UCD 25 watts to 1200 watts using 2 mosfets

Disabled Account
Joined 2005
It's a non obvious way to speak about clarity and fidelity .
It takes 5 minutes to make measurements using a computers soundcard. If you are afraid of showing whese results - I guess the distortion is huge , and there is no use of judging your units as audioamps.[emoji12]

It doent matter people have comments
For me I have no obligation to follow other people request.
Why I should affraid. I dont take any benefit from my works except for myself and to share with other. No.one can force me to do for I want to do
 
Last edited:
Disabled Account
Joined 2005
I did a cliptest with a faseshift type class D, it has low distortion in sim but in real live it needs to be build and tested.

clip was more then 100 Volts peak, and as you see the carrier stays stable quite good, maybe because of the multipole feedback network, who did ask a lot of tryout's..

Last picture,s is distortion with 40 volts peak output.

In real life, clipping in class D self oscillation is very audible. It sounds like "krakkk" "krakkk". Then for live performance rigging limiter is a must. Clip shall be avoided.
 
In real life, clipping in class D self oscillation is very audible. It sounds like "krakkk" "krakkk". Then for live performance rigging limiter is a must. Clip shall be avoided.

You can see a d amp as a transmitter who is fm modulated and right on output demodulated, and when overmodulated you get nasty distortion, so I do now what you mean, but when using a high rail voltage it get be put away on higher output power, but for bass you have very quick high power. For a square wave like transmitter when clip, the mosfets stop switching and rail voltage come o speaker or through mosfets with nasty results, so a one shot protection do well, push power button for reset.

a diode limiter can be used who kicks in and with serie resistor you get also soft clipping (on diodes afcouse) keeping clip voltages away from amp, and when not so high diodes do not have impact, a normal limiter do possible degrade audio quality, I speak afcouse from high end thinking brains.

But I have a circlotron amp who I work on and when let it clip it sounds also bad and blows fuse.

regards
 
In real life, clipping in class D self oscillation is very audible. It sounds like "krakkk" "krakkk". Then for live performance rigging limiter is a must. Clip shall be avoided.

Yes, dropping of switching freq is audible in most cases, sometimes very annoying, sometimes just barely noticeable, depending on the musical material and the freq response of the speaker.

But applying a limiter outside the amp is quite suboptimal solution to this problem, because such limiters typically don't act only when it is needed. And this way your "2 kW" amp will effectively be less than 1 kW.

And I wouldn't say it sounds like "krakk". If it is only the unavoidable freq drop, then it's more like "chirp", and the krakk sound usually indicates something more serious problem like discharging bootstrap cap, incorrect restarting of oscillation, overdriven integrator, etc. You should check the waveform during this behaviour. Overdrive behaviour should not be ignored without checking it in details by saying "I use a limiter".

Shifting switching freq up to 1 MHz as Keen did helps to decrease the natural audible artifact, IF layout is designed very well to avoid self-interferences. But even if somebody manages to do this, the result will be not really efficient, not worth the effort to make it ClassD.
 
I ask myself why mine test design do not oscillate when audio is off in simulation, just set it to 0.01 volts did let it work again, maybe a sim problem.

And for your question about that I do not want to be helped, I am on a technical forum, not a language learn forum, I am dutch and never learn english, only by type it the past years, but keep in mind the tips, if I not forget because designing is also a attack on my mind buffer..

regards
 
hi dimonis please share your working schematic of your approach on class d ;)

I have many working projects , but they are mostly not UcD based . I don't like UcD - due to it's low loopgain it has heightened distortion , and if increase loopgain - it becomes unstable. I only use UcD for high (>1000W) power subamps, where THD is not the main criterion . And due to postfilter feedback can use less expensive output inductor, something like sendust ring cores....

Here is my old UcD bridge , cliplimiter on board.
 

Attachments

  • DSC05611.jpg
    DSC05611.jpg
    431.2 KB · Views: 834
  • DSC05605.jpg
    DSC05605.jpg
    625.2 KB · Views: 778
Last edited:
Disabled Account
Joined 2005
Clipping has every amplifier , of every class ,A, AB , H ,etc. And it sounds the same - a light distortion . If you have a krrrrakkkkk :D - then it is a prolem of your amp .....:wchair:

In class D clipping sounds worse.

As I told before that during momentary clipping for self oscillating class D, PWM will stop for a moment and only pure DC will flow to speaker. And after clipping, it needs time to restart again the PWM. This event makes clipping in class D is worse than class AB.

It seems you don't have experience at all. :D
 
Disabled Account
Joined 2005
I ask myself why mine test design do not oscillate when audio is off in simulation, just set it to 0.01 volts did let it work again, maybe a sim problem.

And for your question about that I do not want to be helped, I am on a technical forum, not a language learn forum, I am dutch and never learn english, only by type it the past years, but keep in mind the tips, if I not forget because designing is also a attack on my mind buffer..

regards

In reality, the actual class D amps will not easy to start with zero input because the amp need high side bootstrap cap to be fully charged and also need very very very small to trigger PWM to start anyway. It does not mean simulation or the amps will not start. It will start but it take longer time.
 
Disabled Account
Joined 2005
I have many working projects , but they are mostly not UcD based . I don't like UcD - due to it's low loopgain it has heightened distortion , and if increase loopgain - it becomes unstable. I only use UcD for high (>1000W) power subamps, where THD is not the main criterion . And due to postfilter feedback can use less expensive output inductor, something like sendust ring cores....

Here is my old UcD bridge , cliplimiter on board.

Aha... low loop gain UcD :D
I can smell it :D
In vice versa, for mid high I use all discrete class D because the amp has very transparent sound. It sounds very similar to class AB.
For low sub I use non UcD just because it is more simple design.
 
Last edited:
You can see a d amp as a transmitter who is fm modulated and right on output demodulated, and when overmodulated you get nasty distortion, so I do now what you mean, but when using a high rail voltage it get be put away on higher output power, but for bass you have very quick high power. For a square wave like transmitter when clip, the mosfets stop switching and rail voltage come o speaker or through mosfets with nasty results, so a one shot protection do well, push power button for reset.

a diode limiter can be used who kicks in and with serie resistor you get also soft clipping (on diodes afcouse) keeping clip voltages away from amp, and when not so high diodes do not have impact, a normal limiter do possible degrade audio quality, I speak afcouse from high end thinking brains.

But I have a circlotron amp who I work on and when let it clip it sounds also bad and blows fuse.

regards

""You can see a d amp as a transmitter ""100% true

PAM8403 IC CLIPPING POINT , TRANSMIT 100 MHZ FM CLEAR SIGNAL
CAN HEAR IT THRUE FM RADIO