Strange glitches

[fmafma]

Hi!

I'm making tests to control a bipolar-modified 28BYJ-48 stepper motor with a DC motor driver DRV8833¹, but I get some glitches in the motor, which I can see on the DRV8833 inputs. So, this looks like a Micropython code issue.

Here is my code:

import time

from machine import Pin

ain1 = Pin(18, Pin.OUT)
ain2 = Pin(19, Pin.OUT)
bin1 = Pin(20, Pin.OUT)
bin2 = Pin(21, Pin.OUT)

pins = [ain1, ain2, bin1, bin2]

SEQUENCE = [
    [1, 0, 0, 0],
    [1, 0, 1, 0],
    [0, 0, 1, 0],
    [0, 1, 1, 0],
    [0, 1, 0, 0],
    [0, 1, 0, 1],
    [0, 0, 0, 1],
    [1, 0, 0, 1]
]

lastStepTiming = time.ticks_us()
delay = 2000

try:
    while True:
        for step in SEQUENCE:
            
            # Set outputs for current step
            for i, level in enumerate(step):
                pins[i].value(level)
            
            # Delay
            while time.ticks_us() - lastStepTiming < delay:
                pass
            if delay > 500:
                delay = delay - 100
            lastStepTiming = time.ticks_us()
            
except KeyboardInterrupt:
    print("Cancel")
    for i in range(4):
        pins[i].value(0)

The glitch seems to occur every 350 steps (very rough estimation) ; do you see anything obvious? I'm using Micropython v1.25.0.

Thanks,

¹ the DRV8833 has current limits, so it can drive the stepper in current mode, allowing to use a higher power supply to overcome the L.di/dt and electromotive counter force, so giving muuuch more tork (I can't stall it by hand!).

[robert-hh]

It may be that automatic gsrbage collect happens. You could try to call gc.collect() manually e.g before the wait loop.

[robert-hh]

On a side note: pinsi runs slightly faster.

[fmafma]

gc.collect() seems to have fixed the problem. Thanks!

[GitHubsSilverBullet]

No, it hasn't. The problem is now well hidden.
There's still time.ticks_us() - lastStepTiming which is wrong because ›ticks‹ wrap around.
One should use ›ticks_diff‹ and/or ›ticks_add‹ to be on the safe side.

[fmafma]

Ah, yes, I forgot that. I didn't see the problem, as I was just running shorts tests.
Thanks!
PS : just for fun, I will try to implement a PWM to mimic micro-steps. But in the end, I will probably use a really stepper driver, like TMC2209.