Stepper motor setup

zjor · February 7, 2021, 3:33pm

Hi, I’m trying to use SimpleFOCLibrary with a stepper motor.
I have L289P-based shield (https://github.com/Luen/Arduino-Motor-Shield-29250)
which has (pwm_a, dir_a, pwm_b, dir_b) pins
or another Chinese board with the similar interface (pwm1/dir1, pwm2/dir2)
How do I initialize the driver, can’t figure out which constructor to use:

StepperDriver2PWM(int ph1PWM,int ph1INA,int ph1INB,int ph2PWM,int ph2INA,int ph2INB, int en1 = NOT_SET, int en2 = NOT_SET);
StepperDriver4PWM(int ph1A,int ph1B,int ph2A,int ph2B, int en1 = NOT_SET, int en2 = NOT_SET);

I guess I should use StepperDriver2PWM, but it has too many parameters.

Any advice appreciated!

Antun_Skuric · February 9, 2021, 8:34pm

Hey @zjor,

I passed throygh the schematic and you’re right that we do not have a good class for this pinout.
It should be StepperMotor2pwm yes but honestly the only way I was able to make l298 work with steppers is with 4pwm signals and this you will not be able to do because the direction pins are inverted in hardware.

I do not have such a board with me so it’s a bit hard to iterate towards the solution

What did you try so far?

zjor · February 10, 2021, 11:48pm

Hi @Antun_Skuric, thanks for getting back on this.
Here is the PR https://github.com/simplefoc/Arduino-FOC/pull/52, I managed my stepper to rotate.
But in the full example code with velocity open loop, the movement was jerky. Probably some delay is needed for transistors to switch (as you explained about dead zone) I’m not a professional in this area.

I’m happy to see your comments!

Antun_Skuric · February 11, 2021, 6:04pm

Hey @zjor,

Thanks a lot for the time you have put into this, I really appreciate it!

So I have few remarks regarding these l298n based drivers. I have one that is really cheap and it has basically two enable (EN1, EN2) pins and 2 direction (INA1,INB1, INA2, INB2) pins for each pair of h-bridges.

I have seen many times on-line that people just connect digital pins to direciton pins (INx1/2) and put PWM to the enable pin. As in the case of your board, you have hardware inverted direction pins so it is clearly meant for this approach. This would be great for this library, because one arduino would be able to run 3 stepper motors with 2x3 pwm pins.

The other approach is to use enable pins as enable pins really and put pwm on each one of the direction pins. Which means that arduino uno can only driver one stepper motor (it has only 6 pwm and it need 4 for one motor)

4PWM mode works well with l298n based drivers, that is something that many people has already tested

Now 2PWM mode, I was not able to get working. I was really not sure why, until I read the datasheet. Here is the image of the driver circuit:

If you look closely on the diagram, you will see that the enable pins EnA and EnB (EN1 and EN2) are connected to the AND blocks on both h-bridges of one phase. And what is more important they are connected to the AND blocks directly, they are not inverted for lower or higher pair of transistors. This basically means, if you put EnA to zero, both your outputs OUT1 and OUT2 will be floating.

You can see that In1 and In2 have been inverted for lower pair of transistors. So when In1 is HIGH the high transistor passes current, when it is LOW, lower transistor conducts. But if the enable EnA is LOW neither of the transistor pairs 1 and 2 does passes current. The motor thinks it is disconnected.

This basically means that L298N was intended to be driven by putting PWM on In1, In2, In3 and In4 not the enable EnA, EnB.

Now, for DC motors, I guess it doesn’t matter that much, even though it should. I am really not sure how it works at all.
Maybe the transistors still pass some current even if they are not enabled because the motor current does not have anywhere else to go. I am not sure, maybe there are people with more experience here who can help us with this.

So for fine controlled Stepper motor it is much more complicated than for the DC motor, and this might be the issue that you are seeing. And that might be the reason why you needed one millisecond delay in order to make the rotation smooth. Since motor does not have proper grounding, only when the enable pin is high ( so some part of the duty cycle ) it might have some strange current waveform and it takes a bit longer to stabilize. I am not sure.

Possible work-arround

So, since you have no way to use all 4 pwm pins separately you cannot use the StepperMotor4PWM class which would not have this problem, we could try a different approach. Let’s try putting the PWM on the direction pins!

We will use enable as an enable (so always HIGH) but control the motor with one PWM pins connected to the direction pins. We will still need 2 pwm pins but the logic that is a bit different.

I’ve checked the schematic of your board and the circuit that is inverting the direction pin should be able to support the high-frequency pwm signals, exciting!

In that case we would have bipolar pwm signal in between positive and negative power supply on the motor. Which means that the 0 volts would be exactly in the middle of the range so pwm of 127.

