Fast forward issue with (dev20110705)

This forum is dedicated to software development related to MultiWii.
It is not the right place to submit a setup problem.
Software download
Post Reply
babelo
Posts: 28
Joined: Wed Jul 06, 2011 12:56 pm

Fast forward issue with (dev20110705)

Post by babelo »

Hi,

Thanks again for previous help; which help me to fix the board freeze :)

I have played with the 1.7 version, it works well. And now i tried version dev20110705.

I ran into an issue with fast forward, when i want to stop, I reduce the pitch then the tricopter go
backward a lot which cause instability.

i ve tried to lower D to 4 also pitch to 1.1 and I also to 0.015. I did not got any effect

Fabrice,

Edit: this problem appear only in fast forward and not in fast backward

PeterPilot
Posts: 19
Joined: Fri Jun 17, 2011 10:08 am

Re: Fast forward issue with (dev20110705)

Post by PeterPilot »

Hi,

I was about to post the same!
I recognised this problem with some dev-versions released in the last six weeks or so.
I was never sure about it but after todays flying with the dev20110710 I am absolutely sure.
I fly with P=3.2 I=0.033 D=11.
My QuadroPlus flys stable with this settings but after a fast forward flight when i pull nick and lower the throttle the Quad behaves like a jumping horse: As soon as the front propeller is raised and the airflow gets under the prop the Quad almost makes a backwards flip.
This behaviour is generally plausible, but this didn't happen with the 1.7 Version of the MultiWii Software so it is definitely software-issue...

Best regards,
Peter

PatrikE
Posts: 1976
Joined: Tue Apr 12, 2011 6:35 pm
Location: Sweden
Contact:

Re: Fast forward issue with (dev20110705)

Post by PatrikE »

+1
I almst thougt the gyro was flying around inside for a split second.
It's almost like the stabilisation shuts off 1 sec. :o
Yaws a 1/4 turn an waggels like crazy.
Then it recovers if you have ice in the stomach and don't lower the throttle. 8-)

/Patrik

babelo
Posts: 28
Joined: Wed Jul 06, 2011 12:56 pm

Re: Fast forward issue with (dev20110705)

Post by babelo »

I noticed a change in SVN repository which seem to affect our issue. It looks like a correction

In config.h

Code: Select all

   /* this parameter defines the maximum correction per axis multiwii can output before mixing the output to different motors
42    this parameter becomes inactive as soon as the stick are far from the center position in order to keep acro abilities
43    reducing this parameter can avoid big wobbles and allows higher PID settings
44   */
45   #define MAX_CORRECTION 100


in ouput.pde

Code: Select all

void mixTable() {
  int16_t maxMotor,a;
  uint8_t i,axis;

  #define PIDMIX(X,Y,Z) rcCommand[THROTTLE] + axisPID[ROLL]*X + axisPID[PITCH]*Y + YAW_DIRECTION * axisPID[YAW]*Z

  //limit big wobble issues
  for(axis=0;axis<2;axis++) {
    a = abs(rcCommand[axis]);
    axisPID[axis] = constrain(axisPID[axis],-MAX_CORRECTION-a,+MAX_CORRECTION+a);
  }
  #if NUMBER_MOTOR > 3
    //prevent "yaw jump" during yaw correction
    axisPID[YAW] = constrain(axisPID[YAW],-100-abs(rcCommand[YAW]),+100+abs(rcCommand[YAW]));
  #endif
  #ifdef TRI
    motor[0] = PIDMIX( 0,+4/3, 0); //REAR
    motor[1] = PIDMIX(-1,-2/3, 0); //RIGHT
    motor[2] = PIDMIX(+1,-2/3, 0); //LEFT
    servo[0] = constrain(TRI_YAW_MIDDLE + YAW_DIRECTION * axisPID[YAW], TRI_YAW_CONSTRAINT_MIN, TRI_YAW_CONSTRAINT_MAX); //REAR
  #endif


back to 4/3 and 2/3 instead of 1/6 and 4/6, that's why there is no force when moving tricopter over pitch axes.
Also a wobble limit appear, which can be set in config.h
So i m going to try this code in the evening :)

Alexinparis
Posts: 1630
Joined: Wed Jan 19, 2011 9:07 pm

Re: Fast forward issue with (dev20110705)

Post by Alexinparis »

Hi,

The PITCH proportion was suggested by Hamburger.
It seems to be more adapted theoretically, but I can't fly with the same confidence with this setting.
(I did not have the chance to test it in FF situation, but the effect you noticed could be tied to this).
That's why I reverted back to the old motor weighted coefficient settings for the TRI.

Do you all confirm it happens only for TRI configs, and is not related specifically to the stable mode ?

babelo
Posts: 28
Joined: Wed Jul 06, 2011 12:56 pm

Re: Fast forward issue with (dev20110705)

Post by babelo »

hi,

I think PeterPilot have same issue with a Quad.

PatrikE
Posts: 1976
Joined: Tue Apr 12, 2011 6:35 pm
Location: Sweden
Contact:

Re: Fast forward issue with (dev20110705)

Post by PatrikE »

I have a quad.
Haven't tried it ön the tri yet...

