Thanks Ken. I will keep a look out for the frames when ready and order a set. I have found USPS usually takes 14 days and no difference if it is first class or priority - in fact I experimented and the first class arrived 4 days before the priority and both were posted together!
BTW I didnt think there was anything odd about ordering 5 sets. Perhaps its because thats what I did
Anyhow... the motors arrived today. I was prepared to wait three weeks but it only took one. Everything is well packed. Each motor comes in it's own little box with prop hub and mount. The motors are very well made. Nicer than I was expecting.
Wow! If that was a HobbyKing order, you're a lucky guy. My last order from them was for two sizes of props to try on my little quad. I paid for EMS shipping ($27) and it took a full month and a half a dozen emails before it arrived. Everything was in stock. They just couldn't get it out the door for some reason. When the package finally arrived, there were no props in the box. I did get two laptop style power supplies! I decided it was hopeless and gave up.
See you ordered the HoverFly Pro. You can't do much better than that! I've been using the HK boards (gyro only). I'm about to build again using the OpenPilot board. It's only $89 and has the accelerometers (6DOF). It has provisions for an altimeter, GPS, and other stuff. Really looking forward to trying it.
Hadn't heard about the OpenPilot board before. It looks promising, the price is sure attractive. I'm really hoping that the HoverFly Pro is worth it. If not, you all will be sure to hear about it.
So I've modeled the Elev8 in SketchUp... using my motors. They don't quite fit. It may be good to make the mount just a little bigger to accommodate other motors. The Park480 motors that I am using have a diameter of 35mm.
Also, the hole pattern to mount the control board appears to only allow mounting it for "+" type flight. It would save a bit of trouble if the "x" type had holes available in the plates - the end user can drill through the tubes.
I have not had any problems with HobbyKing either although I have only used them a few times. I do read the comments on each of their parts before I buy.
That OpenPilot looks interesting.
I have decided to trash my round aluminium tubes and use lightweight 1"square aluminium tubing in an "I" frame layout. That way my motors will mount directly on the tubing and the horizontal I beam between front and rear beams (using X formation) will be the upper tube, so providing a little more protection for the electronics. The motors will all be on the same plane meaning no spacers required. I most likely will followup later with Ken's offer, for a nicer unit. My motors which came with integral mounts on the base will be easy to fasten. Assembly will now be much simpler.
I only need to do the frame and purchase propellers (the airplane variety) because I only have normal propellers and no conter-rotating ones. May have to kludge an angle on one of the motors in the meantime. At least that would get me started.
I have the electronics but will be looking for propeller (chip variety) software. Jason or others... are you around and willing to share MIT licence style???
PM me, or email jason.dorie(AT)gmail and I'll send you what I've got. I have a PASM module that will read an ITG-3200 and ADXL345 and computes angular values for the accelerometer, all on one cog. That's the workhorse of my non-DCM quad (and it doesn't actually use the accel values). I haven't gotten the DCM based code flying yet (see distractions above).
My Quad: I bought the aluminium frame, black annodised 2@ 25.4x25.4x1.2wallx600mm. I also found plastic "T" joiners so I bought 2 as well. That will make my "I" easlier. They also had plastic endstops but I thought better of the weight and didn't get these. (all at Bunnings in Oz). My motors fit nicely on the tube, but not drilled yet. I need to open the mounting holes on the motors as they are less than 3mm and I will be using M3x35mm bolts and locknuts.
I am away interstate tomorrow so will but and mount when I get back and post photos.
Realized I goofed. I need 3 sections for an I design. Then I remembered the hardware store had + joiners so I think when I return I will buy a new joiner and build the traditional x frame.
I picked up a couple lengths of 5/8" towel bar at Home Depot, they reportedly work well for quadcopter booms. A 30" length weighs 75 grams. I designed this motor mount for printing. It is 22 grams at 50% infill. Would probably work just fine with much lower infill.
