Page 7 of 19 FirstFirst ... 3456789101117 ... LastLast
Results 61 to 70 of 181

Thread: Brad's next purchase... (Just kidding!)

  1. #61
    Join Date
    Sep 2008
    Location
    Orlando, Florida
    Posts
    1,023

    Default

    Quote Originally Posted by buffo View Post
    Speaking of quad-copters, did you guys see the new "Lily" drone that was all over Reddit the other day?
    This thing is seriously cool. Check out the video: https://www.youtube.com/watch?v=4vGcH0Bk3hg



    There are two major drawbacks to the thing, in my opinion:

    1) sealed, non-removable battery. Makes sense, since otherwise it would be tough to waterproof the thing, but it means you can't carry extra batteries and swap them out. So you get 20 minutes of flight time and then you're done unless you have a charger handy.

    2) zero obstacle-avoidance. Again, it makes sense, since it's just following the tracking "puck" that you wear, but it limits the usage to wide-open spaces. If you don't consider your surroundings, this thing will happily fly itself into a tree (or a building) while shooting your video.

    With that taken into account, it's still a damned cool idea, and at the pre-order price of just $499, it's tempting. (Rumor is that they'll have them shipping by next February, at which point the price more than doubles.) But I think I'll pass, if only because I'd rather get some FPV gear first. (The camera specs are quite impressive though.)

    Adam
    Nice machine! It seems like the follow me feature is taking off.

    I agree a nonreplicable battery is a major negative.
    I've never seen a quad that has obstacle avoidance. I've heard people talk about it and know someone who wants to develop it and add it to his hexacopter.

    Chris

  2. #62
    Join Date
    Aug 2008
    Location
    UK
    Posts
    5,700

    Default

    If you want waterproof, have you seen these:

    http://www.quadh2o.com/

  3. #63
    Join Date
    May 2014
    Posts
    619

    Default

    I just want to see what happens when the "just throw it" function fails, and it falls to the ground... or better yet, detects it's in mid-air, but too late and it's 1" off the ground and then decides to power up. So many failure modes, too little time.


  4. #64
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,488,805

    Default

    It's got an accelerometer built-in to the flight system, so presumably they use that to detect the free-fall condition that triggers the rotors to start up. (After you've turned it on of course.) If you didn't throw it high enough, I imagine it would just bounce hard off the ground and then take off.

    In the video you can see it land in the water before the rotors spin up. But a few seconds later it takes off. My guess is that it detects the hard landing and waits to see if it stays right side up before it tries to take off.

    Considering that I hand-launch my quads all the time, and I have plenty of time to get my hand back on the controls to take control before they crash, I think a computerized flight controller could do this just as well, if not better.

    Adam

  5. #65
    Join Date
    May 2014
    Posts
    619

    Default

    I get all of that, but I was really talking about failure modes. Not when everything is working as designed.

  6. #66
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,488,805

    Default

    Well, the failure modes for that quad are the same as they would be for any other quad. If the flight control computer fails, it's going to crash. Possibly in spectacular fashion. (Even RC helicopters and quads will crash if you lose the flight computer, with the possible exception of helicopters with a flybar rotor head piloted by very experienced pilots who can fly sans gyro long enough to land them.)

    That being said, I've never had a gyro or flight control computer fail yet, and I own 5 quads and 5 heli's (3 of which are CP, and two of those are flybarless). This represents a variety of radios, flight computers, and gyro/accelerometer combinations, and they're all very reliable despite being stupid-cheap.

    Although I will grant you that the whole "tossing the quad off a bridge" is a bit dramatic, and completely unnecessary. If I had one, I'd toss it out over the concrete, so that if it failed to start up for whatever reason (battery dead?), at least it would only clatter to the ground instead of falling 200 + feet into an inaccessible ravine!

    Really, the only failure that would be a catastrophic problem on initial take-off would be the case of throwing it off a bridge like that and the thing not starting up (forgot to turn it on?). If you were sensible and just tossed it in front of you, having it thud at your feet isn't an issue. (worst case you break a prop, which is an easy fix) And if it merely turned on a few seconds too late, it would likely rebound off the ground and continue flying. (I've "bounced" my quads several times while learning how to gauge the throttle response.)

    Adam

  7. #67
    Join Date
    May 2014
    Posts
    619

    Default

    For instance I'd like to see what happens if you launch it with a lot of random x-y-x spin... how quickly it would be able to compensate for that and fly upright. Can you do that with a regular quad, and it self-rights it self?