Alexinparis
Posts: 1630
Joined: Wed Jan 19, 2011 9:07 pm

Re: Fast forward issue with (dev20110705)

Post by Alexinparis »

Ok, so this issue is not TRI related ;)

It is only the case in level mode, or also in acro mode ?

PeterPilot
Posts: 19
Joined: Fri Jun 17, 2011 10:08 am

Re: Fast forward issue with (dev20110705)

Post by PeterPilot »

In my case it occurs in acro-mode. (I dont use level-mode.)

Alexinparis
Posts: 1630
Joined: Wed Jan 19, 2011 9:07 pm

Re: Fast forward issue with (dev20110705)

Post by Alexinparis »

and which sensor ?

PeterPilot
Posts: 19
Joined: Fri Jun 17, 2011 10:08 am

Re: Fast forward issue with (dev20110705)

Post by PeterPilot »

My config is Quad+ with ITG3200 and a LLC. I use the 3,3V regulator on the Paris V3 Board.

PatrikE
Posts: 1976
Joined: Tue Apr 12, 2011 6:35 pm
Location: Sweden
Contact:

Re: Fast forward issue with (dev20110705)

Post by PatrikE »

WMP,NK both original. power from pin 12.
Only tested in acromode.

Also onboard MAG and BMP085
MAG active at the time.

/Patrik

Alexinparis
Posts: 1630
Joined: Wed Jan 19, 2011 9:07 pm

Re: Fast forward issue with (dev20110705)

Post by Alexinparis »

Ok, thank you.
I managed to isolate a bug which could potentially be the cause of this.
It is gyro related
I've just uploaded a new dev (0714)

PatrikE
Posts: 1976
Joined: Tue Apr 12, 2011 6:35 pm
Location: Sweden
Contact:

Re: Fast forward issue with (dev20110705)

Post by PatrikE »

Great job Alex.
The Bug seems to be solved. ;)

No strange behaivor any more.

/Patrik

Alexinparis
Posts: 1630
Joined: Wed Jan 19, 2011 9:07 pm

Re: Fast forward issue with (dev20110705)

Post by Alexinparis »

good to know :)
I think this bug could explain a lot of strange things

babelo
Posts: 28
Joined: Wed Jul 06, 2011 12:56 pm

Re: Fast forward issue with (dev20110705)

Post by babelo »

hi,

It is ok for me too, tricopter fly very well in accro mode. But in stable mode i have strange issue, when i switch to stable mode in fly i have radom direction to trim.

It last fly today i switched three time in stable mode to check where to trim and last attempt was fatal :(

PeterPilot
Posts: 19
Joined: Fri Jun 17, 2011 10:08 am

Re: Fast forward issue with (dev20110705)

Post by PeterPilot »

I'm not satisfied..
The issue with the gyro "passing out" for a moment when stopping from fast forward flight might be a little better now but I had still problems in flight-situations where the stabilisation has a lot of work to do like during fast direction changes (narrow turns with quad standing almost vertically). I don't have the feeling that the problem has to do with extreme rates of rotation but definitely with the amout of correction needed to maintain a orientation or rotation rate.
So in my opinion there is sill a bug hiding somewhere in the code. :twisted:

ziss_dm
Posts: 529
Joined: Tue Mar 08, 2011 5:26 am

Re: Fast forward issue with (dev20110705)

Post by ziss_dm »

Hi PeterPilot,
Can you increase MAX_CORRECTION in config.h, let's say up to 350?

regards,
ziss_dm

PeterPilot
Posts: 19
Joined: Fri Jun 17, 2011 10:08 am

Re: Fast forward issue with (dev20110705)

Post by PeterPilot »

I am going to try it, but I think it might be tuesday until I have enough time...

User avatar
shikra
Posts: 783
Joined: Wed Mar 30, 2011 7:58 pm

Re: Fast forward issue with (dev20110705)

Post by shikra »

I had my first opportunity to fly at a field today with the new version. Tri with latest dev code + the few patches. I updated last night. wmp+NK

descending at say a 20-30 deg angle and pulling slight back with a bit of throttle to help stop and the rear pitched right back so the tri was almost vertical. I recovered OK, just!

Then pretty much with the same manoeuvre and it had the same result. This time I wasn't so lucky and bust the prop so couldn't verify any more.... The first time I thought I got something wrong - but really it was quite a normal flying - I don't do Warthox stuff!!


Also a minor thing I noticed. If I power up in acro mode, it flies in stable mode! If I first toggle the switch to stable and back to acro it is fine.

Alexinparis
Posts: 1630
Joined: Wed Jan 19, 2011 9:07 pm

Re: Fast forward issue with (dev20110705)

Post by Alexinparis »

@shikra: did you try to increase MAX_CORRECTION to 300 or 400 ?
I'm not sure this parameter is a good idea for acro style movement.

User avatar
shikra
Posts: 783
Joined: Wed Mar 30, 2011 7:58 pm

Re: Fast forward issue with (dev20110705)

Post by shikra »

Thanks Alex.

I haven't tried that yet. Never considered myself an acro flyer - but for sure some of the moments are quite sudden and what the tri would think is acro!
I'll tri and let you know. May be awhile - going away on hols in a couple of days

Cheers,
G

Post Reply