Page 1 of 1

dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 12:30 pm
by marbalon
Hi,
I've jsut uploaded new verwon to my quad and found some problems.

1. There is a little mistake for people with EXT_MOTOR_RANGE

Lines like this:
((motor[0]>>2) - 250) + 2)

should looks like this (missing ; and ) ):

(((motor[3]>>2) - 250) + 2);

But this is smal problem.

2. I've just tested new level mode and it is not usable for me. Level correction is to slow. In old version I need to decrease P for level to 7 (It flyies with 9 but it is to aggressive). Now I have default settings but looks like in old version with about P = 4. I can increase P more then 9 now because get fast oscillations. Just copy old method to this dev and everything works fine.

3. Altitude mode is not working for me with MS5611. Looks better when I increase P to 6, but after hover for a while it drops down and can't enough power to climb to start position.

I have on request for new version. Could you include i2c GPS support. I think this solution will be popular for many people.

Regards,
Marcin.

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 2:33 pm
by Magnetron
I am your some configuration with FreeIMU v0.3.5_MS with MS5611BA barometer. I had tryed v1.9 also with TRUSTED_ACCZ option activated and moving average calculation (that I had implemented) on baro, gyros and accelerometers. The overall stability is improved due to moving average but the altitude hold was poor... I would to try the last dev version like this 20120203.
Could you post you configuration and your gui settings? Many thanks.

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 4:01 pm
by Hamburger
marbalon wrote:1. There is a little mistake for people with EXT_MOTOR_RANGE


corrected in _shared.

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 4:30 pm
by TheX
I can't get the 64 bit Windows version of conf to run. 32 bit starts fine but even trying to run the 64 bit version as administrator does nothing (I do say yes to the UAC prompt) . No errors in the event viewer, nothing in task manager.

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 5:25 pm
by timecop
You need to install 64bit version of Java. (ugh).

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 5:33 pm
by nhadrian
Could anybody explain what has been changed in this dev?
I think for good altitude-hold settings we definitelly need the VEL parameters, that is my experience.

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 5:34 pm
by LuFa
what is chanche in the newest dev =?

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 5:51 pm
by timecop
protip:
make a diff between old dev and new dev.
look at generated patch file.
find out what changed.

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 5:52 pm
by Alexinparis
as I said here:
viewtopic.php?f=8&t=562&start=50#p8508

I released this dev snapshot mainly to have feedbacks on the ALT HOLD feature.

All the changes can be followed here, but it's not summarized:
http://code.google.com/p/multiwii/source/list
As it's a dev version, don't expect to have a full documentation ;)

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 6:19 pm
by nhadrian
Oh, sorry, your post was on the bottom of the actual page so I missed it... ! Sorry again!

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 6:31 pm
by TheX
dongs wrote:You need to install 64bit version of Java. (ugh).


I have it, I use this machine for everything I do.

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 6:34 pm
by shikra
Raise Level D until you get the max speed you like.
Putting Level D = 500 will effectively make it same as before.

Not so nice for flying angle mode, but faster snap to level. You have the choice now!

marbalon wrote:
2. I've just tested new level mode and it is not usable for me. Level correction is to slow. In old version I need to decrease P for level to 7 (It flyies with 9 but it is to aggressive). Now I have default settings but looks like in old version with about P = 4. I can increase P more then 9 now because get fast oscillations. Just copy old method to this dev and everything works fine.


Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 7:32 pm
by Bledi
I have installed the new version + R526 to help PatrickE to integrate I2C-GPS and a version simple version of the gimbal.
I find something stange (but I am not a specialist):
- The output.ino in rev r526 is not correct beacause the declaration of number of motors is missing
- The HMC5883 is looking to be inverted (no problem in the dev version of december)
- The servo outputs for gimbal are not working (no problem in the dev version of december).
- in config.h the #define VTAIL is not present and when you had it you have some simples errors. GUI is not still using this define

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 9:18 pm
by ronco
Bledi wrote:I have installed the new version + R526 to help PatrickE to integrate I2C-GPS and a version simple version of the gimbal.
I find something stange (but I am not a specialist):
- The output.ino in rev r526 is not correct beacause the declaration of number of motors is missing
- The HMC5883 is looking to be inverted (no problem in the dev version of december)
- The servo outputs for gimbal are not working (no problem in the dev version of december).
- in config.h the #define VTAIL is not present and when you had it you have some simples errors. GUI is not still using this define


output r526 misses the first 50 lines.. i fixed this now

regards

felix

Re: dev MultiWii_dev_20120203 test

Posted: Sat Feb 04, 2012 10:27 pm
by Bledi
- The output.ino in rev r526 is not correct beacause the declaration of number of motors is missing => OK it's working
- The HMC5883 is looking to be inverted (no problem in the dev version of december)
- The servo outputs for gimbal are not working (no problem in the dev version of december). => OK it's working (problem of r526)
- in config.h the #define VTAIL is not present and when you had it you have some simples errors. GUI is not still using this define

