This is a theoretical post. Nothing has been made. This is also an uncertain post, which, is based on some myths, which, may or may not be true or may have an insignificant effect. However, some of the ideas may be used ( changed or not ) for Class D amplifiers.
Myths :
Because of lack of information, the proposed controller is made after some myths, made up by the author. Whether these myths are true or false is unknown.
Myths :
1. Continuous is better than discontinuous as far as the noise is concerned.
2. Gradual and slow switching of the primary is better than fast switching as far as the noise is concerned. Instead of the whole, suggested here, transformer, RC filters can be used before the transistor to allow slow switching.
3. Higher resistance ( Rce ) in series to a switching primary is better than low resistance as far as the noise is concerned.
4. Because noise can proliferate from the auxiliary and or primary coils through to the secondary coil ( s ), a controller, made with analogue, audio parts and without any component saturation is better than a controller, made with digital parts or non audio analogue parts or parts which can go in saturation as far as the noise is concerned.
The idea, therefore, is to make a flyback transformer controller with low noise, audio components, which, switches the control transistor of the transformer’s primary gradually and slowly within the period of the switching frequency of 50KHz to > 100KHz. The controller must perform a very gradual switch of the control transistor. This may or may not reduce the noise and may, even, increase the noise. Unknown.
Here is the document :
An Idea for a Controller for a Switched Flyback Transformer Power Supply - Google Drive
The schematics are in the addendum.
Myths :
Because of lack of information, the proposed controller is made after some myths, made up by the author. Whether these myths are true or false is unknown.
Myths :
1. Continuous is better than discontinuous as far as the noise is concerned.
2. Gradual and slow switching of the primary is better than fast switching as far as the noise is concerned. Instead of the whole, suggested here, transformer, RC filters can be used before the transistor to allow slow switching.
3. Higher resistance ( Rce ) in series to a switching primary is better than low resistance as far as the noise is concerned.
4. Because noise can proliferate from the auxiliary and or primary coils through to the secondary coil ( s ), a controller, made with analogue, audio parts and without any component saturation is better than a controller, made with digital parts or non audio analogue parts or parts which can go in saturation as far as the noise is concerned.
The idea, therefore, is to make a flyback transformer controller with low noise, audio components, which, switches the control transistor of the transformer’s primary gradually and slowly within the period of the switching frequency of 50KHz to > 100KHz. The controller must perform a very gradual switch of the control transistor. This may or may not reduce the noise and may, even, increase the noise. Unknown.
Here is the document :
An Idea for a Controller for a Switched Flyback Transformer Power Supply - Google Drive
The schematics are in the addendum.
Where will you get the switching transistors and control ICs?
The designs for SMPS use the reluctance of the ferrites to achieve what you are trying to do.
And sine wave output is not needed.
It seems you have little idea of SMPS and audio in general.
Read up and then decide.
The designs for SMPS use the reluctance of the ferrites to achieve what you are trying to do.
And sine wave output is not needed.
It seems you have little idea of SMPS and audio in general.
Read up and then decide.
Class D amplifiers use switching techniques, thus as long as the supply is less noisy than the amp it supplies, it shouldn't be an issue.
Synchronous switching, with modes allowing it would make the problem mostly insignificant
What counts is the frequency and amplitude of transients generated.
Continuous or discontinuous only changes current transients, which are normally not a problem unless the design/layout is flawed
Softening the switching edges is a good way of reducing EMI pollution.
Using the switching element to achieve it is effective, but costly in terms of switching losses. Reactive solutions do the same, but with ~zero losses
If you go that route, use a linear supply: it will have the same result regarding efficiency, but with no switching at all
I cannot make meaningful comments on that, as it sounds mostly like gibberish to me.
Additional windings and their rectifiers can add to the total noise, but not in a large measure: the main source is the primary switching waveform
All of that is adressed in resonant/ZVS converters. They exist in a variety of topologies, forms, operating modes and can be made extremely quiet when required.
Study the SOTA before trying to improve on it is my best advice
Class D amplifiers do strongly depend on the amplitude of the output voltage, which, in most, usual cases is the power supply. Any noise from the power supply in the audio range is dangerous. Even, away from the audio range as the filter is a simple LC, second order filter, which may not be able to filter all of the noise. Class D amplifiers, the standard ones, depend on the accuracy and NOISE of the triangular wave and the input signal.
In regards to number 4 : This is easy to understand, but, I will try to explain again :
1. Any component, transistor or amplifier, generates MORE NOISE WHEN IN SATURATION.
2. Non audio components generate more noise than audio components.
Therefore, when there are non audio components at the primary side, such as the switching transistor and the controler IC, these may generate more noise, which may be transferred from the primary to the secondary and, thus, to the output. Hopefully, the output capacitors, when of audio quality with low ESR and of high value, will decrease this noise. Yet, better this noise be avoided when possible.
The controller IC also generates noise. In case such is made with audio quality components, which do not go to saturation, the noise is supposed to be lower.
In reality, this noise reduction may or may not be significant. THIS IS WHY THESE POINTS ARE CALLED MYTHS.
In general : whatever power supply one makes, whether switching or linear, this power supply will have a lower noise when the power supply uses audio components only and nothing else. Where such are not available, as for example with voltage references and Zeners, these must be very well filtered, such as component -> R C Filter -> Buffer configuration.
Also in general : transistors and amplifiers generate more noise when in saturation, regardless of how noiseless their power supply is. This is an internal problem with the components, which, cannot be avoided at the manufacturer's side.
Again, how significant these noises are is a different question, which, I cannot answer. Whoever can, please, comment.
In regards to number 4 : This is easy to understand, but, I will try to explain again :
1. Any component, transistor or amplifier, generates MORE NOISE WHEN IN SATURATION.
2. Non audio components generate more noise than audio components.
Therefore, when there are non audio components at the primary side, such as the switching transistor and the controler IC, these may generate more noise, which may be transferred from the primary to the secondary and, thus, to the output. Hopefully, the output capacitors, when of audio quality with low ESR and of high value, will decrease this noise. Yet, better this noise be avoided when possible.
The controller IC also generates noise. In case such is made with audio quality components, which do not go to saturation, the noise is supposed to be lower.
In reality, this noise reduction may or may not be significant. THIS IS WHY THESE POINTS ARE CALLED MYTHS.
In general : whatever power supply one makes, whether switching or linear, this power supply will have a lower noise when the power supply uses audio components only and nothing else. Where such are not available, as for example with voltage references and Zeners, these must be very well filtered, such as component -> R C Filter -> Buffer configuration.
Also in general : transistors and amplifiers generate more noise when in saturation, regardless of how noiseless their power supply is. This is an internal problem with the components, which, cannot be avoided at the manufacturer's side.
Again, how significant these noises are is a different question, which, I cannot answer. Whoever can, please, comment.