Class DA

[ChocoHolic]

Nothing new under the sun, - I guess.
Also this is just another combination of well known topologies.
But I never saw someone building or promoting such a thing.

Coming from the old combo of a class AB amp, which is running with high rails and drives a floating low voltage supply for a class A amp...
I just changed the the AB amp to an Class D amp.
My schematic is massively simplified (biasing, freewheeling diodes, etc...) and just suitable to visualize my idea.
Of course this principle should also work with MosFets.

From my perception such a set up should theoretically allow acceptable
losses and at the same time the pleasant non-crossover-behaviour of
class A amps.
Furtheron you can achieve a perfektly short circuit proof power stage in Class D without excessive transistors.
The class A transistors will just have to carry the full current, but voltage drop across the class A transistors is always low and this limits the generated heat in the silicone also in case of a short circuit.

Did anybody try such a design? Experiences?
...also theoretical comments are welcome.

[sss]

in class ab combined with a the purpose of the amp is to eliminate the crossover distortion generated by class ab .

as far as i know class d amps dont suffer from crossover distortion , so i dont see a point making a circuit like that

[ekaerin]

Hey,
The idea is nice. You may put the class D output filter on the rails
for the A-amp.
So virtually, you will end up with a low distortion filterless output
class D amp. Or, a high efficiency class A amp.

The swedish LAB-gruppen (now owned by Danish TC "electronics" group) have a
patent on what they call class TD which esentially is yours idea but
with a class AB stage.

Works great.

/ Mattias

[darkfenriz]

This can be great, but one thing worries me a bit, namely a high frequency phase shift of class D amplifier. A LC filter on the output of typical class D amp makes nearly 180 degrees phase shift for some dozens of kHz, which may disable the 'linear' (class A) amplifier from working properly at this frequency region. Just guessing, will think about it more. Keep on.
regards

[ChocoHolic]

...thanks to all. I will keep this basic principle in mind.
Unfortunately no chance to work on, -in the moment.


sss:
Class D does not have crossover distorsions, but it
also has several nonlinearities. Also the class D output stage itself
causes distorsions, mostly caused by non perfect dead time adjustment (from what I know).
Currently I am not sure if I would manage to design a class D amp
that matches to the expectations of my ears for the entire audio frequency range. May be after finishing my class D sub woofer amp...,
may then I will become more brave.

[ekaerin]

Go ahead and try it out.
Dont wory about phase shift and stuff. As long as the A-amp don't
hit the rails you will be fine. The caps on the output of the D-stage
will keep the A-amp HF stable.
Virtually you have an audio frequency modulated SMPS with a few
extra volts above your output.
Things may become nasty when the A-amp works to close to its rails.
So putting some protection in might be wise-that shuts the system down
if the A-amp gets to close to the noisy rails.

I can't attach the TD- pdf from LABs homepage couse it is to big.

http://www.labgruppen.com/Default.as...831&AjrThmPg=0

/ Mattias

[destroyer X]

You will produce something that is travelling my head since i bougth a class D system.

I hope you do not give up related the intention to start something alike.

Waiting for you,

Old friend,

Carlos

[ChocoHolic]

I have the intension to try, yes. Sometimes.

But I still did not finish my sub woofers.....
And I am moving, so my current DIY activity is zero

Probably I will win the award: Most Rescheduled DIY

[HarryDymond]

Hi Choco, yes, this has been tried before. I will be trying something similar in a few months as part of my PhD.

A high efficiency class A amplifier accompanied by class D switching amplifier
Jae Hoon Jeong; Gue Hong Kim; Byeong Rok Min; Che Hong Ahn; Gyu Hyeong Cho

28th Annual IEEE Power Electronics Specialists Conference (Cat. No.97CH36043), 1997, pt. 2, p 1210-16 vol.2
ISBN: 0 7803 3840 5
Conference: PESC97. Record 28th Annual IEEE Power Electronics Specialists Conference.

Abstract: A high power class A amplifier has excellent fidelity but dissipates too much power. A class D amplifier has high efficiency but shows poor fidelity. This paper proposes a combination of a high fidelity class A power amplifier with class D power amplifier as variable power supply. This amplifier named as class I has the merits of both class A amplifier and class D. The efficiency is seventy-seven percent at full power rating. The distortion of the proposed amplifier is about the same as that of class A amplifier. The measured 3 dB bandwidth is from 10 Hz to 100 kHz (9 refs.)

The abstract makes it sound superb. The only problem is that the overall amplifier matches the performance of their class A amplifier, which is surpsrisingly poor.

Quite a few different ideas have been proposed for combining switching stages with conventional linear stages. I can send you my PhD literature review if you'd like.

[culcanmar]

Hi Harry,

I would be interested in your paper, so if you are willing to send it to me it would be more than welcome!

Martin