I've run in to this a lot over the years. My solution is tubes. May not work for everyone but it does for me. Sure beats turning up the volume too loud just to get that punch and body that you know is there... |
Yes, tubes and transistors; the jury is out on Class D as it inherently is not Class A.
Tubes generally are usually operated closer to Class A than transistors usually are; most transistor amplifiers are biased at such low idle currents that its driver transistors (not the output transistors) are actually what is driving the speaker at low volumes. Tubes are inherently more linear than transistors, especially triodes, so tube circuits tend to be simpler and operating with less feedback, in fact I think its safe to say that the majority of triode amplifiers don't run loop feedback at all.
So your "punch and body" is coming from the reduced levels of negative feedback and the additional authority gained by operating at or near class A. |
Bob_reynolds, we *do* have some control over that 'configuration' of voltage vs. current. It has to do with the output impedance of the amp.
A higher output impedance can deliver excellent bass if properly matched with the speaker; conversely an amplifier with very low output impedance will deliver poor bass if poorly matched with the speaker.
However it is more than just a conversation about equipment matching; there are two different paradigms that are in common use in high end audio:
http://www.atma-sphere.com/papers/paradigm_paper2.html
IOW, getting bass out of the speaker is not a function of the amplifier **or** the speaker, it is a function of how well they work together. This comes about out of intention- for example, the ability to double power as impedance is halved is not the same thing as saying that the amp will play bass.
Furthermore, amps that are direct-coupled from input to output (and in a very non-intuitive way) are often at a disadvantage in playing bass as they have the ability to modulate their power supplies. As soon as they do this the bass dries up. This is due to the fact that the amplifier LF pole is lower than that of the power supply, when it should be the other way around (something that you can't do with direct-coupled amplifiers unless you use a battery). |
Correct. In the old days, some amplifiers had a 'Damping' control, which was a potentiometer that allowed you to adjust current feedback (not voltage feedback). Essentially, this allowed you to get a better match between the amp and speaker.
This was at the time that the Voltage Paradigm was evolving. Once the ground rules were laid out, the 'Damping' controls disappeared- everybody just used a lot of feedback and called it good. The problem is that feedback itself functions at low frequencies (bass region) as a dynamic compression device. You simply aren't going to get good bass (or **as good** bass) if you are running a lot of feedback, even if your system is matched within the Voltage Paradigm.
This is why a smaller amplifier that has little or no feedback can often seem to play better bass (more impact, more definition, more articulation) than larger 'powerhouse' amplifiers that would seem to have it all over the said smaller amplifier. Of course, such an amplifier itself has to be able to operate properly with the speaker in order to revel these traits, but its not that hard to find speakers that are compatible and capable of delineating the differences between amplifiers that I defined above- the speaker *does not* have to be full range.
I find that class of operation plays a role too- class A amplifiers will have more authority; as you move more towards class B (AB1 and AB2) the authority (especially at low volume levels) goes away.
So- to play bass well, power output is irrelevant, instead the amplifier ideally is low or zero loop feedback and class A, given that such an amplifier also works well with the speaker being used. This of course assumes that the amplifier does not have a low frequency cutoff that interferes with bass reproduction. |
Hi Bob, If you've not looked at
http://www.atma-sphere.com/papers/paradigm_paper2.html
maybe now would be a good time.
Of course it is quite true that if you make 100 watts into a given impedance, it doesn't matter what made the power, the current and voltage will be the same.
Where things get interesting is when you have an amplifier that makes constant power, as opposed to constant voltage. What happens here is that as the load varies, the current and voltage both change with it to that effect. This is what I was referring to in my posts above.
It is **not** true that all speakers are "voltage" driven, in fact before the Voltage Paradigm was developed, tubes (and often zero feedback tubes) were the only game in town. Designers had to accommodate constant power so to get flat response; they had to be a little more careful.
Some examples of where constant voltage is not so helpful:
ESLs- Constant Voltage amplifiers have trouble making bass while too bright for comfortable listening due to the impedance curves typical of ESLs: high impedance in the bass, almost none in the highs. You need constant power to make that sound right.
Full-range drivers/horns- Most high efficiency speakers have tighter voice coil gaps (hence their cost) which creates higher reactivity. The reverse EMF thus generated can wreck havoc upon an amplifier with lots of negative feedback. The result is excess harshness (ringing) at high frequencies; the main source of many listeners opinion that horns are brash, honky and the like.
Its not difficult to choose a speaker for such an amplifier with a high output impedance (also known as a 'current source' amplifier to use Voltage Paradigm vernacular). All you have to do is look at the intention of the designer. Some examples: Merlins, Sound Lab, most horns (except Avantgardes), Lowthers, Fostex, PHY and the like, The original AR-1, Audiokinesis, Coincident Technology, Rogers LS3/5A, headphones, Magnepan, Reference 3A Loudspeakers, to name a few.
