Yes, your comments about outputting the channels in parallel from the same time beginning were exactly my thoughts, but I also want to be able to receive at the same time so I will place instructions inline rather than waitcnt. I am only going for 4 Rx and 4 Escs atm although my pcbs permit 8+8.
I will add the appropriate attributions and post the code here so others may enjoy too.
Thanks to Jason Dorie, here is his code with MIT Licence and attributions.
Note the Licence for the RC_Receiver is yet to be obtained from John (the author) - I have included in the source the thread where it was originally posted.
Here is a simple PASM ESC driver. Its purpose (initially) is to simply test all the motors on the QuadCopter.
It will drive (test) up to 4 ESCs and uses PST and the PC keyboard. It is easy to add more ESCs to the driver.
Timing is not absolutely perfect, but well within what is required in normal radio control.
I use the keyboard to set each ESC with keys 0,1,2,3 for ESC0; q,w,e,r for ESC1; aa,s,d,f for ESC2; and z,x,c,v for ESC3. <space> halts all ESCs. Values are OFF, 10%, 20% 30% but note that 10% will not turn any of my motors on, but 20% works. Also, 3 of my 4 motors will then run at 10% while the fourth will stops.
Note: code removed - see replacement code on Post #250
Here is a simple PASM ESC driver. Its purpose (initially) is to simply test all the motors on the QuadCopter.
It will drive (test) up to 4 ESCs and uses PST and the PC keyboard. It is easy to add more ESCs to the driver.
Timing is not absolutely perfect, but well within what is required in normal radio control.
I use the keyboard to set each ESC with keys 0,1,2,3 for ESC0; q,w,e,r for ESC1; aa,s,d,f for ESC2; and z,x,c,v for ESC3. <space> halts all ESCs. Values are OFF, 10%, 20% 30% but note that 10% will not turn any of my motors on, but 20% works. Also, 3 of my 4 motors will then run at 10% while the fourth will stops. Attachment not found.
Cluso99,
I am really interested at your ESC implementation using the Propeller, for many reasons. One is that it may be possible to build a better controller (faster response) when you have access to the ESC drive code/circuit, which in turn will allow better stability on the qcopter. But out of curiosity, how much current do you anticipate passing through a PCB trace? I can see passing up to 5 to 10 amps, but some of the motors in larger quads can draw 30 or 40 amps, in which case the trace resistance will become significant, to say nothing of cooling the mosfets. Also, by putting them on the main board, they will be away from the prop wash so they will not be cooled by the forced air. Would love to hear your insights and thoughts.
I think you misunderstood my driver. It only drives the ESCs which is low voltage/currrent.
I have and do propose to do a prop based pcb with 4 brushless controls, but a little later on. If you look at an ESC pcb you will see the size of the tracks used. I have yet to work out what current my motors require when providing good lift. I do not expect to deliver anywhere near 20A per motor. I would think 4 @ 10-12A would be fine, but only time will tell.
For heavy current requirements you can also get boards manufactured with heavier copper traces. I believe that most boards usually have 1oz copper plating. There is an open source motor control project OSMC and the boards for that were made with a much thicker 4oz copper plating which could handle higher currents for the same trace width.
Houston.. we have lift-off. Houston.. we have a problem.
Well, I finally have lift-off. But now I have to get counter-rotating props Well, I knew that already, but I had to try to see how bad the rotation was. Since we have a pool closeby, I decided it was not wise to risk lifting off further.
The propellers I was using (8"x6) for previous testing appeared to have too much lift. I have now switched to 10"x4.7 but all normal rotation unfortunately. My receivers throttle only goes between 1.1ms and 1.9ms. I have converted this to 1ms-1.8ms and then divided by 2 to be 1ms-1.4ms (subtract 1.1ms, divide by 2, add 1.0ms). I get lift-off at about 1/2 to 3/4 throttle so I may have to now reduce the scaling a little, or maybe just the offset.
Anyway, here are a few photos - sorry about the focus.
The QuadCopter with the electronics - my propeller BaseBoard and my ServoBoard are on the white page with my Nokia N85 phone. These 2 pcbs stack and I have a 3v3 regulator on my BaseBoard, so power is easily supplied from 1 Esc. The Wii Gyro is not shown here. My Spektrum DX6i is also in the pic. The aluminium arms are 1"sq tubing.
