Thanks for the info Bob. I'll check the local music stores here for that QSC 340 amp and try one out. I think class D is the way to go for max efficiency and the lighter weight is a bonus!
Enjoy the benefits of a premium membership!
Upgrade Account
-
TalkBass has been independent since 1998. Add your voice.
Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
Join free Log in
Want zero display ads or expanded classifieds tools? Compare plans.
Class D popping up everywhere. Innovation?
- Thread starter Volume4
- Start date
Bob Lee (QSC)
In case you missed it, I work for QSC Audio!
- Jul 3, 2001
- 12,668
- 2,036
- 67
- Disclosures
- Former Technical Communications Developer, QSC Audio
I see, unfortunately that's a little more power than I need. Does QSC make a class D amp that puts out up to 1200 to 1600 watts max in mono/stereo mode at 4 ohms? I run 2 Ampeg 810's with a program rating each at 1600 watts at 4 ohms. As you can see I like a lot of power
This has been really good stuff. Thanks to everyone who's posted so far.
So, I'm still feeling like I might want to avoid Class D and/or Switching power supply.
I've put together a list of quotes from this thread (none from me except the first one that I posted from Wiki) that I find most.....concerning (for lack of a better word).....to me and leave me still wondering if my initial concerns are not still a little bit legit.
Here they are:
sometimes mistakenly described as "digital" because the output waveform superficially resembles a pulse-train of digital symbols, but a Class D amplifier merely converts an input waveform into a continuously pulse-width modulated (square wave) analog signal
As these designs move further away from "old school" characteristics, it will become necessary to simulate the tone of historical designs through some sort of analog or digital modeling.
In a class D amp, the pure analog signal is chopped up very, very fine into a pulse train where the width of the pulses varies, and/or the number of pulses varies to reflect an analogy of the original analog signal. This is done at typical line level (~+4dBu) or a little higher THEN this low voltage pulse train is efficiently lefel shifted up to a much higher voltage. This high voltage pulse train is run through a low pass integrating filter to reconstruct the original analog audio but at a higher voltage (and thus power).
Class D breaks apart the signal into smaller chunks than even a push pull amp and places them back together to form the output signal... The newer clases of pure digital use D/A convertors and A/D convertors to convert between
Class D can look analog compared with a true digital signal and digital compared with an purely analog signal.
Class D amps use analog AND digital principals to operate. They are both and yet they are neither...
Seriously, I think I would classify them as being discrete systems rather than analog or digital.
Newer class D amp designs use PDM instead of PWM.
So, I'm still feeling like I might want to avoid Class D and/or Switching power supply.
I've put together a list of quotes from this thread (none from me except the first one that I posted from Wiki) that I find most.....concerning (for lack of a better word).....to me and leave me still wondering if my initial concerns are not still a little bit legit.
Here they are:
sometimes mistakenly described as "digital" because the output waveform superficially resembles a pulse-train of digital symbols, but a Class D amplifier merely converts an input waveform into a continuously pulse-width modulated (square wave) analog signal
As these designs move further away from "old school" characteristics, it will become necessary to simulate the tone of historical designs through some sort of analog or digital modeling.
In a class D amp, the pure analog signal is chopped up very, very fine into a pulse train where the width of the pulses varies, and/or the number of pulses varies to reflect an analogy of the original analog signal. This is done at typical line level (~+4dBu) or a little higher THEN this low voltage pulse train is efficiently lefel shifted up to a much higher voltage. This high voltage pulse train is run through a low pass integrating filter to reconstruct the original analog audio but at a higher voltage (and thus power).
Class D breaks apart the signal into smaller chunks than even a push pull amp and places them back together to form the output signal... The newer clases of pure digital use D/A convertors and A/D convertors to convert between
Class D can look analog compared with a true digital signal and digital compared with an purely analog signal.
Class D amps use analog AND digital principals to operate. They are both and yet they are neither...
Seriously, I think I would classify them as being discrete systems rather than analog or digital.
Newer class D amp designs use PDM instead of PWM.
Volume4 let not your heart be troubled. All that you posted is nonsense as a class D amp is no different other than a more efficient power supply. The signal is analog all the way through and the power supply is not digital either. The manufacturer is more important than the method.
FAQ on PDM versus PWM
Link Removed
In ideal - or actually "real" world the same signal appears at the speaker regardless of the class of the amp.
People really should treat amp class as black box data - you don't need to know what's in the box if the resulting amp meets your needs.
Bruno Putzies puts it: "All Amplifiers are Analogue, but Some Amplifiers are More Analogue than Others"
Link Removed
In ideal - or actually "real" world the same signal appears at the speaker regardless of the class of the amp.
People really should treat amp class as black box data - you don't need to know what's in the box if the resulting amp meets your needs.
Bruno Putzies puts it: "All Amplifiers are Analogue, but Some Amplifiers are More Analogue than Others"
This has been really good stuff. Thanks to everyone who's posted so far.
So, I'm still feeling like I might want to avoid Class D and/or Switching power supply.
I've put together a list of quotes from this thread (none from me except the first one that I posted from Wiki) that I find most.....concerning (for lack of a better word).....to me and leave me still wondering if my initial concerns are not still a little bit legit.
Here they are:
sometimes mistakenly described as "digital" because the output waveform superficially resembles a pulse-train of digital symbols, but a Class D amplifier merely converts an input waveform into a continuously pulse-width modulated (square wave) analog signal
As these designs move further away from "old school" characteristics, it will become necessary to simulate the tone of historical designs through some sort of analog or digital modeling.
In a class D amp, the pure analog signal is chopped up very, very fine into a pulse train where the width of the pulses varies, and/or the number of pulses varies to reflect an analogy of the original analog signal. This is done at typical line level (~+4dBu) or a little higher THEN this low voltage pulse train is efficiently lefel shifted up to a much higher voltage. This high voltage pulse train is run through a low pass integrating filter to reconstruct the original analog audio but at a higher voltage (and thus power).
Class D breaks apart the signal into smaller chunks than even a push pull amp and places them back together to form the output signal... The newer clases of pure digital use D/A convertors and A/D convertors to convert between
Class D can look analog compared with a true digital signal and digital compared with an purely analog signal.
Class D amps use analog AND digital principals to operate. They are both and yet they are neither...
Seriously, I think I would classify them as being discrete systems rather than analog or digital.
Newer class D amp designs use PDM instead of PWM.
Why would any of these concern you? First of all, none of the above quotations are related directly to the fidelity of a Class D amplifier. In fact, many (NOT ALL!) of them are very general, not necessarily correct statements with no actual technical basis.
If those comments scare you, try learning about how AD/DA conversion, DSP, and other modern signal processing "chops up" the signals. These can achieve perfect synthesis of the original signal! For clarity: I am NOT calling Class D digital or likening it to any digital signal processing. I am drawing a related analogy to demonstrate that processing of a signal does not mean destruction of said signal when it is done correctly.
I will put forth a suggestion that Bob Lee (QSC) has offered before to help you clear your mind: Try a double-blind test comparing an amplifier with a Switch Mode Power Supply and Class D output section to an amplifier with a traditional power supply and Class A/B output section. With properly matched gain (arguably the most important part), you will not be able to hear a difference. Unless you are willing to study the actual technology, this is the only other way to clarify the matter.
You should be more clear in separating the POWER SUPPLY from the operating mode of the OUTPUT section. Without that clarity, your above comment is going to be criticized in the same way it was before. It is not correct as you wrote it!
Other than that, I agree with you. The quality of the amplifier, along with its suitability for the job, are much more important factors that the operating mode of the output section.
You keep coming back to this and it's still wrong.
The class of an amplifier has nothing to do with the type of power supply. A switch mode power supply and a class D amp are TWO DIFFERENT THINGS.
Class D amps with conventional power supplies exist.
Class AB amps with switching power supplies exist.
A class D amp is a different type of circuit than a class AB amp, even if they are using the exact same power supply.
As a pro electrical engineer let me see if I can help here.
Both types (linear and switching) of power amps use transistors (or tubes for old-school linear) to deliver the rated power (watts RMS) to the speaker load (OHMS).
It's the "mode" that differs. We speak of linear or switch (class d) mode.
In linear mode the three terminal transitor device accepts a small voltage on one leg and can provide a replicated but larger signal on a different leg that pulls current from the power supply to move the speakers. When the input signal is small the device itself "absorbs" (we say dissapates) lots of the power that could be going to the speaker.
That's where the inefficiency and heat sinking and matching parallel device all come in.
Class A,B,C,AB linears are trade-offs, usually distortion vs efficiency.
In switch mode the device is either ON hard (what we call saturation) or OFF hard to switch current to the speaker. A sample of the output is feedback to control circuits so that the pulse widths are adjusted to replicate the input. Ideally the device would switch in zero time, however it does not..this switching action dissapates power in the device. The output is filtered to smooth the switching humps.
Like amplifiers- the power supply that feeds the power devices (and pre-amps circuits) may be a switch mode or linear design, just that they regulate to a DC reference instead of an audio input.
In each case the switch mode provides better efficiency and thus less massive components.
There are other solutions where the device supply is modulated to match the input, so that it is a very efficient linear design....
Both types (linear and switching) of power amps use transistors (or tubes for old-school linear) to deliver the rated power (watts RMS) to the speaker load (OHMS).
It's the "mode" that differs. We speak of linear or switch (class d) mode.
In linear mode the three terminal transitor device accepts a small voltage on one leg and can provide a replicated but larger signal on a different leg that pulls current from the power supply to move the speakers. When the input signal is small the device itself "absorbs" (we say dissapates) lots of the power that could be going to the speaker.
That's where the inefficiency and heat sinking and matching parallel device all come in.
Class A,B,C,AB linears are trade-offs, usually distortion vs efficiency.
In switch mode the device is either ON hard (what we call saturation) or OFF hard to switch current to the speaker. A sample of the output is feedback to control circuits so that the pulse widths are adjusted to replicate the input. Ideally the device would switch in zero time, however it does not..this switching action dissapates power in the device. The output is filtered to smooth the switching humps.
Like amplifiers- the power supply that feeds the power devices (and pre-amps circuits) may be a switch mode or linear design, just that they regulate to a DC reference instead of an audio input.
In each case the switch mode provides better efficiency and thus less massive components.
There are other solutions where the device supply is modulated to match the input, so that it is a very efficient linear design....
- Dec 18, 2000
- 13,201
- 340
- Disclosures
- greenboy designs: fEARful, bassic, dually, crazy88 etc
So, I'm still feeling like I might want to avoid Class D and/or Switching power supply. I've put together a list of quotes from this thread (none from me except the first one that I posted from Wiki) that I find most.....concerning (for lack of a better word).....to me and leave me still wondering if my initial concerns are not still a little bit legit.
Here they are:
{...}As these designs move further away from "old school" characteristics, it will become necessary to simulate the tone of historical designs through some sort of analog or digital modeling. {...}
I found that particular one especially amusing. I would have at least ten years ago, too - when I bought probably my third or fourth Class D amp after quite a few years using earlier examples. This Class D stuff is practically old hat and "tone" is not the function of an amplifier from the "pro sound sector" anyway: it's merely to amplify faithfully whatever signal is presented at input. Coloration, if desired, can be achieved any number of ways.
Why avoid a switch mode power supply? Every computer you use has one. Most newer wall watt supplies for cell phones and a myriad of other household devices are switch mode units. The technology is mature and highly reliable. How many times have you had to replace the power supply in your computer?? I have only ever had to replace one and that was many years ago.
As for the various classes of amps, it really is immaterial. Others have posted it's how your particular amplifier SOUNDS not what class the output stage operate within. PLAY through the amplifier and if it makes you happy go with it.
I would also like to compliment Bob Lee on his patience in this thread. I would not have been able to be so.
A happy new year to each and all.
Paul
I enjoyed this post quite a bit - sometimes it's hard to tailor detail to a wide audience, but you nailed it. This post deserves to be seen twice.
bongomania
Commercial User
Since "Bellair Audio" sounds like a business, I thought Googling it might turn up some references or credentials. That is all.
- Feb 12, 2006
- 61,051
- 5
- 221,019
- Disclosures
- Development Engineer-Mesa Boogie, Development Engineer-Genzler (pedals), Product Support-Genz Benz
It's a combination of a lot of things actually...
Size, weight, thermal efficiency, power supply efficiency, the ability to recycle power that would be lost in a linear supply due to back EMF in the speaker (potential energy upon change of direction), the more forgiving (typically) nature of driving a lagging reactive load, short term current delivery for sub-nominal impedance loads are some of the potential benefits.
- Feb 12, 2006
- 61,051
- 5
- 221,019
- Disclosures
- Development Engineer-Mesa Boogie, Development Engineer-Genzler (pedals), Product Support-Genz Benz
As someone who has designed amplifiers for almost 30 years, let me assure you that you have this so utterly and completely mixed up that this would probably be a good time to take a break from professing knowledge and really learn something about the topic. First start with simple analog class AB amplifiers and understand how those work first. THEN look at a class D amp and it should become instantly clear.
There is absolutely nothing in common between a class AB/G/H power amplifier and a class D power amplifier other than the analog input and the analog output. Absolutely nothing. It also has absolutely nothing to do with the power supply.
I don't see any reason to avoid Class D, SMPS, or any particular technology............ go for the best combination of sound, practicality, and affordable cost.
Class-D audio was effectively started by John Ulrick and an associate who formed Infinity Systems in the 1970s...... They used bipolar transistors, and after some troubles, made a decent product. He still is around, in a company called Spectron, doing high end audio amplifiers, class-D of course.
It's been getting better ever since. Reliability is always lower with more parts, and class-D or SMPS has more parts. But their basic reliability is high, and in some cases can be higher than alternate designs, due to better fault tolerance. In many ways it is easier to get low distortion in class-D than in analog designs....
Ampeg has done SMPS, Class-D etc, and apparently still is using those technologies in some products. In the class-D units, I don't recall a single field failure of the class-D amplifiers, in any of the 5 different unit series that had them during the SLM time.
As for Bellair Audio*..... most everything he has said is wrong........
I would, however, comment that the tube harmonic issue is not as simple as some engineering types have posted.......
A push-pull tube amplifier normally SUPPRESSES even harmonics, especially low order ones, if the tubes are well matched. harmonic performance when in clipping varies, depending on circuit details.
Virtually NO tube amplifier (there are exceptions) will produce the obnoxious very high order harmonics that the typical clipping SS amp does. This is because most tube amplifiers do not have the feedback and "open loop gain" to produce a very 'hard' square wave output in reasonable clipping conditions.
The harmonic order is the very first issue in "musicality"..... low order are good, high order are bad, sounding like an overlay of white noise or "marbles rattling in a can", especially in situations where distortion is normal, as with guitar amps. So to a certain extent, tubes tend to produce a low order, and hence musical, distortion.
But the even vs odd , at least in the preamp, is very much an issue of the circuit design, the individual tube, and the power supply.... The more symmetrical the clipping, the more odd-order tends to be present in most cases. Most tube amps will actually produce different order harmonics depending on the length of time that clipping continues, and how 'heavy" the clipping is.
I and another SLM engineer have a patent (I think it is now expired) on a technique for producing time-varying harmonics, starting as even order and shifting to odd as clipping continues. This simulates the effect of the coupling capacitors in a typical tube amp as the bias on each stage shifts during clipping. It was pretty successful and was used in a lot of Crate and Ampeg guitar amps over the years.
I see no reason at all that an amplifier with class-D, DSP, and /or SMPS cannot sound as good as a classic tube amplifier. There may need to be some adaptations to prevent bad effects, but that can be done.
Paradoxically, I do NOT think that the most 'linear" amplifiers are the best for musical instrument applications........ to a certain extent that is the 'kiss of death" if done wrong. I suspect that Agedhorse knows very well what I mean...... being that he is with Genzbenz.....
Make your choice of amp on the basis of sound and general practicality, (price will be in there too) and you won't go too far wrong. Actually, DON'T go by "sound", go by how easy it makes your playing..... a good sounding amp is one that makes YOU sound good, and you sound best when you can get what you want easily and without having to work hard to get it.
* There seems to be an AV rental place in Florida under that name.........
Class-D audio was effectively started by John Ulrick and an associate who formed Infinity Systems in the 1970s...... They used bipolar transistors, and after some troubles, made a decent product. He still is around, in a company called Spectron, doing high end audio amplifiers, class-D of course.
It's been getting better ever since. Reliability is always lower with more parts, and class-D or SMPS has more parts. But their basic reliability is high, and in some cases can be higher than alternate designs, due to better fault tolerance. In many ways it is easier to get low distortion in class-D than in analog designs....
Ampeg has done SMPS, Class-D etc, and apparently still is using those technologies in some products. In the class-D units, I don't recall a single field failure of the class-D amplifiers, in any of the 5 different unit series that had them during the SLM time.
As for Bellair Audio*..... most everything he has said is wrong........
I would, however, comment that the tube harmonic issue is not as simple as some engineering types have posted.......
A push-pull tube amplifier normally SUPPRESSES even harmonics, especially low order ones, if the tubes are well matched. harmonic performance when in clipping varies, depending on circuit details.
Virtually NO tube amplifier (there are exceptions) will produce the obnoxious very high order harmonics that the typical clipping SS amp does. This is because most tube amplifiers do not have the feedback and "open loop gain" to produce a very 'hard' square wave output in reasonable clipping conditions.
The harmonic order is the very first issue in "musicality"..... low order are good, high order are bad, sounding like an overlay of white noise or "marbles rattling in a can", especially in situations where distortion is normal, as with guitar amps. So to a certain extent, tubes tend to produce a low order, and hence musical, distortion.
But the even vs odd , at least in the preamp, is very much an issue of the circuit design, the individual tube, and the power supply.... The more symmetrical the clipping, the more odd-order tends to be present in most cases. Most tube amps will actually produce different order harmonics depending on the length of time that clipping continues, and how 'heavy" the clipping is.
I and another SLM engineer have a patent (I think it is now expired) on a technique for producing time-varying harmonics, starting as even order and shifting to odd as clipping continues. This simulates the effect of the coupling capacitors in a typical tube amp as the bias on each stage shifts during clipping. It was pretty successful and was used in a lot of Crate and Ampeg guitar amps over the years.
I see no reason at all that an amplifier with class-D, DSP, and /or SMPS cannot sound as good as a classic tube amplifier. There may need to be some adaptations to prevent bad effects, but that can be done.
Paradoxically, I do NOT think that the most 'linear" amplifiers are the best for musical instrument applications........ to a certain extent that is the 'kiss of death" if done wrong. I suspect that Agedhorse knows very well what I mean...... being that he is with Genzbenz.....
Make your choice of amp on the basis of sound and general practicality, (price will be in there too) and you won't go too far wrong. Actually, DON'T go by "sound", go by how easy it makes your playing..... a good sounding amp is one that makes YOU sound good, and you sound best when you can get what you want easily and without having to work hard to get it.
* There seems to be an AV rental place in Florida under that name.........
- Feb 12, 2006
- 61,051
- 5
- 221,019
- Disclosures
- Development Engineer-Mesa Boogie, Development Engineer-Genzler (pedals), Product Support-Genz Benz
Very good point here Jerrold, the common approach to solid state amps is to go for high open loop gain (because gain is essentially "free") and then use big negative feedback to reduce distortion, lower output impedance, increase linearity etc. In big part, this is for specmanship too. Reminds me of the early Mackie FR1200's, a very well designed amp with really impressive specs (actually measured) but in the real world, they did not hold up all that well in execution and had a difficult time with unusual loads and clipping IIRC. It was quickly replaced with the FR1400 which seems to have done better in this regard.
The problem when the amp clips is that the amp's negative feedback is effectively disconnected from the amplifier and now it's an open loop amp with stupid high gain and looks like a "power comparator" slamming rail to rail and maybe sticking a little bit for added "yuck" flavor. Add to this a reactive load and recovery takes some time, then the signal hits the other rail, and so on. Preventing clipping by various limiting and gain management schemes is helpful here as well as designing an amp that is inherently free from uglyness when clipping does indeed occur.
I have always evaluated the tradeoffs of high gain/feedback with a more gentle, conservative approach and every time I come back to lower gain/feedback designs, along with gain management with creative limiting approaches because their overload behaviour is so much more pleasant (in general).
For Bellair's benefit, this discussion applies only to linear amps, NOT class D amps. There is a difference
^^ Thanks for the many rational and informative posts from both of you's guys.
I've hung a scope on the output of the amps that I've laid my hands on, including two commercial Class-D and a DIY Class-D that I built from a kit. I've always been interested in observing the clipping behavior. The only drawback of my setup is that I haven't ventured into the territory of driving a reactive load balls-to-the-wall because my reactive loads (speakers) all have limited amounts of magic smoke available, not to mention my eardrums. Looking at the waveform before the output filter inductor is eye-opening.
Interestingly, one Class-D amp had a very soft clipping response. That was the Crate Power Block. I didn't know what to make of it, but I suppose for a guitar amp, it might not have been a shortcoming. The Crate used what looked like a Philips chip, or a licensee thereof, based on the pad layout. It was a nifty concept, and I was hoping for a bass version with a bit more oomph.
Two Class-D amps (the Crate and my DIY kit using a Tripath chip) could be driven into protect mode by a heavy attack on low E while driving a speaker with a 5-Ohm DC resistance, despite being within the nominal output power rating. I've never modeled what was actually happening. The built-in limiter of my other Class-D amp (EA Micro300) seems to be unconditionally robust against the same abuse.
One Class-AB amp (names changed to protect the innocent) had an obvious overload recovery glitch. The only redeeming feature was that the speaker was probably not capable of reproducing the higher order harmonics. Another one seemed to have just a bit of extra circuitry to ensure soft clipping of the power amp, possibly by applying a nonlinear function before the power amp.
My DIY power amp built from two National Semiconductor LM12 power op amp chips in bridge mode, was indestructible.
Of course the above studies are strictly for my own curiosity, since I have seldom needed to push my amps to the wall when on the bandstand.
I've hung a scope on the output of the amps that I've laid my hands on, including two commercial Class-D and a DIY Class-D that I built from a kit. I've always been interested in observing the clipping behavior. The only drawback of my setup is that I haven't ventured into the territory of driving a reactive load balls-to-the-wall because my reactive loads (speakers) all have limited amounts of magic smoke available, not to mention my eardrums. Looking at the waveform before the output filter inductor is eye-opening.
Interestingly, one Class-D amp had a very soft clipping response. That was the Crate Power Block. I didn't know what to make of it, but I suppose for a guitar amp, it might not have been a shortcoming. The Crate used what looked like a Philips chip, or a licensee thereof, based on the pad layout. It was a nifty concept, and I was hoping for a bass version with a bit more oomph.
Two Class-D amps (the Crate and my DIY kit using a Tripath chip) could be driven into protect mode by a heavy attack on low E while driving a speaker with a 5-Ohm DC resistance, despite being within the nominal output power rating. I've never modeled what was actually happening. The built-in limiter of my other Class-D amp (EA Micro300) seems to be unconditionally robust against the same abuse.
One Class-AB amp (names changed to protect the innocent) had an obvious overload recovery glitch. The only redeeming feature was that the speaker was probably not capable of reproducing the higher order harmonics.
Another one seemed to have just a bit of extra circuitry to ensure soft clipping of the power amp, possibly by applying a nonlinear function before the power amp.My DIY power amp built from two National Semiconductor LM12 power op amp chips in bridge mode, was indestructible.
Of course the above studies are strictly for my own curiosity, since I have seldom needed to push my amps to the wall when on the bandstand.
Cripes! Where am I? A guitar forum? Nobody here seems to have a clue about bass amps. Allow me to provide some enlightenment.
1. Power supplies. Power for your amp comes from the wall. It is 50 or 60 cycle AC at no more than a couple hundred volts. In the USA it will be 120 volts (typically for small installations) and in the UK it may be higher.
2. Converting line voltage to something useful. To power your speakers you need an electrical voltage (and current) that tracks the output of your bass. Active devices provide that output and usually are tubes or solid state devices of various kinds. They must be powered by a quiet DC source or the 50 or 60 cycle "note" will be heard in your speaker. Thus every amp needs to convert the AC power to a smooth DC voltage. Furthermore, tubes take voltages much higher than line voltage and semiconductors can take voltage lower than line voltages. So a voltage changing action is needed too.
3. The original power supplies of yesteryear do this by stepping voltages up or down with a transformer and then using very large capacitors and often large inductors to filter the "hum" out of the power going to the final amp stages. Needless to say low frequency transformers are very large and heavy. Big chokes are heavy too. This makes "old school amps" weigh a ton just from the power part of it. The transformer also isolates the output power from the line as a safety feature.
4. Switching supplies. So how do we fix that weight problem? The secret is to just feed the AC power through a rectifier into a large cap to get DC. But that just gives us one DC voltage and no isolation. The trick is to take that DC and then chop it to make it a higher frequency than 60 Hz. Back in WWII aircraft used 400 hz instead of 60 hz because the higher frequency made transformers much lighter. Well with todays materials you can go to extremely high frequencies resulting in very tiny transformers that are mostly heavy wire. So the transformer for isolation is still there it is just very small and light (often a toroid). And even better the filters after the transformer are also quite small and light because of the high frequency. Hence by changing the power supply circuit we've cut weight by a huge amount due to the power transformer.
5. Switching supply problems. Well, part of the problem in going with a switching supply is that the higher frequencies while easier to filter tend to get into your audio. So careful shielding is a must. Also in the early days semiconductors tended to not have very high ratings and tended to blow out with some regularity. One of the finest bench fires I ever saw occurred when a friend of mine was showing me how his switching power supply was no longer blowing up when you shorted the output. Happily, modern components have all but eliminated these early problems. Note the power supply in your computer is a switcher. They are quite reliable today and if you tear one apart you'll see some very small light weight components that crank out maybe several hundred watts of power.
6. Given this situation the wonder is why most amps don't use switching power supplies. And in fact why all TUBE amps don't. Half of that weight that is due to the old style 60 Hz power transformer would be gone. And as an added plus you could power the tube heaters with DC as well to reduce hum.
Thus, we conclude that MODERN switching power supplies are half of the answer to reducing amp weight.
1. Amp classes. A class A amp is one where the signal appears whole at all points in the circuit. It is the typical thing for preamps and low level amplification. It's only when you want some power some tricks start to be used. Class B means that one device handles the positive going peaks of a signal while another device handles the negative going peaks. This is more efficient and allows more power. The problem is that "kinks" can occur in the signal near zero where one switches from one device to the other. This makes a lot of distortion at low signal levels. Class C is used only for tuned radio circuits and requires a tuned circuit to act as a "flywheel" it is not used for audio, though is very efficient.
2. Now comes the crux. Class D. What is it? It's NOT a "power supply", though it is related to switching power supplies. Now the idea of an amp is to take a fixed voltage and make it go up and down proportionally to some input signal. We can do that with a tube or transistor in series. But dig. Say our output voltage to speaker is at one half the supply voltage. To do that half the power supply voltage will be across the speaker (which you want) and the other half will be across the controlling device which heats it up (which you don't want as it's wasted power and heat to get rid of) That is the problem.
3. Now say instead of just turning our semiconductor half on and making it hot say we make a rule that says it is either all on or all off. Note that when it's off voltage across it is full but current is zero (Power=IV=0xV=0) and when it's on, current through it maximum but the voltage across the device is nearly zero as it's full on. (Power=IV=Ix0=0). The only time there is any power seen at all is when the device is switching from one state to another. If we switch very fast, power seen will be next to nil!
4. OK. Now how do we use that to get our "half voltage"? What if I apply a square wave to the chopper that makes the output half on and half off. On average the output will be half the power supply voltage. If I make the chopping frequency very high, I not only will not hear the pulses, but also I can filter them with smallish components to produce a nice clean voltage without any trace of chopping! That is basically what class D does.
5. Now there are some problems with this scheme. One is the question can you really filter all traces of the chopping out? Another is that to produce very low voltages you have to ask how narrow can your pulses be made. Device switching speed limits that and means that the amp will tend to not reproduce low levels well (much like the class B problem). Also the maximum output would be when all switches are continuously "on". That represents a "hard limit" for the device.
6 I don't want to get really technical here, but a lot of these problems have kept class D amps from being widely used. However, faster and faster switching devices have greatly improved performance. I own a Crate "power block" 100 watt Class D guitar amp. (now discontinued for some reason). It's a nice little device but you can hear the class D artifacts when you play through it. Yes, they are low level but they are THERE.
7. Enter some "improvements". Some bright engineers came up with a way to much reduce the switching artifacts and improve the accuracy of the output by using some new techniques and filters. There is some Euro outfit selling OEM amp boards using this methodology. For example the class D Galien-Krueger amps seem to use these parts. I've carefully listened to the G-K MB2-500 and I hear zero artifacts! The Markbass class D is likewise excellent. In addition these amps seem to have also added some hard clipping control to prevent the nasty, damaging sounds of running such a rig to the max. Combine the Class D with a Switching power supply and you have a 4 pound 500 watt amp. Too cool for words.
8 The bottom line is that some of these very light weight amps sound wonderful, and produce great sound. But do not forget that the kind of circuits we are talking about are MUCH more complex than old school supplies. And there is the long history of smoke, heat, and melted parts! There is some evidence that on occasion such things still DO occur with even these advanced circuits. But the small size, low heat, low weight, and high power features of these devices really seem to make them the future of bass amps. The real trick is building in the modeling and filters and other stuff that make them emulate the big heavy huge old amps that we all know and love. I really don't think that success in that effort is very far away if not already here.
9. A final word about the word "digital" here. Lets take that to mean that a signal is measured and turned into numbers that represent various points along the signal. Those numbers can at a later time be reconstructed back into an analog signal. That would be a "digital" system. A CD is a "digital" system. Class D amps may or may not be "digital". It depends upon just HOW the signals are created that switch the totally "on" and totally "off" drivers. There are analog circuits that do that and one can also do it digitally. But the method of creating a driving signal has nothing to do with the "class" of the output drivers which in the case of full switching is class D.
OK?