W9GFQ - Do you have a 3D printer? If so, what kind is it? I've been looking hard at those this week. There seem to be a million to choose from and it's hard to know what's good and what isn't. The bottom end seems to be reprap and the top is hundreds of thousands of dollars. Must be something in the middle that's good. I'd be interested in any feedback or suggestions you have about them. Is that pic in red a simulation or a part you made? Also wonder what you mean by "infill".
At least three forum members (myself included) have the Makerbot Thing-O-Matic. Check this thread. You can also look here for more pictures of things that I have printed.
The Thing-O-Matic is far from perfect, but it does work well. There are plenty of things that you can do upgrade-wise to improve it. I only have experience printing with the Cupcakes and Thing-O-Matics.
The red thing is a part that I made.
Infill: If you were to print a cube with infill set to zero, then the cube would be hollow - and the "roof" of the cube may sag depending on the size. With infill set 100, the cube would be solid all the way through. I have mine set up so that the infill is a honeycomb pattern. It adds a great deal of strength to the part.
Shells: this is the outside surface of the part. If set to zero, there will be one shell. I typically print with one or two extra shells. The motor mounts were printed with shells set to one, for a total of two shells.
Rich - I read the entire thread and looked at all the cool stuff you've made. Very neat! Makes me want one of these more than ever. :-) I had been looking at these machines: http://www.dimensionprinting.com/ After reading about your build, I'm wondering if it wouldn't be practical to just put your extruder head in my CNC. You had pics of the extruder, but I think I need to find more info on that. It looks pretty simple. What is the diameter of your nozzle? On the commercial ones, they seem to fall between .005" and .013". Are you in that range? In your thread ReplicatorG was mentioned. I found the site for that and it looks like pretty normal G code for the most part. You'd have to strip out a few commands like nozzle temp, but I think most of it would run on the Haas. I noticed you simply build on a heated sheet. Most of the machines I've been looking at used a support material that had to be removed later. It also requires that there be two extrude heads (complicated). That's something else I need to explore I suppose. What a cool process! I can see I'm going to have to do "something". :-)
Many people do put the extrude heads on their CNC, they call it a RepStrap. The purpose is to print out the parts so that they can build a RepRap Mendel or Prusa. The nozzle I use is .40 mm or about .015". Makerbot has recently made available a .30mm nozzle that I would like to try soon.
The Makerbots and Mendels can use support material but it must be peeled away from the part. I very rarely use any support material and when I do, it is built into the geometry of the part. I am sure that in the near future there will be new Makerbots and Mendels that have two print heads. One for the part and the other for a water soluble support material.
The Dimension machine would certainly be nice to have. Rick G. has one, he brought some parts that he made to UPEW. They were very nice. The downside is that the material cost is quite a bit more.
Thanks Rich! I just looked up "Rick G." in the members list and sent him a PM. I hope I got the right guy. I'd LOVE to talk to someone that actually has one of these things and isn't trying to sell me something. That thing is a lot of money so I sure don't want any surprises. I don't know what the cost of running your machine is, but I suspect it's pretty low. The cartridges for the one I'm looking at are $250. That comes out to about $5 per cubic inch of material. Not great, but not as bad as it might be. I think I could live with that if everything else is really good.
I'll have to go looking for RepStrap now. I actually have two Haas CNC's so if I was doing printing in one, I'd still have the other to use until I was done. I wouldn't want to leave it setup for printing all the time, but if the stuff could be taken in and out pretty easy, it might be an option in the short term. I'm sure if I fall in love with the printing, I'm going to want a dedicated machine to do it on though. We'll see! I'm just exploring options right now. Thanks so much for all the info you have provided.
lol... I was abbreviating his name because I got the impression that he does not want attention drawn. It appears that I accidentally used his actual forum username. Since his only posts to the forum are about a submarine I am reasonably confident that it is the right person but if it isn't, let me know via pm and I will get you his number.
To keep the QUadCopters and others (Tri, Hexa, Octa stc) together, here is my build log...
Cluso's QuadCopter
Frame...
2 @ Aluminium square tube black annodised 600longx25.4x25.4x1.2mm (24"x1"x1"x3/64") cut in half to make 4 @ 300mm (12") long.
