I uploaded a simple PWM example to a Nano and modified it such that I would expect to so a sawtooth (PWM coded of course) at pin A3 (I'm using sigrok pulseview with the D0 input clipped to pin A3 of the Nano) but I don't see a signal. Just an always high level.
This is the code:
// Program B for Nano2 – Changed frequency on Pin 3 :
int i=0;
void setup() {
TCCR2B = TCCR2B & B11111000 | B00000001; // for PWM frequency of 3921.16 Hz
pinMode(3,OUTPUT);
}
void loop() {
analogWrite(3,i);
i+=1;
if(i==256)
i=0;
}
The part will not produce a sawtooth waveform unless there is external circuitry to shape it. I can understand a sawtooth wave form but I cannot fathom a sawtooth PWM, variable frequency yes. By definition PWM is pulse width. You can use a comparator and a sawtooth waveform generator to create a PWM signal.
The problem is the those that reply don't know your level of experience. We try not to assume.
It's my understanding that you are using a Nano. In which case A3 is not a PWM capable pin.
You did confirm that it was a fault, but never mentioned the exact correction that you made. If you're still using A3, see above. Else state that you're now checking pin 3 ( D3) instead of A3.
I (erroneously) said, I was expecting the PWM at A3, but programmed D3 (3). The program as posted, gives a PWM output at D3. Nonetheless the output looks a bit strange at the beginning:
And the end of the loop doesn't look clean, either.
With this being the actual code:
// Program B for Nano2 – Changed frequency on Pin 3 :
int i=0;
void setup() {
TCCR2B = TCCR2B & B11111000 | B00000001; // for PWM frequency of 3921.16 Hz
pinMode(3,OUTPUT);
delayMicroseconds(1000);
}
void loop() {
analogWrite(3,i);
i+=1;
if(i==255)
i=0;
delayMicroseconds(10);
}
Did you want that 10uS delay in every cycle through loop, or only at i==255?
indentation is irrelevant in C++ to the compiler, but your usage suggests you intended it to be part of the if clause.
Did you want that 10uS delay in every cycle through loop, or only at i==255?
indentation is irrelevant in C++ to the compiler, but your usage suggests you intended it to be part of the if clause.
@camsysca : if I had wanted to have a delay only in the "if", I would have written ( ):
if(i==255)
i=0,delayMicroseconds(10);
> i+=1;
As for this, was your intent i++?
Yes.
NB: this is C, not python and indentation may get garbled sometimes in the course of copy/paste. But indentation is irrelevant in C, other than good for legibility.
For the sake of peace I've sent the piece of code through Unix $ indent -br -ci4 t.txt
void
loop()
{
analogWrite(3, i);
i += 1;
if (i == 255)
i = 0;
delayMicroseconds(10);
}
My questions were to ascertain user intent and understanding, not challenge language or impugn the user. Sorry to offend, but your problem lies elsewhere, as does your assistance.
I thought the topic is "project guidance". Anyway, leaving animosities aside, in the pulse diagram I posted, wouldn't one actually expect 255 pulses rather than far less as seen in the picture?
And what missing assistance is concerned, I cannot see me failing in that. I posted that I was using Data pin 3 and I even posted a pulse train.
PWM cycle is not affected by the delayMicroseconds(), but by the fact that you run the analogWrite() statement in a loop with a interval of 10 us between different OCRx values
Check the comments in the data sheet on double buffering of the Timer2 output compare registers. If double buffering is not enabled, I believe you can expect to see glitches when trying to update them at the wrong times, or too quickly.
