I'm trying to turn a Boost into a Flyback and then into an audio amplifier.






Basically I wanted a simple SMPS modulated with an audio signal.

I know, you will say that the class D amplifier is also an SMPS, but the class D amplifier needs at least 2 power transistors + a large electrolytic capacitor or 4 power transistors without a capacitor. And to that will be added the SMPS Up or Down.

Another idea would be:
To make a Boost SMPS with the old good and famous TL494 and to modulate the reaction with an audio signal.
This would mean dividing the audio signal, theoretically a sine wave into positive and negative alternation.

If I use a microcontroller, the dependence of the duty cycle on Vout is not linear at Boot.

 

A boost regulator can only supply current in on direction.
To drive a speaker with an AC signal requires the current to go in both directions. hence the minimum two power transistors in a class D amp.
So how will you get the regulator to do that?

Also switching regulators, by design, have a very low frequency response.

This would mean dividing the audio signal, theoretically a sine wave into positive and negative alternation.

How will that help?

 

Here is a simplified schematic. It drives a motor but could be a speaker.
At about 1980 I had one working with 24V supply and 50A outputs. It was OK just to probe it could be done. Back then we could not switch fast enough to make a real product out of it. We tried again decades later with a 200khz PWM. The problem was the digital audio sampling rate and the PWM frequency beat together and made a slight 8khz tone.
Now days there are IC that make this work very well.

Small D amp from TI

 

I'm trying to turn a Boost into a Flyback and then into an audio amplifier.
View attachment 335401
View attachment 335402

View attachment 335403


Basically I wanted a simple SMPS modulated with an audio signal.

I know, you will say that the class D amplifier is also an SMPS, but the class D amplifier needs at least 2 power transistors + a large electrolytic capacitor or 4 power transistors without a capacitor. And to that will be added the SMPS Up or Down.

Another idea would be:
To make a Boost SMPS with the old good and famous TL494 and to modulate the reaction with an audio signal.
This would mean dividing the audio signal, theoretically a sine wave into positive and negative alternation.

If I use a microcontroller, the dependence of the duty cycle on Vout is not linear at Boot.

Just because you have an idea does not mean that it will work or be practical. Sometimes a sow's ear is just a sow's ear and can never be a purse. If you want purse, then quit messing around: design and build a PURSE.

 

Now days there are IC that make this work very well

I have studied and tested several ICs
The only one that I liked how it sounds is the TDA8920.
It seems simple, it integrates the final transistors, but it needs the differential voltage power supply, other costs.
Best bought as a kit and you have no problems. TDA892X I think similar.

You showed me an H-bridge with bipolar transistors.
It is the simplest method to reverse the polarity or to make a simple voltage power supply audio amplifier.

However, I have read and tested that the MOSFET switches faster than the bipolar transistor.

High power is needed at low frequencies not at high frequencies.
I don't think it would be necessary to reach too high frequencies.

 

Feel free to use MOSFETs if you wish. YOU are in control.

I think Papabravo’s main point is that you need an AC signal with an average voltage of close to zero volts so you don’t cook your speaker.

 

I think Papabravo’s main point is that you need an AC signal with an average voltage of close to zero volts so you don’t cook your speaker.

Certainly. That's why I put an electrolytic capacitor in series with the speaker, to cut the continuous component.
When experimenting there is a risk of burning components. We all know that here.

The problem of how to make an simple SMPS to give selectable either positive voltage or negative voltage without H-bridge is the most difficult.

 

The only one that I liked how it sounds is the TDA8920.

I like it. It has the H bridge with MOSFETs.

You showed me an H-bridge with bipolar transistors.

My first one, MOSFETs were new and would not work.

I don't think it would be necessary to reach too high frequencies.

You need to switch 3 or 4x the highest audio frequency. I suggest 100khz to get 20khz audio.
My first try, we only achieved 9khz audio because switching at 20khz was almost impossible.
The next try I could just make 20khz audio.

I can show you what is inside the IC or you can find out. It is not easy to build from op-amps.


Full bridge. Coilcraft.com makes a single inductor for L1, L2.

Half bridge with caps.

 

The problem of how to make an simple SMPS to give selectable either positive voltage or negative voltage without H-bridge is the most difficult.

I believe that's not just difficult but impossible.
Why do you want to try to make such a strange hybrid of a device?
It certainly won't save you any significant amount of money.

 

capacitor in series with the speaker, to cut the continuous component.

You still need s push pull output to get AC out, unless you want to go class A and make a combo power amp and space heater.

 

and make a combo power amp and space heater.

Yes, the point of a switch mode amplifier is to not generate heat.

 

The first idea is to make a boost smps that can source and sink current.
Boost smps can source current. It is only necessary to create an auxiliary circuit that sinks current.

Then I think of two SMPS placed one after the other. An SMPS Boost that is powered at -12V and provides Vout=+12V....+30V.
After that a Buck-Boost SMPS that will feed from the previous Vout=Vin and will generate Vin,,,,-Vout.
But it's getting too complicated.

 

I will draw the previous attempts with 4 MOSFETs.
I was disappointed with this attempt because distortions appeared at high volume. I didn't have an oscilloscope.


If someone wants to listen to how this test "prototype" sounds, I'll think about how to post a video.

 

The transformer in the previous image is a computer ATX power supply transformer
After the Attini85 comes the MOSFET driver with IR2109/2103
The idea is to use as few expensive parts as possible, such as power MOSFETs.

Why am I looking for an audio amplifier with a microcontroller? I have read more that if the assembly contains a programmed microcontroller it cannot be copied,

 

I like it. It has the H bridge with MOSFETs.

View attachment 335448

From what I can see in the diagram, the TDA8920 even with the bridged outputs still needs a double voltage source. Example +30V GND -30V
It is worth trying to put a simple voltage source +60V and make a table at half the voltage with resistors. Being in deck, it probably won't consume much on those branches.

TDA8920 5$ but IRFZ44 or IRF640/840 is 0.5$

 

TDA8920 5$ but IRFZ44 or IRF640/840 is 0.5$

TDA8920----->$1.80 https://aliexpress.ru/item/10050064...w.productlist.search_results.3.70e373cdrTZTRX

 

the TDA8920 even with the bridged outputs still needs a double voltage source. Example +30V GND -30V

No.
Here is a linear amplifier. C6 blocks the DC. The supply is +12V and 0V. The speaker sees power just like if the amplifier was +/-6V. Just add the capacitor.
Go to post #8 you will see a H-bridge and a half bridge.

 

Now that so many have commented about the need to drive both ways, I wonder if any have ever seen the circuit of the audio output stage of the "All American Five" AM receiver, which uses a single output tube feeding a transformer to drive a speaker???The tube plate current is varied in a class "A" amplifier arrangement. No, it is not very efficient, but the distortion at medium levels is less than 3%.

A switcher power supply with a fast response time, of less than 100 microseconds can work fairly well.

 

receiver, which uses a single output tube feeding a transformer to drive a speaker???The tube plate current is varied in a class "A" amplifier arrangement. .

I guess it is something like this schematic



The transformer is for audio frequencies.
I assume that the transformer in the diagram is iron cored, large and heavy

I intend to use a ferrite transformer on 50-100KHz, because it is smaller and lighter/

I don't think there is a form of signal that you can give at the input of a transformer so that it outputs a constant voltage at the output. Am I wrong?
For example, a PWM with a 90% duty cycle at the input of a transformer would output all 0V direct current on the output.
Am I wrong?

 

"i wanna do something..."
go for it... you may not like what you run into but you will learn a thing or two.