Objektivisten skrev:Hi James, welcome to the forum, what do You think is the secret about tubes and that they almost always sounding more pleasant and true to the ears? And do we really need all that watts, low powered amps seems to gain in transparent sound, may be construction simplicity or cost effectiveness?
Objektivisten skrev:8O Holy macaroni, the significance of my 'statement' is zero (maybe below) James. Can we please go back to the real substance in this thread, your eminent inputs and answers.
Regards
I agree.
In terms of our previous discussion on the causes of sonic differences between power amplifiers, in “Part One” discussed a few of the issues relating to the significance of overload characteristics during normal use of power amplifiers.
As a point of interest, these portions of the total research program that started in the 1970’s and were repeated in the early 1990’s as part of the development program for a new line of amplifiers that were produced in the mid and late 1990’s.
PART TWO: (A brief overview of the impact of wide band output impedance)
Next, in our research on this topic, we uncovered the effects of power amplifier output impedance and the interface with the loudspeaker load.
Output impedance is often referred to as damping factor relative to an 8-ohm load, but as we know, there are is not a single loudspeaker on the market that actually exhibits an 8-ohm, zero phase angle, load over the useful frequency range of the loudspeaker.
Also, damping factor is most often referred to as a low frequency phenomenon, with most amplifier specifications using 100 Hz as a reference frequency, when, in actuality, it is at least as important at high frequencies as it is at low frequencies. Unfortunately, most power amplifiers do not specify their high frequency damping factors… for the reason that most of them are substantially worse than they are at low frequencies.
(NOTE: There can also be dynamic, or non-linear output impedances in power amplifiers, and dynamic, non-linear input impedances in loudspeakers, but for this discussion, we will limit our observations to linear, small signal damping factor/output impedances and input impedances.)
Essentially, as one raises the output impedance (or lowers the damping factor) in a power amplifier, the amplitude error due to the impedance interface can cause the frequency response to take on the shape of the impedance curve more and more as the output impedance is higher.
Nearly all power amplifiers incorporate reactive components between the output stage and the output terminals. Most power amplifiers vary the amount of negative feedback vs. frequency, such that the output impedance increases with frequency due to the reduced feedback and output inductance.
After completing our studies in amplifier overload characteristics, we used the amplifiers in a manner that guaranteed that they operated below overload, and also kept them below the levels that would drive the loudspeaker in test, into nonlinearity, so as to eliminate those two variables in the device under test, as we moved on to isolate remaining parameters that affected sound quality in power amplifiers.
Even though we were operating both the power amplifiers and loudspeakers below and known nonlinear audible thresholds we were still detecting coloration differences between power amplifiers, with greater differences being exposed with some loudspeakers more than with other loudspeakers.
A first causal parameter, that we were able to isolate, was linear amplitude distortion due to the interaction of the amplifier and the loudspeaker impedances.
While there were audible errors in the lower and middle frequencies in certain tube amplifiers and a few specialized, low feedback solid-state amplifiers, (due to higher output impedances - poor damping factor - at low frequencies), this was not too surprising, as this issue had been reported previously, at least anecdotally, if not by way of rigorous controlled testing.
What was more interesting, were the number of amplifier loudspeaker combinations that generated high frequency coloration, due to the amplifier output impedance, at high frequencies, causing high frequency amplitude errors.
Many amplifiers have substantially worse damping factors at the highest frequencies, relative to the lower frequencies (below 1 kHz) and due to phase angle and impedance variations, exhibited in the loudspeaker, substantial response errors were observed. While most varied less than 1.2 dB, in some cases we found high-frequency response variations, between 5 kHz and 20 kHz, of over 2.5 dB!
As one can imagine, these levels of amplitude error, while in many cases subtle, are clearly audible, and repeatable with double blind testing.
It is interesting that there are still no industry standards for minimum impedance and phase angle. (THX and a few private parties have their own guidelines, but that is not recognized or enforced in most realms of audio).
It is easy for loudspeaker designers to be abusive in terms of impedance curves, as it can often serve their purposes to allow the impedance to fall to very low levels in exchange for some other advantage in performance, such as amplitude response.