If the impedance curve operates independently of resonance, you can count on the speaker being a Power Paradigm device rather than Voltage Paradigm. Nearly all planars are Power Paradigm devices. Of course, a smooth impedance curve makes the speaker available to both camps- Avalon is a good example of that.
The two paradigms are responsible for a lot of debate in audio- tubes vs transistors, objectivist vs subjectivist...
What is really going on is that the Power Paradigm nowadays operates around the idea of the rules of human hearing, where the Voltage Paradigm operates around the concept of bench measurement. I hope it is obvious that understanding the rules of these two paradigms creates also a means for using transistors in a musical way- Pass 1st Watt amps and Ridley Audio are good but rare examples.
I am of the opinion that when it is possible to quantify the right measurements on the bench, such as the amount of odd-ordered content generated with a dynamically changing waveform, then a new and encompassing paradigm will emerge. Until then, we are stuck with two schools in competition- and the continuing need to match equipment properly and audition it at home. |
Mapman, no, however the match between the speaker and amplifier is paramount. If the speaker has a low impedance at low frequencies, you are going to have to provide the current that making power into that load demands by law (Ohm's Law- no pun intended). The Ohms were always a difficult load! As you have seen, with less difficult loads the requirement for high power at low impedance is less important.
The flip side of the coin has a question- why go where angels fear to tread? IOW **if** the best sound is your goal, it has been shown that higher impedances favor transistors as well as tubes. A transistor amp driving a 16 ohm speaker will sound better (smoother, more detailed, more impact) than it will driving 4 ohms, all other things being equal. So the argument of current, insofar as the goal of 'best sound' is concerned, would seem to be moot.
|
Saki70, as Bob pointed out above, it takes both voltage and current to make power. The issue is that if we are to make 200 watts into a set of Ohms which might be 3 ohms, the voltage and current can be easily calculated (assuming for the moment that there is no phase angle which there always is in inductive devices).
So you can also look at it as being able to make the Voltage into a load like that- which will only be possible if you can also make the current.
The use of the term 'voltage source' I think can be dangerous without the understanding of the accompanying engineering principles for which the term is a sort of shorthand. You have to keep in mind, IOW, that regardless of the amount of current or voltage that you are making, that the end result is power which is composed of both.
Low impedance amplifiers (Voltage source) *can* make constant voltage with respect to the load. Not all do.
Higher impedance amplifiers (Current source) *can* make constant power with respect to the load. Not all do.
You **must** match such amplifiers with speakers that are designed with intention to work with that particular kind of amplifier that you are using. If you do not tonal aberrations will occur.
It is true that some constant voltage amplifiers can deliver lots of current. It is not true to say that that is the same thing as having lots of bass authority. IOW any kind of amplifier can be perceived as being wimpy in the bass if not set up properly. OTOH, some amps will not play good bass no matter what you do.
In my personal case, because I like amplifiers to be relaxed at all volume levels, I will not use one that has negative feedback, as that design element adds loudness (harshness) cues. So I work with speakers that are designed for amps that have a higher output impedance. So on my speakers there are no 'voltage source' amps that will play the speaker with the authority that my 60 watt triode amps will. But that could be very different on other speakers- my bone of contention is that because the use feedback, as far as I am concerned they will never sound like real music, so who cares :) |
Saki70, thanks for your comments.
What we are talking about is 2 things: 1) output impedance as an actual raw impedance, not complicated by negative feedback (IOW 'open loop') and 2) servo gain- the amount of negative loop feedback employed.
A lot of designers see these two as the same, but they are not. A variety of Voltage Paradigm speakers *require* that the amp have some sort of feedback to accommodate the otherwise improbable impedance curves that have resulted. The feedback is part of mechanism that allows the amplifier to accommodate peaks as well as dips in the curve. You can do this with raw impedance alone, but feedback makes it easier- your amp does not have to have such a low open loop impedance.
With higher impedance amplifiers, in order to get flat frequency response on such speakers, the role of feedback becomes more prodigious. However, many of these amplifiers are probably tubes, and often tube designers will eschew large amounts of feedback as the amplifier will often exhibit some linearity without, something that is rare in the transistor world. Their hope is that you will do the right thing and use these amps on a speaker that has a higher impedance. FWIW the thinking here revolves entirely around sonic performance rather than the ram ability to simply drive a low impedance, something that usually has little to do with overall sound quality.
So we are talking about a spectrum- as output impedance is increased and servo gain decreased, the voltage/current ratio that describes the output of the amplifier changes with it. So there is not a hard and fast rule.
In the past I've seen a lot of DIY hobbyists try to add loop feedback to a 'current source' amplifier with the hopes of getting it to play a four ohm load better- with more power. It does not work. That is because the open loop impedance of the amplifier is too high to be adequate for four ohms. You can reduce distortion and flatten the frequency response using servo gain, but you can't change the power. That is why I say that open loop output impedance and servo gain are different phenomena.
