PWM definitios help ?

[max_imum2000]

hello
since i am new to PWM, i need someone please explain these datails as much as possible.

1. PWM Time base.
2. PWM resolution.
3. PWM frequancy calculation (how to know the best frequancy to drive my dc motor for example?)
4. what is complmentry PWM ?

links will be appricated as wel

thank you very much

 

[dknguyen]

1. PWM Time-base is the clock source around which all other PWM timings are based. Its the clock that clocks all PWM operations.

2. PWM resolution is how accurately you can set the PWM duty cycle. It determines the "minimum increment" of on/off time for the PWM duty cycle. The more bits you have, the more accurately you can set the duty cycle. For a given IC, this available PWM resolution changes with PWM frequency and other factors.

3. You want the PWM pulse to be as fast as possible so that the motor won't pulse and waste lots of energy, but you want it as slow as possible so that the power MOSFETs don't switch back and forth too often or too fast because it takes time for the MOSFET to switch on/off and every time it does, it must pass through an lossy intermediary state between no conduction and efficient conduction. So the more you switch, the more time is spent in this intermediate on/off state and the more power is lost. If you switch at an insanley fast rate, the MOSFET might not be able to react fast enough to change states and may remain locked in an on/off intermediary state. 20kHz is a good number to go by for most things.

4. I think complimentary PWM is when you use two pins and one line pulls low and the other pulls high alternately...or something like that. I am not entirely sure myself.

 

[audioguru]

Some people say that you don't want the PWM frequency too high with a motor because its inductance causes the current to build up slowly for each pulse.

 

[Styx]

Some people say that you don't want the PWM frequency too high with a motor because its inductance causes the current to build up slowly for each pulse.

and some ppl would be right. In fact if the control is such that you do not have pulse-dropping you will never get the current in that you want.

BUT it is more to do with hte inductance. A converter I am working on atm has a 20kHz switching freq at the input stage (it has an active-rectifier front-end) BUT the output is at 10kHz

the input and output are rated for the same power (input takes more current since it actually boosts the DC-link). The reason why the difference in switching freq is because the output motor has a huge inductance and we cannot get any current in at 20kHz.

Well you can you just don't get the benefits of higher-freq. The current will eventually get to the level you require BUT it actually takes longer then using a lower-freq - so might as well use a lower switching freq and get teh benefit of lower switching loss