Hey, that looks really impressive. Great to see the progress on this thread from a design principle to a real flying quadcopter. How 'easy' is that quad to fly? It looks like when you are going at speed and want to stop, you tilt it - is that you doing that or is that software?
I have been dragging the chain, so to speak. I can see I am really going to have to get off my rear and get the pcb completed and order the rest of my motor/esc/propellers and build this quad. My how I can see the hours disappear flying a quad. BTW I think I will be doing 2 pcbs. One will just be the 4 or 5 sensors and the larger one will be a boxed (50x50x20mm) version with the prop and RC connections, etc.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
JasonDorie: Did you switch to the + style instead of X style flight model? Looks like it in that new video.
I've been going down the X style path based on your messages in the Completed projects forum. However, a lot of other people seem to be using + style...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out the Propeller Wiki·and contribute if you can.
Dr Acula: That's me tilting it to stop, yes. I have exactly one 3-axis gyro on board, and that's it - no GPS, no compass, no accelerometer, etc, so the hardware is just trying to keep itself rotating at the rates I give it with the joysticks.
Roy: I did switch the layout from + to X, but the reason is silly: When I modeled the body, I forgot to rotate the channels that I run the arms through (easier to model them vertical / horizontal, then rotate by 45). The body also has mounts for the ProtoBoard, so it was just simpler to change the programming to + style than cut a new body. It's only a couple lines of code.
I have 10mm square carbon rods for the arms, so the next body I make will be built to mount those, and I'll likely flip back to the X layout. I doubt it makes any difference in handling, but I'm curious to find out.
Jason: How are the carbon rods working out? Because of the cost in getting them from HobbyKing (cannot find them in Australia at a reasonable price) I was thinking of going back to aluminium square tube. They would be a little heavier, but easier to get.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
I haven't actually gotten them installed yet - I'm still using aluminum tube which works just fine. My only issue with it is that it's hard to get the motors mounted perfectly vertically and aligned with each other, and the motor mounts I made shift a little. That said, even the aluminum works just fine - I've used 1/2" square aluminum tube in the past, and was going to use it again, but couldn't find it. Home Depot and others now seem only to stock 3/4" or better, and it's thicker (heavier) than the stuff I had before, which is why I bought the carbon.
Well, if that is you doing that tilting to stop, that is very clever flying. I'm saying that as someone who has had helicopters flying but I spend more time fixing than flying *grin*. And it doesn't help that while I'm trying to stabilise a helicopter I'm being buzzed by bright green lorikeets who are zipping between the trees and I'm sure are flying just for the fun of it. Their reflexes are so incredibly fast.
And I guess that was one of Cluso's aims with this project - take something that is hard to fly and move some of the tasks onto a propeller. For example, to me, tilting to make a helicopter stop makes sense. So I can take off, go forward and then stop by tilting and that seems intuitive. Where I find it hard is when you take off, go some distance away so you can't see it so well, and the helicopter rotates 90 degrees. Then all the controls go haywire (as does my brain!).
I've got some micro helicopters and some have gyros and some do not. On a good day, with absolutely no alcohol and not too late in the evening I can get the non-gyro one to land on a coffee table. But the gyro one is so much easier.
So adding spin stabilisation is a huge step forward.
The other thing I find hard as a beginner is what to do when the controls do go haywire. One thing that is nice with some of the small helis is that if you let all the controls go, the helicopter just hovers. So that helps when you get stuck, you can return to a default hover, then work out what to do next in slow deliberate steps.
With a 3 axis gyro, is that enough to keep it in a stable hover? (assuming no wind)
And, thinking of how to make this easier to fly, would a next step be to add in accelerometers to counter any wind?
has anyone in this thread looked at the information at www.aeroquad.com ? I am going to be building one from this page eventually. (read that as when I get the money to buy the parts, roughly 500 dollars total for all the parts to build one)
they use a arduino, but I dont see why the code couldnt be re written for a prop eventually.
FireHopper: Jason and Tim and others are well on the way to writing the code for the Prop. They are flying so it must be partially working. IMHO, the prop would be easier than the arduino.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
Dr Acula: Thanks for the compliments! I've been flying for a while, and I have a PC-based simulator as well, which is a huge help for wiring your brain to handle the many different orientations you can get into with a heli. I find one of the biggest problems with the quad is that there are so few visual cues for "front" that it takes longer to get your brain oriented. I've got a pair of very small bike lights that I mount on the landing gear so I can fly it at night, and I think it's actually easier than flying it during the day, because it's so obvious which way is "forward".
