Class A and A/B vs. Class D

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
This intended to be an open discussion about the acceptance of Class D in High Grade Audio Equipment.

As this technology emerges, with it's obvious benefits (at least for higher power levels) such as :

High Efficiency > 95%
No need for cooling
No need for large power supplies
Small size
Noise instead of harmonic distorsion (less disturbing to ear)
Low price
Ruggedness towards low speaker impedances.

What is the justification really for analog amplifiers anymore?

Are anybody experimenting with diy PWM amplifiers in here, and what are the results? What kind of problems did you run into?

What are the shortcomings of PWM vs. analog amplification or vise versa in your opinion?

(Just adding: i don't intend to say there is no longer a need for analog amplifiers, just wish to start a discussion about analog vs. PWM)
 
Disabled Account
Joined 2003
Lars Clausen said:
What are the shortcomings of PWM vs. analog amplification or vise versa in your opinion?


digital amps are also cheaper, offer a direct digital link to input digital signals, are easier to engineer as well.

analog is easier to put together (for diyers), offers more tweaks.

I think digital is the future, especially for commercial amps.
 
"digital amps are also cheaper, offer a direct digital link to input digital signals, are easier to engineer as well."

Most commercial offerings are class D switching amplifiers, they cannot handle digital data.

"High Efficiency > 95%"

Only if you ignore losses in the output filter and comutating diodes. Large commercial switching amplifiers frequently employ forced air cooling for the output filter components. Driving reactive loads and depending on the body diode is a disaster. External diodes can run hotter than the switching FETs.

"No need for cooling"

The 20 year old Carver M1.5 puts out 1.5KW with no heatsink, no fan, and runs quite cool.

"No need for large power supplies"

Only if you need a large output. The above mentioned amplifier is 85% efficent on reactive loads (real loudpeakers). The 10% or so difference in supply size required is nill.

"Small size"

Six years ago I built a 100W stereo amplifier, pre-amp, image processor, 24dB crossover, headphone amplifier, on a board 4.5" X 5.5", the rectifiers, filter caps, and regulators were also on board. The heatsink was fan cooled. No surface mount parts were used, and the whole package was only 1" tall. The transformer was a desk top unit.

How much smaller do you need it?

"Noise instead of harmonic distorsion (less disturbing to ear)"

Yes, I agree harmonic distortion is less disturbing to the ear than noise.

"Low price"

The delivered to your door price on a Berhinger 2.4KW amplifier is $350. I see no switching amp near that power/price.

"Ruggedness towards low speaker impedances."

A matter of design. Making a switching amplifier short circuit proof is not a non-trivial matter. Its pretty cut-and-dried for analog.

Switching amplifiers could be the future, but as far as I am concerned, the future has not yet arrived.
 
I have no practical experience with class D amplifiers, but I can see at least two disadvantages compared to class AB. Firstly, class D amplifiers with a digital input which are basically power sigma-delta-DAC's have the disadvantage that they need an extremely clean supply, as the supply is basically the DAC's reference voltage. Secondly, both analogue and digital class D amplifiers often have a rather high output inductance due to the output filter, making the frequency response at high audio frequencies more dependent on the loudspeaker impedance.
 
MarcelvdG: You are right about these points, i agree completely with you that a purely digital solution is not optimal (even if it would at first thought appear to be).
Demand for Ultra stable power supply, and delay in feedback loop A/D + DSP handling of error signals, and partly also quantization noise in the feedback loop all contribute to additional distorsions, none of these problems exist in the 'analog PWM' approach.

Today i tested a new 'analog PWM' design, with a +/- 80 V DC supply with 2200 uF capacitance and a transformer of just 55 VA. As you might gather the supply was all but stable when loaded. But it still had a really good bass control and could produce something like 200 Watts in 8 Ohms for a half hour before the 55 VA transformer got too hot. (And the others had left the building because of loud Techno sounds) :cool:

djk: You surely seem to have some serious reservations towards the Class D technology ;) However i will comment on one point:
If you calculate the efficiency to 95% or so, based on Input Power / Output Power, i don't see how you can have: (quote) 'ignore losses in the output filter and comutating diodes'.

Lieven: The TDA8924 looks very interesting! :) Cool that you can get 2 x 120W rms in an SMD package.....!
 
Ex-Moderator
Joined 2002
Please excuse me if I'm out in left field on this not being a digital type person, but doesn't the designation "Class D" pertain specifically to the power supply providing the voltage rails and current for the output. The "amplifier" section itself is still analog isn't it? If that's true than it's misleading to compare A or A/B vs D. Class D amps will most likely have a Class A/B output stage.

This thread is talking about digital amplification:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=22414&pagenumber=1
 
Lupulroz: Sure!

If you play music at a power level of say 200 Watts RMS (clipping level), the music does not contain (like sinewaves) a constant energy. So the amplifier does not put out 200 Watts of power all the time, only at the music's peaks.

You can measure the average energy content of different types of music, for instance by measuring the heat generated over a time period of 30 minutes of music, when dissapated in a closed thermal circuit, like a resistor in a waterbath.

I have performed experiments like tihis, and it turns out the average energy content is something like 30% of the maximum value, measured over 30 minutes. Depending somewhat on the type of music played of course.

Since a normal analog amplifier dissapates a fairly constant amount of power no matter the output power (from 30% to 100%) the benefit of the music's low energy content is un noticable. (If you take a 100W amplifier, it will dissapate 40W, at any output power from 30 to 100W).
However in a PWM amplifier it can be exploited. The PWM dissappates say 8W at zero power and 12W at max power. So the consumption from the transformer is output power + avg. 10W.
Since a well built toroid transformer can deliver much more power than specified, if it is in short periods, you can rely on full RMS power for minutes at a time. Playig music, you can go at clipping level for hours with the tiny transformer.

