David Betz bravely posted some of his early ELEV-8 flights on YouTube. I couldn't keep watching his video, too many repressed memories were beginning to surface.
But I'm convinced that a large part of my problem was faulty assembly. I plan to try this again soon, maybe this weekend. I'm hoping for at least slightly better results. I'll try to get my son to film my attempts again so you can all have a good laugh! :-)
To get back to the original topic of this thread, I think I'm becoming convinced that the extra you pay for Hoverfly products is worth it. First, I believe they are made in the US, second they provide great support, third (maybe this should be first) they use the Propeller! I intend to get a Hoverfly Pro board at some point, maybe when Parallax starts selling them.
I do like the fact that I can reprogram them. A little tongue in cheek, but true - you can just plug in the prop tool and load your own software onto their board. I think the new one that Kevin Cook has mentioned they're working on will be cheaper and will include an accelerometer as well, so the performance could be comparable to the Hubsan. (They = Parallax and HoverFly)
I do like the fact that I can reprogram them. A little tongue in cheek, but true - you can just plug in the prop tool and load your own software onto their board. I think the new one that Kevin Cook has mentioned they're working on will be cheaper and will include an accelerometer as well, so the performance could be comparable to the Hubsan. (They = Parallax and HoverFly)
That new board sounds pretty cool. Where is the description of it?
I interpreted (perhaps incorrectly) Ken's recent post about HoverFly to mean that they weren't moving ahead with that.
Too bad. I guess one could take a HoverflyOpen board and add an accelerometer in the prototypig area. Of course you wouldn't be able to run Hoverfly code on it if you wanted to make use of the extra sensor. I wonder if there would be a way for Hoverfly to adopt some sort of plug-in architecture that would let them keep their code proprietary but still allow customization?
That's right, because this is essentially available in the Hoverfly Pro board, which we will supply soon.
Good point! Do you know if Hoverfly will be supplying some sort of rudimentary flight control software for the HoverflyOpen board or will the buyer have to "bring their own"? In other words, will there be some open source sample code to show how to use the hardware?
Good point! Do you know if Hoverfly will be supplying some sort of rudimentary flight control software for the HoverflyOpen board or will the buyer have to "bring their own"? In other words, will there be some open source sample code to show how to use the hardware?
Jason Dorie has posted code that should work with the Hoverfly Open. Here's the link.
Jason has mentioned he's working on an improved version which I believe he plans to release soon.
Duane / David - That code was for the Sport, and the pinouts for the Open board are slightly different. I do have a version of the code that works for the Open, and over the last little while I've been cleaning up and commenting the code a bit better. It's pretty much ready to release. I want to try to mimic their handling of the gain value for setting the sensitivity, but that's the last piece, and I may release it and then update it retroactively.
The DCM-based code I have that I was flying at the OPC is close, too. I need to add code to handle inversion more gracefully, but I may release it prior to that just so other people can start messing with it too. Currently it doesn't have a clue about being upside down, so it basically crashes. If you're not flying really aggressively you've never see that behavior. This code is based on Sparkfun's 9DOF sensor stick and a Parallax Protoboard. The only other thing on the board is some resistors and the ESC connectors, so anyone here could build it.
My Hubsan flies great. Trouble is in initializing it. The copter and transmitter only bind about one in ten times. Throttle at zero, transmitter on for several seconds, connect copter battery, usually nothing but blinking blue LEDs. Gotta repeat the process 8-10 times consistently before it binds (solid blue LEDs). Bad unit, or am I doing something wrong?
My Hubsan flies great. Trouble is in initializing it. The copter and transmitter only bind about one in ten times. Throttle at zero, transmitter on for several seconds, connect copter battery, usually nothing but blinking blue LEDs. Gotta repeat the process 8-10 times consistently before it binds (solid blue LEDs). Bad unit, or am I doing something wrong?
Only thing I can think of is that it must be at rest for a couple seconds before you can fly it. I believe this is to initialize the yaw gyro. You can initialize it on an angle and it still correctly knows which way is up so it does not have to be level but it does have to be absolutely still.
Well I took my Hubsan apart and it uses brushed DC motors has an Arm processor and uses a MPU3050C gyro with a motion processor that can accept an i2c input for an accelerometer. There is a 3rd chip next to the gyro that looks like it could be an accelerometer but I can't id the numbers on it.
