Andy's Blog : amplifier

Amplifier Classifications - Part II

Andy W — Thu, 04 Oct 2007 18:44:00 GMT

Class D Amplifiers

The output devices are switched on and off at a very high frequency (compared to the desired output signal) and are driven with a PWM signal (Pulse Width Modulation). This creates an output signal that is a square wave with a varying duty cycle. The signal passes through a low-pass filter to remove the high frequency signal. Since the output devices are fully on or fully off, the power dissipation is very low, and efficiency of the output stage can exceed 90% at full power in some cases. Sometimes called a “Digital Amplifier” (usually a misnomer since a square wave is still an analog signal).

Due to continuous advancement in semiconductor technology, class D amplifiers, once an oddity, are now thriving in the market. Driven by the desire for smaller Hi-Fi (and not so Hi-Fi) systems and better efficiency, this segment will continue to grow for the foreseeable future.

Many times class D amplifiers are called “digital amplifiers.” I prefer to call them “switching amplifiers.” There are some varieties that take a digital PCM signal and convert it directly to PWM, so technically it is not inaccurate, though it would in my opinion be technically inaccurate to call an analog input class D amplifier “digital.” Nevertheless people still advertise “digital” amplifiers… because “digital” has been marketed to mean “better than analog” in everything from cell phones to cable to satellite TV to HDTV to HD radio. “Better for whom?” is the question.

Regardless, the current crop of IC class D amplifiers are very good performers, in terms of output power, efficiency, output noise, low distortion (rivaling good Class B amplifiers), and reliability. There are also some high end designs that are even better. With most class D amplifiers though, you have to watch out for changes in high frequency response. This is due to the output filter, generally a second order filter. The output filter is designed to be flat with a particular load impedance. If a lower impedance is used the filter will be “overdamped” and there will be drop in high frequency output. Conversely, if a higher impedance load is used the filter will be “underdamped” and there will be a rise in the high frequency response. The response variation is generally not to severe, but if the amplifier is not loaded the output filter can “ring.” Generally this is planned for, and doesn’t present a hazard to the amplifier.

Class G Amplifiers

This type of amplifier is basically a Class B amplifier with a second voltage rail. When the signal is high enough the higher voltage rail is used, and when it is not needed, the lower voltage rail is used. Due to the dynamic nature of music, this can reduce power dissipation over the Class B amplifier.

Probably more of these amplifiers are used in Pro Audio than home audio, and the improvements in class D technology have pretty much rendered these amplifiers obsolete, in my opinion. Class B (or AB if you must) might be next on the chopping block as Class D designs gain more acceptance.

Class H Amplifiers

This is similar to Class G except only one voltage rail is used, but the voltage of the rail is varied up and down with the input signal. Generally this requires a switch mode power supply (otherwise there would be no point to using Class H, as a linear regulator would dissipate just as much power, if not more, than the Class B amplifier itself).

BASH amplifiers as well as the Carver designs fit into this category.

Bridged Amplifiers

Any of the above amplifier types can be bridged (a BASH amplifier by design is always bridged). To bridge a stereo amplifier the phase of the input to one of the channels is inverted and the speaker load is driven by both amplifier channels simultaneously. The positive speaker lead is driven by the noninverted channel and the negative speaker lead is driven by the inverted amplifier channel. The ground connection (or common connection) is not used. There is more than one way to accomplish this, but the concept behind running a bridged amplifier is to create a “more” powerful amplifier. I put “more” in quotes because the power supply and/or the output devices will ultimately determine how much power you can get. If the power supply can’t deliver the required current, or if the output devices can’t handle the added stress, then the power output will be limited, and it’s possible that the output devices could be damaged.

If under normal circumstances an amplifier is delivering 1V_rms to an 8 ohm loudspeaker, the speaker receives V²/R = 1/8 Watt of power. If on the other hand the same amplifier is connected in a bridged arrangement, and the same 1V_rms is available at the noninverting output, there will be -1V_rms at the inverting output (or 1Vrms with an inverted phase, or 180° out of phase – pardon my nomenclature for the moment, it’s just to make the math easier). The potential difference (voltage) at the speaker is now 1V - (-1V) = 2V. 2Vrms on an 8 Ohm speaker will now deliver V²/R = 4/8 = 1/2 Watt of power, or 4 times the power. The voltage is doubled, and by Ohm’s law the current is doubled; Volt x Amps = Power and 2 x 2 = 4. But notice again that the current is doubled, and unless the amplifier can deliver that current, that level of power will not be achieved. The doubling of current is why many 4-8 Ohm rated amplifiers can only be used with 8 Ohm speakers (or higher) when operating in bridge mode; the doubling of current with a 4 Ohm load would exceed the design rating of the output devices, or the amplifier could overheat.

