Good Afternoon,
I have a tank and regular car APM 2.6 set up. The car is working well doing my grid patters, the tank on the other hand, has gone from weaving all over but hitting the way-points, to having seizures or a dance party and randomly driving into things.
- I have my Sabertooth 2 x 25 set up in RC Mode, Exponential response, Mixing mode. The APM is NOT in skid steer mode.
- The compass is calibrated correctly.
- Manual mode is working pretty well.(I think i may need a new RC transmitter though)
My question is how can I adjust the settings to have the tank, do 90 and 180 degree turns on a dime.
I've been trying to find the balance of:
- STEER2SRV_P
- TURN_MAX_G
- NAVL1_PERIOD
- SPEED_TURN_GAIN
- CRUISE_SPEED
- CRUISE_THROTTLE
and to correct for weaving but still have a tight turn.
- STEER2SRV_P
- STEER2SRV_D
- STEER2SRV_TCONST
I'm following http://rover.ardupilot.com/wiki/tuning-steering-and-navigation-for-... for guidance.
One side questions. Since i'm using an external compass and GPS does the APM 2.6 still have face forward?
Replies
Hi guys,
just wanted to let you know that we didnt forget you and were working on fixing skid steering.
There is going to be a new release possibly the next weeks.
Hi again,
as I wanted to drive my rover around, I thought about a way to solve the skid steering problem without coding.
I only have cheap car escs and no Sabertooth, Roboclaw etc.. So I had a look around and noticed an unused KK2.x board. I flashed the KK2 with openaerovtol, set the Pixhawk to normal steering, connected Pixhawk outputs 1 and 3 to the inputs 1 and 2 of the KK2 and the ESCs to output 1 and 3 of the KK2 (does not have to be 3, but 1 is needed to power the KK2).
I then setup a mixer on the KK2 without any stabilization and now I have a skid steer controller, that can be fully customized.
The whole setup only took about 20min.
The only downside is, pivot turning in auto mode is not possible.
Hi,
I am also trying to setup a n APM Rover for the first time (only planes and multicopter so far).
I am using a Dagu Wild Thumper, brushed car ESC and a Pixhawk
with a Taranis and a D4R-II.
I just did some benchtesting so far, but I noticed I have to set both SKID_STEER_IN and SKID_STEER_OUT to 1 to get my ESCs to arm.
To get the manual mode working, I then have to do the mixing in the Taranis, otherwise I can control the right side motors with channel 1 and the left side motors with channel 3.
The APM:Rover manual sounds like the APM/Pixhawk should do the mixing, once SKID_STEER_OUT is set to 1, or did I get this wrong?
I also noticed, with the manual mode working, switching to another mode causes the motors to spin full speed.
@Sebastian,
Glad to hear of your success:-)
As soon as Tridge is available after the AU Outback Challenge, we will determine what is wrong with the skid steering function in v2.46 and make the necessary corrections.
Regards,
TCIII AVD
Another short question: I have searched the forums, google and the parameter list, but could not find where I can adjust the pivotspeed in manual mode. Right now it seems the Pixhawk is only giving 50% throttle forward/backward with full left/right stick input. While this is enough on a smooth surface indoors, it is not sufficient on rougher surfaces or offroad.
@Sebastian,
There is no parameter for adjusting the pivot rate in the Manual Mode.
Have you maximized your aileron (steering) joystick throw to the maximum range of from 1.0ms to 2.0ms?
This might be an issue with the skid steering function that Tridge needs to take a look at.
Regards,
TCIII AVD
I have setup my transmitter to put out ca. 1100 to 1900us on all channels. I tried to set it to its default values of 990 to 2010, but that did not change anything. It goes full speed forward and backward, but pivot is only half throttle. This is also the case with full forward and full left/right stick. One side goes full speed, but the other only goes to about half speed instead of stopping.
As it is now, it does not equal independent dual stick control of both sides.
I think I found the problem with skid steering in 2.46:
if (g.skid_steer_out) {
// convert the two radio_out values to skid steering values
/*
mixing rule:
steering = motor1 - motor2
throttle = 0.5*(motor1 + motor2)
motor1 = throttle + 0.5*steering
motor2 = throttle - 0.5*steering
*/
float steering_scaled = channel_steer->norm_output();
float throttle_scaled = channel_throttle->norm_output();
float motor1 = throttle_scaled + 0.5*steering_scaled;
float motor2 = throttle_scaled - 0.5*steering_scaled;
channel_steer->servo_out = 4500*motor1;
channel_throttle->servo_out = 100*motor2;
channel_steer->calc_pwm();
channel_throttle->calc_pwm();
}
The 100 and 4500 factors should be equal I think, but I have no idea which one is right.
The slow pivoting/steering is caused by the 0.5 factor on the steering input, but without it the resulting pwm signal would be too high or too low, if both channels are at max/min.
Hi Sebastian,
Nice find. This could definitely be the issue with v2.46. I have passed your analysis on to Linus Penzlien our other ArduRover2 Developer to see if he can confirm your find.
Regards,
TCIII AVD
-
1
-
2
-
3
of 3 Next