What kind of amplifier needs big caps to drive bass?

[Joe Bloggs]

Does this only apply to amplifiers that do not have enough current ability to drive DC through a headphone's impedance load (not actual headphones of course as DC would damage them) for an unlimited amount of time? Are there any headphone amps that don't need caps to drive bass? What about speaker amps?

 

[Wodgy]

The big caps you're referring to are in the power supply. Bass (and any other powerful transient, like an orchestral swell) needs a large amount of power quickly; the caps discharge to provide this. Speaker amps need this even more, because the power required to drive speakers is so much greater. You'll find that quality speaker amplifiers have much larger power supply caps than headphone amps, for this reason. Take a look inside any one of NAD's amps, for example, even their cheapest ones. The power supply caps are huge.

 

[Joe Bloggs]

Right. But I suppose that a power supply made for loudspeakers can handle transients on headphones without even discharging the caps?

BTW, besides the power supply, there's got to be things in the amplifier proper that may have problems with bass response or full frequency response in general, right?

 

[Wodgy]

There are a lot of variables (e.g. it depends on how much power is dissipated in the speaker to headphone resistor network), but in general that's true, the caps will probably never fully discharge when driving an easy load.

Quote:

BTW, besides the power supply, there's got to be things in the amplifier proper that may have problems with bass response or full frequency response in general, right?

Yes, that's the whole science/art of amplifier design. There are tons of issues to think about. For instance, the damping factor of the amplifier will determine how well the amp can control the bass. Similarly, most consumer speaker amps use ICs that run in class D for heat and power reasons. A better sounding amp will run in class AB (or class A up to a point) but at the cost of power consumption and the need for bigger caps.

 

[Joe Bloggs]

Apparently I'd edited the post above while you were replying...

Well, apart from that, if the power supply depends on the caps for a certain output level, then you are depending on the caps not changing in potential difference significantly at the end of discharge, right?

Arrrgh! To go down from Class A down to class D hell!

Somehow I don't notice much difference though

 

[Wodgy]

The main idea is to have caps large enough so that they will never fully discharge under real-world conditions. When this happens, the circuit may become unstable or start clipping.

 

[Wodgy]

An interesting thread would be to compare people's experiences with respect to class A and class AB. I swear I can hear a difference using both headphones and speakers, but it's not clear how many others can. This is one reason people tend to like tube amps; even push-pull designs naturally run in pure class A. Those with Meta42s can easily switch from class A to class AB and see if they detect a difference.

 

[Joe Bloggs]

Well, what does the difference sound like?

Technically, all I know is that as you go down classes low level information is represented less and less well right?

So basically it tends to wash out detail, is that it?
--------------
Hmm, if I'm reading your post above right, you mean that apart from damping factor (which affects the control of bass rather than the amount), there's no compromise decisions in the amplifier proper that would specifically affect bass response? I suppose you'd just not listed it out

I know the input caps (if you have them) can roll off the bass, but that's about it

What I'm trying to figure out is, why people keep on saying that more high end amplifiers have better bass. Is it just because of the power supply or...

And it seems some of the factors affecting bass response have nothing to do with how much power the amp is designed to give and how much it's being asked to give...

 

[Joe Bloggs]

Well, all I know is that the bass on the Onkyo is unreal...

Will building the Grado headphone adapter improve matters by demanding more current of the amp all the time and hence minimize the effect of being er, in a bad class?

Or does it have anything to do with current and power supply at all?

 

[Wodgy]

Quote:

Originally posted by Joe Bloggs Well, what does the difference sound like? Technically, all I know is that as you go down classes low level information is represented less and less well right? So basically it tends to wash out detail, is that it?

Part of it is low level information, but a bigger part is speed and "verve". A class A amp sounds more smooth and dynamic. Also, from an engineering standpoint, there tends to be more distortion in lower classes. This distortion tends to have odd harmonic content, which is unpleasing to the ear. (In contrast, a Stradivarius violin tends to have a lot of even harmonic content.)

Quote:

Hmm, if I'm reading your post above right, you mean that apart from damping factor (which affects the control of bass rather than the amount), there's no compromise decisions in the amplifier proper that would specifically affect bass response? I suppose you'd just not listed it out

No, there's much more than just damping factor to consider. That was just an example.

Quote:

I know the input caps (if you have them) can roll off the bass, but that's about it

Here you're referring to the input coupling caps, not the power supply caps. These aren't really an issue. As long as they're big enough they won't filter out the bass, but this is a separate issue from how much bass the amplifier can produce.

In general, an amplifier may have capacitors in four places:
1) the power supply
2) input coupling
3) output coupling
4) elsewhere for stability.

