Quadrotor Build Thread

[anwar]

So far, I have seen tons of discussion and various threads on this forum about building quad-rotors. But people do not seem to follow up here, and post the results of their work.  So here goes my attempt to build one, based on the design in this post.

http://www.rcindia.org/self-designed-diy-and-college-projects/quad-rotor/msg317/#msg317

This is a very simple quad rotor design (as it does not use micro-controllers, accelerometers, software etc).  It is expected to be a bit challenging to fly, but hopefully my heli flying skills would compensate for that.

Due to the nature of my work/routine, it may take a while to finish this... I hope it is short enough !

The parts list is as follows.

1.  4 x TowerPro Motor/ESC combos (25A)
2.  2 x Normal and 2 x Counter rotating 3-blade props
3.  3 x HK401B gyros
4.  3 x V-Tail mixers.

There are still many decisions to be made, like what battery to use (or rather, how many to use).  My plan is to use a 7 channel receiver, so that I can control the gain of all the different gyros from the TX itself.

The first step is to build a basic frame, it was done with balsa.

[Gaurav]

Wow that looks pretty good!!! Actually I find your design using vtail mixers with gyros a lil complex as oppossed to acceleros and gyros only ...... What props are those? I think I have the same ones....

[anwar]

The counter rotating props are these : http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=5249

BTW, this is absolutely not MY design, it is from someone at rcgroups.com, posted here by Sai.

[anwar]

What a mess

[anwar]

Update :

1 >  The motors were mounted with just cable ties.  Didn't feel comfortable with 3 screws going to the balsa frame, did not want to spend time making another mount plate using thin ply.  This seems to work OK for now.

2 >  Since the first step is only to see if the whole thing works and not build quality, the quad now has "legs" which are just blocks of sponge affixed to the 4 arms of the quad using just fiber tape Planning to change it to something more complicated later on. 

3 >  In testing just one motor (one of the elevator ones), it became apparent that all channels need to work like throttle channels.  On Futaba, the first impact of this is that all channels need to be reversed.  The second impact of this is that other than throttle, all channels need to have a "V" shaped curve so that the quadrotor will have a true "idle".  By default, each channel other than throttle sends out a signal to the receiver that is 50% of its maximum range (so that servos are centered, when the sticks are in their middle positions).  In this case, what we want is that the output should be 0 at middle stick (so the motor does not spin at mid stick for non-throttle channels).  These should then increase to full power on the motors as the sticks (aileron, elevator, rudder) are moved to either side.  This is why a "V" curve is needed on these channels.  On the Futaba 12FG, this is easily done using the "AFR" menu, and choosing a "POINT" curve (instead of an exponential curve).

[saurabhhsrivastavaa]

@ Anwar,

The pics look nice.. It is coming up really well. Though I do not have much experience with helis, but I am sure your heli skills will definately play an important role here.

I guess you would have made one of the arms as master / front, which will guide you to decide the ailerons and elevator during the flight. I suggest you put an high intensity LED to it for demarkation.

How are you doing the CG of this model? I am sure you will take care of it

All the best  and keep us posted. 

[anwar]

Ok, got it into a stage where I can just get it to lift off.  But there are significant issues which I am working through by tweaking the setup, before it can really hover !  I will post details of what has been done so far in a little bit.

[harsh]

way to go anwar....
looks awsome

[rcforall]

Watching this thread carefully , When a heli pro like Anwar does this I guess it makes life for us much easier . once he  gets it to fly well I am going to take all his guidance  to make mine .

Anwar ,
I have supplied a lot of equipment for quads :  but till date have seen successful  completion of so few of them that I am not surprised that the are so many incomplete threads on this  subject.

With you  doing it I think  this could be the first one  that will reach its logical conslusion.

Sai

[VC]

When a heli pro like Anwar does this I guess it makes life for us much easier . one he  get is to fly well I am going to take all his guidance  to make mine

Sai, you took the words out of my mouth. This is going to prove to be the mother of all Quadrotor threads. Anwar Sir, please keep your communication on this thread idiot proof!

I have waited a long time for this to happen.

All the best!

VC

[rohitgupta322]

Wow, it looks awesome. Pretty sure that anwar uncle will get it hovering in the air soon.

[anwar]

Update so far.

4.  Forget everything I said in [3] above   It sounded like a good idea at first while testing on a single motor, but realized later on that this would totally mess up the vtail mixing when multiple motors are in play.  So now we have plain / linear / default response set on all the four channels.   

5.  The next challenge was getting the ESCs to initialize and set the max throttle range.  Because of the gyros and vtail mixers, there was some output signal value being fed to the ESCs all the time when the entire setup is wired up, so one or more ESCs would refuse to arm. After trying various things, including setting up a different "flight condition" on the 12FG which would simulate the controls being pulled to the extremes just for initializing the motors, I finally settled on just setting up each ESC separately, including its throttle range.  One of the ESCs on the elevator still refuses to arm, unless I apply some elevator trim.

6.  The big challenge now is that even though the quad starts off stable, any tilt beyond say 20 degrees seems to get amplified and then it goes into a wild rocking motion which gets progressively worse.  This is regardless of any gain setting for the gyros, any amount of delay and no delay, and any amount of ATV/endpoint settings of the 3 gyro channels.  I also tried another set of props of different size.  I was about to try longer arms, when I came upon this post, which seems to describe the exact situation I am in now.

http://www.rcgroups.com/forums/showpost.php?p=14802789&postcount=31

So the primary suspect now is delay introduced by the V-Tail mixers. Options now are to try getting better v-tail mixers, or go for a slightly different design that avoids the vtail mixers entirely (where the gyros drive the ESCs directly).  It sounds more fun to get the vtail mixer based version to work (the design is so simple), so that is the direction I am going to take, and the hunt for better quality mixers is on.