Re: dev MultiWii_dev_20120203 test

Posted: Sun Feb 05, 2012 1:00 am
by Bledi
It's confirmed, I tested the december version, HMC5883 is working fine. With le last version the north is not ok and the head is changing suddenly

Re: dev MultiWii_dev_20120203 test

Posted: Sun Feb 05, 2012 1:35 am
by Alexinparis
Bledi wrote:It's confirmed, I tested the december version, HMC5883 is working fine. With le last version the north is not ok and the head is changing suddenly

Hi,
Ok, I think I know why.
The code was a little bit changed to include a better mag calibration.

Code: Select all

    magCal[ROLL]  =  1000.0 / abs(magADC[ROLL]);
    magCal[PITCH] =  1000.0 / abs(magADC[PITCH]);
    magCal[YAW]   =  1000.0 / abs(magADC[YAW]);

just replace this part in the Sensors.pde file

Re: dev MultiWii_dev_20120203 test

Posted: Sun Feb 05, 2012 8:25 am
by nhadrian
I tested but level mode is not ok. It has serious drift to front-left, I couldn't take away with acc-calibration. In the gui the maximum level-D allowed is 100 so couldn't set up 500 instead.
Mag calibration with HMC5843 didn't worked,BTW the changing in prev. post would fix it.

Because of these I couldn't test the new alt hold method.
How could I copy this new alt hold into my dev20111220?

Re: dev MultiWii_dev_20120203 test

Posted: Sun Feb 05, 2012 9:20 am
by timecop
Alexinparis wrote:
Bledi wrote:It's confirmed, I tested the december version, HMC5883 is working fine. With le last version the north is not ok and the head is changing suddenly

Hi,
Ok, I think I know why.
The code was a little bit changed to include a better mag calibration.

Code: Select all

    magCal[ROLL]  =  1000.0 / abs(magADC[ROLL]);
    magCal[PITCH] =  1000.0 / abs(magADC[PITCH]);
    magCal[YAW]   =  1000.0 / abs(magADC[YAW]);

just replace this part in the Sensors.pde file


lol!! and I was wondering why the mag was crazy on my board.

Re: dev MultiWii_dev_20120203 test

Posted: Sun Feb 05, 2012 11:42 am
by Bledi
Alexinparis wrote:
Bledi wrote:It's confirmed, I tested the december version, HMC5883 is working fine. With le last version the north is not ok and the head is changing suddenly

Hi,
Ok, I think I know why.
The code was a little bit changed to include a better mag calibration.

Code: Select all

    magCal[ROLL]  =  1000.0 / abs(magADC[ROLL]);
    magCal[PITCH] =  1000.0 / abs(magADC[PITCH]);
    magCal[YAW]   =  1000.0 / abs(magADC[YAW]);

just replace this part in the Sensors.pde file



ok it's working very good.
I had work with Patrick and know he has post a new version with :
- I2C GPS
- The servo mixing for a simple GIMBAL
- The VTAIL

all is ok

Re: dev MultiWii_dev_20120203 test

Posted: Sun Feb 05, 2012 5:29 pm
by crashlander
marbalon wrote:Hi,
I've jsut uploaded new verwon to my quad and found some problems.
1. There is a little mistake for people with EXT_MOTOR_RANGE


It seems that code for EXT_MOTOR_RANGE only exist for PROMINI and not for MEGA, Is that proper way to handle it:

Code: Select all

#if defined(MEGA)
    #if (NUMBER_MOTOR > 0)
    #ifndef EXT_MOTOR_RANGE
      OCR3C = motor[0]>>3; //  pin 3
      #else
        OCR3C = (((motor[0]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 1)
    #ifndef EXT_MOTOR_RANGE
      OCR3A = motor[1]>>3; //  pin 5
      #else
        OCR3A = (((motor[1]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 2)
    #ifndef EXT_MOTOR_RANGE
      OCR4A = motor[2]>>3; //  pin 6
      #else
        OCR4A = (((motor[1]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 3)
    #ifndef EXT_MOTOR_RANGE
      OCR3B = motor[3]>>3; //  pin 2
      #else
        OCR3B = (((motor[1]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 4)
    #ifndef EXT_MOTOR_RANGE
      OCR4B = motor[4]>>3; //  pin 7
      OCR4C = motor[5]>>3; //  pin 8
      #else
        OCR4B = (((motor[4]>>2) - 250) + 2);
        OCR4C = (((motor[5]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 6)
    #ifndef EXT_MOTOR_RANGE
      OCR2B = motor[6]>>3; //  pin 9
      OCR2A = motor[7]>>3; //  pin 10
      #else
        OCR2B = (((motor[6]>>2) - 250) + 2);
        OCR2A = (((motor[7]>>2) - 250) + 2);
      #endif
    #endif
  #endif