1 "+" plastic joiner
M3 nuts & bolts (for mounting motors - need to drill motor mounts out for these as no 2.5mm bolts available). I bought 20 @ M3x15mm metal thread screws and nuts. Locknuts were not available.
Motors etc...
5 @ Emax CF2822 11.1V 1200Kv (1 spare)
4 @ HK-20A ESC
4 @ PropSavers with extra O rings
2 @ LiPo Zippy 1600mAh 3S1P (had these previously)
Charger & monitor
Assortment of propellers but no counter-rotating ones yet
Radio Control...
Spektrum DXi6
Electronics...
Propeller pcb (my own BaseBlade1)
Servo pcb (my own Servo Blade)
3 axis gyro (canabilised from Wii Motion Plus clone)
3 axis accelerometer (canabilised from Wii Nunchuk clone)
Assembling the frame...
I drilled the outer pair of engine mounting holes 240mm from the inner end of the aluminium tubes and then a pair 16mm closer in because my engine mounts are at ~16mm sq . I center punched the holes but with a handdrill the holes did wander a little. I removed the engine mounts by loosening the grub screw and opened the holes slightly to allow for the M3 bolts. Then I assembled the 4 arms to the center plastic "+" piece. Next I screwed the bolts to the engine mounts with the nuts inside the tubing. Fortunately I was just able to hold the nuts to the end of my finger with bluetack and insert my finger far enough into the tubing to just reach the bolts.
Jason Dorie has gratiously sent me his Propeller chip code and I am checking it out later tonight.
Here are the motors mounted with propsavers fitted. You may see that two of the propsavers were too tight to fit the motor shafts so I will have to reem (drill) out the holes slightly. My props are all normal (none counter-rotating) 8x6. I have yet to buy counter-rotating props. At least I can get up to the point of running engines and remote control before I require the correct props.
Initially I will just wire the motors & escs on the outerside of the tube. Once all is functioning correctly I can place wiring inside the tubes.
Happy to see you on this thread, sharing your progress - exactly what I'd hoped for when I started.
Square tubing certainly has the advantages you pointed out. Motors can be directly mounted to the tubing and all kinds of options exist for mounting landing gear on the side of the tube. One big benefit is that the LED tape light will fit nicely on the bottom of square tubes, too.
The only oddity I noticed so far are the propellers you are using. Are these flimsy in some way, or just semi-transparent? I've only purchased two brands of propellers: a no-name eBay Chinese source and the well-known APC props from Woodland, CA (right down the freeway from Parallax!). The Chinese propellers felt quite flimsy to me so I didn't even fly them, so I don't know much about their performance but I'm pretty sure they'll fold or break on power-ups required to slow a vertical descent. The APC props are quite rigid and handle the forces nicely. More info on the propeller source would be appreciated.
I don't know if any suppliers have designed and produced quadcopter-specific propellers.The forces exerted on the propeller are quite unusual during some maneuvers, but perhaps it's no different than hanging your R/C airplane on the propeller when flying around a profile airplane. Somebody must know all about this. I'm just a "try and see" kind of experimenter.
We've been sidetracked on some truly time-demanding projects but we managed to get our big lead-time orders placed with HobbyKing so we never ever experience a shortage of parts again.
I designed this motor mount for printing. It is 22 grams at 50% infill. Would probably work just fine with much lower infill.
First class work, Rich! I'm not sure if you have a CNC milling machine or not, or if the part you designed is even suitable for injection molding, but I can imagine that part or a variation of it being taken to injection molding process.
Thanks Ken, I do have a CNC mini mill but that part was made on the MakerBot, designed in SketchUp. The CNC hasn't seen much use since the MakerBot showed up.
The square tube was a breeze to assemble. I just used an "X" plastic joiner sold with the tubing (at Bunnings for you Oz people). I bought a longer section than required and this provides some protection to the propellers. I can easily remount the motors in or out on the arms further. Currently the motor shafts are 9 3/4" from the centre pivot (i.e. 19 1/2" motor to motor along the arm). Forgot to mention.... The "X" connector does not really give a perfect right angle connection, but I don't think this will matter at all.
