pimotors

A collection of classes to drive motors directly connected to a Raspberry Pi, with smart feedback control.

Assuming your project needs a raspberry pi or similarly capable computer, but is not too heavily loaded, even a Raspberry pi zero can directly control and monitor motors up to several thousand rpm. A polling technique is used.

This updated version (in May '18) uses a small C program that in turn uses pigpio to monitor simple rotary quad encoders. It can easily run 2 motors at up to 13,000 rpm with 12 count per rev encoders. CPU utilisation peaks at about 45% on a Raspberry Pi Zero and there do not appear to be any dropped edges.

This forum thread explains some of the background.

A Raspberry pi is quite fast enough to control and monitor motors running up to around 10000 rpm. This set of classes provides the basic functioanlity to do this with the following modules:

driver and control modules

config_adafruit_dc_sm_hat.py: example config for 2 motors on an adafruit DC and stepper motor hat
config_h_bridge.py: example config for 2 motors driven through H bridge directly from gpio (e.g. pimoroni pHAT
dc_adafruit_dchat.py: driver module for dc motor connected via an adafruit dc and stepper motor hat (used by dcmotorbasic via config file - see config_adafruit_dc_sm_hat.py).
dc_h_bridge_pigpio.py: driver module for dc motor connected via an h-bridge ( such as a pimoroni explorer phat) directly controlled via gpio pins and using pigpio to provide control of duty cycle and pulse frequency (used by dcmotorbasic via config file - see config_h_bridge.py).
dcmotorbasic.py: A basic motor control, which uses helper classes to drive the motor and optionally track and control them using a quad encoder for feedback. It also has functionality to provide linear control of motor speed (duty cycle control is typically close to asymptotic which is not friendly for PID feedback control!
feedback.py: A very simple PID feedback controller.
motoranalyser.py: extends dcmotorbasic with some longer tests which record data for later analysis.
motorset.py: provides a single point of control for multiple motors to co-ordinate them.
quadencoder.py: a quadrature shaft encoder, such as the Pololu Magnetic Encoder Pair Kit for Micro Metal Gearmotors. pigpio is used purely to count the pulses and the quadencoder class polls the counter (typically around 20 times per second) to keep track of the motor. This very is rather CPU intensive and just about saturates a Raspberry pi Zero, use the replacement version described below:
quadfastencoder.py: replacement for quadencoder.py with significantly better performance. It needs quadfast.py running at motorset level.
quadfast.py: interfaces via mmap'ed file with a fast C program that track events from the quad encoders.
quadfeedback.c: I very small C program that runs in its own process and is started by quadfast.py. It tracks multiple rotary quad encoders.

basic logger module used by the drivers above

logger.py: a basic log facility for debug and writing trace files that can easily be analysed later

module to facilitate running motor control in its own process

asprocess.py: a module that allows any class to be instantiated in a new process. This allows the motor feedback and control functions to run independently of whatever controls them (such as a webserver)

modules that provide a console based ui to test motors

keyboardinp.py: simple class to provide asyncronous keyboard input for the console (text) based form system (textdisp)
textdisp.py: a basic form handler that will run over ssh
main.py: a utility to test motors using all the other classes here. Runs on the Raspberry pi and can be access via ssh

setup

This all runs on Raspberry Pi 2x and 3x as well as Pi Zero. It also runs on raspbian lite (i.e. the command line only version)