Recently I was evaluating a new loudspeaker for a company, and I noticed that the impedance curve dipped to well below 2 Ohms at 285 Hz. They had done this because they had a dip in the amplitude response at 300 Hz and the only way to boost the response to flat in the passive-crossover, was to create an under-damped alignment in the high pass filter. The power spectrum in music program tends to be divided approximately equally above and below 270 Hz, and so the greater power demands are right were this loudspeaker had it’s lowest impedance, creating unusual current demands from the power amplifier.
They get the response they wanted and passed the impedance interface problem onto the amplifier designer.
Another popular loudspeaker, from the 1980’s, had a 0.9Ω impedance (and a large phase angle) at about 3.7 kHz. This caused a myriad of problems with amplifier interactions.
Amplifiers with any significant output impedance relative to the input impedance of the loudspeaker don’t operate any longer as an ideal voltage source, and can easily exhibit sonic errors because of this issue.
Depending on the set of conditions, audible problems can be caused by current limiting or by linear amplitude errors, or both.
As an aside, there is a story about our test procedure you may find interesting;
To isolate the audible effects during our testing process we used a number of approaches, such as straight wire bypass tests and double blind testing.
Some folks believe that results derived from blind testing is not valid, particularly when there is a null result.
During the time of our research, and still today, many audiophiles reject double blind testing as a procedure that causes stress that “interferes” with their psychological ability to make accurate assessments.
During our research project, we decided to test this notion. We had two amplifiers that both performed very well, but one sold for seven times more than the other one and was known as a very prestigious amplifier, having received rave reviews. We had two groups of audiophiles that claimed that they could always easily hear the difference between the two amplifiers, speaking rather critically of the lower cost unit, and glowingly of the higher cost amp.
During out double blind tests, each of them failed to consistently pick their favorite amplifier. So, we told them to take the two amplifiers, and listen to them under whatever test conditions they felt they were most comfortable, and to keep notes on what they perceived and which they preferred. We had approximately 20 people in two separate groups involved, all of them diehard audiophiles and critical of DBX testing.
Before we gave them the two amplifiers, we made one change. We exchanged the internal circuitry, so that the “cheap” amplifier was in the “expensive” chassis, and the “expensive” amp was in the “cheap” amp chassis.
We gave them the amplifiers and after they all had a chance to test them in the manner they were most comfortable, we retrieved the amplifiers and their written test results.
It may or may not be a surprise to you, that every one of them chose the amplifier in the expensive amplifier’s chassis as the one that sounded best, and had pages and pages of notes expressing their preference, and also expressing how easy it was to hear the difference and how the difference was quite large.
I learned two things from this experiment: 1) uncontrolled tests, that are not double blind, will produce random or erroneous results, and 2) even when shown this kind of evidence, it is not convincing for most of those that have already made up their mind that double blind testing doesn’t work.
So, I don’t think I made any converts, but it was an illuminating experiment.
Just to be clear, I am not saying that one cannot detect real differences between devices under test without a double blind test. I am only saying that it is important to verify any differences by way of double blind testing.
Actually, we have often been able to identify actual differences with double blind testing that are often not noticed in casual testing… but more on that another time.
Anyhow, so far we have covered two of the elements that can be of significance in explaining audible differences in power amplifiers, both of which are often overlooked;
- Audible clipping characteristics due to partial overload, and
- Wideband damping factor (Output Impedance) interface and interaction with the loudspeaker load
While there are some other issues that can affect amplifier sound quality, we can explain much of your original question with these two forms of distortion.
I can discuss further the specifics of these two forms and how they related to your original inquiry, and we can discuss other issues relative to power amplifiers in the future, if you wish, but I am delinquent on quite a few loudspeaker questions that I should answer first.
All the best,
- James
, eller?
Det är som jag ser det en fullkomlig omöjlighet att både ha en produktion och ha en produktion på hobbybasis.
Holy macaroni, the significance of my 'statement' is zero (maybe below) James. Can we please go back to the real substance in this thread, your eminent inputs and answers. 
