Double Bass What is digital? Amp classification discussed

[Tom Baldwin]

This recent post by jzucker in the classifieds got me to thinking:

D does not stand for "digital"

The letter D used to designate this amplifier class is simply the next letter after C, and does not stand for digital. Class D and Class E amplifiers 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. (A digital waveform would be pulse-code modulated.)

So I wonder, just because I am curious,
1. What are the different classifications?
2. What are their characteristics, electronically and sonically
3. What are some examples of each?

The topic surfaced briefly in one of the MarkBass threads, where it was noted that the MarkBass heads use an A-B type something or other, which is similar to the older MI series Woods heads but not the newer Electracoustic models (class D?) But then the AI stuff is Class D and they seem to be more like the older Woods than the newer. Obviously the voicing of the preamp is a big factor, and perhaps has more to do with the sonic footprint of an amp than its power supply.

Anyhoo, all info is welcome, and thanks for sharing!

 

[fdeck]

Class-A: Typically a single transistor (or tube) controls the output voltage. Positive current might be supplied by the transistor, and negative current by a resistor, or vice versa. Simple to design. When Class-A circuits distort, they do so asymmetrically. Often found in preamps, rarely in power amps above a few Watts.

Class-B: Typically push pull amplifier with an abrupt transition from the positive to negative side of the waveform. Rarely used in audio, by virtue of not really having a reason for existing.

Class-AB: Also typically push-pull (transistors for controlling both the positive and negative current), but with some overlap as the waveform crosses zero to avoid the abrupt transition.

Class-C: Form of Class-A in which there is no control over part of the waveform. Historically, combined with tuned circuits that afforded good thermal efficiency in radio transmitters.

Class-D: Output is switched rapidly between a high and low voltage, or between positive and negative. On average, the output voltage resembles the desired output waveform. The switching occurs at a very high frequency, and is inaudible. The averaging is performed by a passive filter at the output of the amplifier. This is the up-and-coming technology found in the newer "lightweight" amplifiers, because it can achieve thermal efficiency of 90% or more, whereas Class-AB is limited to about 50% efficiency.

Class-H: An extension of push-pull in which the waveform is divided into more than two regions, with different circuits working in each region. This is a move towards the efficiency of Class-D.

Other classes: Some are taken as trademarks, such as the "Class T" of Tripath.

 

[Tom Baldwin]

Oh, gosh! I should have known this would be over my head. Now to further confuse things, I found the post in regarding the Markbass heads here. Apparently there is the power amp and the power supply, which are two different things. You could have an analog amp with a digital power supply.

Fdeck, were you referring to just the power amp? Thanks fo the info!

 

[fdeck]

More: The "digital" term got attached to Class-D because of the on-off nature of the output control, but Class-D amps are usually just a bunch of analog circuitry controlling heavy duty output transistors. Texas Instruments makes an integrated circuit that is an audio power amp with an actual digital input.

Sonically, there is not a strong correlation between class and tone for power amps. Power amps of all types gain their primary characteristics from the use of negative feedback. In preamps, Class-A can provide a musically forgiving overload response. Also, Class-A circuits are often involved in preamps that are intended to distort in desirable ways, such as in guitar amps and effects.

 

[fdeck]

Like you said about the power supply. There are two types. The traditional power supply is a step-down transformer whose AC output is converted to DC by diodes, and filtered (smoothed out) by large capacitors.

A switch mode power supply (SMPS) shares many traits with Class-D power amps. It uses high speed switching to run the power supply at a very high frequency, which allows for a much smaller transformer, and higher efficiency. The result is a smaller and lighter package. Your computer uses a SMPS.

But once again, the class of power supply is not directly related to tone quality of the amp. There are design issues involved with either kind of supply.

 

[drurb]

Great job fdeck! Let's not leave class G out of the list. Discrete rails vs. infinite set of rails. Here's a nice compact reference-- not that fdeck has left much to be desired.

 

[fdeck]

Wow, what a link. I should have done my research before posting!

 

[drurb]

Wow, what a link. I should have done my research before posting!

No, no, no. You did a spectacular and scholarly job! I'm not sure Wikipedia is the best choice but it explains it pretty well. The second link I added is one of the ones I liked the best-- next to yours.

 

[anonymous02282011]

so, perhaps instead of digital, we should be referring to these types of amplifiers as "solid state" as opposed to "tube"?

also, i've heard that solid state amps flip the phase of the signal. is this broad generalization true?

dig the thread.

 

[seanm]

so, perhaps instead of digital, we should be referring to these types of amplifiers as "solid state" as opposed to "tube"?

also, i've heard that solid state amps flip the phase of the signal. is this broad generalization true?

dig the thread.

Class D amps could be made with tubes. Although I think they would be too expensive to be practical.

Any single stage amplifier inverts the phase of the signal. It doesn't matter if they are tube, transistor, or opamp.

 

[drurb]

Any single stage amplifier inverts the phase of the signal. It doesn't matter if they are tube, transistor, or opamp.

Well, not an op-amp, which can easily be configured as a non-inverting amplifier. I've built quite a few circuits that way. For example, look Link Removed.

Of course, given the internal circuit of a typical op-amp, you might not consider it to be a single stage.

 

[fdeck]

so, perhaps instead of digital, we should be referring to these types of amplifiers as "solid state" as opposed to "tube"?

also, i've heard that solid state amps flip the phase of the signal. is this broad generalization true?

dig the thread.

