|   How is amplifier THD measured? | DMK
Nov 4, 2003 5:44 PM | | Is it measured at full power? At only one or two frequencies? I'm curious because someone made the claim that his low powered SET tube amp had less distortion at half power than SS amps. I'm dubious, even though I prefer tube amps. |
| well and some info on tubes. | RGA
Nov 4, 2003 8:39 PM | | I was flipping through a little handbook which is more of a Layman guide. I prefer English to the circular logic I often read.
Here is how this section goes: My comments are in []
How Not To Build an Amplifier:
A) The High Feedback Amplifier. "...A sample of the output is broought back to the input, where it is introoduced out of phase (electrically reversed) so that it cancels some of the amplifier's gain, and also cancels some of its harmonic distortion. High Feedback is popular, because so many manufacturers are hung up on distortion figures: a figure of .00005% looks so good on a spec sheet [I know I got conned into a Pioneer Elite Receiver that bettered that number which was garbage)] and it gives the mid-fi salesman who is striving to to be heard over the cacophony of the TV sets something to talk about. Unfortunately, feedback makes an amplifier slow [my speakers' woofers were flabby and a step behind] and it's easy to figure out why. The signakl gets to the output uncorrected, and it is only when the contents of the feedback loop make it back through the amp once more that correction occurs. It is obvious that such an amplifier will work well only with signals that change slowly over time. Like test tones.
b) the high damping factor factor amplifier (they gone on but I'll type a segmant) "...Simply reducing ONE of those elements - the amplifier impedence - won't improve real-life damping.
This might be a harmless little delusion, except for one thing. The usual method of achieving a high damping factor is by the use of heavy feedback. There's nothing harmless about THAT. We suggest avoiding amplifiers with damping factors much higher than 40.
C) The Wide Bandwidth Amplifier...pass
D) The amplifier with NOTCH. The first transister amplifiers sounded different from tube amps, and many audiophiles didn't like them. But mid-fi salesmen (the only sort there was back then) insisted that the new amps were virtually perfect. Brandishing spec sheets showing very low distortion, they said that we "just weren't used to hearing sound of such purity." (The same argument would later be used to defend the flawed sound of early cd players).
In fact those first transistor amplifiers suffered from a curious problem. Distortion was indeed very low at FULL power, but as power dropped, distortion increased. From a fraction of one percent, it climbed to nearly 10% at low level. No wonder the amplifier didn't sound right. [incidentally cd players have more bits at high level than at low level to work with as well]. What was wrong? In the push-pull" designs which were (and are) used in power amplifier output stage, one transistor (or set of transistors) amplified the positive half of the signal, and another amplified the negative half. The problem came at the "zero crossing" point, when one transistor handed off to the other. There was a small "notch," a discontinuity, which was clearly visible on an oscilloscope. It had a negligeable effect on full power signals (which is what most lab tests are made with), but because the notch was fixed, it became more important as level dropped...if you check amplifier harmonic distortion graphs in mid fi magazines, you'll see that some amplifiers have distortion that increase at low power. Those amplifiershave notch distortion, though the magazines will never tell you that. After all, amplifiers sound alike, don't they? A test at the NRC...
E)The starving amplifier
F)The unstable amplifier
G)The non linear amplifier:
There is a principle sacred to the best amplifier designers: First make an amp which sounds good with NO FEEDBACK applied at all. And then improve it with feedback. Some younger designers have never heard of this principle, or if they have think they know better. Some have hatched clever new designs, in which the amplifier puts out, not sound, but a series of electrical pulses which are controlled by the input signal. |
| 2 | RGA
Nov 4, 2003 8:42 PM | | If you fed this highly non-linear signal to a speaker, it would sound horrible. However, large amounts of invverse feedback are used to make this signal linear once more. What is wrong with this picture?
The class A amp throws away all notions of efficiency. All the output transistors are switched on all the time. There can't possibly be notch distortion.
Tube amplifiers Advantages
2) High Voltage.
Transistors require power supply voltages of 15 to 70 volts, whereas tubes can use voltages of 150-450 volts. it is obvious that you cannot obtain an output voltage greater than the supply voltage, and so a transitor amplifier "runs out of breath" much more easily than a tube amplifier.
3) More graceful overloading:
If you overload a transistor, it becomes pregressively less linear, which means that the output current will no longer be exactly proportional to input current. But at one point, the top of the wave will be "clipped," ... the tube also becomes non-linear, but it doesn't clip the wave. And so a tube amplifier pushed too hard merely sounds increasingly fuzzy. It is interesting to note that some musicians rather like this effect. The fuzz boxes used by some rock guitarists are intennded to simulate the sound of a tube amplifier overloading.
4) Facor X.
There is a "tube sound" which certain audiophiles like. It is characterized by smooth and warm highs (the absence of sudden death overload?) and by richness in the lows(the effect of the output transformer?). There is also fator X: Audiophiles sometimes say, "I don't know why, but I like it."
