UCD 25 watts to 1200 watts using 2 mosfets

Sorry Pafi,

I have seen many transformer sized like that.
Do you think that is 4.5kW transformer? :D

No. This is not a transformer. This is an amplifier. I said it more times. Just read: amp. Simple. I don't omitt important words, I don't rely on smart phone (I use it, but don't let to rule me.) I believe in smart brain instead.

Actually the whole thing is an instrument designed for providing precisely controlled current waveforms on coils, this is why current and voltage is specified instead of power and impedance. And this is why there are additional measurement and control PCBs. But the core of it is a usual bridged amplifier. I also used the whole thing for audio. In this version UcD, since this was readily available for me, later I designed a fix freq PWM optimised for this application.

And for UcD. So you are talking a tiny UcD.
Why only hundred kHz.
Why not MHz?

Which one do you think of? I showed 2 small ones, but I didn't talk about any of them, and I didn't call tiny.

(You mean some hundred I think...)
Finally a relevant technical question, thank you for joining!

Because such high freq (MHz)
- makes switching loss higher
- makes EMI more severe
- is not required for any reason.

I understand now what you are talking about.

You just showed again you don't.
 
Disabled Account
Joined 2005
No. This is not a transformer. This is an amplifier. I said it more times. Just read: amp. Simple. I don't omitt important words, I don't rely on smart phone (I use it, but don't let to rule me.) I believe in smart brain instead.

Actually the whole thing is an instrument designed for providing precisely controlled current waveforms on coils, this is why current and voltage is specified instead of power and impedance. And this is why there are additional measurement and control PCBs. But the core of it is a usual bridged amplifier. I also used the whole thing for audio. In this version UcD, since this was readily available for me, later I designed a fix freq PWM optimised for this application.

As you said from DC to .... (unlimited maybe)
So how does it sound?
I mean the big toroid transformer for the PSU. Does it really have 150V 30A output at that size?
Or do you mean the supply is smaller than the amp itself?

Which one do you think of? I showed 2 small ones, but I didn't talk about any of them, and I didn't call tiny.

(You mean some hundred I think...)
Finally a relevant technical question, thank you for joining!

Because such high freq (MHz)
- makes switching loss higher
- makes EMI more severe
- is not required for any reason.

I really mean that tiny amp
And that is my joke for MHz
That tiny amp can be used for rat repellent maybe :D
I am a builder, already know it
 
Last edited:
Disabled Account
Joined 2005
I am still curious, does this schematic the only shared by Pafi really work?
Does NE5534 can drive IRF530/IRF9530?
 

Attachments

  • switching_amp_by_pafi_2.gif
    switching_amp_by_pafi_2.gif
    8 KB · Views: 858
As you said from DC to .... (unlimited maybe)
So how does it sound?
I mean the big toroid transformer for the PSU. Does it really have 150V 30A output at that size?
Or do you mean the supply is smaller than the amp itself?

Obviously bandwidth is not unlimited. About 30 kHz, but I didn't care, and also for this conversation the exact value is irrelevant. Full range. Release it!

Its sound was perfectly acceptable on full range speakers. But this has been tested just for fun.

The toroid transformer is only 1.6 kVA. This is enough because the intended load is a coil with quite low power factor. So effective power consumption is much lower. Reactive power is maintained by the amp and the capacitor bank (260 V operating voltage, 8x470 uF).

This is true for all ClassD audio amplifiers also. Effective power consumption using speaker is always much lower than the rated output power.

I really mean that tiny amp

And how do you mean I talked about it? It never happened.

And that is my joke for MHz
That tiny amp can be used for rat repellent maybe :D
I am a builder, already know it

If really needed, more than 1 MHz switching freq can be achieved in not only a tiny amp, but at several hundred W. Its just not needed.
 
Anyway that small inductor will be hot like hell for 30A output.
And the other 2kW as well, the core is not big enough. that seems MS130x2 :D

Sometimes you are almost right. The inductor is really hot. Not like hell, but hot. I tested it (and the complete device) with 36 A continuous current for 20 minutes to see if it will be OK, and it heated up to 90 C, which is not too good, but acceptable, since normal operation was limited to 30 A.

The core of 2 kW amp is perfectly enough. You dont know how to design ClassBD output filter, and didnt recognize the common mode core that takes almost all of the AC flux, therefore very low reactive power arises on the diff mode coils.
 
What a theory without practice, and just talking? :D
I have been ten years practice.
I stated that frequency is for FULL RANGE
For sub even 100kHz can be as many big company build sub amp. And make sub amp is much much easier compared to make full range amp.

please can you explain "why ,amplifier has carrier frequncy only 230khz,cannot preform as fullrange amplifier.
 
Sorry Pafi,

I have seen many transformer sized like that.
Do you think that is 4.5kW transformer? :D

And for UcD. So you are talking a tiny UcD.
Why only hundred kHz.
Why not MHz?
I understand now what you are talking about.

for Mhz switching there are not fast enough mosfets for and if excists she have much to high gate capacitance who has to be driven properly.

