MultiWii

Hi,
I want to use the flying wing mode for my small 50cm (20") wing.
Not sure if I should use MWii with a gyro only or with gyro+acc?

I intend to
- add passthrough mode accessible via aux1/aux2 switching (like sensors)
- add throttle-in and throttle out (basic passthrough at the moment)

want to use LCDconfig for config as always.

Any and all pointers welcome on what suitable PID values might be.
Thanks, Hamburger

Update 01.2012 read latest posts for my thoughts on having reached the goals and thedevelopment being complete./Update

When I get back home from my vacation I will get a Swift for FPV use and I will try this with the Flying Wing mode.
I will also try the airplane mode in my Skywalker and see what happens

//UndCon

Hi Hamburger,
please have a look at this thread:
viewtopic.php?f=8&t=364&start=50
There is all the data you may need including a modified version of MWII done by PatrcikE especially for planes.
I've build a VTOL using TRI mode and FlyingWing mode of Multiwii. I'm using a Falcon Flying wing (hobbyking) three motors like a Tricopter and AUX2 to switch horizontally the two front motors.
MWII works wonderfully and quite easy to modify according to the needs you may have.

cob wrote:please have a look at this thread:

cob, thanks for the hint. I had been following that thread and found it inspiring.
I want to progress a little different though. I think having passtrhough as an item one can asign to aux1/aux2 like mag/baro/level/etc. is a more general approach and not too difficult to code.
motor passthrough is merely convenient to use arm/disarm in order to avoid havoc while doing stick-configuring.
Hamburger

Hi Hamburger,

I think this will work fine for you...


AUX1 to switch between Passthru and MWii mode.
Aux2 is free to activate things in GUI.

If you have more cannels on your radio it's possible to use CAMPITCH or CAMROLL instead of AUX1.
Leaving AUX1 free for GUI.


/Patrik

here is the first step towards Flying Wing setup.
done (MWii):
- added passthrough variant (still does mixing of pitch and roll sticks)
- activate passthrough with aux1=low (<1300) to honor the 3-way switching methodology
- added 4 defines for 2 servos 2 directions for pitch and roll (no need to swap servo cables at rx with dx6i)
- added defines for min/mid/max values for servo travels and trim

to do:
- GUI: add checkboxes to asign passthrough dynamically to aux1/aux2 states
- GUI: display current mode (passthrough activated or off)
- MWii: honor checkboxes' settings for passthrough asignment
- MWii: transmit passtrough mode to GUI
- MWii: add variant with one motor (basically 1:1 from rxData)

Hamburger

Cool - a MSComposit Swift II is heading my way.

I will try the new code as well...

So there wasn't some flying wing code already present in the multiwii code, before this new code? Can you explain to me in novice terms this passthrough function?
Is it possible to also have autolevel with the use of bma020? Thanks!

Passthru is a way to disable Mwii an just send the RX-signals direct to the servos.
A handlaunch can be a little wiolent and the gyros will overreact.

Autolevel can be used in Flying wing mode.

/Patrik

Great so for launch one switches off, and then switches stabilization on in flight ..Great,...Have you seen fmkit's videos of wii stabilization (although he seems to use different code and platform),..will that kind of stabilization also be possible with multiwii code ... I'll try to make a depron zephyr for testing, but first have to make my exams

My wing is maidened and ready to fly

This wing will be equipped with FPV gear when I get my camera but until then I can play with some MultiWii on board.

Hope you make some video so we see the stabilization in action

I tried to connect the board and servos went crazy

A few things comes to my mind...

There is no motor support in flying wing code. Does all channels still have to go into the Arduino?
Rudder is needed to start. ailerons/elevators is needed to steer the wing. I also want Auto level ON/OFF on my gear channel (mode) so its 4ch in?

But - motor also needs to be connected or you cannot calibrate sensors - I cannot have my ESC connected while this is performed or I give WOT!
Y-cable does not work= WOT on ESC!

How do I calibrate sensors?

