Quadrotor Project

[Razor Concepts]

Well, I finally decided what I will be working on for the next 1.5 years - a quadrocopter.

Right now I want to focus on naviagation and other stuff, beyond just getting it to hover, so I will be starting with a proven system called UAVX. Right now UAVX has altitude hold, GPS hold, and auto landing, which is great. Unfortunately it uses a PIC18F2620 so I will be spending quite a bit of time adjusting to PIC C    Since the code is already there, I will probably only have to mess around with the gyro code to accommodate the different orientation, so shouldn't be too hard.

I modded the original circuit a little to my style, the PCB is 4" x 3" so that would be $30 just for one from BatchPCB! Looks like I will be using seeedstudio this time.

Specs:

• LPR530AL for pitch and roll gyro
• LY530AL for yaw gyro
• LIS3LV02DQ digital 3-axis accelerometer (features free-fall detection!! interesting)
• HMC6352 compass for long term yaw stabilization
• LS20031 5hz gps
• undetermined barometric pressure sensor

[madsci1016]

Become acquainted with www.diydrones.com community. They specialize in flying robots.

[Razor Concepts]

Unfortunately they focus on airplanes and not helicopters, so I resort to the RC groups community. Big 17 page long thread all about UAVX 

I also wonder why DIY drones is going for a complicated IMU setup for the airplanes. I mean, the UAVX can actually fly stabilized with just 2 gyros for pitch and roll, the yaw gyro and the accelerometer is not actually needed! Don't see how a simple UAV airplane needs a full 6dof IMU.

[madsci1016]

All electronic (or otherwise) gyros suffer from drift, that's a fact. I don't know how UAVX accounts for this, but the common way to fix drift is by combining data with an accelerometer.

And while the main tutorials are airplanes, there are a lot of member pages on copters.

[Razor Concepts]

Hmm, I would guess the gyro-only mode is designed for RC use, so the quad keeps relatively level and the user has to manually come in and actually fly it.

[madsci1016]

Hmm, I would guess the gyro-only mode is designed for RC use, so the quad keeps relatively level and the user has to manually come in and actually fly it.

Then it sounds like they are counting on a user to 'bump' the level every once in a while to fix the drift. That would be a problem if you want to go full autonomous, hence why the IMU has a 3 axis accelerometer.

[corrado33]

Hey Razor, I saw this on TV today, it's not a quadrotor, but it's a trirotor (I think), and it has 6 blades and motors. (EDIT: I was wrong, it's a six rotor, so a hexarotor?)  It relies solely on changes in RPM to move around.  It's very interesting.

http://www.draganfly.com/uav-helicopter/draganflyer-x6/

Oh, it also features auto hold (I think), basically it'll simply level off if you take your thumbs off of the controls, and it can land itself in an emergency.  And it can lose an engine and still land...

[Razor Concepts]

Yup, unfortunately it comes with a heavy price tag and closed source    I would really want some blades like that though, they might be more efficient than these normal RC airplane props.

[corrado33]

Yup, unfortunately it comes with a heavy price tag and closed source    I would really want some blades like that though, they might be more efficient than these normal RC airplane props.

Interesting...  I don't think it'd be too hard to at least build something like that.  Now programming it to work correctly... that's a different problem...

[Razor Concepts]

Updated board - the tradeoff between using the pitch-roll gyro is that it is slightly worse than the single-axis (and expensive at $50) ADXRS gyros! So my solution is to make a board that tests them all     I incorporated enough space for my gyros and the originally intended ones.


Right now this is actually looking pretty complete, I'll think it through for a bit until I use Seeedstudios fusion service. $40 for 10 10cm by 10cm boards is pretty good!

[dunk]

hey Razor,
out of interest, have you considered using QFN or TQFN packages for your ICs to bring the board size down?
as well as being smaller a package that does not use through hole can be mounted on only one side of the board so you can have something else on the other side.

on a flying platform weight is the enemy. i think you could shrink the board a lot with less through hole parts.
also if you make it small enough it will be cheaper (or you can get multiple boards for the same price as your big one).

out of interest, what breakout boards are you using for your gyros and accelerometer?


dunk.

[Razor Concepts]

Right now I am using "big" parts so that I can easily swap stuff in/out and make adjustments. This is basically my test board where I can modify stuff along the way. I plan on using this for a few months, and later make a final board that goes SMD for most of the ICs. This is my first foray into PICs so I think the final board will have a mega328 on board, connected through I2C for some stuff I want to do later.

Weight is a problem, but having flown RC helis, the extra few ounces of this board won't make a huge difference for the size quad I am making, the only thing that I don't like is the size of the board, but I need the space to change stuff out. And seeedstudio only has two PCB options now, 5cm x 5cm which would be impossible to fit all of this in, and a 10cm x 10cm. So might as well get the most out of my money  

The LPR/LY gyros are from sparkfun, the accelerometer and the ADXRS gyros (which I dont have yet) are sparkfun designs, but I believe are being made by a third party since the PCBs are green.

All the parts can be purchased from quadroufo.com. They actually sell a fully assembled and tested flight controller with all gyros, accelerometer and compass for $400, however they solder everything in for strength, and that means I can't remove any of the sensors  

[chelmi]

You could probably use linear voltage regulator in a smaller package, like TO92.
And if you have "unused" space, why not turn it into a "prototyping area" with a grid of .1" space holes?