I think 800 Khz is nice, and far away from the audio, making even a 50 Khz wide amp.
 
Disabled Account
Joined 2005
Sometimes you are almost right. The inductor is really hot. Not like hell, but hot. I tested it (and the complete device) with 36 A continuous current for 20 minutes to see if it will be OK, and it heated up to 90 C, which is not too good, but acceptable, since normal operation was limited to 30 A.

The core of 2 kW amp is perfectly enough. You dont know how to design ClassBD output filter, and didnt recognize the common mode core that takes almost all of the AC flux, therefore very low reactive power arises on the diff mode coils.

I know not sometimes!
It is very simple. Losses! Even I can smell it from here! :D

You dont know how to design....
This is your bad behaviour!!!
 
Disabled Account
Joined 2005
Obviously bandwidth is not unlimited. About 30 kHz, but I didn't care, and also for this conversation the exact value is irrelevant. Full range. Release it!

Its sound was perfectly acceptable on full range speakers. But this has been tested just for fun.

The toroid transformer is only 1.6 kVA. This is enough because the intended load is a coil with quite low power factor. So effective power consumption is much lower. Reactive power is maintained by the amp and the capacitor bank (260 V operating voltage, 8x470 uF).

This is true for all ClassD audio amplifiers also. Effective power consumption using speaker is always much lower than the rated output power.



And how do you mean I talked about it? It never happened.



If really needed, more than 1 MHz switching freq can be achieved in not only a tiny amp, but at several hundred W. Its just not needed.

Now it is more revealed!
You have never built a single amp for live performance!

Again I can smell it from here! :)

The toroid transformer is only 1.6 kVA

Yes, that is what I smell
I dont smell an 150V 30A amp.
But a for fun amplifier with inductor hot like hell.
Hot core means bad sound, always!
Right Pafi?
 
Last edited:
Disabled Account
Joined 2005
please can you explain "why ,amplifier has carrier frequncy only 230khz,cannot preform as fullrange amplifier.

230khz not enough to slice audio input and reproduce back to be a good sonic quality. It will sound like a compressed sampling in digital sound.
I am not a master like Pafi. But trust me. I have been doing class D more than 10 years. I have very good ears as well because actually I am a musician
 
Disabled Account
Joined 2005
for Mhz switching there are not fast enough mosfets for and if excists she have much to high gate capacitance who has to be driven properly.

I think 800 Khz is nice, and far away from the audio, making even a 50 Khz wide amp.

You right, but not neccesary.
A very good Hi Fi class D amp can be with only 400kHz pwm.
In this case, UcD much better from clarity because of better feedback and sound profile similar to class AB.
 
230khz not enough to slice audio input and reproduce back to be a good sonic quality. It will sound like a compressed sampling in digital sound.
I am not a master like Pafi. But trust me. I have been doing class D more than 10 years. I have very good ears as well because actually I am a musician

Pafi please can you explain why 230khz switching frequncy amplifire cannot preform as fullrange amplifier

and I dont know lot about class D

I lern a somthing at theory class
minium switching frequncy== (9x output signel frequncy)
(bla bla bla........... lot of equations and explanations)
so to produce 20hz-20khz max output frequncy 20khz
20khz * 9 =180khz is enough
it means
9 sampling points inside 20khz signal pulse
 
Disabled Account
Joined 2005
Pafi please can you explain why 230khz switching frequncy amplifire cannot preform as fullrange amplifier

and I dont know lot about class D

I lern a somthing at theory class
minium switching frequncy== (9x output signel frequncy)
(bla bla bla........... lot of equations and explanations)
so to produce 20hz-20khz max output frequncy 20khz
20khz * 9 =180khz is enough
it means
9 sampling points inside 20khz signal pulse

It is very simple, no need master Pafi, but I can tell you.
Your cut off amp is not 20kHz but 30kHz, like other class AB for example, so:

30khz * 9 =270khz

That is minimum
 
You right, but not neccesary.
A very good Hi Fi class D amp can be with only 400kHz pwm.
In this case, UcD much better from clarity because of better feedback and sound profile similar to class AB.

Problem with UCD when you get close to rails, the carrier drops. But reading did reveil it is the way to get the best specs within simplicity, but this is also with a triangle based version and here we keep carrier stable, however it have not to be a big problem if carrier is higher as normal.

As for the coil heating, when go in Kw area you have to choose a good core because in fact the core draws output with the demodulation, wire thickness also have in count the skin effect, and use twisted wire (if this work on a class D amp).
 
Last edited:
Disabled Account
Joined 2005
Problem with UCD when you get close to rails, the carrier drops.

As for the coil heating, when go in Kw area you have to choose a good core because in fact the core draws output with the demodulation, wire thickness also have in count the skin effect, and use twisted wire (if this work on a class D amp).

Exactly!
You are on track!
I always use double size for UcD compared to non UcD.
That is practice make you better! :)