These props are black. They were spinning when I took the pic so perhaps that has something to do with them looking transparent (flash perhaps). I have a number of props and they are all from Hobbyking. But I dont have any counterrotating ones yet (because they were all out of stock when I bought them). I thought I would have a play to see what sizes suit. I am truly amazed at the power of these motors and props (8x6).
There is no way I will get these motors to full speed so I am going to divide (scale) my radio by 2.
W9GFO: Yes, I thought those parts were made with some form of RepRap. They are nice parts. My motors come with the included base mount which is slightly different to yours.
Here is my Quad with wiring. You will note I have not buried my wires and Escs inside the arms (at least not for now).
All 4 engines now work. The program is just a kludged version programmed into the Eeprom. From power up, there is a short delay, then all motors spin at 12.5%, then spin faster at 25% and then they turn off, and delay and repeat again.
Jason Dorie so kindly sent me his code and I have taken a good look at it. He has used RC_Receiver_4.spin which has no author listed, and Beau Schwabe's Servo32v3 Driver v1.3 with modifications by Jason. The main.spin code is by Jason, as is the ITG-Gyro module (excepting the I2C code from sdspiFemto by Michael Green). The IntPID routine (no author although I think this might be Jason's) is loaded in 3 cogs.
I need to do an alternative ITG-Gyro code to read the Wii gyro. The Wii gyro and nunchuck basics have already been done by Pat Daderko (DogP) and others, so this will be the starting point for this module.
Thinking about what we are doing here, and for the future, I think I may write a single pasm module to do both read the Receiver channels and control the servo (ESC) output channels. All processing will then be done by other cogs. My reasoning is simple... Make this code simple and place it in a single cog.
Another cog will be used to read the gyro, and later, also the accelerometer, compass and pressure, and possibly even a GPS. Currently this will be either the ITG3200 gyro module or the Wii gyro module. This will still fit within the framework of Jason's work so as to use his calculations.
My understanding is that an RC Receiver only outputs a single channel at a time, meaning 1 channel will have a pulse width between 1-2ms to indicate 0-100% while all other channels will be idle. Then the next channel is sent a signal, and so forth. Now, we can drive the motors effectively in parallel, and because we are calculating the motor revs dynamically within the QuadCopter (ie the Propeller chip), we may be able to take advantage of a higher resolution (repeating the pulse width) on all channels simultaneously.
I just have this really stupid idea that there may be enough time to calculate and vary motors such that all normal rotating props could be used and the quad could be allowed to rotate/spin while flying.. a real flying saucer!!
Anyway, I am interested in any feedback or suggestions.
BTW All my code will be MIT. I do need to check with Jason as there is no notice on his code.
I think the IntPID is mine, though it's likely based on someone else's code. The changes to Servo32 are to make it update 8 servos at 250 Hz instead of 32 servos at 50Hz. (At least I think it's 250Hz...) The reason being that you can get significantly more responsive control with a higher update rate.
On an older Microchip processor I was able to output 8 channels in 2ms by creating a list of "off events" - Sort the output pulses from shortest to longest, merge any with the same lengths, and then the output loop just becomes wait, mask, wait, mask, etc..
Turn on 1111
@1500, turn off 0010
@1650, turn off 1100
@1700, turn off 0001
RC receivers usually only output one channel at a time, but they vary by brand as to the order of the channels, and some higher end receivers, like Futaba, will actually output all channels in parallel, or groups of channels in parallel. That's why I used this RC receiver code instead of the ones in the Obex - The ones I've looked at there assume sequentially output channels.
Comments
BTW I didnt think there was anything odd about ordering 5 sets. Perhaps its because thats what I did
http://www.ecalc.ch/xcoptercalc_e.htm
A bit late, here's a video of a tri copter flying. If you watch close, especially right before liftoff, you can see the tail rotor tilting.
