Nice link, thanks. I've read from several sources that dead time in the switches are the biggest source of distortion. Why not do this and prevent this type of distortion in the first place?
Some dead time is necessary to prevent cross-conduction (shoot-through). It can be minimized by careful design, but not eliminated.
I know, that's why my above schematic removes cross-conduction from the picture. Why not do this?
That circuit isn't suitable for a class D amplifier. Here are some more tutorials:
https://www.infineon.com/dgdl/an-1071.pdf?fileId=5546d462533600a40153559538eb0ff1
http://www.irf.com/product-info/audio/classdtutorial.pdf
https://www.infineon.com/dgdl/an-1071.pdf?fileId=5546d462533600a40153559538eb0ff1
http://www.irf.com/product-info/audio/classdtutorial.pdf
Last edited:
Can't I just wrap the class D amp within the feedback loop of an opamp or something to eliminate the class D linearity issues?
Last edited:
It doesn't say anything about the effects of an opamp used for error correction. I also don't understand why the example circuit shown is using the PWM signal instead of the analog output signal as feedback. It's not going to error correct the absolute signal output that way.
The op amp is within the nfb loop. The output filter isn't included in that loop, but there is an RC low pass filter before the nfb resistor.
Yeah I know but wouldn't it make more sense to put the feedback at the output of the amplifier? The filter has non linearity of its own.
The output filter is passive, and ideally has only linear distortion (aka controlled frequency response). Phase response must be nonlinear in a filter. A first order filter has an arctangent phase curve. Linear phase only delays the original waveform, and an all-pass filter can be used to alter the phase response of a filter without altering its amplitude response.
Last edited:
I need something that does not cause a delay between the input and the output. Can this be practically done in class D?
Other than trying to get class A sound out of it by wrapping it inside of the feedback loop of something else I also want to build a class D tracking power supply to reduce the power dissipation of class A circuits.
Here is some background.
New supply tracking technology for Class D
DSpace@MIT:
A class-D-tracking-rail class-A audio power amplifier
Welcome to AUDERA
Novel Tracking Power Supply for Linear Power Amplifiers - IEEE Journals & Magazine
Improving the efficiency of class-D audio amplifier systems using envelope tracking DC-DC power supplies - Semantic Scholar
http://download.yamaha.com/api/asse...ite-master.prod.exp.yamaha.com&asset_id=53052
New supply tracking technology for Class D
DSpace@MIT:
A class-D-tracking-rail class-A audio power amplifier
Welcome to AUDERA
Novel Tracking Power Supply for Linear Power Amplifiers - IEEE Journals & Magazine
Improving the efficiency of class-D audio amplifier systems using envelope tracking DC-DC power supplies - Semantic Scholar
http://download.yamaha.com/api/asse...ite-master.prod.exp.yamaha.com&asset_id=53052
Those systems rely upon putting a delay on the main amplifier signal to compensate for the class D delay. I want to do the opposite.
Can you answer my previous question as it pertains to my purposes?
Can you answer my previous question as it pertains to my purposes?
I would see what has previously been done. It's likely that you would want a fast rise and a slow decay.
They use the output filter as part of the design, so that is an advantage. This is one commercial approach.
(PDF) Simple Self-Oscillating Class D Amplifier with Full Output Filter Control
(PDF) Simple Self-Oscillating Class D Amplifier with Full Output Filter Control
- Status
- Not open for further replies.