AVS Forum | Home Theater Discussions And Reviews

AVS Forum | Home Theater Discussions And Reviews (http://www.avsforum.com/forum/)
-   Receivers, Amps, and Processors (http://www.avsforum.com/forum/90-receivers-amps-processors/)
-   -   Amplifier FAQ (http://www.avsforum.com/forum/90-receivers-amps-processors/1294455-amplifier-faq.html)

MichaelJHuman 11-30-2010 02:54 PM

Amplifier FAQ

* How much power do I need?

A typical speaker might produce 85 to 90 dB SPL at one meter from the speaker with one watt of power ( as supplied by an amplifier.) As 85 to 90 dB is loud, I hope it's obvious it takes very little power to get subjectively loud sound.

As you increase volume, power needs will rapdily increase. You may have heard that you need to double power for each 3 dB increase in SPL. How much louder is 3 dB more? Some sources say "just" noticable. I think that's a fair explantion. You will notice a 3 dB increase, but it won't be dramatic.
To DOUBLE perceived loudness you will need a 10 dB increase in SPL. An increase in 10 dB SPL requires 10 times the power. In other words you need TEN TIMES the power for TWICE the perceived loudness.

Summarizing, you need very little power to get a loud music or movie, but as you increase volume you very quickly run out of power.

Speaker sensitivity, as mentioned above can be quite helpful in being able to play your system more loudly. The higher the sensitvity, the less power you will need to play at the same loudness. To some extent, sensitive speakers greater than 90 dB will cost more money than lower sensitivity speakers, so there is a financial trade off between amplifier power and speaker sensitivity. This is only a general observation.

In open space, for each doubling of distance away from a speaker, SPL will drop by 6 dB DPL. At 2 meters, SPL will be down 6 dB (from what it was at one meter). At 4 meters, 12 dB. In a room, the situation changes, as sound is reflected. With an SPL meter, you can get some idea of what the drop would be in your own room.

There are handy online calculators which can be help give you an idea (e.g. http://www.crownaudio.com/apps_htm/d...ct-pwr-req.htm
Your SPL will be higher with seven channels playing at the same time, than it would would say, two channels. I compared one channel playing to seven channels once, and got an increase of 3 dB SPL. I can't say if this is typical though.

You may ask why you need 100 or more Watts if you can play loudly with one Watt of power. I would say it's because sound is very dynamic, especially movies. If a movie's sound adheres to the THX standard, it's average level would be 20 dB below peak level. This means you need 100 times the power at peak level than average level. You may not need this much peak power for every channel at the same time. For movies, it's clear that the loud scenes in the movie are the scenes you need a lot of power for.

Using one or more more powered subwoofers will reduce the power your amplifier needs to power the other channels. Which is why it's typical to use the bass management available in most audio video receivers.

It might be helpful to have more power if you do a lot of music listening and are critical about sound quality. Distortion due to running out of power may be more obvious for music. This is because a lot of music has a lower dynamic range than movies, so the average level is higher and if distortion is happening, it might be more obvious.

Different individuals will have different sensitivity or interest in ensuring inaudible distortion. Therefore their interest in spending more money to ensure they are not hearing distortion due to running out of power will vary.

A number people like to throw around is THX reference level. This means your system can play EACH (excepting the subwoofer) channel at 105 dB. The subwoofer would have to play at 115 dB, and that's just for LFE - it does not include any bass signals crossed over from other speakers via bass management. I personally don't think this is an important goal unless you actually want to listen to movies this loud. If you can live with less SPL, you can likely save money.

I think that about covers the basics. In summary -
* Your power needs will vary a lot depending on listening tastes, your speakers, and how your speakers are positioned in your room
* You don't need a lot of power to play loudly
* You may need a lot of power if THX reference level is a goal

* How can I assure the best sound for my budget?

There are two answers to this.

One answer is to buy as much power as you think you will need, budget permitting. Compromising between power and budget is a good idea, as you can accept a 3 dB drop in SPL, which is not much of a drop in loudness and reduce power needs by half.

The other answer is that amplifiers can sound different, and you will have to take this into account. Quite a large number of blind listening tests have shown that amplifiers sound very similar. To take one extreme case, the most budget of Pioneer receivers was once involved in a listening test with much more expensive amplifiers. No one could do better than chance in identifying the Pioneer (statistically speaking). If someone tells you they heard a dramatic or a night and day difference by switching amplifiers, you may want to consider the possibility they are overstating the situation.

ALL amplifiers distort. None is perfect.

Distortion could be grouped into three general categories - noise, harmonic distortion, and non linear frequency response (these are my own broad definitions rather than a strictly academic set of defintions.)

