asdf44
Advanced Member level 4
I've seen examples of two different ways to approach a similar design problem
Say I'm desiging a SMPS and my requirement is for about 5khz of bandwidth.
Example 1 places the output LC poles close to 5khz and uses zeros in the control compensation to "step over" the LC crossover/resonance point (standard Type III compensation strategy) and maintain phase margin. The LC filter drives the system through zero at about 5khz and the control itself crosses 0db later, at say 30khz.
Example 2 reverses the above placing the LC crossover frequency at about 30khz and setting the control crossover at about 5khz. The control in this case drives the system through 0db and the filter follows later.
In theory, both can achieve the same bode plot of performance, but I'm not clear on what the other tradeoff might be?
All else being equal example 1 has simpler control it seems, while example 2 would have less output ripple due to the lower filter frequency (though parts might be larger to achieve that).
Anything else?
Say I'm desiging a SMPS and my requirement is for about 5khz of bandwidth.
Example 1 places the output LC poles close to 5khz and uses zeros in the control compensation to "step over" the LC crossover/resonance point (standard Type III compensation strategy) and maintain phase margin. The LC filter drives the system through zero at about 5khz and the control itself crosses 0db later, at say 30khz.
Example 2 reverses the above placing the LC crossover frequency at about 30khz and setting the control crossover at about 5khz. The control in this case drives the system through 0db and the filter follows later.
In theory, both can achieve the same bode plot of performance, but I'm not clear on what the other tradeoff might be?
All else being equal example 1 has simpler control it seems, while example 2 would have less output ripple due to the lower filter frequency (though parts might be larger to achieve that).
Anything else?