How To Write Servo Code And Teach You Step By Step How To Adjust PWM And Avoid Pitfalls

When it comes to writingservocode, many friends who are just getting started get stuck here. It seems that the rudder is easy to operate, but when you actually write the code to control it, you will find that various problems such as inaccurate rotation, jitter, and slow response appear. In fact,servocontrol is not that mysterious. If you understand the principles and common pitfalls, you can also write silky-smooth control code.

What is the principle of steering gear control?

The core secret of the steering gear's operation lies in the 20 millisecond periodic signal. Simply put, there is a reference circuit inside theservo, which is always waiting for a pulse signal. The width of this pulse determines where the servo shaft stops.

Usually we use a pulse of 1.5 milliseconds corresponding to the middle position of 90 degrees, 0.5 milliseconds corresponding to 0 degrees, and 2.5 milliseconds corresponding to 180 degrees. You may ask, why does it have to be sent every 20 milliseconds? This is actually an industry standard, allowing enough response and stabilization time for the servo.

After understanding this, you will understand the essence of writing code: isn't it just to send a pulse of a specific width to the servo every 20 milliseconds. Don't think too complicated, this is the delicate work of a timer.

How to generate that pulse in the code

The most direct way is to use the PWM module that comes with the microcontroller. This thing was born to do this. You only need to set the timer's frequency division and counting period so that it overflows exactly every 20 milliseconds.

Then set the value of the comparison register to the pulse width of the corresponding angle. For example, if you want the servo to turn to 0 degrees, set the PWM high level time to 0.5 milliseconds. This process is like setting an alarm clock, setting when to sound and how long to sound.

️ The specific operation steps are actually three steps:

1. Initialize the timer and let it run at a frequency of 50Hz

2. Configure the PWM output channel and set the polarity

3. Write an angle conversion function to map 0 to 180 degrees into 0.5 to 2.5 milliseconds

If your microcontroller doesn't have a PWM module, don't panic. It can be done by using an ordinary IO port and adding a delay function, but the CPU has to be busy all the time and can't do other work.

How to solve the problem that the servo keeps vibrating?

When you encounter a servo that shakes like a sieve, don't doubt your life. Most likely, the pulse in your code is unstable. The most common pitfall is that interrupts interfere with PWM generation.

For example, if you turn on the serial port interrupt and receive data for a few microseconds, the pulse that should be sent for 1.5 milliseconds becomes 1.6 milliseconds, and the servo will shake. The solution is to give PWM the highest priority, or use hardware PWM to be completely managed by peripherals and not rely on the CPU.

Insufficient power supply will also cause shaking. The starting current of the servo is quite large, especially when it is under load. If the microcontroller resets as soon as your servo rotates, or the servo trembles on its own, check the power supply quickly and replace it with a sufficient power adapter.

What should I do if the steering gear cannot turn in place?

This problem is actually quite common. You clearly calculated that 1.5 milliseconds corresponds to 90 degrees, but the servo only turned to 85 degrees and stopped moving. There is a small detail here. For different brands of servos, the angle corresponding to the same pulse width may be different.

This is especially obvious for cheap servos, with large gear gaps and mediocre potentiometer accuracy. The solution is to add tweaks to your code. Write a calibration function that allows the pulse width value of each angle to be fine-tuned in the program.

There is also the issue of mechanical limits. The servo has not reached the designated position, maybe it is stuck by something. At this time, forcibly increasing the pulse width to push it through will only burn out the servo. Remember to limit the pulse width in the code, and do not let the command value exceed the physical limit of the servo.

How to write multi-channel steering gear control code

If you want to make a robot or robotic arm, you must have more than one steering gear. If each servo uses a separate PWM channel, the microcontroller pins will not be enough. At this time, the servo control panel or software simulation method must be used.

It is most worry-free to use this kind of specialized servo driver chip. One I2C interface can control 16-channel servos. The code is also simple. Just write a library function to directly adjust the angle. If you use a timer to simulate it yourself, you can set up a timed interrupt to process one servo pulse at a time, taking turns.

Note that when multiple servos operate at the same time, the power supply pressure will be great. You can design a starting sequence in the code so that the high-current servos start at staggered times so that they don't all turn at the same time, which can effectively avoid voltage drops.

What are the techniques for debugging steering gear code?

I put the written code in and the rudder worked, but I still felt something was wrong. At this time, it is useless to just look at it, you have to use data to speak. The simplest way is to output the serial port and print out the pulse width value sent each time to see how much it differs from the theoretical value.

You can also use a logic analyzer to capture the PWM waveform. You can tell at a glance whether the pulse width is accurate and whether the period is stable. Many problems can be understood at a glance by looking at the waveform, such as pulse glitches or periods that are long and short.

Another stupid way to debug the servo is to test it in sections. First test the 0 degree position separately, then 90 degrees, and finally 180 degrees to see if the position of each point is correct. If a certain point is inaccurate, there is basically something wrong with the calculation corresponding to that pulse width.

Speaking of which, have you ever encountered a situation where the forward and reverse speeds of the steering gear are different? Welcome to share your pitfall experience in the comment area. If you find it useful, give it a like and support it so that more friends who play servos can see this article~