Noise is an inescapable fact of the physics of electronics. Even the best amps can't eliminate noise. The good news is that noise can be very low when compared to the signal. A specification called Signal to Noise Ratio (SNR) is often given for amps. Say an amp has a SNR of 80 dB. That means the signal is 80 dB higher than the noise (100,000,000 times higher.) This is pretty good if you consider that an average room might have an ambient sound level somewhere in the range of 10 dB to 20 dB SPL. 80 dB above 10 dB (ambient) is 90 dB, which I would call loud. And 80 dB is not a particularly good SNR for an amp. With highly sensitive speakers, you may want an amp with better than average SNR.

Harmonic distortion means that if you feed the amp a sine wave signal, for example at 1 khz, unwanted waveforms at other frequencies will appear on the output. Harmonic distortion is measured as THD (Total Harmonic Distortion.) This measures what percentage of the amplified signal is made up of harmonic distortion. Most amps will have a very low THD spec. THD is not absolute, it varies depending on power output. An amp with what looks like a high THD (> .1% THD) will almost certainly have a lower THD at lower power output. THD levels on good amps are well below the point of audibility, which is 1% THD according to some studies. 1% is not an absolute though. Humans are more sensitive if the harmonic distortion is odd order and/or higher order. Higher order just means higher multiples of the base frequency. We notice distortion more if the harmonics are higher multiples of the base frequency. As an example, a 7 khz harmonic distortion component from a 1 khz test signal would be odd order and would probably considered to be high order.

The last type of distortion mentioned is frequency response non linearity. In other words, the amplifier does not respond identically to all frequencies. Ideally, if we want a perfect amplifier all frequencies would be treated equally. This is not the case in reality, but most amplifiers have excellent response linearity over the audio frequency range of 20 hz to 20 khz.

If you run out of power, the amplifier will distort the signal. If it's severe enough, it will be audible.

Now that I have mentioned these various types of distortion, I think it's safe to say a well designed amplifier will reduce these effects to the point of inaudibility. Because of this, and because of the conclusions from blind listening tests, it's not clear why some people hear differences in how amplifiers sound.

In spite of some common beliefs, I think it's safe to say the job of an amplifier is to amplify the input signal while adding as little as possible distortion. This is not always true in reality, but it should be the goal. The preamplifier, pre processor or other device hooked into the amplifier should apply tone controls, room correction and such and the amplifier should just do what it sounds like it does - amplify.

Some people will spend a lot of time and money finding that perfect amp. I can't think of any guidance here. Make sure you can return the amplifier if you don't like it, with small or no restocking fees, and try to minimize the cost of shipping where possible. Reviews may be helpful, but some of them seem of questionable value given that humans have been shown to have bias and that bias effects their judgement.

* Does too little power damage speakers?

You may see statements like this in various sources. "It's not too much power that damages speakers. Too little power damages speakers."

I feel this sort of statement deserves clarification.

In one sense, it's 100% wrong. Too much power damages speakers. This is not debatable.

There's two ways to damage a speaker - exceeding it's mechanical limits, or exceeding it's thermal limits. Both situations require "too much power." If you are pushing so much power to a speaker that it's woofer is experiencing overcursion, that's too much power ( it's not a pretty sound.) If you push sufficient power to a speaker for long enough, you can damage it's drivers. Voice coils in a speaker's drivers will heat up, and if they heat up too much, they will fail (this seems more common with tweeters than woofers.)

* Why do people say that too little power damages speakers?

If you play your system too loud, the amplifier will clip. Clipping means that the amplifier is trying to amplify the signal, but hits a limit. That limit is it's supply voltage. This supply voltage could be reduced by load on the power supply or the amplifier's limiter circuits kicking in to try to save the amplifer from the abuse you are trying to inflict on it.

When an amplifier clips the result is that average power increases more quickly than it would if the amplifier is not clipping. This topic is non trivial, and Rodd Eliott explains it better than I ever could so if you want the gory details, read this (http://sound.westhost.com/clipping.htm)

This increase in average power is the most likely reason for speaker damage.

Modern music is highly compressed which makes the situation worse. Average signal level is so high in compressed music, that the speaker drivers don't have as much time to cool down.

If you have sufficent amplifier power to avoid clipping, your chance of doing damage to your speakers will be lower.

Can you buy an amp with too much power? This seems possible. While many sources of online info say you can't have too much power, that's clearly an oversimplification. If you had a 1000 watt amp, and you connected one of those cheap portable 8 ohm speakers to it, which is designed for less than a watt of power, I am 100% sure you could damage it, and with no clipping present. This does seems like an unlikely scenario with normal speakers and with normal music/movie sources though.

* What specs are important?

Not many, except for power. And power is measured in different ways, so direct comparisons should be taken lightly.

Hopefully, power will be measured from 20hz to 20khz into 8 ohms (ideally also into 4 ohms,) at some low THD figure, such as .01 % THD. And hopefully this is a continuous power measurement as dictated by the FTC rules (in the US.) A multi-channel (more than two) amplifier may include an all channels driven measurement. If not, it will almost certainly put out less power into all channels. Some people have criticized the all channels driven measurement. The main reason is that when playing movies, the most common use for mult-channel amplifiers, it's said that max power will never be demanded from all channels at the same time. Personally, I think it's nice to know the ACD spec, but realize that in some cases, it might not be continuous (it may have been measured with a short duration "sweep".)

