This would call for a dedicated NFB loop around each output device IMO (controlling drain current).
But I doubt whether this would improve things that much. Keep in mind that you still can't switch as fast as you like specially the switching off is dangerous since high dv/dt can lead to destruction of the output devices.
A better thing IMO would be to model the inverse of the nonlinearities introduced by both deadtime and finite dv/dt and introduce the proper correction in the forward path of the amp's NFB loop. This does not mean that deadtime shouldn't be optimised of course.
Regarding fast switching amplifiers: A fast conventional amp in parallel with a switching current dumper could produce high output bandwidth at the cost of efficiency.
Regards
Charles
Class D byproduct reduction methods
Yes the goal would be to eliminate dead time altogether or at least control the current during the dead time no matter how short the dead time can be made. I will run some simulations on some of the ideas I have been kicking around, some of which I have mentioned in my prior post for open general discussion. I can’t do this right away since I have other obligations that have priority. My simulations and actual tests would be directed toward the output and driver stage and the output filter network to reduce or control kickback currents from reentering the output stage or power supply. This would involve some kind of active snubber circuit around the output filter which may also have other benefits.
Whether any of these ideas are viable or worthwhile will need to be determined by experimentation. I’m sure there are other individuals that have similar or other ideas to help increase power bandwidth and thus potentially reduce the level of unwanted byproducts. Computer simulations may work great while the real thing when built is unreliable and may self destruct at unexpected and thus random times.
As I stated in one of my prior post I would be happy to work with anyone to explore all the possibilities with the goal of developing a good working circuit. I have all test equipment required to evaluate the technical aspects of any developed circuit and use MultiSim for simulation work.
Johannes Fassotte
Fairbanks, Alaska
Alaskan Audio
Yes the goal would be to eliminate dead time altogether or at least control the current during the dead time no matter how short the dead time can be made. I will run some simulations on some of the ideas I have been kicking around, some of which I have mentioned in my prior post for open general discussion. I can’t do this right away since I have other obligations that have priority. My simulations and actual tests would be directed toward the output and driver stage and the output filter network to reduce or control kickback currents from reentering the output stage or power supply. This would involve some kind of active snubber circuit around the output filter which may also have other benefits.
Whether any of these ideas are viable or worthwhile will need to be determined by experimentation. I’m sure there are other individuals that have similar or other ideas to help increase power bandwidth and thus potentially reduce the level of unwanted byproducts. Computer simulations may work great while the real thing when built is unreliable and may self destruct at unexpected and thus random times.
As I stated in one of my prior post I would be happy to work with anyone to explore all the possibilities with the goal of developing a good working circuit. I have all test equipment required to evaluate the technical aspects of any developed circuit and use MultiSim for simulation work.
Johannes Fassotte
Fairbanks, Alaska
Alaskan Audio
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.