Amplifier Classifications - Part I

Andy W — Fri, 28 Sep 2007 19:40:00 GMT

Here is a brief look at the classes of solid-state audio amplifiers. These classifications do not include tube amps (valves, for you Brits).

Class A Amplifiers

The output device(s) operate over the full 360° of the waveform. The output devices never turn off. Efficiency of the output stage is very low, with either 12.5%, 25%, or 50% theoretical maximum efficiency.

Class A amplifiers are highly touted for their linearity (low distortion). Most often this is true, but sometimes it isn’t. A class A amplifier can be as simple as a single transistor and a few biasing components, or as complex as any other. One thing that is always true about class A is that it is horribly inefficient, since the full output current capability of the amplifier is always flowing in the amplifier. Current that is not delivered to the load is wasted as heat. This fact alone usually relegates class A designs to lower output levels.

Still, a well-designed class A amplifier can put up some very impressive numbers. And a poorly designed (or should I say conceived?) one will do no better than a tube amp. (Oops, did I say that?)

Class B Amplifiers

The output device operates for 180° of the waveform. A complementary pair of output devices is used to deliver the full 360° of the waveform. Efficiency is much improved over Class A operation, since the quiescent current is very low compared to Class A. Most people call Class B amps “Class AB” (a misnomer, in my opinion) due to the presence of a small bias voltage. If the bias voltage were not present, then the output devices would deliver power to the load for less than 180° of the waveform and you would have crossover distortion and technically you are now in Class C operation.

Class B amplifiers make up the bulk of commercial and professional power amplifiers. Although in my opinion it is technically incorrect to call them Class AB amplifiers, most people do (even engineers) and most people will say, “Class B? No, class B amps will have crossover distortion and aren’t any good for Hi-Fi.” This is again one of those nomenclature things that has been around so long that it’s almost of no use trying to convince people otherwise. In fact the term “Class B” has such a negative connotation that the sales and marketing departments would probably still refuse to use it even if it were technically correct. Each output device in a correctly designed Class B will deliver current to the load for exactly 180° of a sine wave cycle. A pair of output devices, precisely timed, take turns delivering current to the load. If it does not deliver current to the load for a full 180° the result will be crossover distortion which is a very clearly audible form of distortion due the presence of high-order harmonics. A well designed class B amplifier will have no audible crossover distortion. This is done by not allowing the output transistors to fully turn off at the crossover point (zero volts on the output). A small bias voltage on the output devices allows a small amount of quiescent current (aka bias current) to flow in the output devices even though there is no output voltage (the complementary device is sinking the current), and therefore no output current to the loudspeaker.

To explain it any further requires a picture… so let me know if this is of any interest and I will try to get it done.

Class AB Amplifiers

Most Hi-Fi amplifiers are marketed as Class AB amplifiers, but are truly Class B. Please read the aforementioned stuff in the Class B section for my explanation.

An otherwise Class B amplifier that is heavily biased can also be termed a Class AB amplifier. It will run in class A for a (usually) small part of the output power then revert to Class B (sort of) operation at higher power levels. The other case of Class AB operation is when a class A amplifier drives a heavy (low impedance) load slips into class B operation when it can’t deliver the required current in class A mode. This type of operation could be intentional or unintentional (or perhaps not designed for). One curious type of amplifier is a “trimodal” design that can operate in Class A, Class B or Class AB at the flick of a switch.

Class AB (as described here) will actually have more distortion than an optimally designed (and optimally biased) class B amplifier. Not enough bias voltage causes crossover distortion, but too much can cause distortion too (though it is not as readily audible), and you get to say “2W of Pure Class A” in the marketing copy. Whoopdy-doo.

The technical case used for calling a “class B” amplifier a “class AB” amplifier is this: at idle the output devices aren’t really “off” but conduct a small quiescent current so it’s Class AB. Whatever. 99% of the people agree with you.

Maybe it's just bias. (pun intended)

...to be continued