The QuadCopter underside showing the battery ziptied in place. I intend to use wide velcro.
The QuadCopter just off the ground. Note the electronics is currently in pink bubblewrap for easy access. The Quad is ratating due to all normal rotating blades
BTW I used some old pool cleaner hose that had deteriorated for the landing pads. Seems to be excellent and easy to mount on each arm with a cable tie.
Yes I know it is antistatic, and no it was what I had lying around. Ultimately it will be in a small Hammond 2"x2"x1" box, but for now I need to be able to access the pcb to plugin my propplug to update code, and to see what is happening.
I had another few small flights today. Not far off the ground because my counterrotating props have not arrived yet. Took a couple of short videos so will post on you tube later:)
Yes, I have to say that it is the first landing I've made
I never got that far with my model planes, although at least I was able to get enough altitude to clip my undercarriage on the football light posts at the park
The counter-rotating props were shipped over a week ago so they should be appearing anytime soon. When I ordered the originals and motors, etc, there were no counter-rotating pairs in stock.
BTW I have not yet added the gyro into the equation.
I see the thread title has changed.. so an update on my quad:
I have nearly all the parts- the motors, props, ESCs, HoverFLy Pro board, wires, connectors etc... I even picked up a Flip MinoHD camera from Woot! for $39 that I will mount on it.
I haven't worked on the quad lately because I want to laser cut some frame parts for it. I can't use my laser until I finish the remodeling in my shop because it is buried beneath a pile of... stuff. I hope to be finishing up in a week or so.
Meanwhile, my counter-rotating props arrived today. The are 8x4 and it takes a bit more throttle to lift-off. I adjusted the scaling back and have lift-off. Now I see that the gyro is required as the quad tips meaning not all motors are spinning at the same rpm, or the props are giving slightly different lift - surprise! surprise!
So the w/e will be hopefully getting the gyro providing feedback
I have a question for anyone of you that is writing code for your quadcopter. I have a real interest in seeing how your code is sensing the gyro's/accellerometer's/IMU's whatever your using, and reacting with the motor control. It's not that I'm too cheap to buy a Hoverfly unit, but instead would like to understand how your accomplishing your "balancing act"?
I have posted the code Jason Dorie uses.
To get off the ground firstly without counter-rotating props I just scaled the throttle to 50% and sent it to all 4 motors. This was only a few inches, so no surprises here. With the counter-rotating props they appear to be not so well balanced (motor/prop mix). I have not yet enabled the gyro into the mix although I do have the gyro module (a Wii Motion Plus clone from eBay) working.
I also wanted to say I have an IMU from Sparkfun and there is an object in the obex to use it. Maybe when Ken comes out with the Elev8 kit, I might try my hand at writing a program to control it. And if I don't destroy it first, there is always the Hoverfly.
Cluso99 I actually went back and found his original thread. He has modified Beau's H48C object to use an IDG300 dual axis gyro and a MMA7260 tri-axis accelerometer to create an IMU. Jason is one smart cookie! He has also included a kalman filter and calls his program "WolfSpyder". This will take me some time to digest while Ken works on Elev8. I had some ideas but after looking at Jason's code that gives me a good starting point. I am glad Ken is working on putting this Elev8 kit together. Thanks for this info Cluso99.
Would like to provide some updated information on the ELEV-8 platform we've had in testing for the last couple of months. Several posts ago I provided a list of changes to the ELEV-8 platform which have now been made and tested
The motor mounts had some vibration potential but it was eliminated by moving the spacers around.
The battery was moved up top, underneath the Hoverfly board, creating space for a camera (or robot) to be carried underneath.
The chassis was improved by using thinner Delrin and some spacers to create more space, without the loss of rigidity.
A variety of landing gear configurations were created, tested, revised, tested before we very happily accepted what is shown below. The landing gear configuration allows for sideways sliding on pavement, open view for carrying cameras, and breaks at the right time (before bending booms!).
Some design decisions were yours, others were made by us after testing. Keeping in mind that this is a low-volume, community-oriented project the parts had to be low-cost, easy to build or repair, and minimize the use of anything custom. Fully loaded (4000 mAh 3S battery, Hoverfly board, ESCs, receiver, motors - everything) we weigh in at under 3.2 lbs. This leaves a pound or two for payload.