Only 1) is truly necessary, the rest are optional, especially 3) and 4). As you move to higher end designs, the optional ones dissapear. For instance, the Gilmore amp uses a servo instead of 3). The Meta42 avoids 3) also. The perceived amount of bass output by the amp is primarily affected by 1).

Quote:

What I'm trying to figure out is, why people keep on saying that more high end amplifiers have better bass. Is it just because of the power supply or... And it seems some of the factors affecting bass response have nothing to do with how much power the amp is designed to give and how much it's being asked to give...

It's not only just because of the power supply, but the power supply is a big part of it. Visit a store and listen to some high end amps driving speakers with large woofers and you'll see that there are often dramatic differences in bass response. You're absolutely right on the second point, of course, the output power rating of the amp (e.g. 30wpc) has virtually nothing to do with the amount of bass it puts out. There are crappy Sony subwoofers rated at 100wpc that have absolutely flabby bass. I used to own one before I knew any better.

 

[Wodgy]

Quote:

Originally posted by Joe Bloggs Well, all I know is that the bass on the Onkyo is unreal...

Really? Are you talking about driving speakers directly, or are you talking about driving a powered subwoofer?

I suspect the Onkyo uses a class D amplifier IC like the TDA1521 or TDA2003. These are essentially power op-amps (you can think of them that way). I doubt they can drive speakers directly with the same kind off bass authority you'd get from a discrete power amp, but I'll reserve judgment since I haven't heard it.

Note that I'm not "dissing" these kinds of amps. They're nicely designed for mini units. It would be very difficult to build a class AB amplifier into a minisystem like the Onkyo. The heat dissipation and power supplies would both be a real problem. You'd never get class A in that size at all. Remember that FET transistors running in class A are only about 5-10% efficient, meaning that to output 10 watts in class A you'd need a heat sink capable of dissipating approximately 100 watts. These kinds of heat sinks are big and very expensive (about $150-200 here in the US).

Quote:

Will building the Grado headphone adapter improve matters by demanding more current of the amp all the time and hence minimize the effect of being er, in a bad class?

My guess is that it would be the opposite. You want an adapter that puts as small a load on the amp as possible. This is the idea behind Joe Lau's speaker amp-to-headphones transformers.

 

[thomas]

Class D amps are very different from power opamps. Class D means digital amplification, which operates in a manner similar to switching power supplies. They are known for their high efficiency (>90%) ... But there aren't very many designs that use it, as the technology is still quite new. Its more likely that the amp uses a power opamp, which is just normal analog opamps with higher output capabilities- almost all will have a class AB output stage. I am currently using a DIY power amplifier using a power opamp, and the sound is very good, certainly better than many descrete designs. (the famous gaincard amp which sells for $$$$ is also based on a power opamp) Many cheap amplifiers such as those in the 500watt minisystems use cheap discrete designs because it is cheaper and more powerful than power opamps...


Coupling caps have a large role in bass response... Some designs require output caps, and most of those amps improve by increasing the size of those caps. Input coupling caps may also have an effect, but they are usually large enough not to affect bass response...

btw, #4 should be manditory in a proper design, and can actually have an effect on the bass response- an amp without proper bypassing will be less stable and may oscillate in the MHz range- although this is inaudiable, it will rob the amplifier of power needed for audiable frequencies.

 

[Wodgy]

Quote:

Originally posted by thomas Class D amps are very different from power opamps. Class D means digital amplification, which operates in a manner similar to switching power supplies. They are known for their high efficiency (>90%) ... But there aren't very many designs that use it, as the technology is still quite new. Its more likely that the amp uses a power opamp, which is just normal analog opamps with higher output capabilities- almost all will have a class AB output stage.

You're right about the efficiency advantages, but I think you might be mistaken about class D not being used. I was under the impression that most minisystems these days use class D amplification. Otherwise they'd need larger heat sinks than most have space for.

Check out TI's directory of "Audio Power Amplifiers":
http://focus.ti.com/docs/search/para...sociated=false
Almost all of these ICs are class D.

 

[thomas]

Nope, less than half of those chips are class D, and pretty well all of those class D chips were released in the last 2 years. Most consumer electronics use the exact same circuity year to year, and only the cosmetics are changed every year. Most of those minisystems today were probably designed in the late 80's or 90's, when this technology was not available. There is no doubt that these class D amplifiers will become much more popular in the future, but it will take several years for it to become common....


BTW those small microsystems usually only output 15-30 watts in class AB mode, (or even class B) so there isn't too much heat that needs to be removed...

 

[OtterMarc]

Joe and Wodgy,

Though far less knowledgable than you 2, I thought I'd point out the (interfering) role a crossover can have on bass response in speaker systems. Anotter place to consider the role of capacitors.
Marc