I was going to get the sport version but then discovered that it had no pressure sensor or accelerometers.
Anyhow... the motors arrived today. I was prepared to wait three weeks but it only took one. Everything is well packed. Each motor comes in it's own little box with prop hub and mount. The motors are very well made. Nicer than I was expecting.
See you ordered the HoverFly Pro. You can't do much better than that! I've been using the HK boards (gyro only). I'm about to build again using the OpenPilot board. It's only $89 and has the accelerometers (6DOF). It has provisions for an altimeter, GPS, and other stuff. Really looking forward to trying it.
Hadn't heard about the OpenPilot board before. It looks promising, the price is sure attractive. I'm really hoping that the HoverFly Pro is worth it. If not, you all will be sure to hear about it.
Also, the hole pattern to mount the control board appears to only allow mounting it for "+" type flight. It would save a bit of trouble if the "x" type had holes available in the plates - the end user can drill through the tubes.
That OpenPilot looks interesting.
I have decided to trash my round aluminium tubes and use lightweight 1"square aluminium tubing in an "I" frame layout. That way my motors will mount directly on the tubing and the horizontal I beam between front and rear beams (using X formation) will be the upper tube, so providing a little more protection for the electronics. The motors will all be on the same plane meaning no spacers required. I most likely will followup later with Ken's offer, for a nicer unit. My motors which came with integral mounts on the base will be easy to fasten. Assembly will now be much simpler.
I only need to do the frame and purchase propellers (the airplane variety) because I only have normal propellers and no conter-rotating ones. May have to kludge an angle on one of the motors in the meantime. At least that would get me started.
I have the electronics but will be looking for propeller (chip variety) software. Jason or others... are you around and willing to share MIT licence style???
PM me, or email jason.dorie(AT)gmail and I'll send you what I've got. I have a PASM module that will read an ITG-3200 and ADXL345 and computes angular values for the accelerometer, all on one cog. That's the workhorse of my non-DCM quad (and it doesn't actually use the accel values). I haven't gotten the DCM based code flying yet (see distractions above).
Jason
I am really impressed with that CNC.
My Quad: I bought the aluminium frame, black annodised 2@ 25.4x25.4x1.2wallx600mm. I also found plastic "T" joiners so I bought 2 as well. That will make my "I" easlier. They also had plastic endstops but I thought better of the weight and didn't get these. (all at Bunnings in Oz). My motors fit nicely on the tube, but not drilled yet. I need to open the mounting holes on the motors as they are less than 3mm and I will be using M3x35mm bolts and locknuts.
I am away interstate tomorrow so will but and mount when I get back and post photos.
The Thing-O-Matic is far from perfect, but it does work well. There are plenty of things that you can do upgrade-wise to improve it. I only have experience printing with the Cupcakes and Thing-O-Matics.
The red thing is a part that I made.
Infill: If you were to print a cube with infill set to zero, then the cube would be hollow - and the "roof" of the cube may sag depending on the size. With infill set 100, the cube would be solid all the way through. I have mine set up so that the infill is a honeycomb pattern. It adds a great deal of strength to the part.
Shells: this is the outside surface of the part. If set to zero, there will be one shell. I typically print with one or two extra shells. The motor mounts were printed with shells set to one, for a total of two shells.
The Makerbots and Mendels can use support material but it must be peeled away from the part. I very rarely use any support material and when I do, it is built into the geometry of the part. I am sure that in the near future there will be new Makerbots and Mendels that have two print heads. One for the part and the other for a water soluble support material.
The Dimension machine would certainly be nice to have. Rick G. has one, he brought some parts that he made to UPEW. They were very nice. The downside is that the material cost is quite a bit more.
I'll have to go looking for RepStrap now. I actually have two Haas CNC's so if I was doing printing in one, I'd still have the other to use until I was done. I wouldn't want to leave it setup for printing all the time, but if the stuff could be taken in and out pretty easy, it might be an option in the short term. I'm sure if I fall in love with the printing, I'm going to want a dedicated machine to do it on though. We'll see! I'm just exploring options right now. Thanks so much for all the info you have provided.