Well I took my Hubsan apart and it uses brushed DC motors has an Arm processor and uses a MPU3050C gyro with a motion processor that can accept an i2c input for an accelerometer. There is a 3rd chip next to the gyro that looks like it could be an accelerometer but I can't id the numbers on it.
I have 3 pcbs that I have not found the time to build. First one is a prop 1.75"sq that has 8 servo inputs (typically from receiver) and 8 servo outputs (typically to drive escs). it has provision for daughter pcb with 10DOF. While I made the daughter pcb, it is now cheaper to buy this pcb assembled on ebay.
I also have a 1.75"sq pcb that distributes the power to up to 8 escs.
Iirc (i am not home) the prop has a microsd also - for logging data.
I have not done any worthwhile coding, but expect jasonscode would work fine with minor mods. I used a wii plus to obtain the gyro data and a wii for the accel data.
If anyone is interested, please pm me. The 1.75 pcbs fit in a hammond box-es. I anticipate making these pcbs on 2"sq form factor together with some other pcbs I am making in the next month or two.
I got my Hubsan H107 yesterday and played with it a little. If this is easier to fly than the Elev-8 then I need a lot more practice. I was able to fly it around a bit but I certainly wasn't very good at it! Has anyone taken one of these apart to see what they use for electronics? The box says it has a 6 axis gyro. I didn't know there was such a thing. Do they mean a 3 axis gyro and a 3 axis accelerometer?
I think I'm ready to try my Elev-8 again now that I broke my Hubsan H107 the second time I tried flying it. I did get it into the air and was able to control it a little. Unfortunately, it landed too hard once and one of the arms broke off. It looks like it would be pretty easy to superglue it back together and I'll probably try that but first I'm going to try the Elev-8 again. I'm afraid I'm not very good at flying these things but maybe I'll improve with practice.
...I broke my Hubsan H107 the second time I tried flying it. ...
That's why I've avoided the quadcopter scene altogether. It's the eye/hand coordination thing, by which I am sorely challenged. I can probably handle two dimensions okay (e.g. cars, boats, etc.), since they stay at ground/water level. But add that third dimension, and the potential for draining my hobby budget via expensive crashes looms. <sigh>
I have real bad eye/hand coordination because of some prob's awhile back and I am surprised I can even get the Elev-8 in the air. It maybe that your Husban needs to be trimmed in. I haven't trimmed mine yet and it always slides to one side after lift off. Maybe Rich will chime in about how he went about trimming his. I haven't trimmed mine yet as it is almost put back together. As far as flying your Elev-8 again you might want to try these things if your using the Hoverfly Sport or Open board - since these boards use a gyro only they have no reference for earth level, only the level they are started up at. When I first started flying I would take a small pocket bubble level bar with me and make sure my craft is at earth level in all directions before powering up.
Also make sure the trims on your aileron, elevator, and rudder on your transmitter are trimmed as close to 1500us output as you can get them when those sticks are not being touched. The main reason for making sure it is level before power up is that when you give throttle and take off the less corrective action will be needed on the aileron and elevator sticks as it will start out at earth level. Unless there is a problem you can ignore the rudder stick (make sure it stays centered) on your first flights.
EDIT: Make sure to give your self plenty of room like a wide open grass field.
When I first started flying I would take a small pocket bubble level bar with me and make sure my craft is at earth level in all directions before powering up.
Does that help? I thought the quad just needed to be still. The gyros just keep the quad from rotating without input. I would think gyros would drift too much to remember a horizontal orientation.
About needing to be coordinated to fly these things, I'm sure it helps to have good hand eye coordination but some activities require input much fast than the brain can consciously decide in the input needed. It doesn't matter how smart or coordinated someone is, if they haven't practiced the piano, they aren't going to be able to play from sheet music on their first try. It takes times for your fingers to automatically find the appropriate keys to correspond with the notes the eye sees.
Unless you have an exceptionally stable quadcopter, you're not going to be able to make the corrective inputs required fast enough to fly it. It takes time for your brain to be able to move the control inputs from conscious thought to unconscious action. I've lost count of how many times I've recommended a flight simulator to forum members. It makes all the difference in the world to be able to practice without breaking your aircraft. The flight simulator will pay for itself in crash packs you don't have to buy.