Regards
Andrej

Re: dev MultiWii_dev_20120203 test

Posted: Sun Feb 05, 2012 5:39 pm
by Hamburger
ronco wrote:
Bledi wrote:I have installed the new version + R526 to help PatrickE to integrate I2C-GPS and a version simple version of the gimbal.
I find something stange (but I am not a specialist):
- The output.ino in rev r526 is not correct beacause the declaration of number of motors is missing


output r526 misses the first 50 lines.. i fixed this now

Bledi, Felix,

you shall not mix files from different revisions - that may work, or not.
The code you declare missing was moved for good reason from Output.pde to def.h, which is part of that same revision (r525).
If you check out all files from that revision (or a later one) all the code is there.

Re: dev MultiWii_dev_20120203 test

Posted: Sun Feb 05, 2012 5:47 pm
by ronco
Oh.. i see :(

sorry for that .. Bledi said its not working and i just checked the output -.-

regards felix

Re: dev MultiWii_dev_20120203 test

Posted: Tue Feb 07, 2012 5:01 am
by kataventos
Hi,

In this dev (jumped from 1.9) my ACC (Nunchuck) is present but not recognized, led is blinking after cal and no ACC values! My debug 4 is giving an error (never before), jumping random from 0 to -4. :|

Does anyone know what´s going on? Everything was working on 1.9.


UPDATE: found the "problem"... I change some things like low pass filter value, now it is not possible! why? changing from 4 to 8 or 16 til 64 was possible in 1.9.

Cheers,
Henrique

Re: dev MultiWii_dev_20120203 test

Posted: Tue Feb 07, 2012 11:42 pm
by wilco1967
Can't get the tail servo to move on a Tri with MultiWii_dev_20120203 version.
With earlier software version all is working fine :?

Running a Flyduino MEGA.
Tail servo output seems to move correctly in the GUI, but no response from servo. The rest (motors, sensors) seems to work fine...

Wilco

Re: dev MultiWii_dev_20120203 test

Posted: Wed Feb 08, 2012 1:15 am
by Alexinparis
wilco1967 wrote:Can't get the tail servo to move on a Tri with MultiWii_dev_20120203 version.
With earlier software version all is working fine :?

Running a Flyduino MEGA.
Tail servo output seems to move correctly in the GUI, but no response from servo. The rest (motors, sensors) seems to work fine...

Wilco

Hi,
right, there is a bug in the servo assignment
a small fix:
replace
#define SERVO_5_PINMODE pinMode(6,OUTPUT);
#define SERVO_5_PIN_HIGH PORTH |= 1<<3;
#define SERVO_5_PIN_LOW PORTH &= ~(1<<3);


with
#define SERVO_5_PINMODE pinMode(2,OUTPUT);
#define SERVO_5_PIN_HIGH PORTE |= 1<<4;
#define SERVO_5_PIN_LOW PORTE &= ~(1<<4);

Re: dev MultiWii_dev_20120203 test

Posted: Wed Feb 08, 2012 1:17 am
by Alexinparis
crashlander wrote:
marbalon wrote:Hi,
I've jsut uploaded new verwon to my quad and found some problems.
1. There is a little mistake for people with EXT_MOTOR_RANGE


It seems that code for EXT_MOTOR_RANGE only exist for PROMINI and not for MEGA, Is that proper way to handle it:

Code: Select all

#if defined(MEGA)
    #if (NUMBER_MOTOR > 0)
    #ifndef EXT_MOTOR_RANGE
      OCR3C = motor[0]>>3; //  pin 3
      #else
        OCR3C = (((motor[0]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 1)
    #ifndef EXT_MOTOR_RANGE
      OCR3A = motor[1]>>3; //  pin 5
      #else
        OCR3A = (((motor[1]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 2)
    #ifndef EXT_MOTOR_RANGE
      OCR4A = motor[2]>>3; //  pin 6
      #else
        OCR4A = (((motor[1]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 3)
    #ifndef EXT_MOTOR_RANGE
      OCR3B = motor[3]>>3; //  pin 2
      #else
        OCR3B = (((motor[1]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 4)
    #ifndef EXT_MOTOR_RANGE
      OCR4B = motor[4]>>3; //  pin 7
      OCR4C = motor[5]>>3; //  pin 8
      #else
        OCR4B = (((motor[4]>>2) - 250) + 2);
        OCR4C = (((motor[5]>>2) - 250) + 2);
      #endif
    #endif
    #if (NUMBER_MOTOR > 6)
    #ifndef EXT_MOTOR_RANGE
      OCR2B = motor[6]>>3; //  pin 9
      OCR2A = motor[7]>>3; //  pin 10
      #else
        OCR2B = (((motor[6]>>2) - 250) + 2);
        OCR2A = (((motor[7]>>2) - 250) + 2);
      #endif
    #endif
  #endif