Cluso's QuadCopter
Frame...
- 2 @ Aluminium square tube black annodised 600longx25.4x25.4x1.2mm (24"x1"x1"x3/64") cut in half to make 4 @ 300mm (12") long.
- 1 "+" plastic joiner
- M3 nuts & bolts (for mounting motors - need to drill motor mounts out for these as no 2.5mm bolts available). I bought 20 @ M3x15mm metal thread screws and nuts. Locknuts were not available.
Motors etc...- 5 @ Emax CF2822 11.1V 1200Kv (1 spare)
- 4 @ HK-20A ESC
- 4 @ PropSavers with extra O rings
- 2 @ LiPo Zippy 1600mAh 3S1P (had these previously)
- Charger & monitor
- Assortment of propellers but no counter-rotating ones yet
Radio Control...- Spektrum DXi6
Electronics...- Propeller pcb (my own BaseBlade1)
- Servo pcb (my own Servo Blade)
- 3 axis gyro (canabilised from Wii Motion Plus clone)
- 3 axis accelerometer (canabilised from Wii Nunchuk clone)
Assembling the frame...I drilled the outer pair of engine mounting holes 240mm from the inner end of the aluminium tubes and then a pair 16mm closer in because my engine mounts are at ~16mm sq . I center punched the holes but with a handdrill the holes did wander a little. I removed the engine mounts by loosening the grub screw and opened the holes slightly to allow for the M3 bolts. Then I assembled the 4 arms to the center plastic "+" piece. Next I screwed the bolts to the engine mounts with the nuts inside the tubing. Fortunately I was just able to hold the nuts to the end of my finger with bluetack and insert my finger far enough into the tubing to just reach the bolts.
Jason Dorie has gratiously sent me his Propeller chip code and I am checking it out later tonight.
Here are the motors mounted with propsavers fitted. You may see that two of the propsavers were too tight to fit the motor shafts so I will have to reem (drill) out the holes slightly. My props are all normal (none counter-rotating) 8x6. I have yet to buy counter-rotating props. At least I can get up to the point of running engines and remote control before I require the correct props.
Initially I will just wire the motors & escs on the outerside of the tube. Once all is functioning correctly I can place wiring inside the tubes.
All 4 motors wired to their escs and each one tested to my prop hardware.
I have wired 2 of the motors and escs together, and controlled them with the prop (see the second pic - the propellers are actually spinning slowly)
Once I get to 30% speed I really have to hold the quad down and I only have 2 motors running!
Happy to see you on this thread, sharing your progress - exactly what I'd hoped for when I started.
Square tubing certainly has the advantages you pointed out. Motors can be directly mounted to the tubing and all kinds of options exist for mounting landing gear on the side of the tube. One big benefit is that the LED tape light will fit nicely on the bottom of square tubes, too.
The only oddity I noticed so far are the propellers you are using. Are these flimsy in some way, or just semi-transparent? I've only purchased two brands of propellers: a no-name eBay Chinese source and the well-known APC props from Woodland, CA (right down the freeway from Parallax!). The Chinese propellers felt quite flimsy to me so I didn't even fly them, so I don't know much about their performance but I'm pretty sure they'll fold or break on power-ups required to slow a vertical descent. The APC props are quite rigid and handle the forces nicely. More info on the propeller source would be appreciated.
I don't know if any suppliers have designed and produced quadcopter-specific propellers.The forces exerted on the propeller are quite unusual during some maneuvers, but perhaps it's no different than hanging your R/C airplane on the propeller when flying around a profile airplane. Somebody must know all about this. I'm just a "try and see" kind of experimenter.
We've been sidetracked on some truly time-demanding projects but we managed to get our big lead-time orders placed with HobbyKing so we never ever experience a shortage of parts again.
Ken Gracey
First class work, Rich! I'm not sure if you have a CNC milling machine or not, or if the part you designed is even suitable for injection molding, but I can imagine that part or a variation of it being taken to injection molding process.