Another big help is if you can learn with someone who knows how to fly already. They can make sure your quad is set up and trimmed correctly before you try to fly it.
I don't think I've mentioned my list of ELEV-8 links in this thread yet. I have an incomplete list of ELEV-8 related threads in post #10 of my index.
I do not know how Hoverfly does it but Jason's program integrates all of the axis X,Y,Z and compares them with the stick positions. I know for the PID settings he came up with for the Elev-8 the gyro has to drift in number by 4096 before it even starts affecting the output from the proportional section.
And depending on the circumstances the gyro drifts that far in a second or 2. In Jason's program as soon as you start to give throttle it starts the gyro integration and also re-zero's all the axis integration when the throttle is off. So what ever attitude the craft is sitting at when you first give throttle is the starting attitude reference coming from the gyro to compare against the stick positions (thru the PID).
EDIT: It was the Hoverfly sport board I first used a bubble level with.
It maybe that your Husban needs to be trimmed in. I haven't trimmed mine yet and it always slides to one side after lift off. Maybe Rich will chime in about how he went about trimming his
Take off and hover a few feet off the ground, take note of control inputs required to stay in place, land and trim - repeat. Don't expect it to fly hands off, it will always need constant small corrections.
As far as flying your Elev-8 again you might want to try these things if your using the Hoverfly Sport or Open board - since these boards use a gyro only they have no reference for earth level, only the level they are started up at. When I first started flying I would take a small pocket bubble level bar with me and make sure my craft is at earth level in all directions before powering up.
The only "remembering" that the HoverFly Sport or Open does is it's orientation (attitude) when the sticks are centered. It is neutrally stable - that means that it will maintain its current attitude when there are no control inputs. The Hubsan X4 has positive stability, When control inputs are stopped (release the right stick) it will level itself. A helicopter can have negative stability, when you let go of the stick it's attitude changes at an accelerating rate.
If you take the trouble to level the ELEV-8 before liftoff the benefit will be lost the instant you make a control input.
The reason I leveled it before flight is the less interaction I had to do with aileron and elevator sticks to get it to hover. In other words it comes up almost in a hover because it is earth level to begin with. That helps compensate for my coordination problem.
I flew my Elev-8 again today. This attempt was much more successful than the previous one so I think my big problem last time was a propeller that wasn't screwed down tightly enough. I was able to get it off the ground and hovering and I flew it around a bit. Unfortunately, I eventually crashed it and bent one of the motor shafts. Since I don't have another one of those motors I would have to replace all of them with a different model of motor. Instead I'm going to give the whole thing and another unbuilt Elev-8 kit to my son to see if he has better luck with it.
David - what kind of motor was it? One of the blue ones? I have spares... I could send you a replacement shaft or even a whole motor. The shafts are pretty straightforward to replace. In fact, if you still have the packages, the motors usually come with spare shafts.
David - what kind of motor was it? One of the blue ones? I have spares... I could send you a replacement shaft or even a whole motor. The shafts are pretty straightforward to replace. In fact, if you still have the packages, the motors usually come with spare shafts.
It is one of the red motors. Ken Gracey already generously sent me two spares but I seem to have broken both of them. I successfully replaced the shaft on one. I guess I can try that again. I'm pretty sure I have another shaft. I've definitely decided that I need to take advise given to me by a number of people here and learn to fly on a simulator before trying an actual Elev-8 again though.
Edit: I just ordered RealFlight 6.5 which comes with a quadcopter model. Maybe if I practice with that and the Hubsan H107 I'll be more successful with the Elev-8 if and when I get it repaired again.
Practice with any of the helicopters and it'll help, too. There are a couple of larger ones that are pretty stable, and the Elev-8 controls are identical to those on a heli. I actually didn't like the quadrotor model in RealFlight - I found the control response too different from the real-life quads I've flown.
Comments
I interpreted (perhaps incorrectly) Ken's recent post about HoverFly to mean that they weren't moving ahead with that.
That's right, because this is essentially available in the Hoverfly Pro board, which we will supply soon.
Jason Dorie has posted code that should work with the Hoverfly Open. Here's the link.
Jason has mentioned he's working on an improved version which I believe he plans to release soon.