Regards
Andrej


yes, not everything is properly coded at the moment.
output code was changed and still needs some check

Re: dev MultiWii_dev_20120203 test

Posted: Wed Feb 08, 2012 3:21 pm
by ronco
Hi,

ther is a mistake in my motor soft PWM code .. at line 132

Code: Select all

for(uint8_t i=0;i<min(NUMBER_MOTOR,4);i++)

will set only the first 4 motor pins to output .. not motor 5-8

so is it ok if i change it to

Code: Select all

for(uint8_t i=0;i<NUMBER_MOTOR;i++)

?

regards

felix

Re: dev MultiWii_dev_20120203 test

Posted: Thu Feb 09, 2012 11:25 am
by crashlander
Alexinparis wrote:
crashlander wrote:
marbalon wrote:Hi,
I've jsut uploaded new verwon to my quad and found some problems.
1. There is a little mistake for people with EXT_MOTOR_RANGE


It seems that code for EXT_MOTOR_RANGE only exist for PROMINI and not for MEGA, Is that proper way to handle it:
...


yes, not everything is properly coded at the moment.
output code was changed and still needs some check

I made tipo in my first code correction (motors 2 and 3 were driven with motor 1 impulse) and is now corrected.

Code: Select all

#if defined(MEGA)
    #if (NUMBER_MOTOR > 2)
...
          #else
        OCR4A = (((motor[2]>>2) - 250) + 2);
 ...
    #if (NUMBER_MOTOR > 3)
 ...
      #else
        OCR3B = (((motor[3]>>2) - 250) + 2);
 ...
  #endif

But with that setup my servo on Three is not working (anymore), it worked with previous snapshots?!

Regards Andrej

Re: dev MultiWii_dev_20120203 test

Posted: Thu Feb 09, 2012 2:07 pm
by neverfree
hello
my tricopter flies very well with 1.8 patch2 except altotude hold i try to installe the 20120203 dev but i don't understand how to configure configurable parameters for my tricopter and what arduino or prrocessing version?
thank you for your answer

best regards

Re: dev MultiWii_dev_20120203 test

Posted: Thu Feb 09, 2012 7:23 pm
by kataventos
Hi,

OK, :) I tryed this dev without my changes to low pass filter etc, etc.
I´m just delighted! this is something a lot different, level Mode is perfect, what did I miss in the last month?
Just one question about something that bothers me almost from the beginning, at first the D value was negative and increase D was closer to zero, after the change from negative to positive, increasing is a lower value or higher? :oops:

I ask this because I see a D on level and it is 100 so, that means no D?
What are the random values on debug4?

Cheers,
Henrique

Update: Can anyone be so kind to answer this questions!

R557 level fault ?

Posted: Fri Feb 10, 2012 11:42 pm
by wilco1967
Seems like something went wrong with the level mode in R557
It was working fine in dev20120203.

In R557, there is a slow oscillation (approx 3 seconds cycle), which does not disappear when I change the level PID parameters (just becomes more/less pronounced, but always remains present). It's a circular type of oscillation. It used to fly pretty good with default PID parameters on earlier revisions.

I'm no expert (far from that) :oops: ... just reporting what I see, but could there be a mixup between pitch and roll somewhere (that would be my gutt feeling based on how it is behaving) ?

this is on a Quad-X, promini, 9DOF Drotek IMU. Haven't tried the Tri yet....

BTW: the alt hold seems to work a lot better on dev20120203.... Not properly tuned yet, but definitely moving in the right direction. I only have a baro on the Tri, not for the quad.

--------
EDIT.

Seems like it is solved in R569..... With the addition of a new default level D setting of 100, it levels nice again..... If I put the D back to 0 (earlier standard), the problem is back....
Thanks !

Re: dev MultiWii_dev_20120203 test

Posted: Thu Feb 16, 2012 9:41 pm
by KeesvR
Today my T-Copter flipped 2 times with this version. The right motor stopped.
Gone back to Dev20111220 and I haven't seen it anymore.

Re: dev MultiWii_dev_20120203 test

Posted: Fri Feb 17, 2012 5:07 pm
by shikra
To alex

If you decide to go with the LEvel D method, can you set default param in EEPROM checkfirst time ?
Also it would be real nice to have it on the LCD for tuning at the field ..... :P