My current code has no idea where "down" is, so it will counter the inherent instability of a quad, but it's unlikely that it could hover hands-off - it's not self-leveling. Adding an accelerometer and modifying the code to use the accel / gyro combination to model the current orientation would fix that, and should allow for hands-off control. Even in wind, it should be able to keep itself level, it'd just drift laterally.
That said, I wanted something that flew like a heli, not a drone. I'll probably try to make it "fully stabilized" as an intellectual challenge, but I'm pretty close to what I want already. I might even make that a "mode" - flip a switch and it's self-stable so other people can try it, or if it gets a little too far away and I get nervous. [noparse]:)[/noparse]
Jason: Yes, I like the switch idea. My RC gear (DX6i) has a number of switches. I haven't worked out how they work - they must multiplex somehow as it is a 6 ch tx/rx. Anyway, it would be good to just flick a switch and the copter goes to a hover mode. Perhaps we could use something like mercury switches for levelling as it would be easier than trying to work out the levelling from accel + gyro. Anyway, I will have the works available - I have the chips. Just need to finish the pcb!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
Mercury switches are just binary though - there's no gradient to tell you how close you are to level, and they're subject to inertia & centrifugal forces as much, or more, than gyro / accel (at least, I would imagine them to be worse).
A lot of the switches on radios are internal - The idle-up / stunt1 / stunt2 switch and dual rate switches, for example, just swap between different banks of settings for end point adjustment, exponential, and power curves. The only "real" switch on a 6-channel radio is the landing gear switch, often called "aux2". In airplane mode, the 5th channel is usually mapped to the 3-position switch, and emits pulses of either 1100, 1500, or 1900us in the three switch positions. In heli mode that channel is the pitch servo, and the mapping from throttle stick to pitch is controlled by that same switch.
I'm using a 7 channel that has analog dials available - I eventually want the motor pods to transition forward, like an Osprey, for normal (winged) flight, and I want landing gear. 6-channels total for that, and I'll have one left for random stuff, like lights perhaps, or a flight mode switch. [noparse]:)[/noparse]
Re " I eventually want the motor pods to transition forward, like an Osprey, for normal (winged) flight"
That sounds fun. I tried lifting a 'foamie' plane with my chinook helicopter, and it is close to lifting, so with a very light wing design it would be great to convert to forward flight. The extra lift from a wing ought to extend the flight time.
Your switch idea sounds a great idea. Stable=easy to fly. Or more unstable, but able to do more acrobatics. You choose the amount the propeller helps you.
"I eventually want the motor pods to transition forward" - There is a video showing a different way of doing this (which I can't find now). The quad is flow as a + with the 2 side arms having a wing built around, but pointing upwards. To fly using the wing, the whole quad is tilted forward 90 degrees so that the wings come into play and all 4 props act to move it forward. No need to rotate the motors and you would already have had to fit wings to the arms anyway.
A quad that takes off and then turns into a bi-plane.
That is just a beautiful piece of engineering. And flying - I see it landed back on the same place it took off from.
I guess my question is - how much of that is due to the skill of the pilot, and how much due to (any) onboard software, and how much could the propeller help make flying such a thing easier?
I have this simulator, and flew this model in it once before deciding it needed to exist. The motor pods transition like an Osprey, which adds complication to the design, but it's just so pretty. [noparse]:)[/noparse]
I have to admit though, that biplane idea looks very elegant.
How would you rotate the pods? I guess that might be 4 servos. There will be a weight cost there.
I guess many flying things try to replicate bigger flying things that humans fly, and humans like to stay 'upright' so rather than rotate the plane, we rotate the wings or propellers.
With a model like that biplane or a UAV, you can think differently. Rotate the whole craft.
I've been pondering what it is like for the controller as you do that rotation. You are flying a quad and balancing it with the power to 4 engines. Increase the power to two engines and it starts to flip over and travel forward. Get it up to speed and then execute a turn - I presume that involves both increasing the power to the propellers on the outside of the turn, and also banking (which is the same control as rotation/yaw when hovering). Yikes - the beginner's brain is definitely going to need some help here!
How is everyone doing the radio communications - and in particular, how many channels are needed?
To turn, its the same as going forward, slow the motor on the side you want to go, i.e. tilt it in the direction you want to do.
You actually dont need to rotate it at all, it can go in any direction without rotating. If you have driven a omniwheel robot or a mecanum robot its similar and takes some learning not to drive it like a car, i.e. not to worry about turning it in the direction you want to go. You do need to know the front. I have status leds mounted on the front arm - tells me the status of the system as well as where is front.
