so use all motors rotating the same way.
Maybe it's doable but it's just too complex - if so, then hey, we have seen the impossible here before. Who said you couldn't emulate a Z80 fast enough with a prop??? Or USB???
It seemed natural to me that opposite motors spin the same way and adjacent motors spin the opposite way, as you say. Thus canceling out rotational reaction.
I said:
Ignoring UFO spinning effects, isn't that what you have to do anyway to change heading, that is if you want your "front" motor at the front as you move?
Now, there are also some who fly "X" instead of "+".
However, if the craft is rotating fast, I am not sure of the effects of speeding and slowing the motors as the craft rotates. Need to get of the ground and back safely first, then master the flying. Maybe then we can think about alternatives. I think flashing the leds to simulate spinning will be easier :smilewinkgrin:
PS I never did refine the art of landing my RC planes, but I did learn how to build them to save destruction from crashing :idea:
This is the code I use to calculate the motor speeds both for + and x, it will handle upto 8 motors.
mptr := MotorPtr mixptr := motortable m1 := long[SpeedPtr+FOFFSET] ' pitch m2 := long[SpeedPtr+SOFFSET] ' roll m3 := long[SpeedPtr+TOFFSET] ' throttle m4 := long[SpeedPtr+YOFFSET] ' yaw 'Mix using a configurable table repeat i from 0 to NoMotors-1 long[mptr] := m3 + (m1*~byte[mixptr])~>7 + (m2*~byte[mixptr+1])~>7 + (m4*~byte[mixptr+2])~>7 mptr += 4 mixptr += 3What rate are you reading the sensors (gyro, accel, compass)?
I wonder how much rotation there would be with all motors rotating in the same direction without angling the motors to counteract the spin?
Certainly this is for further investigation.
I wouldn't have all the motors spinning in 1 direction. The motors are spining at a few 1000s rpm even allowing for the difference in angular inertia of the props/motors to the quad thats still a fast spin. very rough approx - motors + props x 4 ~100gr, quad is ~1000gr, so if props spin at 10K rpm, the quad would spin at ~1000rpm.
That's probably more than Cluso wanted for his UFO effect:)
Now, excuse my ignorance here, having defined the "front" of the copter and painted it up to make it obvious when flying what is the normal way of navigating the thing?
That is, say it is heading north and you want to then head east. Does one turn the machine so that the front points east or does one just "slide" sideways in an easterly direction.
I've always assumed the former, which implies the thing can be turned easily already.
tim: 1000rpm =16.7 rps so that could perhaps be controlled. Now, if the saucer had it's own propeller or rotor shape maybe we could take advantage of this too. The mind boggles. Probably the gyro/accel/compass would not be able to keep up with the changes.
I better just get the basics working first.
[/quote]
I better just get the basics working first.[/QUOTE]
K.I.S.S. ...!
Back from a loooooong time waiting on the ordered parts, but finally they arrived yesterdag!
While I was trying to figure out how to solder everything together, I came across a problem: The Gyroscope ITG 3200 manual says that 2 "pull up" resistors must be connected to Vdd and to the I2C channels.
What they don't tell you is what value the resistors should have, so I'm stuck...
Can any of you guys tell me how you did it? (I saw you use the same gyro :smilewinkgrin:)
I read these resistors change the data rate some how, can anyone explain that?
Thanks in advance!
Just had my first flight (No gyro nor accelerometer compensation) and lifted almost 5 kg at full throttle!
Now comes the hard part for me: Stabilizing the flight with the onboard gyro and accelerometer.
(I'll try to use the Kalman Filtering technique to calculate the corrections and see how that works out.)
say me welcome back! :-)
Sorry for leaving you for such a long time. I was involved in some projects (through my company) and was traveling a lot....so I had no chance to follow the forum.
Now that I spent several minutes to read back I hope that I can catch up with you guys and to contribute as much as I can.
Meanwhile my ordered parts from HK arrived after 2 or 3 months (true that I placed order for some out of stock items too)
Now I'm on holiday for a week but I will post you some pictures about my parts hereafter.
@Cluso, if I remember right you asked me to let you know about the quality of the carbon-fibre roads ...this will follow too.
Did I missed something really important?
Cheers,
Szabi
That sounds good !
DIY Drone announcement:
http://diydrones.com/profiles/blogs/arducopter-development-amp?id=705844%3ABlogPost%3A212135&xg_source=msg_mes_network
Purchase kits:
http://store.fahpah.com/
Frame kit price looks in order of pricing individual parts out. Payload is not mention, or I missed it.
Jim
I found some cheap but really good quality AL materials in a local store. This frame is really cheap and can be built in just a few hours (I did it in 5hr). My total material cost was about 5 USD.
This is a 400mm sized frame calculated for 9-10'' props. The frame is made out 10.5 x 8mm U and 15x1.5mm I profile. It weights 245 grams (without motors).
Attached are my blender plans and a blurred picture of the ready frame (I will post new ones as I'm progressing)
Looking at DIY Drones it seems they made a very good job with there Arducopter project .
Hover1 said they sell kits for frames or full blown quads.
Have you a project using this frames and motors with a Propeller in place of the Arduino ?
It will be better and much easier for Propeller users but I dont have the skills to make such adaptation.
I had I look at separate pieces of projects from Cluso99 , JasonDorie and other : very nice work but it seems that nobody tried to put this pieces together to make a full coherent project like the Arducopter.
Any hope to see a more complete "Propellercopter" project for beginners in a more compact document (the first post of this tread was the beginning of such work , after that there are a lot of very good but separate ideas )?
Cats92