If you want to see this exact version demonstrated, find the video from the October 27th Propeller Meetup Group (I don't think it's been posted yet).
People are usually concerned about crashing something that costs so much money. This concern isn't the same with a quadcopter. Due to the small number of moving parts and relatively simple design, our crashes have usually required the replacement of a propeller, sometimes a motor mount, and once in a while a motor. Only one intentional crash bent a boom and no body plates have been damaged. The motor problems were a result of omitting Loctite on metal to metal connections (motor set screws, for example). Since we've made hundreds of flights we will be able to provide a concise list of assembly tips to maximize your potential for success. But be advised: failure is part of this process, and a big part too!
Inside Parallax we have had tremendous help from Jessica, MattG, Nick, and Kevin C. It's a perfect crew for this project.
Below I'll detail a few more considerations for you. Please comment on anything, anyway you want.
Open-source project - 100% without restrictions
The drawings, design, and everything about the ELEV-8 are free, totally free (Creative Commons 3.0 Share Alike). You can make a kit too, or anybody can make one for any other controller. You can improve the design and sell them. You won't owe us a penny or any open-source license fees. We undertook this project because we want to further the development and use of the Propeller in aerial flight - it's the perfect controller for this kind of application and deserves more recognition. Much like we back Gadget Gangster as a Propeller platform, we also support Hoverfly Technologies in this application. Use the ELEV-8 in R/C or as autonomous with your own code (add a GPS, accelerometer). It's a good step to put any UAV under R/C control first to verify the airworthiness. Chances are you'll be flying this where you have no particular insurance, so liability is very high (and we'll accept none of that). And finally, other electronic designs are welcome. Cluso99's boards or anything else you wish to make can be put on the ELEV-8.
Drawing Formats: DXF, PDF and SolidWorks
The design is now in SolidWorks. We will offer E-drawing compatible files, DXFs, PDFs and native SolidWorks. The solid models and native formats will be available next week. These will be useful for anybody wishing to make their own, or to make modifications to our design.
+ or X Configuration
Both flight styles are supported. For R/C flight the + configuration is preferred because you only need to spot a single, bright boom denoting the front (or rear) of the craft. For first-person view you will probably prefer the X configuration because the landing gear is out of the way of the camera. Mounting holes will allow the Hoverfly board to be placed in both ways.
ELEV-8 Package Choices
We think you'll be happiest with three choices, but you should confirm:
Airframe with mounting hardware (target price of around $100)
Airframe with mounting hardware, with motors, ESCs, props, wires, connectors (target price around $325)
Airframe with mounting hardware, with motors, ESCs, props, wires, connectors and Hoverfly SPORT (target price around $550)
A couple of crash packages will be available too. All of the individual parts will be available. You still need a 5-channel or higher radio, costing $160 to $500.
The whole BOM with sources, manufacturer part numbers and costs will be available so you can compare it to building one on your own. No secrets! The SPORT board will be used until the OPEN is available from Hoverfly. This will be a full XLS set of data to make your life easier.
Does all of this sound okay to you?
Manufacturing Timing
Hobbyking is the longest lead time. We've had orders in with them since the middle of summer. Assuming we get confirmation that their motors and ESCs are on their way we can promise inventory no later than December 1st. Does that sound okay by you? Figure we all need a most awesome activity to undertake around Christmas. This lets the parts fit right behind a major Eddie build and avoids any manufacturing disruption association with the PropBOE. Any earlier or later is a problem and would push this till Christmas. Our initial build will be about 25 kits - a fairly small run but enough to satisfy.
Some photos of the latest revision follow. I've made some comments along the way. We'd appreciate any comments you wish to make.
Sincerely,
Ken Gracey
Motor mounts One more change will be made to the bottom motor mount, eliminating (16) 4-40 x 3/8" screws and (16) 4-40 nylon locknuts. We will attach the motors to the bottom motor mount with four screws straight into the motor housing. The only risk with making this change is that he user omits Loctite, in which case we can guarantee that vibration will shake the motor loose.