Ken Gracey
The square tube was a breeze to assemble. I just used an "X" plastic joiner sold with the tubing (at Bunnings for you Oz people). I bought a longer section than required and this provides some protection to the propellers. I can easily remount the motors in or out on the arms further. Currently the motor shafts are 9 3/4" from the centre pivot (i.e. 19 1/2" motor to motor along the arm). Forgot to mention.... The "X" connector does not really give a perfect right angle connection, but I don't think this will matter at all.
These props are black. They were spinning when I took the pic so perhaps that has something to do with them looking transparent (flash perhaps). I have a number of props and they are all from Hobbyking. But I dont have any counterrotating ones yet (because they were all out of stock when I bought them). I thought I would have a play to see what sizes suit. I am truly amazed at the power of these motors and props (8x6).
There is no way I will get these motors to full speed so I am going to divide (scale) my radio by 2.
W9GFO: Yes, I thought those parts were made with some form of RepRap. They are nice parts. My motors come with the included base mount which is slightly different to yours.
All 4 engines now work. The program is just a kludged version programmed into the Eeprom. From power up, there is a short delay, then all motors spin at 12.5%, then spin faster at 25% and then they turn off, and delay and repeat again.
ESC_Motor_Control_rr006.spin
Jason Dorie so kindly sent me his code and I have taken a good look at it. He has used RC_Receiver_4.spin which has no author listed, and Beau Schwabe's Servo32v3 Driver v1.3 with modifications by Jason. The main.spin code is by Jason, as is the ITG-Gyro module (excepting the I2C code from sdspiFemto by Michael Green). The IntPID routine (no author although I think this might be Jason's) is loaded in 3 cogs.
I need to do an alternative ITG-Gyro code to read the Wii gyro. The Wii gyro and nunchuck basics have already been done by Pat Daderko (DogP) and others, so this will be the starting point for this module.
Thinking about what we are doing here, and for the future, I think I may write a single pasm module to do both read the Receiver channels and control the servo (ESC) output channels. All processing will then be done by other cogs. My reasoning is simple... Make this code simple and place it in a single cog.
Another cog will be used to read the gyro, and later, also the accelerometer, compass and pressure, and possibly even a GPS. Currently this will be either the ITG3200 gyro module or the Wii gyro module. This will still fit within the framework of Jason's work so as to use his calculations.
My understanding is that an RC Receiver only outputs a single channel at a time, meaning 1 channel will have a pulse width between 1-2ms to indicate 0-100% while all other channels will be idle. Then the next channel is sent a signal, and so forth. Now, we can drive the motors effectively in parallel, and because we are calculating the motor revs dynamically within the QuadCopter (ie the Propeller chip), we may be able to take advantage of a higher resolution (repeating the pulse width) on all channels simultaneously.
I just have this really stupid idea that there may be enough time to calculate and vary motors such that all normal rotating props could be used and the quad could be allowed to rotate/spin while flying.. a real flying saucer!!
Anyway, I am interested in any feedback or suggestions.
BTW All my code will be MIT. I do need to check with Jason as there is no notice on his code.
I modified it to read only 4 channels so I'd get slightly better potential resolution, though it's not really necessary.
I think the IntPID is mine, though it's likely based on someone else's code. The changes to Servo32 are to make it update 8 servos at 250 Hz instead of 32 servos at 50Hz. (At least I think it's 250Hz...) The reason being that you can get significantly more responsive control with a higher update rate.
On an older Microchip processor I was able to output 8 channels in 2ms by creating a list of "off events" - Sort the output pulses from shortest to longest, merge any with the same lengths, and then the output loop just becomes wait, mask, wait, mask, etc..
Example:
Becomes:
RC receivers usually only output one channel at a time, but they vary by brand as to the order of the channels, and some higher end receivers, like Futaba, will actually output all channels in parallel, or groups of channels in parallel. That's why I used this RC receiver code instead of the ones in the Obex - The ones I've looked at there assume sequentially output channels.
...and yes, my code can be shared MIT style.