THD is considered important, but consider that any good amp designed for high fidelity will have very low THD. THD is not an absolute for a given amplifier. It will vary with power, and manufacturers may squeeze out a bit more power by measuring at a higher THD than another manufacturer. A commonly stated theory is that THD of below 1 % THD is inaudible. This is also not an absolute as human sensitivity to harmonic distortion varies depending on the type of disortion. For example, if you consider a musical chord, and consider the 2nd and 3rd notes as disortion, they are actually pleasing to listen to. But if you try to play an off note, it will be perceived of as discordant. On paper, it looks like the harmonic distortion for most amps should be inaudible.

The SNR (noise) spec could be important in some cases. It's measured as Signal to Noise ratio. It will usually be very high, at least in amps I have seen. If you consider that noise in a quite room may be no lower than 20 dB, and breathing is 10 dB, a SNR of more than 80 dB seems pretty good. If you have very high efficiency speakers, you might want to pay some attention to the SNR.

There are various other measurements, none of which are of real use. Slew rate does not tell you much, unless it was abnormally low (in which case, they are not going to list it http://www.avsforum.com/avs-vb/images/smilies/smile.gif Damping factor has been critcized as being pointless in a number of articles by knowledgeable people. Occasionally you will see some max current output measurement. Not sure what that tells us with no other info. For all I know, they shorted the all the amp's output terminals and measured the peak output amperage before it shut down or blew up.

* Do I need to match an amplifier's impedance (such as 8 ohms or 4 ohms) to my speakers?

A solid state amplifier does not have an impedance as such (tube amps are a bit different in this respect.)

Almost any audio amplifier will work with 8 ohm speakers (with some exceptions such as fixed voltage amps.) As you lower the impedance of speakers you connect to a amp, you place more demand on it. Lower impedance means more current will flow through it's output devices (usually transistors). More current means more heat.

4 ohm loads present a more difficult load for amplifiers. The lower impedance means electricity will flow more easily. That means the current flow will be higher. Higher current means more heat will develop in the amplifier, especially in the output devices. Amps typically have themal protection circuits. If you your amp is shutting down due to thermal overload, you are driving it too hard, not giving it enough ventilation, or the amp is not working properly.

Amplifiers sold for pro audio applications, and for home audio would usually be able to handle 4 ohm loads with no issues. Receivers may shut down under heavy load. Or perhaps their longevity could be reduced, because they were not really designed for 4 ohm loads. Lower impedance is not suggested, such as connecting two 4 ohm speakers in parallel.

* How do I connect an amplifier?

For connecting an amplifier to a source such as a pre processor, there are four standard connectons -

* Unbalanced RCA jacks
* Unbalanced 1/4" jack (aka phono jack)
* Balanced 1/4" (aka TRS)
* Balanced XLR

An unbalanced connection has two wires. One is a ground, the other carries the audio signal.

A balanced connection needs three wires. One carries the audio signal, another an inverted "copy" of the audio signal, and the third is a ground.

The cheapest method is probably RCA. This should work fine when the cable length will be short, and the device being connected lacks balanced outputs.

If an amplifier lacks RCA inputs, and your device has only RCA output, such as an Audio/Video Receiver, use a cable with RCA jacks on one side, and 1/4" cables on the other side. This will be an unbalanced connection.

If the device being connected to the amp has balanced outputs, it makes sense to use them. For home audio, this is often an XLR connector requiring use of an XLR cable. You may also have the option for a 1/4" connection. For balanced 1/4" connections, you must use a cable with TRS jacks on both sides. These will have two insulators on the jack, giving it 3 contact points. Hence the acronym TRS for Tip, Ring, Sleeve. This is also known as a stereo 1/4" jack (as opposed to a mono 1/4" jack with only two contacts.)

* What do input sensitivity and gain mean?

They measure the same thing.

Input sensitvity is simple to understand. If the input sensitivity is 1 volt, the amp will be able to hit full power with an input of 1 volt. A higher input sensitivty means you need more voltage to drive the amp to full power. Many receivers have a pre amp rated for 1 volt RMS, which means they may not be able to drive all amps to full power.

Gain is measuring the same thing as input sensitvity in a different way. I will give a handy formula -

Gain (as a multiple) = 10 ^ (gain / 20)
Divide gain (dB) by 20, and use inverse logarithm (base 10) on your calculator

For 32 dB, we would get a gain of 40 times.

To convert this to input sensitivity, we would use this formula -
Input sensitvity = sqrt(max_amp_power * 8) / gain_multiple

Assuming a 200 watt / channel amp into 8 ohms with a 32 dB sensitivity -

sqrt (200 watts * 8 ohms ) / 40 = 1 volt

So a 32 dB gain would mean a 1 volt input sensitvity. Higher gain means an amp will match better with processors, preamps and receivers with lower pre amp output.