If the control system is setup right it speeds up the opperside motor when it slows the other down, so the total lift is unchanged (sort of since some of it is now not vertical but it could adjust that as well). The stablization software I am working on does that, if you dont change the throttle the altitude doesn't change, if you tilt, it will speed up the motors to compensate (doesn't work yet)
I use a 2.4GHz RC system. 4 channels for throttle, forwards, sideways and yaw, +2 for IMU control - altitude hold, IMU on/off. the radio I use has 8 channels, so I have 2 spare probably 1 for gps control on/off.
To bump the thread again (and to show off) I'm posting video of my latest flight. I crashed it for the first time, but nothing broke - I just bent the motor arms back and flew it again immediately (lucky). I also show that it will hold level completely hands off with almost no drift, and this is still using just the ITG-3200 gyro with no additional sensors.
Do you think an altitude sensor would have prevented the crash? I noticed the on board camera video was missing post-crash. Did the camera not make it?
Can't wait to get unpacked in a couple of weeks to get started on mine. For now this is the last thing to get packed:
JasonDorie said...
To bump the thread again (and to show off) I'm posting video of my latest flight. I crashed it for the first time, but nothing broke - I just bent the motor arms back and flew it again immediately (lucky). I also show that it will hold level completely hands off with almost no drift, and this is still using just the ITG-3200 gyro with no additional sensors.
Jim - That crash was all me - The quad did exactly what I told it to, which unfortunately was "go into the ground really fast". [noparse]:)[/noparse] A lot more sensors and some clever programming might prevent things like that (stay within 30 degrees of level, don't descend faster than X, etc) but that would also make it much less agile. I want to fly it aggressively, and I accept the consequences.
Normally I can see if the white motor pods are in front of or behind the rotors - your brain instantly interprets that as "angled away from / towards me". In this case, since the sky was white, all I could see was the silhouette. Even though it was MOVING toward me, I still somehow interpreted it as angled away from me, and tried to fix it, making it worse. Oops.
The camera survived just fine (it was still recording after the crash) but the double-sided tape it's held on with got covered with dirt and didn't stick any more.
As a fellow heli pilot, I know the meaning of " It's not IF you crash, it's WHEN you crash"
JasonDorie said...
Jim - That crash was all me - The quad did exactly what I told it to, which unfortunately was "go into the ground really fast". [noparse]:)[/noparse]
WOW! That's impressive with just a gyro. Of course you can fly which definately helps.
I only ever learnt how to quickly rebuild my planes after I crashed them - never got to land. That was years ago :-(
I noticed it drifted slightly during hands off. Is that because of gyro drift? If so, I guess this could be determined (perhaps on the ground before flight) and factored in. The pressure sensor would help to maintain altitude under hands off but then would need a switch to place into hover mode with hands off.
Certainly the prop is going to make it easy to do other things in parallel.
Those 3 blade props work well. Have you tried 2 blades? If so, what can you say about the differences.
Would you mind posting a photo or two closeup of the electronics? You have probably said before, but what motors, blades and esc are you using?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
Cluso - It drifts "hands off" because it has no idea what "level" is, and even if it did, a slight breeze would be enough to keep it moving even if it were perfectly level. The code is written to hold a set level and heading, and it manages to do that very well given that it has no external frame of reference. To actually stay in one place would require a GPS, an accelerometer, and a pressure sensor. That said, I want to fly it, not give it coordinates to go to, so I'm quite happy with it the way it is. [noparse]:)[/noparse]
I've added temperature drift compensation to the gyro code, so the gyro itself basically doesn't drift any more. The bulk of errors now are due to accumulation inaccuracy or noise. When I get more time with it, I'm going to try mucking with the analog filter settings in the ITG to see how they affect stability, and I'm going to try putting the gyro in a small box with some mass to it to physically dampen the vibration even more.
I chose these particular 3-blade props because they were as light as the typical 2-blade versions, and they come with a variety of shaft size inserts for mounting them. The typical 10x4.7 counter-rotating props that most people use have a very small hole that requires drilling, and if you don't center them perfectly when you drill them, they're impossible to balance. The mechanical vibration from unbalanced props makes the gyros wonky.
I have some proper 2-blade APC props on the way. They should produce nearly identical lift and be more efficient (lower blade count props are more efficient by nature, as they produce less turbulence).
