PWM start and stop

[jjude]

How I can start and stop PWM? When stop PWM_OUT_PIN is low (0).
When start duty is 50%. My device is PIC16F616.

[mbradley]

Here is a tiny snippet from a project I did, to stop the motors from moving. (This was part of a touchpad lens controller)

ZOOM_A is a pwm output pin, _B is just a GPIO
FOCUS_A is a pwm output pin, _B is just a GPIO
STOP is 0, so both motor pins get a logic low

A&B pins went to mosfets, depending on the state of _B I can controll motor direction

[Ttelmah]

As a further comment, on the original posted code, do a search on the forum about the numbers used to control the PWM duty cycle.
For 50% duty cycle, you need to use 200L. Notice the 'L'.
The search will reveal 'why'.

Best Wishes

[jjude]

[Ttelmah]

The PWM, counts from '0', to your (Timer2 reset value+1)*4 and resets.
In your case, 0 to 399, and resets on the 400th count.
Any value greater than this, up to 1023 (the maximum that can be loaded into the registers), will leave the output high.

Best Wishes

[languer]

Depending on your oscillator frequency and the PWM frequency, the number inside the pwm function will change (set_pwm1_duty(XX)).

What Ttelmah is referring to is that the PWM register is 10-bits; so the set_pwm1_duty function can use either a 8-bit byte or 16-bit long. So set_pwm1_duty(150L) is not the same as set_pwm1_duty(150). The former is casted to a long value and the later is a byte value (so they give different duty cycles).

After you determine your PWM frequency, and TMR2 div-ratio, you can tell what number you need to pre-load the PWM register for 100% (and anything below). If the calculations you used for the PWM register were based on 10-bits (this would be normally the case) then you need the "L" suffix after the number so the set_pwm1_duty works as you expect.

As for the set_pwm1_duty(0L) giving you 0% or setting the output LOW; you can try it yourself. set_pwm1_duty(0L) still keeps the PWM operational. In theory, you are still pulsing the output; but nothing should ever come out. I have seen instances where there is some leakage and very narrow spikes are present. If you require the output to go low and stay low, without worries; I would follow something like what mbradley suggested.

[Ttelmah]

The reason for leakage, is that people use:

set_pwm1_duty(0);

Rather than '0L'.

If you use a 'short' value, as opposed to a long, the low two bits of the 10bit register don't get updated. So, if there is anything other than '00' in these two bits, when you try to set '0', a small pulse will be generated. The same happens with the 'full on' value, where if there is anything other than '11' already loaded in these two bits, again a small pulse will be seen.....
Using the long value throughout, avoids the problem.

Best Wishes

[languer]

Learned something new. That makes perfect sense. Since I learned about the "L", I have never seen any glitches happen; still I figured why take a chance. However, what you say fully explains it.

Thanks for the piece of info...