* How do I bridge an amplifier

Some amplifiers can be bridged. Don't try to bridge an amplifier not designed to be bridged. See the manual for detailed instructions.

Bridging is pretty simple. Two of the amp's outputs are used for one speaker. This will double voltage, and quadruple the power. This requires that one signal uis inverted.

In a typical bridgeable amplifier you need to set a switch for bridge mode. You will then carefully note the wiring diagram in the manual and/or on the back of the amp, and connect your speaker to the indicated outputs on the amp.

* How important is an all channels driven power measurement

This is discussed better than I can manage by other sources, such as Audioholics.com. I do have some thoughts on it though.
There's some factors to think about when considering how useful an all channels driven power rating is.

* Test signal correlation vs real audio correlation
* Test signal crest factor vs real audio crest factor
* Real world average power usage
* Law of diminishing returns

It's helpful to understand how an all channels driven test is conducted. It's done using test signals. The test signals are also likely correlated. By that I mean that they all peak at the exact same time. Music signals may tend to peak in the same spot, but are certainly not test tones. We don't expect all the channels in a movie soundtrack to peak at the same time ( based on various expert sources.)

Because test signals are likely correlated, by that I mean they peak at the same time this means the AVR/amplifier is being asked to provide peak power to all channels at the same time. Real audio is likely not like that especially movies (peaks in stereo music may happen at the same point in time.)

Crest factor is a measure of peak vs average level. A rough crest factor for movies might be 20 dB. That works out to require 100 times peak vs average power. All channels driven testing is very different than movie audio unless it's some sort of burst testing. As such it's relation to reality is limited.

Average power level may be quite low for a typical listening situation ( typical being very poorly defined, but it's the best I can do.) A typical speaker has a sensitivity close to 90 dB. That means it will put out 90 dB SPL with one watt of input. That's loud. There is some loss for distance. There's no great rule of thumb here, I use 6 dB. To output 90 dB factoring in the 6 dB loss for distance we need four times the power or 4 watts. A typical listener may listen a bit lower than that. For argument's sake let's say a typical listener listens 84 dB (which makes math simple too.) We are back to one watt. If peaks take 100 times the power that's 100 watts.

Using the above logic, we expect even budget receivers to manage this or manage it with minimal sonic impact. Consider that people usually use subwoofers taking a load off the receiver's power needs. Consider that every channel is NOT peaking at the same point in time according to various sources. Consider that even if the receiver runs out of power it's only for some brief moments. I once test Yamaha's cheapest receiver and did not find it wanting for movie use and moderate volume music use.

The final point is about the law of diminishing returns. Say we have an AVR which manages 100 watts peak per channel under real world listening. Well, say you don't believe that due to looking at all channels driven numbers. Let's call it 50 watts for argument's sake. So you buy an amp that can do 200 watts. Four times the power. Each doubling of power is a gain of 6 dB SPL ( sound pressure level.) That's louder but note that it takes 10 dB to double the perceived loudness. So it's not unreasonable to think we can't do much better than 6 dB more SPL even after spending a lot more money. My point is a person could save some money accepting whatever SPL their AVR could safely deliver without compromising a lot.

* What's the difference between solid state and tube amps

Perhaps the number one reason people state for buying a tube amp is that they say tubes have a warmer sound.

Tubes are less efficient, and potentially more fragile than solid state amps. To get a tube amp with the same power as a solid state amp, you would usually pay more more. High power tube amps are uncommon. Very few people would choose tube amps in a set up designed mainly for surround sound use, as the expense would be high, and they would take up more room.

Tube amplifiers may not measure as well as solid state amplifiers, Randy Slone mentions this in his book on high power amplifiers. There's a class of tube amps known as SET (single ended triode) amplfiiers. In bench tests I have seen for this sort of amplifier, they have measured very badly when compared to solid state amplifiers.

From an audio fidelity perspective, I could find no evidence that tube amplifiers are more accurate than solid state amplifiers. A number of sources said that tube amps can be more linear and thus would need less negative feedback. This would be an advantage if negative feedback was a bad thing. A number of audio designers have specifically addressed criticisms of negative feedback, and have rejected the criticisms. There are audio designers who seem to dislike negative feedback however.

Is there a reason the sound of tube amps is preferred to solid state amplifiers by some people?

One often stated reason is that tube amplifiers sound better due to second order hamonic distortion. Rather than being discordant, second order harmonic distortion is said to sound pleasant. So even though it's distortion, it's pleasant sounding distortion.

Vaccum tube amps are said to sound better then they are overdriven than solid state amplifiers. This would help explain their popularity as guitar amplifiers.

For more information, I suggest doing a search online for 'tube sound.' You should be able to read about studies that have been done which better explain the technical details on how solid state amplifiers and tube amplifiers can sound different.

* What are amplifier classes?

This is covered in numerous places online in better detail. And in general, it's not that important. If an amplifier is lighter than other amplifiers of the same power output, it is helpful to know what design it uses. If the lighter amp is more efficient, such as class D, the difference in weight can be understood.