I had to solder a bunch of male->male servo cables to connect this one...
I will keep testing today.
Might need a new RX so I can configure a separate memory in the radio...need 125% on all channels to get full PPM.

//UndCon

In my test on the desk it seems like YAW and PITCH are swapped in code (unmodified 1.8)

ELEVATOR

When I pull up YAW value goes down
When I pull stick forward YAW goes back up

RUDDER
left turn = PITCH value goes UP as well as LEFT & RIGHT servo bars
right turn = PITCH value goes down as well as LEFT & RIGHT servo bars

using a MODE2 radio...

And I think values has to be toned down a bit to get a smooth ride...
1 more thought about this...Motor should be added to compensate for climbs/descends - but that will solve itself once wing is stable in air


UPDATE
New test
In RX I modified 1st line to this
static uint8_t pinRcChannel[8] = {ROLLPIN, YAWPIN, PITCHPIN, THROTTLEPIN, AUX1PIN,AUX2PIN,CAM1PIN,CAM2PIN};

(yaw before pitch)

This solved stick input vs meter movement in GUI.
I know this is not the right way to solve this...a static should not be changed - must be changed in code elsewhere

Next issue is the fact that servos increase a bit when I apply throttle
Some calculations seems off...

Another issue:

When giving full ailerons(roll) servo goes to full travel.
This cannot be combined with full elevator
Some travel limitations is needed or you end up flying funny cork-screws.

//UndCon

UndCon,

which software version did you test with?
I addressed some of the points you mention in an earlier post - still need some coding to do.
You can find my latest open code in the branches tree of the MWi google rep.

Hamburger

HI Hamburger!

I used a fresh copy of 1.8 - Ill see if I can find your modified code.
If I only knew how to get the files...do I really have to download them 1 by 1?

//UndCon

get an svn / subversion client for your OS. That makes dealing with such compilations a breeze.

I think I have turtoise svn - at least I used to...

http://tortoisesvn.net/

@UndCon
Try to limit the throws per channel.



/Patrik

Cool - I downloaded Hamburgers dev

Ill update Output.pde and try again

1st I must try Hamburgers code - then Ill continue to next part.
//UndCon

Now I think my wing is ready for a test



Small adjustments have been made and now I must get the code running properly, my test-rig only allows 5 channels

How did the test go?
Anybody have any success with this?
I am buiding a wing for AP and fpv and am hoping to stabilise it using MultiWii

Thanks..!!

HI!
I did not have time to test this yet'

//UndCon

Hi first of all one great Arduino project well done to all!! So my small bit of input to flying wings. I did a flight on Sunday using my Maxi Swift wing and code 1.8 patch 2 with MP only and I can report It's working but I could not test in in full.

Hamburger I'm with you on passthrough and all the other functionality's of current code is what I would like to see remain.
OUTPUT CHANNELS:
A0 & A1 for pitch and roll if I ever add a camera I will try something like Y lead mmmm wonder if that will go down well with mixing?
Motor on D9 had to define number motor from 0 to 1 like this :
#if defined(GIMBAL) || defined(FLYING_WING)
#define NUMBER_MOTOR 1
INPUT
six channels from RX this to have full functions from code: Pitch, Roll ,Yaw, Throttle & Aux1 on my RX used 5 on toggle switch for passthru.
I used a serial LCD (LCD117) that was lying around and it worked good for me

I added mainly this code to OUTPUT tab from every one's input here thanks people:

and some others small changes please see file attached sorry but I can't recall what I did in detail but think was general changes to CONFIG.H .

Warning I have not tested yaw servo on D3 at all. O and I made a video where can I post it?

Greetings

Video http://www.youtube.com/watch?v=AO_6NAbVUDw. Not much but the idea was to adjust PID's at the field with LCD but engine mount problems after first flight put a halt to the day's flying.

One more note: I had problems on the bench were elavons would move up as throttle was increased to full and got it sorted by setting all the PIDS to 0 with only Roll & Pitch P=3 , Level P=7 and RC RATE = 1. expo and others = 0 from here I'll adjust it to suit. Still need to investigate the culprit as I get time.