It is obvious that supurb amplifiers CAN be built with transistors, and there are numerous examples. But the tube, for all its disadvantages, is far from dead. |
| A lot of half truths and outright errors from a propagandist | skeptic
Nov 5, 2003 4:28 AM | | This is what happens when somebody who doesn't know what he's talking about shoots his mouth off. I don't know where this gobbledegook came from but it is more than misleading, it takes you to the wrong conclusion. It would have you believe that after 4 years of engineering school and a lot of hard work on lab benches electronics engineers are almost all idiots. That they don't know what they are doing or talking about. Clearly this comes from someone who is not an electronics engineer. Maybe a salesman who decided to go into the business and read a little looking for ways to rationalize punching holes in the prevalent technology which has reigned for over 50 years. Why has it done that? Because it is better. The world could go back if it wanted to. It doesn't want to. At least more than 99 percent of it.
OK, here are a few corrections to the most obvious mistakes. When solid state amplifiers first made their appearance in the early 1960s, THERE WAS a lot of crossover notch distortion. And so manufacturers with superior solid state products began to specify harmonic distortion in words to the effect that the distortion figures were at maximum power and that noise and distortion decreased as power output decreased. This effectively told customers that their products did not have crossover notch distortion. When power FETs arrived on the scene, the thing was practically a dead issue because FETs have similar characteristic curves to vacuum tubes and it became just as easy to design quality class AB amplifiers around them. Early solid state amplifiers were a problem for many engineers who couldn't make the transition from voltage controlled devices (tubes) to current controlled devices (tranisitors.) They quickly fell by the wayside and went into sales or selling insurance or something else. There was a lot of marketing pressure on many companies to get solid state models on the market quickly. If they succumbed to this pressure before they were ready, they suffered not only with poor sounding equipment but with unreliable models. Pilot Radio had amplifiers blowing up after 8 hours use and at one point they were coming back faster than they were going out.
If you don't know how to correctly apply the complex formulae to design and evaluate negative feedback circuits, using it haphazardly can be a disaster. If you do know how and apply it skillfully, it can make a very good amplifier excellent. Without negative feedback, many circuits, including most vacuum tube circuits are at the very least, unstable because small changes in filament temperature can result in radical changes in electron output and radical changes in gain. If a power supply isn't regulated, all operating parameters can drift. Regulation requires negative feedback in the power supply. My advice, never buy an amplifier where the manufacturer tells you has to be on for at least half an hour to an hour to sound good. Properly designed solid state amplifiers stabalize within seconds and tube amplifiers within a few minutes at most.
Damping factor is defined as the speaker impedence divided by the source impedence (including the wire impedence.) The "ideal voltage source" has zero impedence and the ideal wire also has zero impedence making the damping factor for any load infinite. What this would mean is that if the speaker has a tendency to spurious resonance, that is not follow the amplifier voltage but expend built up momentum by generating back emf, it would have no effect on the amplifier output voltage. In fact, the amplifier would act to shunt it out and surpress such spurious resonances. You can always lower the effective damping factor by inserting a resistor in series with the speaker. This will increase the audible bass on some speakers but it is not good bass, it is false resonance. The higher the damping factor, the better. Because tubes have an inherently high plate impedence, it is very difficult to get the |
| A lot of half truths and outright errors from a propagandist | skeptic
Nov 5, 2003 4:44 AM | | Because tubes have an inherently high plate impedence, it is very difficult to get them to even deliver power to a loudspeaker let alone keep it under tight control. The answer for most tube amplifiers it the output transformer, the weakest link in most vacuum tube amplifiers.
70.7 volts applied across an 8 ohm load, will deliver 625 watts. Applied across a 4 ohm load, it will deliver 1250 watts. Audio amplifier power output is invariably limited by the amount of current it can deliver, not voltage. High fidelity amplifiers should not be driven to clipping whether they are vacuum tube or solid state. When vacuum tube amplifiers are overloaded, damage to the tubes can result. When solid state amplifiers are overloaded, they usually go into a protection mode. There is no such thing as graceful overload. (BTW, fuzz boxes used by musicians who play electronic instruments are virtually all solid state today.) That vacuum tube amplifiers sound better because of "soft clipping" is a silly myth.
If you like the sound of a particular tube amplifier, buy it. But you don't always get a lot of performance just because you paid a lot of money. In high end audio, high price seems to be a way to make you think you are getting far more than you actually are. |
| A lot of half truths and outright errors from a propagandist | RGA
Nov 5, 2003 7:33 PM | | High Damping Factor under the bad: This is a Layman text.
The Damping Factor is a measure of an amplifiers ability to control the cone of the speaker, and of course that is a good thing.
As the amplifier signal flows through the speaker's voice coil the speaker acts like a linear(back and forth) electric motor and the cone moves. But at the same time the speaker also acts like an electric generator, and it generates a "back Voltage" which it sends back to the amplifier. If the amplifier has a very low internal impedence, it will virtually short out the back voltage. That will tend to brake the cone motion, and the cone will be less likely to go on flapping after the end of the signal which set in motion.