The pity of this whole thing is the idea that the ability to drive 4 ohm loads is a sign of being 'beefy' or 'gutsy' (somehow better anyway) in the amp. The fact of the matter is no transistor amplifier sounds right on 4 ohms, nor does any tube amp. If you want to really see what either one is really capable of, you need a higher impedance -16 ohms is nice- to do that.
At higher impedances speaker cables are far less critical in the overall sound and all amplifiers will exhibit less colorations due to reduced distortion. Transistor coloration BTW is the harshness caused by odd-ordered harmonics at very low levels. Tube coloration is the added 'warmth' or 'bloom' that is a product of even-ordered harmonics that are at a more pronounced level. So transistors will sound smoother with more detail and tubes will sound more neutral with more detail. Win win.
Sorry there was not a simple answer to your question! |
Saki70, in the old days you did. Then transistors came along, and the industry figured out that it could charge almost the same money for the transistor amps, while in fact they were costing only about 1/10th as much. Suddenly power was cheap.
It was not long before the speaker manufacturers realized that they could take advantage of this, because they could make lower efficiency speakers that also cost about 1/10th as much to make. To make them seem more efficient, 4 ohm speakers began to appear.
Its all about money IOW. But- if sound quality is your goal, then 4 ohms is right out. |
Saki70, I don't like the word 'synergy' as it suggests that we are going to balance one defect against another, and I think that in the case that we are talking about, the word is 'compatible' and the answer is 'yes'.
Loop and global feedback are the same thing. Zonal is usually the same as 'local' and may or may not be loop feedback, it could also be degenerative feedback. They are all forms of negative feedback, and there is a lot of verbal shorthand around, so you have to be careful.
I think what is important if you use feedback is that a) the amplifier must be fast and low distortion to begin with and b) the feedback be low enough that the odd-ordered harmonic enhancement is kept out-of-band. That really does not allow for very much feedback in practice.
The ZERO is the only autoformer I know of (we used to make a similar product called the Z-Music autoformer years ago). The concept is all about problem solving- this or that amp can't drive a speaker load that low so here's a solution to that problem. Given that that is the case, they work very well indeed. But if you had a 16 ohm speaker to start with that would be better :) |
Saki70, FWIW Steve McCormick, who has been making solid state amps for years, reports that his amps sound better driving a 4 ohm load though the ZERO, such that the amp is loaded at 16 ohms. His amps easily double power- they have no problem driving 4 ohms. I interviewed several solid state amp manufacturers at CES this year, and independently of each other they all concurred that while their amps easily **drive** 4 ohms, that due to effects withing the transistors themselves (which are exacerbated by more current) the amps **sound** better on 16. So much better, that using a set of ZEROs to do that is effective.
Transformers are inductors and so they ring if improperly loaded. Putting a 16 ohm speaker on an 8 ohm tap will mean that the transformer is going to add some distortion due to ringing, but the tubes otherwise will see a higher impedance load. Quite often, although they will make slightly less power, the distortion will go down too.
Mapman, you bring up a good point- feedback as a design issue can be tricky to sort out. Its the sort of thing that a designer or DIY person might tinker with and compare. There are so many variables that go into an amplifier/preamplifier design that it would be hard to ascribe something you hear in such a product to any one thing. You do have to work with the designs a bit in order to begin to sort out the sonic artifacts that different design considerations bring.
However we do already know that negative loop feedback contributes to odd-ordered harmonics in the regions that the ear uses as loudness cues. The contribution is slight- 100ths of a percent- but human ears are very sensitive to that sort of thing. We perceive this as 'brightness' or 'hardness'.Conversely, even ordered harmonics can be several orders of magnitude higher and our ears don't seem to mind.
I feel we are getting OT. If you wish to continue in this direction, let's start another thread :) |
Saki70, with regards to efficiency vs impedance, to me they seem to carry about the same weight. However if the speaker is really low efficiency, you are going to have power issues driving it (BTW 'really low' is 87db or less). Given that that is the case, you are facing compromise- with tubes, 16 ohms would be favored, with transistors I might be tempted with 4, just because of the power issue.
I am a fan of higher efficiency. Good Quality tube power is expensive and I have yet to hear a transistor amp that is better, although being who I am you have to expect that from me :) I **have** heard some excellent transistor amps though (that I liked as much or more that many tube amps), however none of them were what I would call high power. So it seems that the high efficiency/high impedance rule is a good one to follow (so long as those qualities do not interfere with resolution or bandwidth).
Note: Sensitivity and Efficiency are two different ratings. Sensitivity is 2.83V @ 1 meter, Efficiency is 1 watt @ 1 meter. 2.83V into 8 ohms works out to 1 watt, but into 4 ohms its 2 watts. So a 4 ohm speaker can have the same Sensitivity as an 8 ohm speaker but actually be 3db less Efficient.
|