The DCM-based code I have that I was flying at the OPC is close, too. I need to add code to handle inversion more gracefully, but I may release it prior to that just so other people can start messing with it too. Currently it doesn't have a clue about being upside down, so it basically crashes. If you're not flying really aggressively you've never see that behavior. This code is based on Sparkfun's 9DOF sensor stick and a Parallax Protoboard. The only other thing on the board is some resistors and the ESC connectors, so anyone here could build it.
My Hubsan flies great. Trouble is in initializing it. The copter and transmitter only bind about one in ten times. Throttle at zero, transmitter on for several seconds, connect copter battery, usually nothing but blinking blue LEDs. Gotta repeat the process 8-10 times consistently before it binds (solid blue LEDs). Bad unit, or am I doing something wrong?
...Tiger
Only thing I can think of is that it must be at rest for a couple seconds before you can fly it. I believe this is to initialize the yaw gyro. You can initialize it on an angle and it still correctly knows which way is up so it does not have to be level but it does have to be absolutely still.
http://forums.parallax.com/showthread.php/148190-QuadX-source-code-for-Hoverfly-Open-Elev-8
I also have a 1.75"sq pcb that distributes the power to up to 8 escs.
Iirc (i am not home) the prop has a microsd also - for logging data.
I have not done any worthwhile coding, but expect jasonscode would work fine with minor mods. I used a wii plus to obtain the gyro data and a wii for the accel data.
If anyone is interested, please pm me. The 1.75 pcbs fit in a hammond box-es. I anticipate making these pcbs on 2"sq form factor together with some other pcbs I am making in the next month or two.
-Phil
Also make sure the trims on your aileron, elevator, and rudder on your transmitter are trimmed as close to 1500us output as you can get them when those sticks are not being touched. The main reason for making sure it is level before power up is that when you give throttle and take off the less corrective action will be needed on the aileron and elevator sticks as it will start out at earth level. Unless there is a problem you can ignore the rudder stick (make sure it stays centered) on your first flights.
EDIT: Make sure to give your self plenty of room like a wide open grass field.
Does that help? I thought the quad just needed to be still. The gyros just keep the quad from rotating without input. I would think gyros would drift too much to remember a horizontal orientation.
About needing to be coordinated to fly these things, I'm sure it helps to have good hand eye coordination but some activities require input much fast than the brain can consciously decide in the input needed. It doesn't matter how smart or coordinated someone is, if they haven't practiced the piano, they aren't going to be able to play from sheet music on their first try. It takes times for your fingers to automatically find the appropriate keys to correspond with the notes the eye sees.
Unless you have an exceptionally stable quadcopter, you're not going to be able to make the corrective inputs required fast enough to fly it. It takes time for your brain to be able to move the control inputs from conscious thought to unconscious action. I've lost count of how many times I've recommended a flight simulator to forum members. It makes all the difference in the world to be able to practice without breaking your aircraft. The flight simulator will pay for itself in crash packs you don't have to buy.
Another big help is if you can learn with someone who knows how to fly already. They can make sure your quad is set up and trimmed correctly before you try to fly it.
I don't think I've mentioned my list of ELEV-8 links in this thread yet. I have an incomplete list of ELEV-8 related threads in post #10 of my index.
And depending on the circumstances the gyro drifts that far in a second or 2. In Jason's program as soon as you start to give throttle it starts the gyro integration and also re-zero's all the axis integration when the throttle is off. So what ever attitude the craft is sitting at when you first give throttle is the starting attitude reference coming from the gyro to compare against the stick positions (thru the PID).
EDIT: It was the Hoverfly sport board I first used a bubble level with.
Take off and hover a few feet off the ground, take note of control inputs required to stay in place, land and trim - repeat. Don't expect it to fly hands off, it will always need constant small corrections.
The only "remembering" that the HoverFly Sport or Open does is it's orientation (attitude) when the sticks are centered. It is neutrally stable - that means that it will maintain its current attitude when there are no control inputs. The Hubsan X4 has positive stability, When control inputs are stopped (release the right stick) it will level itself. A helicopter can have negative stability, when you let go of the stick it's attitude changes at an accelerating rate.
If you take the trouble to level the ELEV-8 before liftoff the benefit will be lost the instant you make a control input.
Edit: I just ordered RealFlight 6.5 which comes with a quadcopter model. Maybe if I practice with that and the Hubsan H107 I'll be more successful with the Elev-8 if and when I get it repaired again.