Class A - Theoretically the amp with the least distortion. The downside is that it will be more expensive, heavier, run hotter, and be less efficient than a class AB amp. Two writers of two books on amplifier design, Self and Slone downplay the real world advantage of this design. Put simply, the design is inefficient, because the transistors are "fed" a bias current such that they are always on when the amp is turned on. This means that with no signal, the transistors still consume a lot of power, which is wasted power, and dissipated as heat.

Class B or AB - Used in receivers, and most other power amps. The downside of class B or AB amps is crossover distortion. There are two sets of output devices (e.g. transistors.) If the audio signal is positive, one set is used for amplification, and if negative the other set is used. Both sets of output devices are not always on, making this design more efficient than class A. When the audio signal passes through 0 volts from positive to negative, or vice-versa a "glitch" occurs in the point of crossover. There are ways of dealing with this issue, which is good news as this is the most common design in amplifiers. The only difference between pure class B and class AB is that the class AB biases the output devices such that they conduct for more of the cycle. Some people extol the virtues of class AB, but not everyone agrees ( See Audio Power Amplifier Design Handbook by Douglas Self for the details.)

Class D - An more efficient when compared to other all other designs. The design is interesting. The incoming signal is converted to a square wave, where the duration of the pulses in the signal carry the audio signal. This signal is then used to switch on and off the power transistors in the output stage leading to a very high efficiency. This is called switch mode, and class D amplifiers are sometimes called switch mode. Another name for class D is PWM because the audio signal is converted to a Pulse Width Modulated signal. Along with the efficiency comes a decrease in weight. These amps are very common in powered subwoofers and the pro sound amp market. They seem to be slowly becoming more common. There's one line of Pioneer receivers now using them. Time will tell if they overcome the popularity of class AB. Note that class D does not mean digital. The term digital is often applied when the class D amplifier's input is a PCM signal rather than an analog signal (the amp can directly convert the PCM signal to the PWM signal used to drive the output devices.)

Class G/H - Amplifiers which improve efficiency by using more than one supply voltage. They otherwise work like class B amplifiers. If an amplifier's power supply voltage was exactly what was needed at any instant in time, it's more efficient. Class G and H differ in how this is done, neither achieves the goal of suppling the exact voltage needed. A class H amplifier has an infinitely variable power supply voltage. A class G amplifier has two or more supply voltages.

Switching power supply amps - This is not an amplifier class. But an efficiency can be gained by using switching rather than linear power supplies. Transistors in circuits are said to operated in linear mode or switch mode. In switch mode, they are turned on or off, there's no intermediate state. In linear mode, they are fed varying amounts of current. Switching power supplies are used in computers. Switch mode power supplies are seeing more use in pro audio amplifiers. They are uncommon in home audio at this point in time. Your typical home audio amplifier uses an unregulated power supply (see the section on "How do amplifiers work")

* How do amplifiers work?

A brief explanation of how amplifiers work might be helpful.

This section will explain the most common amplifier design - a three stage class B amplifier using an unregulated power supply.

The amplifier's "heart" is the power supply. The power supply transformer puts a limit on how much power the amplifier can produce. Bigger is better in this case, and very high power amplifiers of the type we are discussing here can weigh over 100 pounds.

A typical unregulated power supply is really simple. The incoming AC voltage (120 Volts in the US,) would typically be lowered to a smaller voltage. Higher amps would need a higher voltage. A bridge rectifier then converts this smaller voltage into a postive and negative DC voltage. An audio signal will alternate between positive to negative (it's alternating current.) Class B (or AB) amps need two voltage "rails" as they have two "sides" one of which amplifies the positive part of the signal, and the other the negative part.

The typical three stage amplifier has the following stages -

* Input stage; converts the alternating voltage input to an alternating current signal
* Voltage amplifier stage; the signal will be increased to a much higher voltage
* Output stage; This could be considered to be a current amplifying stage - or to put it another way perhaps, it buffers the voltage amplifier stage from the speakers, and allows the amplifier to provide the potentially large current needed by the speakers

The goal of the input stage is to convert the input signal where a changing voltage is converted to a changing current while introducing little additional noise and distortion. It also needs to cope with ripple from the power supply, which is not providing a perfectly constant DC signal. The input stage also takes feedback from the output to lineraize (reduce distortion.) There are quite a few sources of distortion in an amplifier, and feedback is used to reduce them.

The voltage amplifier stage (VAS) is doing the real amplifying of the input signal. The input signal might be one or two volts at maximum, and the output from the VAS will be tens of Volts. For example in my receiver, it's max voltage is going to be close to 70 Volts.

The output stage does the heavy lifting. The output devices, commonly bipolar junction transistors (MOSFETs are also used,) are connected directly to your speakers. The current flowing through the transistors can be very high. Consider a 100 Watt amp connected to a 4 ohm speaker. Voltage would peak around 24 Volts. Current would be be pushing 6 Amperes peak. If you don't think 100 Watts is a lot of power, consider how you almost certainly have no desire to touch a hot 100 Watt light bulb. Don't do it, you will get a painful burn.