The electronics could be described as "boring". [noparse]:)[/noparse] I have a ProtoBoard with an ITG-3200 breakout on it, pin headers for the speed controls and R/C inputs. The only added components on the protoboard are the in-line resistors for the R/C inputs. I'll post pics, but there's really very little to it, which was my goal - I wanted to produce something reasonably stable without $400 worth of hand-holding hardware. [noparse]:)[/noparse]
For the motors, ~1000Kv, or even lower, is best for a quad. You want low speed so you can run a larger, lower pitch prop. You need lift, not airspeed.
For the ESC, it's important to find one that has (or can be set to) fixed endpoint throttle curve. Most are set to "adapt" their input range based on the range of values they see. On an airplane, you throttle up to pretty much full on takeoff, so it works well. On a quad, this will drive you insane since it becomes a feedback loop.
Jason
Post Edited (JasonDorie) : 7/13/2010 3:40:15 AM GMT
Jason, How much temp drift do you see? Curious if its similar across devices or the same for different itg-3200's.
It looks like I see a change of a bit less than 0.1deg/sec per degree temp (C). Interesting though is the Z axis is positive and the X and Y is -ve for the same temp change. I have attached a graph of the drift, the temp is the top line in degrees C, the bottom 3 lines are the gyros (the scale is 10x deg/sec).
Time - I posted a graph toward the top of page 15. Mine looks different than yours, partially because I plotted the X/Y/Z values based on the temp value (instead of time) so it was easier to see correlation.
Mine doesn't quite fit your pattern though - All three axis on mine move in the same direction. I was able to model the drift amount by using the equation:
drift := (temp + 15000) / 100
That drift value gets subtracted from all samples (X/Y/Z) after they're read. All the values on my chart are raw - not converted to degrees (angles or temperature). What do yours look like unconverted?
It is·fantastic that you can achieve this stability with just the gyro. An accelerometer would help the drift, although I thought it may have been possible to counteract some of it. What you have achieved should make it fairly easy for a relative newbie like me to fly it slowly while learning.
For your circuit connection to the esc(s), I presume you do not connect the center pin (5v)? Do you have 1 of these connected to the input of the prop regulators? What about the input from the RC receiver and how is that powered?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔ Links to other interesting threads:
Comments
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.smarthome.viviti.com/propeller
Post Edited (Dr_Acula) : 6/28/2010 6:53:24 AM GMT
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
I've been going down the X style path based on your messages in the Completed projects forum. However, a lot of other people seem to be using + style...
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Check out the Propeller Wiki·and contribute if you can.
Roy: I did switch the layout from + to X, but the reason is silly: When I modeled the body, I forgot to rotate the channels that I run the arms through (easier to model them vertical / horizontal, then rotate by 45). The body also has mounts for the ProtoBoard, so it was just simpler to change the programming to + style than cut a new body. It's only a couple lines of code.
I have 10mm square carbon rods for the arms, so the next body I make will be built to mount those, and I'll likely flip back to the X layout. I doubt it makes any difference in handling, but I'm curious to find out.
J
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
J
And I guess that was one of Cluso's aims with this project - take something that is hard to fly and move some of the tasks onto a propeller. For example, to me, tilting to make a helicopter stop makes sense. So I can take off, go forward and then stop by tilting and that seems intuitive. Where I find it hard is when you take off, go some distance away so you can't see it so well, and the helicopter rotates 90 degrees. Then all the controls go haywire (as does my brain!).
I've got some micro helicopters and some have gyros and some do not. On a good day, with absolutely no alcohol and not too late in the evening I can get the non-gyro one to land on a coffee table. But the gyro one is so much easier.
So adding spin stabilisation is a huge step forward.
The other thing I find hard as a beginner is what to do when the controls do go haywire. One thing that is nice with some of the small helis is that if you let all the controls go, the helicopter just hovers. So that helps when you get stuck, you can return to a default hover, then work out what to do next in slow deliberate steps.
With a 3 axis gyro, is that enough to keep it in a stable hover? (assuming no wind)
And, thinking of how to make this easier to fly, would a next step be to add in accelerometers to counter any wind?
Can I say again, very impressive flying.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.smarthome.viviti.com/propeller
they use a arduino, but I dont see why the code couldnt be re written for a prop eventually.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
My current code has no idea where "down" is, so it will counter the inherent instability of a quad, but it's unlikely that it could hover hands-off - it's not self-leveling. Adding an accelerometer and modifying the code to use the accel / gyro combination to model the current orientation would fix that, and should allow for hands-off control. Even in wind, it should be able to keep itself level, it'd just drift laterally.