Top View You'll see that the battery can now fit underneath the Hoverfly SPORT board. This leaves room underneath for a camera. Additionally, the top and bottom chassis are spaced 1" apart. This provides lots of space for rearranging ESCs, adding wiring harnesses for LEDs, etc.
Upside Down You can see that there's room for a GoPro, a Flip (or several of them) or even an S2 robot. My personal goal is to fly with first-person view (FPV). Also, you'll see that we're using LED tape attached to the booms with heat-shrink tubing. We won't be supplying the LED tape in the kit, at least for now. It's available for a low cost off of eBay and a variety of web sites, and it's one of those personal touches that we all make to the kits we assemble. We will likely include the heat-shrink tubing, though, as it's pretty cheap and a bit of a procurement hassle to get exactly the right size. With LEDs it is actually easier to fly at night because you can keep the orientation under control.
Comments
Yes, your comments about outputting the channels in parallel from the same time beginning were exactly my thoughts, but I also want to be able to receive at the same time so I will place instructions inline rather than waitcnt. I am only going for 4 Rx and 4 Escs atm although my pcbs permit 8+8.
I will add the appropriate attributions and post the code here so others may enjoy too.
Note the Licence for the RC_Receiver is yet to be obtained from John (the author) - I have included in the source the thread where it was originally posted.
Jasons_QUadCopter.zip
It will drive (test) up to 4 ESCs and uses PST and the PC keyboard. It is easy to add more ESCs to the driver.
Timing is not absolutely perfect, but well within what is required in normal radio control.
I use the keyboard to set each ESC with keys 0,1,2,3 for ESC0; q,w,e,r for ESC1; aa,s,d,f for ESC2; and z,x,c,v for ESC3. <space> halts all ESCs. Values are OFF, 10%, 20% 30% but note that 10% will not turn any of my motors on, but 20% works. Also, 3 of my 4 motors will then run at 10% while the fourth will stops.
Note: code removed - see replacement code on Post #250
Cluso99,
I am really interested at your ESC implementation using the Propeller, for many reasons. One is that it may be possible to build a better controller (faster response) when you have access to the ESC drive code/circuit, which in turn will allow better stability on the qcopter. But out of curiosity, how much current do you anticipate passing through a PCB trace? I can see passing up to 5 to 10 amps, but some of the motors in larger quads can draw 30 or 40 amps, in which case the trace resistance will become significant, to say nothing of cooling the mosfets. Also, by putting them on the main board, they will be away from the prop wash so they will not be cooled by the forced air. Would love to hear your insights and thoughts.
I have and do propose to do a prop based pcb with 4 brushless controls, but a little later on. If you look at an ESC pcb you will see the size of the tracks used. I have yet to work out what current my motors require when providing good lift. I do not expect to deliver anywhere near 20A per motor. I would think 4 @ 10-12A would be fine, but only time will tell.
This code is derived from John Birkeland's RC_Receiver.spin
Note: code removed - see replacement code on Post #250
NOTE: Please change the _xinfreq = 5_000_000 ... I use 6_500_000 overclocked!
The same code performs the following tests...
- Tests an RC Receiver (4 channels)
- Tests the ESC Motor Drivers, or servo outputs (4 channels)
- Tests the RC Receiver and ESC Motor Drivers (set for QuadCopter - all channels initially off)
- Can sample any ESC output channel
With this code I can run the 4 brushless motors from the RC gear, with 1 channel for each motor.RC4_ESC4_Demo - Archive [Date 2011.10.08 Time 16.14].zip
Now to get the Wii Gyro working with this, and then incorporate Jasons PID code.
Well, I finally have lift-off. But now I have to get counter-rotating props
The propellers I was using (8"x6) for previous testing appeared to have too much lift. I have now switched to 10"x4.7 but all normal rotation unfortunately. My receivers throttle only goes between 1.1ms and 1.9ms. I have converted this to 1ms-1.8ms and then divided by 2 to be 1ms-1.4ms (subtract 1.1ms, divide by 2, add 1.0ms). I get lift-off at about 1/2 to 3/4 throttle so I may have to now reduce the scaling a little, or maybe just the offset.
Anyway, here are a few photos - sorry about the focus.
