So you have already done such an amp, which do has SR limit, and tries to falsify me when I tell you it happens sometimes. Interesting. 
I repeat: the discussion is not about 'optimal' ratios (whatever it means), but about SR limits of ClassD generally. Whatever you say, a designer can have different aims, possibilities and abilities then you assume.
OFF:
Actually the concept of phase margin is good only for estimation of the idle freq, but absolutely improper for calculation of freq drop at large signal, so please don't refer to it! With no load these circuits are still stable (check it, if you don't believe!), but you can place an RC damper to the output if you are afraid.
These circuits have a little more freq drop. So what? Does it make them unusable? No. Does the extra loop gain hurt you?
So you would like to hide the limited abilities of the amp by explicitely band-limiting the input signal. And everybody should do the same. Great!
Yes, you can see into the head of every ClassD designers on the world, and you saw that none of them have (or had) such a reason. For example nobody can decide using a cheap iron-powder core for output filter, and nobody can try to minimize the idle loss by increasing the inductance. It's impossible that somebody can reach only 1 uF output filter cap in SMT, so nobody is forced to use lower Fres then ideal. Everybody designes ClassD exactly like you, so if you haven't seen some kind of ClassD amp, then it's impossible to build.
I repeat: the discussion is not about 'optimal' ratios (whatever it means), but about SR limits of ClassD generally. Whatever you say, a designer can have different aims, possibilities and abilities then you assume.
OFF:
Actually the concept of phase margin is good only for estimation of the idle freq, but absolutely improper for calculation of freq drop at large signal, so please don't refer to it! With no load these circuits are still stable (check it, if you don't believe!), but you can place an RC damper to the output if you are afraid.
These circuits have a little more freq drop. So what? Does it make them unusable? No. Does the extra loop gain hurt you?
So you would like to hide the limited abilities of the amp by explicitely band-limiting the input signal. And everybody should do the same. Great!
Yes, you can see into the head of every ClassD designers on the world, and you saw that none of them have (or had) such a reason. For example nobody can decide using a cheap iron-powder core for output filter, and nobody can try to minimize the idle loss by increasing the inductance. It's impossible that somebody can reach only 1 uF output filter cap in SMT, so nobody is forced to use lower Fres then ideal. Everybody designes ClassD exactly like you, so if you haven't seen some kind of ClassD amp, then it's impossible to build.
I would say that generally a class d-amp is prone to problems with slew-rate limiting if:
1.) It has a post-filter NFB topology
AND
2.) its small - signal bandwidth is significantly higher than its power bandwith
AND
3.) one is actually misusing it as an amp having power-bandwidth = small-signal-bandwidth
When used sensibly (and there is maybe some pre-filter employed as well) it shouldn't be much of a problem at all.
Regards
Charles
1.) It has a post-filter NFB topology
AND
2.) its small - signal bandwidth is significantly higher than its power bandwith
AND
3.) one is actually misusing it as an amp having power-bandwidth = small-signal-bandwidth
When used sensibly (and there is maybe some pre-filter employed as well) it shouldn't be much of a problem at all.
Regards
Charles
As long as the feedback factor is greater than one (i.e. for frequencies below the unity-gain point of the feedback loop) an amplifier shows a frequency response which is approximately the inverse of the feedback network's frequency response.
An RC Parallel network for instance will force the overall amp to behave like a first order lowpass. But we usually have a second- or higher- order output filter. If we have high input levels at high frequencies the NFB will force the modulator/output stage into transient saturation while trying to emulate the first order lowpass response of the overall amplifier.
If an amp has a purely resistive feedback path the problem is further exagerated and proper input filtering is definitley required for such a toplogy.
If NFB is taken before the output filter transient overloading wouldn't happen.
But if used properly and reasonably post-filter NFB is still the way to go IMO.
Regards
Charles
An RC Parallel network for instance will force the overall amp to behave like a first order lowpass. But we usually have a second- or higher- order output filter. If we have high input levels at high frequencies the NFB will force the modulator/output stage into transient saturation while trying to emulate the first order lowpass response of the overall amplifier.
If an amp has a purely resistive feedback path the problem is further exagerated and proper input filtering is definitley required for such a toplogy.
If NFB is taken before the output filter transient overloading wouldn't happen.
But if used properly and reasonably post-filter NFB is still the way to go IMO.
Regards
Charles
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.