MOTOR CONTROL NOT RUNNING

[thanhtruongphan]

I run with command line roslaunch diffbot_bringup bringup.launch
then I turn 2 wheels and the encoder-ticks woking proberly.
then I rosrun rqt_robot_steering rqt_robot_steering and the wheel_cmd_velocities read ok, too.

but the two wheels not move.
@fjp hiii... did i do something wrong or is there a mistake in the repositories code...? can you have any ideas,
thank you so much and have a nice day~ 👍 :D

[thanhtruongphan]

here is the code, I only edited the base_controller part to suit the use with L298, the rest is kept the same as the original code.

[fjp]

Hi @thanhtruongphan thanks for the update using a teensy. Can you please post another link or edit the first one? I only get 404 with the current repo link

[thanhtruongphan]

yeah, here is the new update link.

I just modified only base_controller

thanks for the update using a teensy
hah i feel low mood while it's still not running and have a lot of error, I'll come back to stm32 when I feel better... 😂

[fjp]

Thanks for the new code link. I've reviewed it and everything looks like it should work, so nothing obviously wrong with it in my opinion. What's strange is that your log output shows motor_cmd_left/right_=0 even though `pid_left/righ_errors are not equal to zero.

It seems that motor_pid_left/right.compute() returns zero in the write method:

template <typename L298MotorController>
void diffbot::BaseController<L298MotorController>::write()
{
    motor_cmd_left_ = motor_pid_left_.compute(wheel_cmd_velocity_left_, joint_state_left_.angular_velocity_);
    motor_cmd_right_ = motor_pid_right_.compute(wheel_cmd_velocity_right_, joint_state_right_.angular_velocity_);
    
    p_motor_controller_left_->setSpeed(motor_cmd_left_);
    p_motor_controller_right_->setSpeed(motor_cmd_right_);
    
}

Looking at your debug output, I can see:

1. wheel_cmd_velocity_left_ = 7.67 (this would be your setpoint coming from the high level hardware interface from the Raspi SBC, which is good)
2. measured_ang_vel_left = 0.00 (this is your measured_value)
3. pid_left_errors (p, i, d) shows 7.67 which means:
   • proportional_ = 7.67 (this is correct since error = setpoint - measured = 7.67 - 0.00)
4. motor_cmd_left_ = 0 (this is your final pid output after constrain())

With kp_ = 10 and proportional_ = 7.67, the PID calculation should be:

pid = 10 * 7.67 = 76.7

Your min_val_, max_val_ (PWM_MIN, PWM_MAX) as well as your PID values seem to be good as well. Although I think K_P of 10 will be too high. I guess you have tried already the defaults?

#define K_P 0.6 // P constant
#define K_I 0.3 // I constant
#define K_D 0.5 // D constant

I'd suggest to add more logging in the PID class by either

1. [preferred logging approach] pass the ROS node handle to the PID class here after adding it as a constructor argument and having it as a member (similar to how its done in Encoder class). Then use it to log pid just before the Arduino constrain() method and also log the k_p value (which should be 10.0). You could also add another local variable, e.g. constrain_value = constrain(pid, min_val_, max_val_) and log it as well before returning it.
2. [less preferred logging approach] or turn the local pid variable into a member variable in pid.h and then log its value in the printDebug() method. I'd suggest to also log k_p_ and you could also do the same for additional intermediate variables, like constrain_value_.

Looking forward to your new log output 😊

[thanhtruongphan]

thank you, @fjp 🥇

i probably won't be able to use teensy anymore, my board just broke this morning 🥹 I was still using it yesterday but this morning, when i plugged in the USB power source, the laptop didn't recognize the USB port, i suddenly realized that the power light wasn't on anymore. maybe it's broken. i don't know, why did such an expensive board break, while Arduino and STM32 were fine. hah first time i had a circuit break like that. maybe I'll study the integration with stm32 again soon... and, I also try to so something with the andino that you have mentioned before.

🥇 😊

[fjp]

That's sad to hear that your Teensy broke 😢 If it's within the warranty period, it's definitely worth reaching out to where you purchased it for a potential replacement. Or you can contact PJRC through their forum (https://forum.pjrc.com/) or email Paul Stoffregen (the creator) directly - they're known for good customer service. Even if it's outside the warranty period, you might want to contact them anyway - sometimes they can offer advice or discounted replacements, especially if it's an unexpected failure.

You might want to try a few troubleshooting steps:

• Try a different USB cable - sometimes it's just a cable issue
• Check if there are any visible damage to the USB connector or board
• Try a different USB port on your laptop
• If possible, test with a different computer

maybe I'll study the integration with stm32 again soon

If it's truly dead, it would be great to see ROS serial working on STM32! But I agree it has a bit steeper learning curve than Teensy/Arduino.

and, I also try to so something with the andino that you have mentioned before.

Yes, the Andino software should also work with your Ardunio board. In case you plan to stick with ROS 1 and want to use diffbot software you'd have to edit the diffbot_base high level hardware interface running on the Raspi to use libserial, as explained previously.

[thanhtruongphan]

thanks for your advice, @fjp

I'm come from Vietnam and no PJRC official on my region. The shop where I bought, it doesn't waranty for electronic components, electronic circuits like Teensy 😄

It would be a hassle if I contacted Teensy service and got it under warranty, if I lucky...but never mind, I'll deal with it later.

[thanhtruongphan]

but, I wonder, will Teensy 3.2 work well with this diffbot?

If yes, why does stm32f411 often get "lost sync..."
I tested, if comment out all "debug info" and "loginfo" then there is no "lost sync..." error.

I thought it was because the stm32 hardware is much smaller than the Teensy 4.0, but in "list-parts" there are two options "Teensy 4.0 or Teensy 3.2"

STM32F411CEU6 (128kb SRAM, 512kb FLASH, 100MHz)
Teensy 3.2 (64kb SRAM, 256kb FLASH, 72MHz)

"wrong checksum for topic id..." appears on Teensy 4.0 but the system still works fine. So temporarily don't worry about it.

when launching "diffbot_bringup", the first few seconds will "Emmergency STOP" until the connection to the ROS system is recognized. It will print out the "debug" parameters and the system will start working.

Next is the "log output shows motor_cmd_left/right_=0 even though `pid_left/righ_errors are not equal to zero" issue as we mentioned above. I will need to deal with it when the above part (stm32 instead Teensy) is ok.