So hook a toroid up to a PWM amplifier, and you can play 100 Watts RMS at clipping level, only using a 30 VA transformer. I know it may sound crazy to some, but i have actually demonstrated this at a danish audio convention this last spring.
You may think the sound is flat or bass control is weak. But try it, you might be surprised!
 

TNT

Member
Joined 2003
Paid Member
I have now owned a pair of monoblocks based on Ice-power (semi-diy) and I think they are superior in certain areas. I would never buy anything else actually.

The trancparency and directness is stunning. Everything alse just sound dull or artificial compared.

/
 
Just wanted to show you this, i placed a 500 VA to the right of the 55VA one playing.

The bottom is powder painted, so i had to use thermal grease to get a good cooling for the amplifier. Note the main caps are 2200 uF 100V. Not much to find in a 200W Amp. Just to test the sound anyway.
 

Attachments

  • 200w pwm.jpg
    200w pwm.jpg
    51.6 KB · Views: 2,126
2KW CLASS_D SUBWOOFER AMPLIFIER

Hello Guys Iam AbHiShEk from "InDiA"and Listen to me About my LOW COST CLASS-D PWM 2KW SUBWOOFER AUDIO AMPLIFIER
Features
1 No mains transformer needed i.e. operated directly from 220V AC mains @ 340V DC
2 Power output 2000 W RMS with lots of "SUBSONIC THUNDERBOLTS"
3 Frequency response 10Hz to 200Hz +-1dB
4 Switching Frequency @ 5KHz only "hahahahahahahaha" but too much for subsonic Class-D amp
5 Light weight Fan cooled Heatsink used because the efficiency is > 90 %
6 Dampening Factor > 500 @ 4 OHMS
7 No Feedback used ThereFore RealTime Amplification
8 THD < 0.1% @ full power
9 0 % Crossover Distortion
10 FINALLY Fully TESTED WITH 2 X 18" Low Frequency Drivers @ 8ohms 1000 W RMS each
Major Components used are SG3524 PWM Modulator/Oscillator/Error amp, OptoIsolated Gate drivers,2 X 75A IGBTs in pull manner, 8 X 550 mFD 400V Capacitors for Filtering,1000 V /35A Bridge Rectifier, 12dB/octave Linkwitz-Riley Low-Pass passive output filter and ofcourse a little bit of resistors and small capacitors etcetra.

WANNA BUILD IT YOURSELF or You Think I must be "KIDDING"
THEN MAY I HELP U PLEASE @ FACTOR_TEN@HOTMAIL.COM
 
I really don't subscribe to this eternal discussion about the ultra exact definition of classes. Just don't find it interesting. For some people it's the only interesting discussion to have...
But it seems that no two people have the same perception of what those classes really mean, so the discussion can go on and on. I know i have referred to class terms in the past, but i have decided in the future not to. (Except for class A and A/B).

My amp is a PWM (Pulse Width Modulation) amplifier, the modulator is not a Sigma Delta type, but i actually got the inspiration for it in the book 'Sigma Delta Converters' by James Candy. (Brilliant book!).

I think the Pulse Width Modulation term covers just about cover every amplifier of this type, be it digital, analog switching etc.
Even if some vendors try to confuse the customer by saying: Our technology is much better than the old crappy PWM. It's pretty much like saying:

If you eat our pill, you will only breathe clean air.

So to win the customer over, these same vendors sometimes invent all new 'classes' as 'proof' that their PWM amplifier is different from the rest. Total confusion yes.. But true product information? Not so much!

I will supply measurement results very soon. THD is around 0.02% in the range 100 - 5000 Hz, going slightly up above 5000 Hz.
 
Lars, what about radiation?

I had some terrible experiences in the past with a Tripath evaluation board, which was inherently unstable and gave out massive amounts of RF, to the point of triggering self-oscillation!

I'm toying with the idea of building a 6-channel power amp and I'm concerned that it might give out a substantial amount of radiation, to the point of disturbing my tuner and/or other equipment.
Keeping the speaker cables short wouldn't be an option, since my equipment rack is located 3+m. away from the speakers.

How would you approach and possibly solve this problem?
 
Hi Lars

So you are definitely talking about a classic PWM, do we agree upon this ?
In this case I'd like to know how you come to the conclusion that one does trade noise against distortion within such an amp ?
As far as I experienced do the same mechanisms that reduce distortion also reduce noise.

Regards

Charles
 
phase_accurate: PWM stands for Pulse Width Modulation. It is the widest denominator for all amps that give out 1's and 0's.

It is unrelated to the specific mechanisms you refer to that control the signal conversion, amplification, noise THD etc. So you may say 'Classic PWM' or 'Modern PWM', it's like saying 'Classic Weather' or 'Modern Weather'. PWM is PWM. And you can put a whole bunch of 'classes' in that box .. :)

Different methods are usable to generate the PWM signal, and some of them don't have harmonic distortion, (since they don't have any analog audio signal processing), only a noise floor.
Of course you can also have PWM modulators that have analog signal processing and harmonic distortion. So it was not a general term. The one i use is the integrator type, with no analog signal processing, and therefore no THD.
It has distortions, but the harmonics of the distortions are random nonharmonics of the signal, (or the same as: noise).
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.