Good job! I am planning to mount a diy board with atmega and WMP on my easy glider.

Cool - I have not had the time to tinker with my Swift II so much as I would like.

I did some tests and it seemed to act as supposed to, but I think it was very sensitive to inputs making servos work a lot (over reacting)

//UndCon

I tested on a funflymodel once and there is a lot of adjusting on the PID.s
It overcompensated massively.
P must be lowered quite much.
I plan to test on my EZ Hawk but haven't done it yet.

PatrikE wrote:I tested on a funflymodel once and there is a lot of adjusting on the PID.s
It overcompensated massively.
P must be lowered quite much.
I plan to test on my EZ Hawk but haven't done it yet.

That most likely explains what mine did...servos was twitching a lot

Gagerien,
thanks for the feedback. Glad it works for you. Would you care to post your PID settings?
I will continue with code and my wing later this month after return to desk. Alex did not integrate my latest patch yet, not sure why.
Cheers, Hamburger

Hamburger I only had these PID's and wanted to adjust the rest at the field but my engine mount had other ideas:

One more note: I had problems on the bench were elavons would move up as throttle was increased to full and got it sorted by setting all the PIDS to 0 with only Roll & Pitch P=3 , Level P=7 and RC RATE = 1. expo and others = 0 from here I'll adjust it to suit. Still need to investigate the culprit as I get time.Gagarien

What I'm going to do (or thats my under standing of adjusting PID's) is to adjust Roll & Pitch P intill it starts to oscillate in flight then back a touch followed by I and then D same method and then increase P again. I recon because rc aircraft differ much more in design they fly differently more than quad's and thus more teadious adjusting of PID's.

UndCon is right the servo's is very active and I started with these PID's on the bench tunning it down so my motor and prop does not have such a aggresive input on the MP I did wrap it in foam though but would seem to be a different beast while flying.

I will post more results after this comming weekend.

Cheers

Hi,
I uploaded my working setup to the google repository here http://code.google.com/p/multiwii/source/browse/#svn%2Fbranches%2FHamburger%2Fr252_FlyingWing.

base: r252 (v1.8p1 with some extra features)
added flying wing:
- one motor on pin 9 (optional) with support in GUI
- working mixing options for typical 2 servo mirrored setup (configurable)
- no need to swap servo connectors (as with dumb dx6i elevon mixing)
- servos: mid- and endpoints configurable
- working passthrough option with aux1<1300 - no visual feedback in GUI yet
- works (but difficult to tune PID parameters)

My settings are in wing500.h. Make your changes here or (in config.h and remove include of wing500.h in multiwii.pde). This file layout is my way to safely deal with Alex' changing defaults in development versions.

Cheers, Hamburger

Hi Guys
I'm still a bit of a newbie at this ...
Need a little help please, I am trying to download code using Tortoise svn and keep getting the following error -

Checkout from http://code.google.com/p/multiwii/sourc ... 253Dclosed, revision HEAD, Fully recursive, Externals included
Unable to connect to a repository at URL
'http://code.google.com/p/multiwii/source/browse/%23svn/branches/Hamburger/r252_FlyingWing/MultiWiiConf_1_8_patch1%253Fstate%253Dclosed'
Server sent unexpected return value (405 Method Not Allowed) in response to
OPTIONS request for
'http://code.google.com/p/multiwii/source/browse/%23svn/branches/Hamburger/r252_FlyingWing/MultiWiiConf_1_8_patch1%253Fstate%253Dclosed'

Can someone point me in the right direction to download the code please?

Fran

is the link you posted the one you use for checkout? does not look correct. But I do not use nor know the win* GUI.

Thanks, I'll try it soon

Frank

http://www.youtube.com/watch?v=4aP1kXDGEow
Hi guys did some flights on Sunday with a friend, again you can look at the video sorry for the quality but I reduced to 10mb. I have more video’s but need to edit them and I’m like UndCon just running short of time always but at least you can get the idea.