Most power amplifiers do in fact have low impedences. If an amp has an internal impedence of .2ohm annd the speaker has an impedence of of 8ohms, the damping factor will be 8 / 0.2 = 40
This is a typical value, but some amplifiers have a damping factor of around 80 or more. You might suppose that a higher figure is better, but in fact it makes little difference. A glance at the diagram opposite reveals why.
Impedence of amplifier: .02
Resistance of connectors and cable: .05
Resistance of speaker voice coil: .03
Inductance of speaker and crossover: Impedence varies with Frequency.
As you can see, there is morethan the amplifier's own impedence between it and the speaker. There is also the electrical resistance of the connectors and cables, the impedence of the elements in the speaker's crossover network, and the impedence of the many turns of fine wire in the speaker's voice coil. Simply reducing ONE of these elements - the amplifier impedence won't improve real life damping.
This might be a harmless little delusion, except for one thing. The usual method of achieving a high damping factor is by the use of heavy feedback. There's nothing harmless about that. We Suggest avoiding amplifiers with damping factors much higher than 40."
But hey...none of it really matters - because you obviously think all tubes make crappy sounding music and SS is bliss.
I'm thankful I'm listening to my system and not yours = and I'm sure you feel the same way.
Ohh well that is why there are so many SS fans and so many tube fans. Though more ss amp owners who hear tubes make the switch rather than the reverse. Deluded? IMO no they have finally heard an instrument not a Fax copy. But you feel the opposite and you're entitled. |
| Correction typo | RGA
Nov 5, 2003 7:36 PM | | Impedence of amplifier: .02
Resistance of connectors and cable: .05
***Resistance of speaker voice coil: .3 *** .3 not .03
Inductance of speaker and crossover: Impedence varies with Frequency. |
| A lot of half truths and outright errors from a propagandist | skeptic
Nov 6, 2003 6:30 AM | | "I'm thankful I'm listening to my system and not yours = and I'm sure you feel the same way."
I have no way of knowing what your system sounds like and you have no way of knowing what mine sounds like either so that statement is absurd.
"Impedence of amplifier: .02
Resistance of connectors and cable: .05
Resistance of speaker voice coil: .03
Inductance of speaker and crossover: Impedence varies with Frequency."
The internal impedence of the woofer is 4 to 8 ohms and therefore fails to electrically damp itself internally. It requries the external electrical damping of the amplifier.
Not only does the amplifier lose control over the woofer voice coil motion and fail to brake spurious resonance but the back emf adds to the bias voltage across the transistors causing their quiescent operating points to change until the transient is filtered out by the power supply.
Irrational railing against negative feedback doesn't change the fact that in the hands of highly competent designers, it is one of the most important ways to improve amplifier performance developed in the last 65 years. And of course in the hands of tinkerers and incompetents, it results in mediocre to awful performance. This is not my opinion, it was the opinion of ALL of the professors of electrical engineering who taught me circuit design theory while I was in school. BTW, some of them wrote the textbooks. And while that was many years ago, the mathematics hasn't changed. And yes it is still very hard and very complex. |
| You don't get it, skep | RobotCzar
Nov 15, 2003 9:56 AM | | Negative feedback is "bad" because it is negative. If you use postitive feedback you would get a better result.
I am absolutely amazed that people who don't know ohm's law feel they can deride negavtive feedback. Their reasons sound as if they were experts about something, I just don't know what that something is. The high end objection to negative feedback is not based on any sound concept, so why do people like RGA keep repeating it?
What's wrong with negative feedback if it lowers overall distortion, RGA? Don't start with the "it only lowers distortion at low power" stuff without some evidence. All reviews of amps these days indicate that they have very low distortion at all power outputs into typical loads AND at all frequencies, though cheaper amps display higher dynamic distortion at high frequencies.
Tubes suck. Not because they tend to make amps that are not flat (as they do), but they deteriorate quickly and most non-new tube amps have higher distortion that the already high distortion they have when new. Why not admit that you have just been won over by the tube cult and you are trying to justify that with rationalizations.
You like the sound of tubes. Great. Some people probably prefer Edison cylinders, so what? |
| You don't get it, skep | RGA
Nov 21, 2003 1:57 PM | | I buy Solid State amps...My three favorite amps under 3k are solid state amps.
the difference is that I, Unlike you obviously, go out and listen to products for an informed opinion as to what they sound like. I have no doubt Tubes measure worse, for whatever that's worth. They can and often do sound better than SS amps. It's called a preference. http://www.stereophile.com//reference/70/index.html |
| Damping Factor | Norm Strong
Nov 15, 2003 12:00 PM | | When you see an amplifier specification called "Damping Factor", it is obtained by dividing the output impedance at low frequencies into 8. How could it be otherwise? To divide the speaker impedance by the output Z of the amp, it would be necessary to know the exact impedance of the speaker. Since that is unknown, damping factor is standardized to 8 ohms. |
| |