[Razor Concepts]

Thanks for the tips! That big space was for the receiver, but I could use the prototyping area.

As for the voltage regulator, I may be powering additional items (there are two extra servo ports for pan/tilt of a camera), so I figured the TO-220 package would give me some room to use big regulators.

[dunk]

one thing i have on my UAV board which takes up minimal room is a ZXCT1009 current monitor and a voltage divider across the supply battery so i can monitor current draw and battery voltage.

this is more useful on my platform because i have 2 way radio but is still might be useful on your test platform to know your power status and it's cheap to implement. (providing you have spare a2d inputs.)


dunk.

[Razor Concepts]

Right now I have a voltage divider on the board power input (small 2cell lipo), unfortunately no more ADCs left for the motor power supply. That will be for later and the mega328 

[Razor Concepts]

While electronics are in progress I'm planning out the mechanics (considering it takes 3 weeks for stuff to arrive). Hopefully I should have a decent mechanical platform before school ends in late june or else I won't be able to use the awesome laser cutter.

Big motors! Funny how china pricing works out. Big scale power system totals at ~300, small scale at 250 and in between at 340+. So I just went for the huge scale.

150 gram brushless motor on 4s lipos, should put out 1500-2000 grams of thrust... times 4 motors. To power it all I am getting 3x 4s 5aH lipos for a total of 15aH at 14.8v nominal voltage! I wish I could have gone higher to 6s but that would have been significantly more expensive.

For frame material, I'm a fan of polycarbonate (lexan) so I may get some 3/16" thick sheet of it, make some clamping mounts to fit onto aluminum rods as the booms, and another lexan centerstructure.

[Razor Concepts]

Designing motor mount:


3/16" lexan (polycarbonate) using the cool bolt-nut gripping system to clamp it together. Will need to add some side plates to keep it from spinning around the boom but not bad for an hours worth of work. The boom will be an aluminum tail boom for a T-Rex 600 helicopter.

On the electronics front, I've begun making the circuit on some protoboard.

[Razor Concepts]

[Razor Concepts]

Here it is so far. Unfortunately, it will need to go through a complete redesign     I emailed Universal, the people who made the laser cutter my school has. They said the 50W laser could cut 3/16 polycarb, but it actually can't. So I wasted lots of time designing that, and $70 on a bunch of 3/16" polycarb. Great.....

[corrado33]

Here it is so far. Unfortunately, it will need to go through a complete redesign     I emailed Universal, the people who made the laser cutter my school has. They said the 50W laser could cut 3/16 polycarb, but it actually can't. So I wasted lots of time designing that, and $70 on a bunch of 3/16" polycarb. Great.....

Can't you just... you know... do it the old fashion way?  With a bandsaw or jigsaw and some sanding?  Just because you HAVE a very expensive tool doesn't mean you HAVE to use it.  Plus, it's plastic.  Plastic could be SANDED to any shape you want.  It's easy to work with.  Just don't get it too hot.  Heck, you might even be able to cut it with a hot wire.  Will it take less time on a laser cutter?  Yes, less time than a complete redesign... maybe, maybe not.  Still looks good though.

[Razor Concepts]

It's easy to work with.

Polycarbonate is a whole new deal    Not the stuff you can stick on a bandsaw and make nice rounded cuts or anything. You can't even work with that stuff with a hand saw. The accuracy of laser cutting is also crucial - if I cut by hand I will end up with a very crooked quadrocopter  

Anyways, a redesign won't take too long, most of it will be just adjusting the thickness of the pieces and shifting some things over a bit. I just got finished redoing the motor mounts in ~20 minutes.

[corrado33]

Polycarbonate is a whole new deal    

Eh I googled polycarbonate and saw plastic, therefore I assumed "softish".  Guess not!  I figured you need the laser for accuracy reasons.  Well, good luck then!  It is looking good.

[dunk]

Polycarbonate is a whole new deal    Not the stuff you can stick on a bandsaw and make nice rounded cuts or anything. You can't even work with that stuff with a hand saw.

not true.
very easy to cut with a hand saw.
also easy to sand.
when using power tools you have to be careful not to cut too fast otherwise it will melt rather than cut.

my favourite method of "cutting" thinnish polycarbonate is to simply score it with a sharp knife then snap it along the score mark.
this method isn't really practical for over 5mm thick though.

when you are working polycarbonate it is important not to leave any scratches as it can crack quite easily along these.
if you polish out any scratches though, it's really strong.


any reason not to do multiple cuts with the laser cutter or does the gunk from the first cut get in the way of the 2nd?


dunk.

[Razor Concepts]

any reason not to do multiple cuts with the laser cutter or does the gunk from the first cut get in the way of the 2nd?


dunk.

Hmm that could be a possibility, although it will take ~10 passes to cut it through. There may be some crazy melting/beveling going on, but I guess I'll try it and see how it turns out

[Razor Concepts]

The new ABS plastic version is here  
Only the motor mount so far.
Note to self: .12" wide -.15mm top and bottom, stock circle

[Razor Concepts]

Most of the cutting is done 

[TK]

Looks very nice.   How bad was it in the end to cut it all?

[Razor Concepts]

Oops haven't updated in a bit. I still need to get around taking pics of the electronics...


Fully deployed


Storage mode

[Joker94]

coming along nicely.

It is quite a sizable platform. I like it 

good luck and keep us updated on this awesome project.