Summary: M+ & NK very windy day I recon at least 25kt gusting , take off in pass thru (Although I will test launch with gyro’s next) switched to gyro’s did a couple of circuits then switched to stable mode more circuits and landed like that. Adjusted the PID’s and more flights in that order. I will post PID settings later. To do = adjust PID to my final liking.

But at least good results there was a circuit in pass through that had me on my toes at all times battling to maintain a constant altitude and directional position but work load gets dramatically reduced to almost just navigation with MultiWii. I cannot compare this to anything else on the market simple reason being that I have only flown with FMA co-pilot’s before.

Couple of notes that spring to mind:
1.In stable mode I cannot get the swift to loop or roll I wanted to see how it would recover from such attitudes maybe in later test = fly in pass thru then activate stable from awkward position. Simulating it on the ground I get confused to how it will recover actually.
2.On my friends wing with digital servo’s I had random servo movements but cannot report anything now as I need to investigate this before I can make an conclusion.
3.In my first reply here I saw elavon movements with throttle input but I could not replicate it again (also did not try hard enough).
4.Now I might get this wrong I and will need to sit down and explain this again with everything on the bench but essentially what happens during PID settings is when Level P, Roll P and PITCH P( I will confirm in what order) is adjusted there is an offset position on the servo’s selecting from Pass thru to Gyro and Stable, playing around with the settings I manage then get stable elavon movements during selection. What will happen is in flight when selecting Gyro etc. The A/C will say roll a bit or pitch up or down if not setup correctly.
5. Gyro’s once again needs to be tune down quite a bit otherwise the servos work hard mainly due to induced vibrations on my wing setup and at half throttle it is the worst that can be seen on the bench and even with motor not running the elevons can induce a negatively unstable scenario just statically starting from a little bump. Initial set up was adjusting only P’s so that there is little movement from the servo’s with engines running at half power.
6. SOMETHING ELSE VERY IMPORTANT During calibration of sensors ensure to compensate for angle of attack, if you place the acc tru to the cord line you will have to trim up allot.
Other than that Bestbuy BT works great with an Parani SD100 on the receiving end I think +/- 40 meters maybe more. Parani is very expensive for the range and is on loan only to me the APC220 would be a better buy.

Cheers

I saw your flights - they are cool!

And yes - Its very annoying not having the time for this hobby - this week I have been at home with sick kids...trying to sort out my Quadcopter problems meanwhile...

Hi Gagarien,
flight is looking good. From my experience pusher wings tend to have lots of vibration (and noise).

It was not possible for me to identify from your video which version/variant/implementation of multiwii for flying wing are you using?
From your experience any ideas on what feature/functionality should be implemented?

Currently I am not sure if we might need separate min/max values for left/right aileron/elevator each? As of now, we can only limit the maximum servo travels (min/mid/max) separately. But have no means to e.g. set travel adjust for elevon to +100/-50 and aileron to +200/-150 as would be possible if we used the delta mixing of modern TXs.

If I get you right Hamburger you want more mixes added into the software?
(more tweaking/fine tuning parameters)

If we take a look at the camera gimbal setup - can we use similar approach?

//UndCon

UndCon wrote:If I get you right Hamburger you want more mixes added into the software?
(more tweaking/fine tuning parameters)
If we take a look at the camera gimbal setup - can we use similar approach?
//UndCon

if you look at my current implementation the two are already similar.
I am not sure whether we need more mixing, but adding separate limits for two axis inputs would get us closer to what people have come to know from their computer radio mixing (though dx6i leaves a lot to be desired on that part).
Basically I consider something like

for now these MID/MIN/MAX values would be #defines but could be converted to parameters to be set via GUI/LCD to equal TX capabilities.

Hi Hamburger/UndCon
Thanks, As per my first post here I used again the same 1.8p2 with modified output tab and Config1.8p2 see my file attached there.

My approach here was to use the standard code from PatrikE and modify it as little as possible hoping that other advancements made like GPS, Sensors etc. can be retained easily for aircraft use and thus having a all-rounder. For myself I doubt using it for any other aircraft plat forms.