At the moment the pwm setting function is unipolar, one h-bridge of each pair is always set to 0 and the other one is alternating with the pwm frequency. This is the code:

setPwm(float Ua, float Ub) {  
  float duty_cycle1(0.0),duty_cycle2(0.0);

  // hardware specific writing
  duty_cycle1 = _constrain(abs(Ua)/voltage_power_supply,0.0,1.0);
  duty_cycle2 = _constrain(abs(Ub)/voltage_power_supply,0.0,1.0);
  
  // phase 1 direction
  digitalWrite(dir1, Ua >= 0 ? LOW : HIGH);
  // phase 2 direction
  digitalWrite(dir2, Ub >= 0 ? LOW : HIGH);
  
  // write to hardware
  _writeDutyCycle2PWM(duty_cycle1, duty_cycle2, pwm1, pwm2);

For bi-polar code we would need to do this;

setPwm(float Ua, float Ub) {  
  float duty_cycle1(0.0),duty_cycle2(0.0);

  // hardware specific writing
  duty_cycle1 = _constrain( Ua/voltage_power_supply + 0.5,0.0,1.0);
  duty_cycle2 = _constrain( Ub/voltage_power_supply + 0.5,0.0,1.0);
    
  // write to hardware
  _writeDutyCycle2PWM(duty_cycle1, duty_cycle2, dir1, dir2);

What do you think about this approach, would you be in to try?

zjor · February 11, 2021, 9:31pm

Hi, @Antun_Skuric, thank you for such in-depth research! So basically you suggest using PWM to control not the voltage but the direction? This is a beautiful idea!
Meaning that 127 will make it push backwards and forwards the same amount of time resulting in zero-force while having both coils enabled.
I tried it, here is my code in the modified version of StepperDriver2PWM:

int StepperDriver2PWM2Pin::init() {
  // a bit of separation
  _delay(1000);

  // PWM pins
  pinMode(pwm1, OUTPUT);
  pinMode(pwm2, OUTPUT);
  pinMode(dir1, OUTPUT);
  pinMode(dir2, OUTPUT);

  digitalWrite(pwm1, HIGH);
  digitalWrite(pwm2, HIGH);


  if(enable_pin1 != NOT_SET) pinMode(enable_pin1, OUTPUT);
  if(enable_pin2 != NOT_SET) pinMode(enable_pin2, OUTPUT);

  // sanity check for the voltage limit configuration
  if(voltage_limit == NOT_SET || voltage_limit > voltage_power_supply) voltage_limit =  voltage_power_supply;

  // Set the pwm frequency to the pins
  // hardware specific function - depending on driver and mcu
  _configure2PWM(pwm_frequency, dir1, dir2);
  return 0;
}

// Set voltage to the pwm pin
void StepperDriver2PWM2Pin::setPwm(float Ualpha, float Ubeta) {  
  float duty_cycle1(0.0),duty_cycle2(0.0);
  // limit the voltage in driver
  Ualpha = _constrain(Ualpha, -voltage_limit, voltage_limit);
  Ubeta = _constrain(Ubeta, -voltage_limit, voltage_limit);
  // hardware specific writing
  duty_cycle1 = _constrain(abs(Ualpha)/voltage_power_supply + 0.5, 0.0, 1.0);
  duty_cycle2 = _constrain(abs(Ubeta)/voltage_power_supply + 0.5, 0.0, 1.0);

  // write to hardware
  _writeDutyCycle2PWM(duty_cycle1, duty_cycle2, dir1, dir2);
}

but the motor was just whistling and not moving…
it feels like the motor is fighting itself, maybe the second coil should be disabled when pwm=127

Hope I got your idea correctly

Antun_Skuric · February 11, 2021, 9:53pm

Hey @zjor,

You don’t need the voltage_limit constrain at least for this moment.

And remove absolute value when calculating duty cycle

Let me know how it goes!

manuelbaum · December 18, 2022, 9:35am

Hi,

is there any news on using this shield with simpleFOC? @zjor did you make progress on this?

Best,
Manuel

manuelbaum · December 30, 2022, 10:24am

In case anybody is wondering about this in the future, I got it to work with:

#define E1_PWM  3
#define M1_DIR  12
#define E2_PWM  11
#define M2_DIR  13

StepperDriver2PWM driver(E1_PWM, M1_DIR, E2_PWM, M2_DIR);

and

// Set voltage to the pwm pin
void StepperDriver2PWM::setPwm(float Ua, float Ub) {
  float duty_cycle1(0.0f),duty_cycle2(0.0f);
  // limit the voltage in driver
  Ua = _constrain(Ua, -voltage_limit, voltage_limit);
  Ub = _constrain(Ub, -voltage_limit, voltage_limit);

  // hardware specific writing
  duty_cycle1 = _constrain(fabs((Ua)/voltage_power_supply), 0.0f,1.0f);
  duty_cycle2 = _constrain(fabs((Ub)/voltage_power_supply), 0.0f,1.0f);

  // phase 1 direction
  digitalWrite(dir1a, Ua >= 0 ? LOW : HIGH);
  // phase 2 direction
  digitalWrite(dir2a, Ub >= 0 ? LOW : HIGH);
  // write to hardware
  _writeDutyCycle2PWM(duty_cycle1, duty_cycle2, params);
}

It’s still a little “steppy”, but i guess that is cause i’m just doing open loop right now