An amplifier is not that complex a device in some ways, and there has been little change in it's basic design for a long time (I believe the basic 3 stage amp design has been around since an RCA engineer developed it in the 1950s)

MichaelJHuman 12-01-2010 09:46 AM

Thought I would bump once for comments. I am thinking I need a few more sections. A section on connections would be good I think (balanced v unbalanced.)

Maybe a section on input senstivity/gain with an explanation of how they are calculated from each other, and what they mean in terms of a pre-amp.

And possible a section on bridging, but I would probably need to clarify the finer points in my own mind before adding that one.

t6902wf 12-01-2010 03:44 PM

I am surprised there are no comments! My comment would be thanks for taking the time. It is a very informative post.

dunan 12-01-2010 04:02 PM

good info

mods, can we get a sticky on this?

M Code 12-01-2010 04:23 PM

Just a clarification..
About linear power supplies and power output ratings..
Some brands can use different amounts of regulation, the more regulation the stiffer the power supply including higher capacity electrolytic capacitors, and this design will typically have lower peak power..
Thats why certain lower priced AVR brands use very little regulation and can have very high peak power for a short time for 1 or 2 channels.

However if ALL channels are being driven simultaneously the power supply will fold back accordingly. Today the majority of AVRs whose SRP is <$999 rate their power with (2) channels driven, and if (2) more channels are driven the power supply capacity (and power output) drops by 50%. So it is very EZ to calculate power output of all channels driven.. Thats why a brand's 7.1 AVR rated @ 75W/CH with 2 channels driven will only output 21W x7 with ALL channels are driven. Simply take the total power capability of 150W and divide by the number of channels being driven, for example if 5 channels being driven each channel will output 30W (150W divided by 5 = 30W).

Just my $0.02..

MichaelJHuman 12-01-2010 04:51 PM

Right, I know the situation changes with a stiff power supply. I was probably not clear on why amps use unregulated power supplies. One of the reasons was the one you gave according to the literature I have read, and that makes sense. Makers like Krell will use a regulated supply, which is basically another amplifier, and it will allow stable operation into low impedances. But the downside is less dynamic capability.

I think your calculation method of all channels driven may run into some issues. Such as that Onkyo model who's ACD rating was very low due to it's limiter circuits kicking in and limiting power more than would have been expected.

M Code 12-01-2010 06:26 PM

Originally Posted by MichaelJHuman View Post

Right, I know the situation changes with a stiff power supply. I was probably not clear on why amps use unregulated power supplies. One of the reasons was the one you gave according to the literature I have read, and that makes sense. Makers like Krell will use a regulated supply, which is basically another amplifier, and it will allow stable operation into low impedances. But the downside is less dynamic capability.

I think your calculation method of all channels driven may run into some issues. Such as that Onkyo model who's ACD rating was very low due to it's limiter circuits kicking in and limiting power more than would have been expected.

A lower cost AVR (SRP <$999) may use different schemes to protect itself or else it with self-destruct, especially as the load impedance decreases and bandwidth increases..
These include current limiters, thermal protection and even lower voltage power supplies as in Yamaha products but this is another discussion.

Your post explains well the amplifier basics, thanks for posting this..

Just my $0.02..

MichaelJHuman 12-02-2010 09:12 AM

Attempted to clarify some wording. Added sections on connections and bridging.

Dark Rider 12-02-2010 09:33 AM

Excellent write-up! +1 for sticky!

palehorse 12-02-2010 10:02 AM

Great write-up!! This definitely needs to be made sticky...

jake227 12-02-2010 11:26 AM

I found this to be a very interesting read for a slow afternoon at work

Is there any difference in sound quality when comparing class AB and class D amps? I'm guessing this has been covered before, but as long as we are on the subject I thought I'd ask. My guess is that there is little to no difference, but I'm not 100% on that.

MichaelJHuman 12-02-2010 12:46 PM

Originally Posted by jake227 View Post

I found this to be a very interesting read for a slow afternoon at work

Is there any difference in sound quality when comparing class AB and class D amps? I'm guessing this has been covered before, but as long as we are on the subject I thought I'd ask. My guess is that there is little to no difference, but I'm not 100% on that.

I don't know enough to evalulate that. We know that there's a big difference between class D and non switching amplifiers. They completely transform the signal, and then reconstruct it using a filter. I would not say this is inherently bad. Transformation of audio signals is common these days. Analog signals are converted to digital. Some audiophiles love SACD, and SACD audio is not a lot different than the PWM audio used in class D amps (see delta-sigma modulation.)

I am reading through Randy Slone's book on high power amps, copyrighted back in 1998 or so, and he seems pretty down on class D amps back then. Douglas Self, who had revised his amplifier design book more recently is more positive about them, and goes into much more detail.

Spec-wise, modern class D amps seem to do pretty well. D-Sonic's specs (they build ICE based amps,) have pretty good THD and a rated bandwidth who's top freq is well above 20 khz. ICE based amps, in other words, don't seem to suffer from high freq. response issues.

Crown's ultra-light line with class D amps AND switching power supply, has a spec'd bandwidth to 20 khz, and is .1 dB down at that point. A very good damping factor. Lowest model does 215x2 Watts into 8 ohms, with a 1 khz signal, and .5% THD. One assumes THD is even lower with less power output (most manufacturers rate their power on the part of the curve where THD is already increasing.) From these specs, it looks pretty good. Perhaps some people can tell the difference between this amp, and some comparably powered class AB amp. But as usual, I could not say why with these specs.

M Code 12-02-2010 03:06 PM

Originally Posted by jake227 View Post

I found this to be a very interesting read for a slow afternoon at work

Is there any difference in sound quality when comparing class AB and class D amps? I'm guessing this has been covered before, but as long as we are on the subject I thought I'd ask. My guess is that there is little to no difference, but I'm not 100% on that.

The latest Class D solutions are excellant, and sonically they are superb.
Yet today still are many who are sceptical of their electrical performance.

I will recap for you a personal experience for this subject..
About 3 years back, I was asked by the president of a very large US based audio brand to organize an audio comparison demo of the latest Class D amplifiers vs. high quality Class A-B amplifiers. I took 4 of the latest Class D solutions (Ice, Zetex, Intersil, TI) and a high-end Class A-B amplifier of the client's brand. I level matched each driving a pair of well-known high-end loudspeakers that sold for about $6,000 and used a low-loss speaker relay switching system.

The listening panel was made up of the brand's golden-ears, their marketing guys and president, our objective was to compare each amplifier and our basic question was Which amplifier sounded the best?
Note that the listening panel did not know which amplifier was playing, after 3hours of playing these each member of the listening panel could not tell any difference between each amplifier.

The end-conclusion was that all of the latest Class D solutions are capable of delivering a high-quality, neutral, uncolored response, equal to a comparable, quality Class A-B amplifier. There remains some biases over the smaller footprint of the Class D solutions rather than massive heat sinks. But in this face-off there was no discernable audible differences.

Just my $0.02...

MichaelJHuman 12-02-2010 03:25 PM

I really want to buy one of the Crown Drive Core series (like the XLS 1000,) and run that for my front channel amp, and use my Emotiva XPA-3 for my center, and side surround and use my Z7 for the rear surround.

And my sole reason for doing this is for the clipping indicator, to see if it ever goes on. Money and a lack of open space in my race is really all that's preventing the purchase.

MichaelJHuman 12-03-2010 09:08 AM

I rewrote some stuff for attempted clarity. I especially wanted better organization of the section on amp sound, as it's the most hotly debated topic in amplifiers. I added a section on gain and input sensitvity.

lizrussspike 12-03-2010 01:28 PM

Originally Posted by MichaelJHuman View Post

I rewrote some stuff for attempted clarity. I especially wanted better organization of the section on amp sound, as it's the most hotly debated topic in amplifiers. I added a section on gain and input sensitvity.

Thanks a ton for your time and effort. Your details answer quite a few questions, as I am on the border of looking/ purchasing my first amp. Am on the border of the Emo XPA-3 or 5. Once again I found the information very informative, and should get a lot more activity.

MichaelJHuman 12-03-2010 02:01 PM

Added section on too little vs too much power.

ninja12 12-03-2010 02:59 PM


That was a great read and another +1 for a sticky.

Call me Ugly 12-04-2010 12:41 PM

Well as a noob to the nth degree, this helped me out alot.

I do have one followup question though, for anyone willing to answer, regarding an impedance mismatch. The FAQ mentions heat from increased load, and it's easy to understand the basic ramifications of that in an electronic device, but are there any other side effects? Specifically sound quality or risk to the speakers?

I've got a pair of 6 ohm Pioneer BS41 speakers on the way and I'm probably going to drive them with an old Technics SU-v78 stereo amp rated for 8-16 ohm. The manual for the speakers recommends a 4-16 ohm amp and that has the noob in me unnerved. I've got a Phillips HTIB receiver that was made to work with 6 ohm front speakers, but "Phillips" and "HTIB" should explain why I don't want to use it unless I have to.


MichaelJHuman 12-04-2010 01:02 PM

Many receivers are not specifically designed for 4 ohm loads. That does not mean anything bad will happen if you use them with 4 ohm speakers, though. Some HTIB systems do seem to be designed for lower impedance speakers, though.

If you push the receiver too hard, it may protect itself and shut down. It's lifespan could be shortened. Or it could fail. But if you play at moderate levels there is little chance of problems. Manufacturers don't have any desire to have products come back because the owner abused them, so they put in various protection mechanisms such as thermal protection, over voltage and over current detection and circuits which limit power.

If you are not the type to really crank your volume levels up, and you give the receiver plenty of ventilation, you may be fine (just be aware of the possible issues.)

whoaru99 12-04-2010 06:28 PM

Originally Posted by MichaelJHuman View Post

There's also the amusing situation of amplifiers being rated for more power than could be pulled from the wall. Take a 400x7 Watt amplifier. That's 2800 Watts. Even a 20 amp circuit is not going to manage to supply it with enough power to meet it's specification. There's little point to an all channels driven measurement of this sort. The amp is likely limited by wall power. They may as well just list the specification as 'Puts out as much power as you wall outlet can supply minus losses.' Maybe that's overstating the situation, but not by much I suspect.

This amp really isn't, IMO, limited by wall power when playing actual movies or music. Playing test tones OTOH? OK, I'd agree.

A 20A breaker can handily support full output of this amp considering average PAR/crest factor of music and movie.

ratm 12-04-2010 06:37 PM

Great read. Another vote for sticky.

Perhaps a section on different kinds of amps? SS, Tube, etc. (unless I missed it?)

MichaelJHuman 12-04-2010 08:33 PM

Originally Posted by whoaru99 View Post

This amp really isn't, IMO, limited by wall power when playing actual movies or music. Playing test tones OTOH? OK, I'd agree.

A 20A breaker can handily support full output of this amp considering average PAR/crest factor of music and movie.

Good point, perhaps.

However, when I say it can't put out 400 watts continuous per channel, I am not sure how crest factor enters into it. By definition, average voltage will be about 57 volts. Were I to play music with a peak voltage of 57 volts, but an average voltage of far less, say 11 volts (one fifth,) the continuous wattage would no longer be 400 watts. Instantaneous draw from the wall socket could be higher than rated for short periods of time, and filter capacitors would also help. But now we are not talking about an amp with continuous 400x7 watt power, we are talking about an amp with instantaneous 400 watt power, right?

MichaelJHuman 12-04-2010 08:34 PM

Originally Posted by ratm View Post

Great read. Another vote for sticky.

Perhaps a section on different kinds of amps? SS, Tube, etc. (unless I missed it?)

Good suggestion. That will take some time to research it, and think about what best to say about it. Certainly tube amps are popular, and an important segment of the amp market.

whoaru99 12-05-2010 09:11 AM

To me, the whole thing about amplifiers being rated "continuous" power is somewhat misleading since, for the most part, that's not how they work in a music/movie system.

However, I don't want to belabor the point. In the context of continuous output, you're right, a 20A circuit wouldn't handle the example amp.

MichaelJHuman 12-05-2010 10:18 AM

Originally Posted by whoaru99 View Post

To me, the whole thing about amplifiers being rated "continuous" power is somewhat misleading since, for the most part, that's not how they work in a music/movie system.

However, I don't want to belabor the point. In the context of continuous output, you're right, a 20A circuit wouldn't handle the example amp.

A valid point, of course.

A number of articles I have read, have questioned how amplifier power should be measured to better match real world usage.

THX has some method of using burst tones which are supposed to better match the patterns of movies.

For highly compressed music, I would think the continuous sine wave thing comes closer to reality, but that would usually be two channels only (unless you are using 7-channel stereo.)

IMO, 400x7 watts is overkill for the vast majority of people. They will simply not hit it's limits, so perhaps it's (mostly?) irrelevant to talk about it's real world limits.

Thanks for bringing up the topic. Maybe I should just remove that whole paragraph

jake227 12-06-2010 08:41 AM

I don't really care how an amp's power is measured - I just wish all manufacturers were required to measure the exact same way. That way you know you are comparing apples to apples.

MichaelJHuman 12-06-2010 09:50 AM

Originally Posted by jake227 View Post

I don't really care how an amp's power is measured - I just wish all manufacturers were required to measure the exact same way. That way you know you are comparing apples to apples.

To some extent they are at least if they follow the FTC rule for amps sold in the US. But that rule does not work as well with multi-channel amplifiers.

There's some CEA burst test which might be more helpful to evaluate multi-channel amplifiers. Audioholics tested the Emotiva UPA-7 using that test.

Maybe there's no perfect test. I would not mind seeing these tests for every receiver -
* Standard FTC test with two channels driven; 20hz to 20khz at no worse than 1% THD, ideally both 1% THD and .1% THD
* A consistent all channels driven test; 1% THD showing the worst power at 20hz, 1khz and 20khz. Duration could be very short, like 10 seconds. Long enough so that the filter capacitance is irrelevant

Mike_WI 12-12-2010 01:48 PM

A great FAQ to have to "point" people to when questions come up.

Thanks for starting it.


Bidniz 12-13-2010 06:31 AM

Just wanted to say thanks for taking the time to write all that up. It was a great read for me and very interesting!

All times are GMT -7. The time now is 11:07 PM.

Powered by vBulletin® Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
User Alert System provided by Advanced User Tagging (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.