Other implementations would be the yaw servo for people with 3 axis aircraft and I’m sure I will come up with other nice to have as I progress.

As to your min/max values ranges and not sure if it needs to be implemented or not I would like to make an suggestion from what I understand U are saying, why not using your code as is for now and look at it at a later stage when the need arises from flying issues? It would sure be handy if you have a complicated aircraft that needs special control throws in roll or pitch or whatever.
I can set the sensitivity to roll and pitch separate in the GUI as my wing respond deferent in R & P to the amount of control input and of course I can program that in my radio for manual flying. But this is not your issue here and from the bit of flying I’ve done I don’t think I need separate servo min/max as the aircraft stabilize nicely to horizontal from say 70 deg bank without any pitching or so.

Regards

ported and submitted to trunk/*_shared in google rep.

Some more flights eventually between everything else and some thoughts:

When adjusting I on Roll Pitch and Level results in elevon movement with throttle input so I left it at 0 again D seems no problem as with P.
If Level P above 7 elevons move up when selecting Level.

Acc position relative to angle of attack is important if aircraft is level when calibrating it will result in a pitch down condition when it is selected so you need to calibrate it with angle of attack so it will fly level in that attitude. Maybe # define angle of attack 12 // degrees in the code so you can calibrate the aircraft level.

Jerky servo problem was resolved by a faulty speed controller it would work flawlessly with Futaba receiver but when you plug it in to Multiwii board the servos go crazy. It tryed all the filtering tricks and no luck.

I am still planning to give your code a go Hamburger hopefully before Christmas, UndCon any progress?

next step is accomplished - passThru mode is user configurable in GUI.

turn passthrough mode into a checkboxes item so user can asign to preferred aux1/aux2 low/mid/high state(s).
Currently in use for Flying Wing. Aiplane should adapt easily. Status: done.

Example: Wing with gyro and acc
in GUI turn checkboxes on for Level - high and PassThru - low, both for aux1
This yields ->
aux1=low : no multiwii sensors interference, only delta mixing
aux1=mid: gyro only acro mode
aux1=high: gyro+acc level mode

Please, test and enjoy. Hamburger

Hi Hamburger

I think the work you and others are doing is fantastic, I appreciate it very much - thanks to you all

I have a small problem, how can I download the code? EDIT - Figured it out now
Thanks
Frank

turn servos' midpoints into LCD configurable items for trim (needed because the TX trims are occupied (together with TX subtrim) to center stick outputs to 1500 for multiwii); save in eeprom [status: done]

caveat: in my case the eeprom cells used for the storing the servo trim values was preset with some rubbish, so after loading sketch first time the midpoints were set to nonsense values. This required to use the LCD once and set meaningful values (1500 for me). In a release version this problem did not happen because Alex would/could increment the version number and force writing a complete set of sensible presets.

Flying Wing is working for me for some time now after I removed a faulty nunchuk (only 2 axes working, no roll axis values).
I could successfully fly my 50cm (20") wing in pretty strong winds with nasty gusts. Eventually I crashed the wing when I switched to passthru mode to test my limits of flying skills - found those all right. Had to add some more glue to the foamie.

For me the endevour of development of Flying Wing is feature complete (for the time being), so:
I would like to hear other success stories and possibly what features you consider missing.
Anyone using the servos' midpoint trim via LCD already?

Now that I have enabled night-mode on mine I can continue testing...

http://www.youtube.com/watch?v=-DNV0TeYFtU


I need to cut up a PET bottle or similar for protection of the electronics...

I did add 3 axis Gyro Smoothing for my FlyingWing (wmp+bma020) with [10, 10, 3] and it seems to really help.
The Wing is smallish 50cm wingspan foamie pusher, so sensors see a lot of noise. Now can raise roll/pitch P from 1.8 to default 4.

PatrikE described the Smoothing and code in the CP Heli thread viewtopic.php?f=7&t=611&start=30#p9056.