Not in general, unless it is some industry convention that I am unaware of. The most common power amp circuit topologies are non-inverting. But an amp can have any number of stages for different purposes, thus potentially ending up with either phase.

If a power amp has a balanced input, e.g., XLR, then it's just a matter of how you name the two input terminals.

The only time it matters is if you have multiple amps driving multiple speakers, where you want to avoid cancellation.

 

[Phil Redman]

Class D seems to be gaining a reputation as having a high degree of transparency, along with some of the warmth associated with tube amps, but not the flabby bass that tubes are sometimes known for. Also Class D doesn't seem to suffer from the high-frequency sibilance associated with many transistor-based products.

A peak into the audiophile world might be helpful: http://www.tnt-audio.com/ampli/t-amp_e.html

Cheers,
Phil

 

[drurb]

Class D seems to be gaining a reputation as having a high degree of transparency, along with some of the warmth associated with tube amps, but not the flabby bass that tubes are sometimes known for. Also Class D doesn't seem to suffer from the high-frequency sibilance associated with many transistor-based products.

No offense, but associating high-order, subjective sonic qualities with amplifier classes is largely folly. The particular execution of each design is likely to account for more of the variance than the particular class, per se.

For example, the "warmth" associated with tubes stems largely from characteristic frequency response anomalies resulting from interactions between the output transformers typically used in tube amps and the load (i.e., the loudspeaker). One would hope that a solid-state amplifier designed to have a flat frequency response would not exhibit such warmth. By the same token, the "flabby" bass response often associated with tube circuits can also be traced to the poor damping factor that is a result, again, of the impedance matching output transformer.

Bob Carver demonstrated long ago what many of us knew and/or suspected. The sonic differences heard stem from the transfer function of the amplifier and not from the particular type of amplifying elements employed. In English, one can intentionally build a solid-state amplifier to have the transfer function (frequency response) of a typical or even venerable tube amplifier such that the two cannot be distinguished under normal (non-laboratory) listening conditions.

To be sure, certain amplifier classes and topologies have their specific drawbacks and advantages. Early class-B solid-state amps employing large amounts of feedback were subject to high levels of TIM (tranisent intermodulation distortion, also called Otala distortion). We are long past those days.

Again, I am not picking on Phil. He reported what reputations are being gained. My point is to caution folks against being taken in by such claims. By the way, modern class D circuits are, themselves, transistor-based products.

 

[Tom Baldwin]

Thanks for all the excellent info. Clearly, most of it is beyond my understadning. I should probably take a beginner electronics course. Anyone know of anything like that online, or other resources?

Also, I'm still wondering forif there is an example of a digital amplifier that is on the market for bass.

 

[satchmo]

Digital just means that at somepoint the waveform is converted into a digital signal (binary).

A line 6 pod or any modelling amp could be 'classified' as digital, because the analog signal from the bass is converted to an digital signal for computerized processing before being reconverted to analog at the end.

 

[drurb]

Digital just means that at somepoint the waveform is converted into a digital signal (binary).

This is a crucial point as there are "sampling" systems that are not digital at all but which sample the waveform along a continuously variable (analog) continuum.

 

[Ric Vice]

Thanks for all the excellent info. Clearly, most of it is beyond my understadning. I should probably take a beginner electronics course. Anyone know of anything like that online, or other resources?

Also, I'm still wondering forif there is an example of a digital amplifier that is on the market for bass.

T-Bal
Several years ago a company built an amp with some kind of processor in the preamp section of the unit. Victor Wooten endorsed them for a while whether he actually used them or not I don't know. When the Line Pod hit the market they disappeared. I'm not sure what kind of power amp and power supply those units used.I do know that they aren't around anymore.

Ric

 

[fdeck]

Ashdown made a digital amp called the Superfly, and Yamaha had a digital amp as well for a while. There are some power amps that come with DSP (digital signal processing) and rack mounted DSP controllers by DBX and others for managing PA speaker systems. The latter are quite sophisticated, and could be used for bass.

In my way of thinking, a completely digital system would limit analog circuitry to a bare minimum: Some sort of buffer stage to receive the bass signal, then directly into an analog-to-digital converter, digital processor, and a digital-to-analog power stage to drive the speaker. One such power stage is the [Invalid or Expired Link Removed] chip, but I doubt that it has made it into a music product yet. It's also a fair bet that the final power stage in that chip is analog.

You could also have a conventional digital-to-analog converter feeding an analog power amp. Such choices seem to be mainly driven by practical concerns.

Where I would define "digital" is if the amplifier uses digital processing to implement the controls. Everything else is rather mundane, but digital controls are interesting. Virtually any control function could become a matter of writing software instead of designing circuitry. The knobs are simply digitized and used to compute control parameters for the software. This is really where I think the benefits of digital would accrue.

I think the Line6 Pod, and Behringer V-Amp products do it this way. Also, there is a software product that uses a PC as a digital signal controller, sold under the Ampeg brand IIRC.

In a moment of insanity, I ordered a chip set for a digital audio controller, but have not gotten up the gumption to do anything with it.

Note that if "beyond insanity" is desired, this page sketches what a truly digital power amp would consist of. But the author mentions some real drawbacks to such a system.

 

[drurb]

ANote that if "beyond insanity" is desired, this page sketches what a truly digital power amp would consist of. But the author mentions some real drawbacks to such a system.

Nice article! The author demonstrates the drawback for a DC input. If I understand correctly, drawbacks of such converters can be alleviated via proper dithering techniques. For anyone insane enough, Google "sigma-delta dither".