I have a software implementation for the pic 12's... As the pic10 has no interrupts if you are going to struggle... Plus what are you going to use to change the pulse width?? There are no adc ports...
What other jobs will the pic do while it is making the PWM signal?
There is very little room on the stack and only a very simple timer.
If there is only one job, to make 1khz at 10% then yes. Like NE555 IC can do.
The PIC12F1571 has stack=16 so you can do subroutines. AND it has 3 PWMs in hardware. You can start them and they run with out software help. Hardware PWM
You might also look at the PIC12F320. Stack=8 and two PWMs.
The compiler I use; uses 100% of the CPU time to "delay". You might have to make a delay function that allows you to do another job. BUT With no/little stack it will be hard to do anything inside the 90mS delay. If the duty cycle varies it makes the timing harder.
The PIC10F320_2 can be had in a 6-Pin SMT (SOT) configuration. It is remarkable for what it is, far surpassing the PIC10F2XX offerings. Plus the 320_2 has ADC. What input or trigger would you use to determine the desired PWM DC%?
According to Microchip one of that chip's features is "a PWM with a fixed 50% duty cycle and an increased frequency resolution function". How can it be PWM if the duty cycle is fixed? I call that misleading if the only variable is the frequency.
According to Microchip one of that chip's features is "a PWM with a fixed 50% duty cycle and an increased frequency resolution function". How can it be PWM if the duty cycle is fixed? I call that misleading if the only variable is the frequency.
Looking at the datasheet, it appears to have 2 fully functioning PWM modules. Are you referring to the Numerically Controlled Oscillator as this has a fixed duty cycle mode. Very powerful chip for the size.
It's clearly stated in the page 1 Summary of the post #14 link that the NCO has two functions, the first of which is (and I quote) "a PWM with a fixed 50% duty cycle".