The QuadCopter with the electronics - my propeller BaseBoard and my ServoBoard are on the white page with my Nokia N85 phone. These 2 pcbs stack and I have a 3v3 regulator on my BaseBoard, so power is easily supplied from 1 Esc. The Wii Gyro is not shown here. My Spektrum DX6i is also in the pic. The aluminium arms are 1"sq tubing.
The QuadCopter underside showing the battery ziptied in place. I intend to use wide velcro.
The QuadCopter just off the ground. Note the electronics is currently in pink bubblewrap for easy access. The Quad is ratating due to all normal rotating blades
And here is the code
RC4_ESC4_Demo002 - Archive [Date 2011.10.09 Time 19.35].zip
BTW I used some old pool cleaner hose that had deteriorated for the landing pads. Seems to be excellent and easy to mount on each arm with a cable tie.
<iframe width="425" height="349" src="http://www.youtube.com/embed/chtIRE6rulc?hl=en&fs=1" frameborder="0" allowfullscreen></iframe>
Excellent progress!!
I never got that far with my model planes, although at least I was able to get enough altitude to clip my undercarriage on the football light posts at the park
The counter-rotating props were shipped over a week ago so they should be appearing anytime soon. When I ordered the originals and motors, etc, there were no counter-rotating pairs in stock.
BTW I have not yet added the gyro into the equation.
A first try with an alu frame, Turnigy motors, ESC 25A plush, 3S battery.
It lifts off but is unstable.
Jean Paul
Picture attached
Keep the info coming here please
And tested two cards : HoverflySport and a card using an ITG 3200 gyro with Jason Dorie program (modified )
I have nearly all the parts- the motors, props, ESCs, HoverFLy Pro board, wires, connectors etc... I even picked up a Flip MinoHD camera from Woot! for $39 that I will mount on it.
I haven't worked on the quad lately because I want to laser cut some frame parts for it. I can't use my laser until I finish the remodeling in my shop because it is buried beneath a pile of... stuff. I hope to be finishing up in a week or so.
Meanwhile, my counter-rotating props arrived today. The are 8x4 and it takes a bit more throttle to lift-off. I adjusted the scaling back and have lift-off. Now I see that the gyro is required as the quad tips meaning not all motors are spinning at the same rpm, or the props are giving slightly different lift - surprise! surprise!
So the w/e will be hopefully getting the gyro providing feedback
To get off the ground firstly without counter-rotating props I just scaled the throttle to 50% and sent it to all 4 motors. This was only a few inches, so no surprises here. With the counter-rotating props they appear to be not so well balanced (motor/prop mix). I have not yet enabled the gyro into the mix although I do have the gyro module (a Wii Motion Plus clone from eBay) working.
Would like to provide some updated information on the ELEV-8 platform we've had in testing for the last couple of months. Several posts ago I provided a list of changes to the ELEV-8 platform which have now been made and tested
- The motor mounts had some vibration potential but it was eliminated by moving the spacers around.
- The battery was moved up top, underneath the Hoverfly board, creating space for a camera (or robot) to be carried underneath.
- The chassis was improved by using thinner Delrin and some spacers to create more space, without the loss of rigidity.
- A variety of landing gear configurations were created, tested, revised, tested before we very happily accepted what is shown below. The landing gear configuration allows for sideways sliding on pavement, open view for carrying cameras, and breaks at the right time (before bending booms!).
Some design decisions were yours, others were made by us after testing. Keeping in mind that this is a low-volume, community-oriented project the parts had to be low-cost, easy to build or repair, and minimize the use of anything custom. Fully loaded (4000 mAh 3S battery, Hoverfly board, ESCs, receiver, motors - everything) we weigh in at under 3.2 lbs. This leaves a pound or two for payload.If you want to see this exact version demonstrated, find the video from the October 27th Propeller Meetup Group (I don't think it's been posted yet).
People are usually concerned about crashing something that costs so much money. This concern isn't the same with a quadcopter. Due to the small number of moving parts and relatively simple design, our crashes have usually required the replacement of a propeller, sometimes a motor mount, and once in a while a motor. Only one intentional crash bent a boom and no body plates have been damaged. The motor problems were a result of omitting Loctite on metal to metal connections (motor set screws, for example). Since we've made hundreds of flights we will be able to provide a concise list of assembly tips to maximize your potential for success. But be advised: failure is part of this process, and a big part too!