  8. #68
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,488,805

    Default

    Quote Originally Posted by HankLloydRight View Post
    For instance I'd like to see what happens if you launch it with a lot of random x-y-x spin... how quickly it would be able to compensate for that and fly upright. Can you do that with a regular quad, and it self-rights it self?
    Sometimes, yes. It depends on the quad and the flight mode I have selected. If I select "beginner" and "limited control response", then yes, the bird will right itself even if you launch it in a tumble, so long as you have enough throttle on. (The board only controls the speed split between the rotors, so if you have the throttle at or near zero nothing will happen because the blades don't have enough lift and the board will never increase the throttle setting by itself.)

    Granted, I haven't tried tossing it like a frisbee or intentionally trying to send it tumbling like a kicked football though! At some point I'm sure I could overload the abilities of the board, but with maximum throttle these things are amazingly maneuverable.

    The caveat is that it has to have enough altitude (for a given throttle setting) to right itself before it hits the ground. If not, then it will crash. (Though such a crash would be relatively mild.) That wouldn't be an issue for Lily, since the computer has direct control of the throttle as well as the speed split. So it should have maximum maneuverability on-demand at all times.

    Also bear in mind that the flight computer on my quads is quite different from what you would have on an autonomous bird like the Lily. The board on an RC quad is really only there to translate the RC controls into speed settings for the rotors. It will help you maintain a hover, and it has a heading-lock mode that helps prevent yawing, but that's about it. It's not actually capable of autonomous flight; it only has basic flight stabilization, and even that can be dialed back if you want to allow the bird to do flips or high-speed maneuvers (which require the quad to pitch and roll quite a lot in order to get the thrust aligned in the direction you want to go).

    And yes, a quad would be virtually un-flyable without a board like this, but if you gave the board full control of everything, it could force the bird to do some amazing tricks (like stabilize itself after being thrown like a frisbee, for example). I've seen videos on YouTube where researchers were using computer-controlled quads to balance a broomstick vertically on one quad, and then "tossing" the stick to another quad that would catch it. That sort of automation is probably what they have in the Lily. (Albeit perhaps not quite as complex, since it doesn't need to be all THAT maneuverable, nor does it need to worry about balancing a stick on it's hood.)

    Really though, one of the main functions of the board is to slow things down so that a human can react in time. Take the human pilot out of the loop and replace it with an electronic autopilot, and you can now vary the rotor speeds at a rate limited only by the on-board electronics. That's where the crazy maneuverability comes from.

    Edit: Here's that video: https://www.youtube.com/watch?v=pp89tTDxXuI



    Adam
    Last edited by buffo; 05-18-2015 at 14:43. Reason: added video

  9. #69
    Join Date
    Sep 2008
    Location
    Orlando, Florida
    Posts
    1,023

    Default

    Quote Originally Posted by HankLloydRight View Post
    For instance I'd like to see what happens if you launch it with a lot of random x-y-x spin... how quickly it would be able to compensate for that and fly upright. Can you do that with a regular quad, and it self-rights it self?
    I've actually done this type of testing with one of my micro quacopters. I can toss it into the air with a random x-y-z spin and as long as it isn't perfectly inverted it will recover. Multiple times I was able to "fool" it and get it inverted. When you throttle up it shoots straight into the ground. This was done over grass so there was no damage. Even over a hard surface though it won't be damaged since there is so little mass.

    Another test I did was to spin it like a top on a table then throttle up. If spun below the flight controllers threshold it will stop spinning and stabilize. If spun too quickly the flight controller can't react and continues to let the quad spin. It is uncontrollable but as long as you don't touch the controls it will hover. I would love to know what the controller thinks is going on. It's really amazing to see.

    Chris

  10. #70
    Join Date
    Jan 2006
    Location
    Charleston, SC
    Posts
    2,147,488,805

    Default

    Quote Originally Posted by Lazerjock View Post
    If spun too quickly the flight controller can't react and continues to let the quad spin. It is uncontrollable but as long as you don't touch the controls it will hover. I would love to know what the controller thinks is going on. It's really amazing to see.
    I need to try this! I would love to see it hovering in mid-air while spinning around super-fast!

    How did you set this up? Did you have it sitting on some kind of rotating base (lazy susan?) and spin it up that way?

    Adam

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •