Class amplifiers

[kostas]

People one question.

What is the difference of a A class, a B class and a C class amplifier?
I think i have also seen an AB class amplifier. Is it true?

?

[kam]

In a few words:

Class A is the HiFi, both positive and negative side of the input waveform is amplified and driven to output

Class B, Only one side (+ or -) of the input waveform is amplified and driven to output. With this way you have with the same amplifier double the output amplify than the HiFi, but half the waveform is clipped.

Class C, only o portion of one side is amplified. You can achieve great amplify gains but the output is a mess. Used for cheap (really cheap) radio players etc, that needs with less power more amplification.

Class AB, or push pull. This is something like B but different in the way that it works. It might amplify one side or one side plus a small portion of the other side. This is a very common amplifier.

[Johnny2Bad]

There are both technical and practical answers to this question.

On a technical basis, already covered by kam briefly, these are "classes" of design. They are not indications of quality so much as designations used in the history of electronics ... Class A, for example, is the simplest form and the first used in the early development of technology in the early 20th Century.

Another aspect of understanding Amplifier Class which is helpful is the concept of "linearity". Electronic devices are "linear" over a portion of their operation, and "non-linear" over portions of the possible operation. Linearity can be simply described as the area where the device operates in a useful fashion. We typically design electronics based on formulas. During the linear portion, we offer a value for input and run it through a formula, and obtain a certain output.

Lets say we desire an amplifier that increases voltage by a factor of two ... a typical amplifier circuit, in other words.

During the linear portion of the device's operation, input value 1 results in output value 2; input value 0.5 results in output value 1; input value 3 results in output value 6, and so on.

During the non-linear portion, this relationship cannot be relied upon. So, outside the linear range, our design's intended factor of two does not work. It may be higher, it may be lower, it may be no change at all. Regardless, we cannot rely on this area of operation for predictable results.

As it turns out, electronics typically are linear in the middle portion of it's possible operation and non-linear at the edges ... very low or very high areas. This is not really limited to electronics, either ... to a large extent everything in the ordinary world is like this. It is easier to pick up a marble than a butterfly without damaging either. We could say the nerves and muscles in your finger are more linear at the sizes and density of the marble, and non linear when dealing with the size and fragility of butterfly wings. You cannot pick up a very heavy weight at all with your fingers ... another non-linear area where no amount of input (muscles and nerves) will result in anything but zero output (the object will not move at all).

A device (transistor, vacuum tube) built as a Class A amplifier conducts over the entire linear range of operation. This is simple ... one device can operate the entire sine wave, which is what amplifiers must do. This also means it is very inefficient ... a lot of the work is lost as heat. A Class A amplifier will be roughly 25% efficient. But it operates entirely in it's linear area ... it therefore is extremely accurate and has low inherent distortion.

If you listen to any pre-recorded music, microphone preamplifiers are typically Class A devices. Vocals and acoustic instruments often have a natural, realistic quality and that partly can be attributed to the use of a Class A device for un-amplified instruments during the recording session.

The Class B amplifier divides the two halves of the sine wave into two parts ... one device works on the positive wave and the other works on the negative wave. During the time the "other" device is working, a device is turned off entirely. This is significantly more efficient ... more power is used to do the work of the amplifier and less is lost as wasted heat. However, there is a problem. The area where the device goes from conducting to turning off entirely is in the non-linear area of operation.

There is also the problem of precisely timing the point where one device hands off the work to the other. These are both difficult to solve, and broadly speaking, a Class B amplifier is not suitable for accurate sound reproduction. Distortion is high and not easily managed or minimized. It is somewhat more complex than Class A and requires two devices to do what one can in Class A.

An example of Class B amplifiers in use are electronic megaphones ... all of which sound terrible, if you've ever heard one. They take advantage of the high efficiency to make best use of a combination of limited battery power and high audio output (further efficiency is achieved with the horn shape).

Class C is not suitable for audio use. It is very efficient but the distortion products are not manageable. It is similar to Class B but each device conducts over less than the full 1/2 of the sine wave. Obviously, there is no possibility of one device handing off to another cleanly, since the first device is no longer even conducting (doing work) at the point where the two halves of the sine wave meet. It is used in Radio Frequency applications, where the necessity of tuning to a specific frequency means the distortion components can be managed by filtering. The audio band covers too wide a frequency range for this type of filtering to be effective.

There are no examples of Class C devices used in an audio application. But, if you would like to imagine one, it would sound like the electronic megaphone, only much worse.

Before we move on, there is the case of Class AB. After the above discussion regarding Class A, where the entire operation is in the linear range, and Class B, where there is a problem with linearity and timing during portions of it's operation, we can look at Class AB, which is a solution to the quality of Class B reproduction issues. The two devices in a  Class AB amplifier do not turn off at the handover point ... they both conduct somewhat. This is less efficient ... you lose the advantage of turning off entirely that Class B has.

However, this can be designed so that they are now in the linear area during the hand off from one device to the other. This reduces significantly some of the distortion of a Class B design. Because each device still conducts much less power during 1/2 of each cycle, it still has greater efficiency than a true Class A design. In essence, the devices operate at Class A at very low signal levels and at Class B at higher power levels, thus the designation Class AB.

There is still the issue of crossover distortion ... exactly timing the point where each device hands off to the other. Class AB amplifiers typically incorporate a Negative Feedback Loop to reduce this form of distortion. Like all correction circuitry, it is not perfect, but it does work well enough to make the Class AB amplifier the most popular design available today.

[kam]

excellent info Johnny2Bad, thanks for sharing.

By the way, it seems that you have quite a god knowledge in audio electronics. Has this to do with your work or is it a hobby? Do you make amplifiers and radios?

[Johnny2Bad]

excellent info Johnny2Bad, thanks for sharing.

By the way, it seems that you have quite a god knowledge in audio electronics. Has this to do with your work or is it a hobby? Do you make amplifiers and radios?

At one time I owned a retail store selling audio equipment. That was quite some time ago ... I've been in my new career for 25 years, which is about as far away from retail and audio, or electronics in any form, as you can get.

We focused on three primary markets: the serious home audio enthusiast, the serious home recording studio enthusiast, and the Professional market with emphasis on clubs (they were called 'disco's" then), small to medium ( up to concert hall) live music sound reinforcement, FM radio stations, and professional recording studios. We had a service tech, obtained factory repair authorization from our suppliers, and handled warranty and other repairs in-house for gear we sold only (ie we were not a "walk-in" repair shop). I've worked with and peered into the guts of a lot of gear that many people know only by reputation; a lot of which is in the "collector status" today.

We also began work on advanced car audio systems, at first just for our own vehicles but later to retail. We imported Nakamichi 250 and 350 DC-powered decks from Nakamichi Japan, imported Jim Fosgate's amps at a time when he had less than 5 dealers worldwide and built them essentially in his garage, and used selected drivers sourced directly from the manufacturer. We used SON-Audax, Peerless, Philips and KEF drivers. I still miss the KEF T-27 tweeter ... an outstanding HF dome with excellent, lifelike HF response, but you could cross it over at 1500 Hz/24 dB per octave (although KEF never did in their manufactured speakers, similarly the famous BBC LS3/5a monitor crossed it over at 3 Khz).

An example of the Nak 250 can be found here: http://www.nakamichi-schenk.nl/SpecialNaks_content/Nak250E.htm

I came to this site via the excellent summary of the 555 timer found elsewhere.

My interest these days in electronics is purely personal and for my own use. My current project is a car system for my personal vehicle. It has two design goals ... to maintain the stock appearance and functionality of the vehicle and to obtain the highest sound quality. That combination results in a bit of a challenge, which i'm enjoying.

I'm not interested in SPL for it's own sake, or bass levels that belong in a movie theatre rather than a serious sound reproduction system. I have always installed some kind of audio system in all my vehicles.

I've done quite a bit more with this install than most. Briefly, the system consists  of 6 channels of amplification for full range speakers and a true subwoofer. The main amplifier is a DIY project with 3x 2020 Tripath Class-D chips. I am going to try an OEM full-range Peerless driver in the dash tweeter location (200~22Khz 2"metal cone) that is commonly found in flat screen TVs. The balance will consist of Peerless drivers with the exception of a VIFA dome tweeter originally used in a ProAc speaker (ProAc Studio 110, 130, 140 and centre).

The OEM deck will be retained, with two channels of transformer-based impedance matching to convert the power-amp level signal to line level impedance (although it will be at a few volts level, so not truly "line level" of 0.775V) to a ECC82/12AU7 vacuum tube powered buffer stage, and then to the three Tripath chips. Another two channels are available from the deck (so the F-R fader control will work) and they will also be transformer coupled, to a dedicated LF amplifier. The buffer stage will not be in the LF path.

At the typical voltage levels of an automobile power supply and driver impedances I'm using, I expect 6 channels of clean, wideband power of about 13.5 W RMS per channel (eg all channels driven, 20~20,000 Hz at less than 0.1% THD, 4 ohm purely resistive load) for a total of about 80 watts RMS.

I KHz distortion, 1 channel driven at 13.5W 4 ohms, measured, varies from 0.031 to 0.036%, depending on which of the six channels you test.

By "car stereo" standard specification methods, most manufacturers would call this a 150 Watt amplifier  (eg 25 Watts RMS, measured via one channel driven, 1KHz, 4 ohms, 10% THD and then the product multiplied by six).

I will need to build a 555-based timer circuit to manage the turn-on and turn-off delays of the buffer stage and the following amplifier stages, so that the system is well behaved. As I said earlier, this should be, to casual observation, an ordinary, unmodified, stock vehicle. This means the sound system will be turned off and on by the power/volume knob on the factory deck.

The final system will acommodate a different source, such as an iPod, along with the factory radio/disk player. Radios for this vehicle sell for about $150 used, and along with someone else's idea of a quality interface for about $60, certain models would swap in, retain the stock appearance, and allow this aux input. I'm going to save the $200+ and go with what is installed and my own mods, as there is no significant difference otherwise between the stock head units available for this vehicle.

I expect decent performance from the stock head unit. Many people are of a different opinion, but I've actually measured a lot of aftermarket and factory decks. The factory sound system is actually quite listenable at moderate power levels, suggesting decent overall performance and reasonable compromises for the 6 factory speaker locations. At moderate power levels the factory units are up to the job, so all we need to do is insure the amplifier operates below the onset of clipping in the factory head unit (the "knee" of the THD +N curve). The buffer stage has the ability for a gain of 27x, with sub- 0.1% THD at gains of 6x or less, so this will not be a problem.

I expect improvements in overall SPL, of course, but also in detail, soundstaging and depth. The Tripath chips have extremely low overall noise. Bass performance will be significantly enhanced, but I have not spent much time there yet aside from insuring it will have a line level output available from the deck ... it will be a "stage 2" project. Because the factory deck has 4 channels available I will be able to adjust LF balance via the F/R fader control. For the 6 channels of main amplification, each pair of channels will have a level control for system balance using a ladder type 23-position switch.

I have no intention of entering any sound competitions, but the rulebook of the IASCA is a fundamentally sound blueprint regarding safety and allowable installation procedures. Their requirements differ somewhat from my standard practice, but only in the details (eg the size of grommet required for a through-firewall power cable) so I've decided to follow their rules for the installation. I recommend it to anyone contemplating a car audio project, as they are all sensible and can be considered known good practice.

Interestingly because I'm not using off-the-shelf store-bought electronics, I'd be forced to compete in a "Pro" Class with IASCA. I'd probably come in last, as this will not be a system capable of SPLs above perhaps 105~110 dB.

[circuitfella11]

hi,

on basic technical side, Class A, B, AB, and D are differentiated with their efficiency and output waveforms.

Class A- most linear, 100% of the input is in the output
           - efficiency is around 20% most likely
           -when used, it needs too much ventilation and cooling because it heats most parts of the circuit

Class B - input is larger than output so would likely have an input with gain to increase its input
            - opposite of Class A

Class AB - combination of Class A/B, it gives better operation than both classes.
               -signal outputs are specific.
               - 50% efficiency
               - best for audio applications
Class D   -used as digital amplifiers
               -switching amplifiers
               -80-90% efficiency


---regards