That said, I wanted something that flew like a heli, not a drone. I'll probably try to make it "fully stabilized" as an intellectual challenge, but I'm pretty close to what I want already. I might even make that a "mode" - flip a switch and it's self-stable so other people can try it, or if it gets a little too far away and I get nervous. [noparse]:)[/noparse]
J
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
A lot of the switches on radios are internal - The idle-up / stunt1 / stunt2 switch and dual rate switches, for example, just swap between different banks of settings for end point adjustment, exponential, and power curves. The only "real" switch on a 6-channel radio is the landing gear switch, often called "aux2". In airplane mode, the 5th channel is usually mapped to the 3-position switch, and emits pulses of either 1100, 1500, or 1900us in the three switch positions. In heli mode that channel is the pitch servo, and the mapping from throttle stick to pitch is controlled by that same switch.
I'm using a 7 channel that has analog dials available - I eventually want the motor pods to transition forward, like an Osprey, for normal (winged) flight, and I want landing gear. 6-channels total for that, and I'll have one left for random stuff, like lights perhaps, or a flight mode switch. [noparse]:)[/noparse]
J
That sounds fun. I tried lifting a 'foamie' plane with my chinook helicopter, and it is close to lifting, so with a very light wing design it would be great to convert to forward flight. The extra lift from a wing ought to extend the flight time.
Your switch idea sounds a great idea. Stable=easy to fly. Or more unstable, but able to do more acrobatics. You choose the amount the propeller helps you.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.smarthome.viviti.com/propeller
www.youtube.com/watch?v=LWTi2Dgh0Fw&feature=player_embedded#!
A quad that takes off and then turns into a bi-plane.
That is just a beautiful piece of engineering. And flying - I see it landed back on the same place it took off from.
I guess my question is - how much of that is due to the skill of the pilot, and how much due to (any) onboard software, and how much could the propeller help make flying such a thing easier?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.smarthome.viviti.com/propeller
www.knifeedge.com/forums/showpost.php?p=56116&postcount=4
I have this simulator, and flew this model in it once before deciding it needed to exist. The motor pods transition like an Osprey, which adds complication to the design, but it's just so pretty. [noparse]:)[/noparse]
I have to admit though, that biplane idea looks very elegant.
J
In some ways, it reminds me of the harrier jump-jets.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
How would you rotate the pods? I guess that might be 4 servos. There will be a weight cost there.
I guess many flying things try to replicate bigger flying things that humans fly, and humans like to stay 'upright' so rather than rotate the plane, we rotate the wings or propellers.
With a model like that biplane or a UAV, you can think differently. Rotate the whole craft.
I've been pondering what it is like for the controller as you do that rotation. You are flying a quad and balancing it with the power to 4 engines. Increase the power to two engines and it starts to flip over and travel forward. Get it up to speed and then execute a turn - I presume that involves both increasing the power to the propellers on the outside of the turn, and also banking (which is the same control as rotation/yaw when hovering). Yikes - the beginner's brain is definitely going to need some help here!
How is everyone doing the radio communications - and in particular, how many channels are needed?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.smarthome.viviti.com/propeller
You actually dont need to rotate it at all, it can go in any direction without rotating. If you have driven a omniwheel robot or a mecanum robot its similar and takes some learning not to drive it like a car, i.e. not to worry about turning it in the direction you want to go. You do need to know the front. I have status leds mounted on the front arm - tells me the status of the system as well as where is front.
If the control system is setup right it speeds up the opperside motor when it slows the other down, so the total lift is unchanged (sort of since some of it is now not vertical but it could adjust that as well). The stablization software I am working on does that, if you dont change the throttle the altitude doesn't change, if you tilt, it will speed up the motors to compensate (doesn't work yet)
I use a 2.4GHz RC system. 4 channels for throttle, forwards, sideways and yaw, +2 for IMU control - altitude hold, IMU on/off. the radio I use has 8 channels, so I have 2 spare probably 1 for gps control on/off.
Enjoy:
www.youtube.com/watch?v=EjvJLm725h0
Great video!
Do you think an altitude sensor would have prevented the crash? I noticed the on board camera video was missing post-crash. Did the camera not make it?