Inside Parallax we have had tremendous help from Jessica, MattG, Nick, and Kevin C. It's a perfect crew for this project.
Below I'll detail a few more considerations for you. Please comment on anything, anyway you want.
Open-source project - 100% without restrictions
The drawings, design, and everything about the ELEV-8 are free, totally free (Creative Commons 3.0 Share Alike). You can make a kit too, or anybody can make one for any other controller. You can improve the design and sell them. You won't owe us a penny or any open-source license fees. We undertook this project because we want to further the development and use of the Propeller in aerial flight - it's the perfect controller for this kind of application and deserves more recognition. Much like we back Gadget Gangster as a Propeller platform, we also support Hoverfly Technologies in this application. Use the ELEV-8 in R/C or as autonomous with your own code (add a GPS, accelerometer). It's a good step to put any UAV under R/C control first to verify the airworthiness. Chances are you'll be flying this where you have no particular insurance, so liability is very high (and we'll accept none of that). And finally, other electronic designs are welcome. Cluso99's boards or anything else you wish to make can be put on the ELEV-8.
Drawing Formats: DXF, PDF and SolidWorks
The design is now in SolidWorks. We will offer E-drawing compatible files, DXFs, PDFs and native SolidWorks. The solid models and native formats will be available next week. These will be useful for anybody wishing to make their own, or to make modifications to our design.
+ or X Configuration
Both flight styles are supported. For R/C flight the + configuration is preferred because you only need to spot a single, bright boom denoting the front (or rear) of the craft. For first-person view you will probably prefer the X configuration because the landing gear is out of the way of the camera. Mounting holes will allow the Hoverfly board to be placed in both ways.
ELEV-8 Package Choices
We think you'll be happiest with three choices, but you should confirm:
- Airframe with mounting hardware (target price of around $100)
- Airframe with mounting hardware, with motors, ESCs, props, wires, connectors (target price around $325)
- Airframe with mounting hardware, with motors, ESCs, props, wires, connectors and Hoverfly SPORT (target price around $550)
A couple of crash packages will be available too. All of the individual parts will be available. You still need a 5-channel or higher radio, costing $160 to $500.The whole BOM with sources, manufacturer part numbers and costs will be available so you can compare it to building one on your own. No secrets! The SPORT board will be used until the OPEN is available from Hoverfly. This will be a full XLS set of data to make your life easier.
Does all of this sound okay to you?
Manufacturing Timing
Hobbyking is the longest lead time. We've had orders in with them since the middle of summer. Assuming we get confirmation that their motors and ESCs are on their way we can promise inventory no later than December 1st. Does that sound okay by you? Figure we all need a most awesome activity to undertake around Christmas. This lets the parts fit right behind a major Eddie build and avoids any manufacturing disruption association with the PropBOE. Any earlier or later is a problem and would push this till Christmas. Our initial build will be about 25 kits - a fairly small run but enough to satisfy.
Some photos of the latest revision follow. I've made some comments along the way. We'd appreciate any comments you wish to make.
Sincerely,
Ken Gracey
Motor mounts One more change will be made to the bottom motor mount, eliminating (16) 4-40 x 3/8" screws and (16) 4-40 nylon locknuts. We will attach the motors to the bottom motor mount with four screws straight into the motor housing. The only risk with making this change is that he user omits Loctite, in which case we can guarantee that vibration will shake the motor loose.
Top View You'll see that the battery can now fit underneath the Hoverfly SPORT board. This leaves room underneath for a camera. Additionally, the top and bottom chassis are spaced 1" apart. This provides lots of space for rearranging ESCs, adding wiring harnesses for LEDs, etc.
Upside Down You can see that there's room for a GoPro, a Flip (or several of them) or even an S2 robot. My personal goal is to fly with first-person view (FPV). Also, you'll see that we're using LED tape attached to the booms with heat-shrink tubing. We won't be supplying the LED tape in the kit, at least for now. It's available for a low cost off of eBay and a variety of web sites, and it's one of those personal touches that we all make to the kits we assemble. We will likely include the heat-shrink tubing, though, as it's pretty cheap and a bit of a procurement hassle to get exactly the right size. With LEDs it is actually easier to fly at night because you can keep the orientation under control.