I wonder how Sahil's mixers would work ! 

And I wish I had better testing tools, at least something like this :

http://www.hobbyking.com/hobbyking/store/uh_viewItem.asp?idProduct=10572

[anwar]

Another guy who has run into the same issue.

http://www.rcgroups.com/forums/showpost.php?p=14055152&postcount=49

He seems to have ditched the project, and went for a mC based one.

[anwar]

I have been chipping away at this as time permits.

7.  Got the ESC arming issues solved by powering up the ESCs and rest of electronics separately.  For  now, just added another receiver battery to power the electronics first, then power up the ESCs later on using a big lipo. The converse also seems to work (ie, arm the ESCs first, then power up the receiver etc).

8.  The hunt for a better mixer is still on. Our local mixer king Sahil confirmed that his mixers are also low res.  Further investigation has lead me to this thread http://www.rcgroups.com/forums/showthread.php?t=1218835 .  So I have ordered 3 Digimix IIs.  They are sort of expensive at $60 for all 3 including shipping, and could have got a mC based setup working for similar prices.  But there is certain challenge in getting the simple design to work, so I am pushing it.  Plus, our club (Qatar Scientic Club) has expressed interest in the quad if I can get it working well, so I may just sell it to them when I am done.

9.  Some people have suggested moving the gyros after the mixers (the above thread has got that working on a tricopter), but I don't see how it can work on the quad design being used by me.  Comments ? 

10.  I can always cheat the design by getting rid of the vtail mixers and just going with 5 gyros and 1 or 2 servos to tilt the motors that serve as "rudder".  But then I have to prove Sai's version as viable

[anwar]

Apparently, people who have attempted this have run into similar trouble before, as yet another thread reveals.

http://www.rcgroups.com/forums/showthread.php?t=1203378

This particular post from the above thread is very telling about the mixer quality issue : http://www.rcgroups.com/forums/showpost.php?p=14585139&postcount=6

[rcforall]

Anwar ,
This not my version , this is what some one tried on RCG , I wonder  how he got the video of it functioning I have not tried it as well hence  no way my version.
The way I think this should be looked at is  as a platform to start from then  what ever is needed  to get it working with your expertise is what  should be added or deleted  so that  you can finally have a quad design on a low budget.


sai

[anwar]

Sai - It is clear that it is from RCG, I only meant that you publicized it here

Getting a working version is not a problem, as long as you use the right components.

I am tempted to try this, as time permits : 

http://www.embedds.com/mixer-of-two-servo-signals-for-v-tailed-airplanes/
http://www.gadgetparadise.com/121901/vtailmixer.pdf
http://www.gadgetparadise.com/121901/VTailMixer.asm

Rich Smith from RCG who confirmed that he has it working with a $1 AVR chip responded saying that he will forward the code when he is back from his present situation that is keeping from away from home. He said that he tweaked one of the commonly available mixer codes for the Tiny13s. This would keep the cost to a minimum, at the expense of added effort.

[anwar]

Answering a related question from another thread : http://www.rcindia.org/radios-and-receivers/gyro-output-digital-mode-and-escs/msg21977/#msg21977

Yes, I tried different soft start settings.  I was planning to try different timing settings too.  It was while testing the soft-start settings that I tried hooking up two servos to the output, and that is when I noticed the big delay in response to gyro induced movements (the delays on the gyros were set to minimum, and limits and gain were set to maximum).  So I started reading up on delays introduced by mixers, and did not get back to testing various ESC timing settings.  Will test them out anyways, but I seriously doubt if they will help.  Now that I am waiting for the new mixers, I have some time to play with !

[anwar]

Since our club is interested in a quad build, I have ordered this also :

https://store.diydrones.com/ProductDetails.asp?ProductCode=KT-ArduIMU-20

Basically, this has everything one needs to build a quad.  It is more fun to assemble such a board together with individual gyro components, accelerometers, play with some code etc; but I know myself too well to go that route now. Life is fairly busy, soldering skills are sub-optimal, etc etc  The software part can still be tweaked, so a lot of  fun still remains.  

This board is very cost effective too, for the amount of stuff that is on there.

The plan is to get the basic system working with better mixers, but this is a side effort to get the quad stable at the earliest.

http://diydrones.com/profiles/blogs/arduimu-quadcopter

http://www.diydrones.com/photo/arduimu-quadcopter-2?

[Gaurav]

Hey Anwar I have almost the exact same setup.... Its just my Arduino is separate, and the IMU board is separate... Do you know Arduino [rogramming? I'd recommend you look up www.Arduino.cc if you dont..

[anwar]

If your quad is flying, please post pictures/videos

[Gaurav]

Oh havent assembled it yet too busy with the software part, will do so ASAP...

[rcforall]

I have the ardu imu board  can you point me to the codes  for using it as a quad

[anwar]

The ardu-imu codebase is here : http://code.google.com/p/ardu-imu/downloads/list

One quad specific version is here : http://diydrones.com/profiles/blogs/arduimu-quadcopter (look for a download link named "Quad1_15.zip")

[sunk?]

hi guys.
not into electronics............so what is the 'ardu-imu '??
a complete software package that talks to esc,gyro etc ?
regards
don