Can't wait to get unpacked in a couple of weeks to get started on mine. For now this is the last thing to get packed:
http://www.art-tech.cn/english/ArticleShow.asp?ArticleID=151
Jim
Normally I can see if the white motor pods are in front of or behind the rotors - your brain instantly interprets that as "angled away from / towards me". In this case, since the sky was white, all I could see was the silhouette. Even though it was MOVING toward me, I still somehow interpreted it as angled away from me, and tried to fix it, making it worse. Oops.
The camera survived just fine (it was still recording after the crash) but the double-sided tape it's held on with got covered with dirt and didn't stick any more.
Jason
As a fellow heli pilot, I know the meaning of " It's not IF you crash, it's WHEN you crash"
I only ever learnt how to quickly rebuild my planes after I crashed them - never got to land. That was years ago :-(
I noticed it drifted slightly during hands off. Is that because of gyro drift? If so, I guess this could be determined (perhaps on the ground before flight) and factored in. The pressure sensor would help to maintain altitude under hands off but then would need a switch to place into hover mode with hands off.
Certainly the prop is going to make it easy to do other things in parallel.
Those 3 blade props work well. Have you tried 2 blades? If so, what can you say about the differences.
Would you mind posting a photo or two closeup of the electronics? You have probably said before, but what motors, blades and esc are you using?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
I've added temperature drift compensation to the gyro code, so the gyro itself basically doesn't drift any more. The bulk of errors now are due to accumulation inaccuracy or noise. When I get more time with it, I'm going to try mucking with the analog filter settings in the ITG to see how they affect stability, and I'm going to try putting the gyro in a small box with some mass to it to physically dampen the vibration even more.
I chose these particular 3-blade props because they were as light as the typical 2-blade versions, and they come with a variety of shaft size inserts for mounting them. The typical 10x4.7 counter-rotating props that most people use have a very small hole that requires drilling, and if you don't center them perfectly when you drill them, they're impossible to balance. The mechanical vibration from unbalanced props makes the gyros wonky.
I have some proper 2-blade APC props on the way. They should produce nearly identical lift and be more efficient (lower blade count props are more efficient by nature, as they produce less turbulence).
The electronics could be described as "boring". [noparse]:)[/noparse] I have a ProtoBoard with an ITG-3200 breakout on it, pin headers for the speed controls and R/C inputs. The only added components on the protoboard are the in-line resistors for the R/C inputs. I'll post pics, but there's really very little to it, which was my goal - I wanted to produce something reasonably stable without $400 worth of hand-holding hardware. [noparse]:)[/noparse]
Motors: www.hobby-lobby.com/brushless_skatty.htm
Props: www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=5248&Product_Name=GWS_HD9050_3_Blade_Prop_(2pc_pack) (and the counter rotating versions)
ESC: www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=5449&Product_Name=TowerPro_n20_Acro_Brushless_Speed_Controller
For the motors, ~1000Kv, or even lower, is best for a quad. You want low speed so you can run a larger, lower pitch prop. You need lift, not airspeed.
For the ESC, it's important to find one that has (or can be set to) fixed endpoint throttle curve. Most are set to "adapt" their input range based on the range of values they see. On an airplane, you throttle up to pretty much full on takeoff, so it works well. On a quad, this will drive you insane since it becomes a feedback loop.
Jason
Post Edited (JasonDorie) : 7/13/2010 3:40:15 AM GMT
It looks like I see a change of a bit less than 0.1deg/sec per degree temp (C). Interesting though is the Z axis is positive and the X and Y is -ve for the same temp change. I have attached a graph of the drift, the temp is the top line in degrees C, the bottom 3 lines are the gyros (the scale is 10x deg/sec).
Post Edited (Timmoore) : 7/13/2010 7:16:43 AM GMT
Mine doesn't quite fit your pattern though - All three axis on mine move in the same direction. I was able to model the drift amount by using the equation:
drift := (temp + 15000) / 100
That drift value gets subtracted from all samples (X/Y/Z) after they're read. All the values on my chart are raw - not converted to degrees (angles or temperature). What do yours look like unconverted?
Jason
Just remembered my gyro is mounted upside down, thats why the difference.
It is·fantastic that you can achieve this stability with just the gyro. An accelerometer would help the drift, although I thought it may have been possible to counteract some of it. What you have achieved should make it fairly easy for a relative newbie like me to fly it slowly while learning.
For your circuit connection to the esc(s), I presume you do not connect the center pin (5v)? Do you have 1 of these connected to the input of the prop regulators? What about the input from the RC receiver and how is that powered?
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Links to other interesting threads:
· Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
· Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
· Prop Tools under Development or Completed (Index)
· Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
· Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz