Digital Amplifiers?

[merlin3791]

I guess all theses company's claiming they use digital amps need to rewrite their descriptions, heres one on the RSW-15 straight from the Klipsch site:

Powered by a high-output BASH® "digital" high-efficiency hybrid amplifier, the drivers deliver earthshaking sound without detail-robbing distortion. The RSW-15 is available in only a 120 volt version and comes in furniture-grade blond maple, medium cherry or jet ash wood veneers

DJK maybe you should let Klipsch know there is no such thing as a digital amp yet, so they can correct this description.

[merlin3791]

Heres a car amp on the crutchfield website that claims to be digital? Its the BLAUPUNKT PA2100 check out this excerpt from their description.

Tripath Class "T" Amplifier Design: Class "T" digital amplifier design combines the high fidelity sound reproduction of class "AB" amplifiers with the low heat dissipation benefits typically found in class "D" amplifiers. Although class "D" amplifiers generate little heat, they are almost exclusively used as subwoofer amps, because they are not known for high fidelity, full range sound reproduction. Class "AB" amplifiers generate much more heat than class "T" and class "D" amplifiers, so large heat sinks are needed for cooling. The class "T" design allows this amplifier to be made smaller than class "AB" amplifiers of a similar power rating, without the sound quality sacrifice that comes with a class "D" design

If these arent really digital Im thinking I can probably sue Klipsch and Blaupunkt for false advertisement! DJK would you be willing to come to court as my resident expert on AMPS?

[DrWho]

here's a good article that should clear things up if y'all bother to read it

http://www.prosoundweb.com/live/articles/biner/digitalamps.shtml

and allow me to take out a few quotes for those too lazy (like myself)

===================================================

"Historically, power amplifiers have been of so-called Class AB design. But there is an alternative design that uses digital techniques called among other things, Class D, PWM or just plain digital amplification."

"Digital amplifiers (also known also as class D, or PWM) use the MOSFET as a switch: its either in the ON state or the OFF state, and is only in the active state for a very brief period of time while it is switching between ON and OFF.

When it is ON, the voltage across the MOSFET output is zero. Like a switch that is closed, it looks like a piece of wire and dissipates very little heat. When it is OFF, the current through it is zero (like an open switch), also a low-power-dissipation state.

So the key point here is that if we only put the MOSFET in the ON or OFF states, it is most of the time in a state that dissipates no power."

"Since the voltage and current on the MOSFETs is near zero most of the time, theoretically they can dissipate zero power, and the efficiency of the amplifier will approach 100%. Practically, no design is perfect and overall efficiencies in the low 90%s are achievable in practice."

"Digital amplifiers use something called pulse-width modulation."

"Advantages of PWM amplification:

Less power dissipated by the MOSFETs means smaller heat sinks and power supplies which means smaller, cheaper, cooler, less weight.

Most of the amp circuitry is now digital. Digital electronics have huge advantages like repeatability, immunity to noise, and reprogrammability. You can also include DSP functions like effects, crossovers, equalization, and limiting in the control chip cheaply.

The input can be digital, not analog, so we can save a stage of data conversion. Your whole signal chain can be more completely digital.

Disadvantages of PWM amplification:

So everybody loves and uses PWM amplification, holding hands with their PWM amplifier and singing Kumbayah, right? Wrong: there are drawbacks:

Noise: The high-frequency switching events make the circuit a potential radiator of noise. This noise gets into things like sensitive radio and TV tuner stages, power supplies of other circuits and it radiates from the speaker cable.

Distortion: Historically, it has been harder to equal the distortion performance of Class AB amps, which perform well. This has placed a lot of digital amps in subwoofer applications, where distortion is not so noticeable.

Expense: The components required to build a digital amp are generally more expensive than Class AB."

===================================================

There's another article that addresses why "digital" amps are misnamed and all that, but i can't seem to find it...it is on the prosoundweb website though. (they recently changed the layout and i dunno where to find it anymore) I did want to point out that what we call "digital" amps are indeed entirely analog (despite using DSP technology in other parts of the circuit...the actual amplifying of the signal is analog). the PWM, aka turning the MOSFET on and off is still an analog process, even though it's easier to think about it as being "digital" because of all the ons and offs (aka 1s and 0s).

The fact that the entire industry calls it a digital amp holds as much ground as when everyone thought the world was flat.

[ZAKO]

DrWho ....I agree these PWM amps are not true digital.

[Ray Garrison]

Can someone describe, in terms a layperson with some understanding of audio electronics can understand, what a hypothetical truly "digital" amplifier would look like? How would it amplify a signal?

[merlin3791]

its kinda funny where this thread is going. Started out asking about the sound quality of a so called digital amp, not whether or not a digital amp is truly digital. Or how a digital amp functions. Anyhow I'd like to hear your thoughts on DIGITAL Cable! LOL

[ZAKO]

digital cable.. is that the same as the one hanging off my Direct TV antenna????

[jnorv]

I have a pair of TriPath 2020 eval boards running my 828/1005s in biamped mode. I have not made the suggested mods to the input/output caps. For $100 each, they are hard to beat. For casual listening they are hard to beat. Idle draw is just about zero amps. 10 watts on the bass is a little light for 100Db+ listening but for back ground listening they are just fine. When I get serious I fire off my Rowland 5 (150 watts) and 8 watt SET.

Jim N

[DrWho]

----------------

On 2/12/2004 12:57:23 PM Ray Garrison wrote:

Can someone describe, in terms a layperson with some understanding of audio electronics can understand, what a hypothetical truly "digital" amplifier would look like? How would it amplify a signal?

----------------

a "hypothetical truly digital amp" does exist...one common place you might see this is with computers. the cable that we all connect our monitors to our computers is carrying digital signals between the two. there are instances when one might want to have the monitor in another room (aka "far away"). There are all sorts of cool little boxes that you can buy that will amplify this digital signal so that it can travel the farther difference. unlike an analog-signal-carrying-cable (we'll call it an "analog-cable"), a "digital cable" (a digital-signal-carrying-cable) suffers greatly from the slightest changes in impedance (and a longer cable changes the impedance). anyways, that's a totally different subject. the point here is that this little box acts is a digital amplifier, making the computer monitor signal "big enough" to travel down the cable.

the problem with having a hypothetically truly "digital amplifier" in the audio realm is that the mechanism that allows our speakers to function requires an analog waveform. if you'd like to hear what a "hypothetical truly digital amp" sounds like...connect the digital out from a dvd player into an analog in on your amp. you will hear some very pretty white noise/static (this isn't guaranteed to work as some components output differently). granted this would be a digital signal being amplified by an analog speaker, but it's still the same thing you would hear. either that or you will hear lots of pops and clicks.

all that to say, if you want a true digital amplifier, they do exist. but it would require a new kind of speaker (one that turns on and off into the super complex waveform that music entails). this is a process that i would think very impossible, let alone at all feasible.

anyways, gotsa go and sorry for the unorganized post...i hope it made sense.

as far as current "digital amps," i've read lots of good things about them, but the technology is far from developed. I would have to do some intense listening and comparing before i would consider buying one. I do have to say that I have never listened to one though, so my thoughts are purely speculation as to how they sound (which was the initial intent of this thread)

[djk]

"DJK maybe you should let Klipsch know there is no such thing as a digital amp yet, so they can correct this description."

They don't care.

Advertising people are allowed to change the laws of physics, its part of their job description.

The BASH amplifier is just like the Sunfire sub amplifier. It is an analog amplifier that is cascoded by a switching power supply.

This is the heart of the original 'magnetic field ' amplifier from Bob Carver. He could not make it work on a production basis (this was 25 years ago) because the power FETs were not really 'there' yet.

"If these arent really digital Im thinking I can probably sue Klipsch and Blaupunkt for false advertisement! DJK would you be willing to come to court as my resident expert on AMPS?"

Can you prove a loss? Damages are bassed on losses.

In terms of what is truly a digital amp, the Tact would be one if you could find the outputs that could switch at 2.8Mhz, the frequency required by the multi-bit to 1-bit conversion for PWM amplification.

As it is, everything stays in the digtal domain, so it is truly digital (but then, so is your telephone with its 10-bit 300hz~3Khz response).

The question I have is what happened to all the data?

The final clock frequency is only 1/8th the data rate, 7/8ths of the data has to be 'lost' (reduced).

Maybe they have a good way of doing this, so far all I have been able to find is a bunch of 'handwaving', no explanation.

Think about all the 8-bit length words you can construct, 00000000, 01010101, 11111111, etc.

You have to reduce these to only 1-bit, 0, or 1.

00000000 is easy, it equals 0.

11111111 is easy, it equals 1.

How do you represent the other 2.5 million that have to be lost?

In analog; interpolation is easy and accurate, in digital; the equivalent is decimation (to kill every 10th man, or to destroy a large part of...).

The $$$ Tact is software driven, it can be updated as the decimation algorithms improve.

Texas Instrument (TI) is selling cheap chip sets with EquiBit technology, no provision for updates here. A 100W X 6 evaluation board is available for about $500 for those that would like to try it without laying out $5K for the 150W X 2 TacT.

[minn_male42]

djk,

everything that you have stated can also be said of the cd.....

maybe cd's aren't digital either..... just switching devices.... 1 or 0....

[michael hurd]

The way I understand, a cd is a just a digital representation of an analog waveform. The output from a cd player can be in the analog or digital form. It must pass through a dac (digital to analog convertor) and then a low pass filter to remove noise, and is then finally a varying amplitude and voltage to drive the inputs of the next device in the audio chain, be it a preamp or amplifier. A digital output either uses an optical cable or a coaxial cable to output a PCM signal to an outboard device with a dac.

[Marvel]

dennis,

Stewart Audio has never claimed their amps to be digital, just the high speed switch mode design (class H). They don't get much press either, but seems that having a 120 watt plus per channel amp with no fans and weighing in under 10 pounds could be a real plus. Do they sound awful or are they 'okay'?

Also, our school is getting some crestron automation equipment, which has a 30 wpc amp at under 3 pounds. I figure it is also a switch mode amp as well, but they don't say a whole lot on their web site.

Marvel

[michael hurd]

I think a class h is just a class ab amplifer that kept the rail voltage just slightly above what is required by the output transistors - like the sunfire amps, where the rail voltage is determined by the input signal and the voltage across the transistors is kept as small as possible. ( somebody correct me if I am wrong )

[djk]

"everything that you have stated can also be said of the cd.....

maybe cd's aren't digital either..... just switching devices.... 1 or 0...."

I'm sorry you don't understand the issues.

I said in advance that it was technical, some are just not going to get it.

Having said that, I should also point out that I have spent about 15~20 hours over the last three days, and downloaded over 40 pages of technical data on the subject of EquiBit.

We have a MMIC lab at work, and also build PHEMTs (Pseudomorphic High Electron Mobility Transistor) up to 40W (that would be good enough for 800W in class D). The PHEMTs are in short supply right now, but I could probably use rejects (if we are not salvaging the header and case). I'll have to see if they will go slow enough (2.8Mhz) to be of use, we normlly run them in the L-band (1.2Ghz).

I might be able to put something together that would run straight off the Delta-Sigma bit-stream.

"Stewart Audio has never claimed their amps to be digital, just the high speed switch mode design (class H). They don't get much press either, but seems that having a 120 watt plus per channel amp with no fans and weighing in under 10 pounds could be a real plus. Do they sound awful or are they 'okay'?"

In general I would have to think they sound OK, but the only one I have heard was driving a subwoofer and didn't cut it (compared to a 'normal' amplifier).

"Also, our school is getting some crestron automation equipment, which has a 30 wpc amp at under 3 pounds. I figure it is also a switch mode amp as well, but they don't say a whole lot on their web site."

Not seen this amp at all.

One of the all time space/weight/power champs has to be the Carver M1.5T; only 16lbs for 1200W, 60hz supply, no fan. The PM2.0 with a switching supply only lost 5 lbs for the same power out (11lbs for 1200W).

[chickey]

Michael,

You are absolutely correct. The Carver gear produced in the 90's, as well as the current Sunfire designs, are all class H amps. There's actually some interesting history regarding the evolution of Carver's earlier topologies (the "infamous" magnetic field amps of the 80's)to today's models as embodied in the Sunfire product line.

Jim Croft, one of Carver's lead engineers, led a design team in the early 90's which radically altered the engineering topology of Carver amps. Most of the breakthroughs came as the team engineered the legendary Lightstar amp. The Lightstar was based on the principle of a tracking power supply, which provided high current and could drive almost any load. Carver had tried to build such an amp in the 70's and 80's, but eventually gave up and opted for the magnetic field design instead. The result of Croft's work was the Carver Lightstar, which every reviewer who auditioned it remarked at how absolutely neutral it was. The problem with the Lightstar was that it had to be built by hand, and employed some very sophisticated circuitry in order to achieve a signal tracking voltage that was infinitely variable. The result carried a price tag of $4,000, which took it out of mass market reach.

Croft and his team took the learnings from the Lightstar and created a series of Carver amps that utilized a similar topology to the Lightstar, but with design tradeoffs to cut cost. The biggest tradeoff was the application of three discrete rail voltages as opposed to an infinitely variable implementation. The result, as many reviewers observed, was an amp that was very smooth, could drive difficult loads, and sounded remarkably like the Lightstar. The TFM 35x and 55x amps are based on this design, as are the later A-series amps.

By this point, Bob Carver had left Carver and founded Sunfire (actually, the board booted old Bob and then proceeded to run the company into bankruptcy). Bob took the original Lightstar design and created Sunfire. The Sunfire amps also feature tracking power supplies. As a matter of fact, the tracking power supply is at the core of Sunfire's subwoofers as well...they are the reason why such prodigious power can emanate from such a small footprint. Bob stated that he "voiced" the Sunfire amps on the side of being slightly warmer than the Lightstar (I believe through the insertion of 1 ohm resistors at the output stages). The Lightstar, on the other hand, has been touted as dead on neutral (nothing added, nothing subtracted). Tube-o-philes who have heard it are amazed at how...shudder to saw..."tubelike" it sounds.

If you inspect my signature, you'll see that I own both the Carver Lighstar and the 35x's. The Lightstar powers my mains, and provides a superb 2 channel listening experience. However, it would have been financially and physically impractical to stack these 89 pound beasts to achieve 6 channels of amplication. As a result, I use Carver TFM 35x's for center and surround duty. They match up well with the Lightstar from a sonic perspective, and more importantly, sound excellent with my Klipsch home theater. If anything, the 35x's tend to gently roll-off the high end, and have a bit of a smooth and laid back tendency. Their sonic signature seems to match up well with horns that can be oh-so-foward from time-to-time.

BTW, I have really enjoyed this whole thread on digital amps. It's really intersting to see where class D amp design has been, and where it's potentially going.

Colin

[analogman]

I guess no one remembers when "the industry" and the "press" proclaimed "PERFECT SOUND FOREVER" I'll stick with my KT88s.

Just my two cents worth

Analogman