OK, we now have a term "muffled" to describe a negative quality in the subjective experience - could you possibly be more precise in what you mean by that, Dejan, and perhaps a few examples of recordings that fall into that category,?
Muffled as drawn back, unclear, lacking resolution, as if someone put something over the mike befre recording, making voices hard to understand.
(Side note: I searched for what could be put over the mike, and originally thought of a nylon stocking. I didn't say that because I expected you to argue logically and note that only one stocking was over the mike, where was the other? Why would a woman take off only one stocking? Was she hinting at something, and perhaps this caused the engineer's mind (possibly not only mind) to wander off, causing the recording to be muffled, although the session may have been sheer stardust? And if it was a stocking, what color was it? Did it perhaps have that sexy stripe on it, which was off center, cusing anomalies in the pure stereo effect? But I didn't, I thought it would be unfair to rattle your mind like that.
)
Last edited:
Quite so, despite the fact that both use internal combustion engines and wheels.
A 100 cc Honda motorcycle taught me that in very practical terms.
I was very astonished to read SE valve amps modify the sound so badly. The big shock is they don't. In fact I would argue that most hi fi removes the real life experience. That is Beethoven sounding as in real life. Like a garden of beautiful birds in song. Hi fi Beethoven is a soup.
Output transformers are bad news. Not so bad as to make it impossible. 1% THD is possible without feedback if SE. 8 watts is realistic as a maximum. You can go higher. My ears say it is not worth bothering as 20 watts is still low.
If you copy what is typical in SE you are possibly wasting your time. Maybe this is the problem? Those who could build good SE amps are too narrow minded to try. Those who would too limited in thought power to succeed.
My own experience of SE valve design is that 90% of what valve people say is complete nonsense. I was forced to build mine with good cheap parts of an industrial type. The PSU if I say it myself is cute. Although cheaply made it was apparent it considerably outperformed some giants of the SE type ( Ongaku in my opinion although I never had one at home ) . What point to change the modest BC 400 V caps if they work. What my amp has is virtually no components and even less in the signal path. Most of it's great tricks are from early transistor amps. They really work and suit purists as they seem not to be active transistorized solutions. I used active CCS to show me where to get improvements, then used the lost arts to get what I wanted.
What I learned is loudspeakers are mostly awful. SE amps must not be blamed for the awful speakers. Anyone who says differently I have to say I don't believe.
A/ You have tied this at home with speakers that suit.
B/ You have a love of acoustic music in a concert hall.
Output transformers are bad news. Not so bad as to make it impossible. 1% THD is possible without feedback if SE. 8 watts is realistic as a maximum. You can go higher. My ears say it is not worth bothering as 20 watts is still low.
If you copy what is typical in SE you are possibly wasting your time. Maybe this is the problem? Those who could build good SE amps are too narrow minded to try. Those who would too limited in thought power to succeed.
My own experience of SE valve design is that 90% of what valve people say is complete nonsense. I was forced to build mine with good cheap parts of an industrial type. The PSU if I say it myself is cute. Although cheaply made it was apparent it considerably outperformed some giants of the SE type ( Ongaku in my opinion although I never had one at home ) . What point to change the modest BC 400 V caps if they work. What my amp has is virtually no components and even less in the signal path. Most of it's great tricks are from early transistor amps. They really work and suit purists as they seem not to be active transistorized solutions. I used active CCS to show me where to get improvements, then used the lost arts to get what I wanted.
What I learned is loudspeakers are mostly awful. SE amps must not be blamed for the awful speakers. Anyone who says differently I have to say I don't believe.
A/ You have tied this at home with speakers that suit.
B/ You have a love of acoustic music in a concert hall.
I do, and to me it is a reference sound to compare reproduced sound with, though it is difficult to achieve anything close to live orchestra sound.
SE ended amps have low "damping" power. That's why they are more suitable for crossover-less full-range drivers. It is understandable that many have said that "no-feedback is bliss". Feedback is only a way to gain some objectives (many may have said power and distortion but I prefer "damping"), we don't need the feedback itself.
I would have thought feedback was very useful in improving stabiity, especially with awkward loads?
And yes, NFB does lower the output impedance. thus improving the damping factor.
I do not believe that simply using global NFB is a problem in itself, the problems start when one uses GNFB to cover up his shortcomings elsewhere. I own integrated and power amps which use unusually low NFB factors, just 12 dB, or just 3:1, while still having very good classic specs, such as THD, IM, etc, and still have very reasonable voltage slew rates. Obviously, they (Harman/Kardon) are using GNFB only to polish up what is already good, while keeping excellent current delivery capabilities.
Oh yes, and they don't need output transformers.
Good question thank you. I was meaning THD measure of the amp.
I think single ended amps and speakers are best considered a system as well as the music you play on them. The radiotron book recommends as even a load as possible for single ended pentodes. They don't recommend using them but they make recommendations if you do. So a mated pair of amp/speaker seems best there as an extreme example of non interchangeability of speaker and amp.
I agree with what you say about the speaker. I think the speaker (placement, enclosure, and driver) has the greatest potential variables in sound quality given the same source material.
People should try experimenting with damping factor by inserting a 1ohm resistor, (for example) in series with a speaker and alternatively short it out with a clip lead to compare the difference. You might find an IMPROVEMENT in sound quality. I found that out about 50 years ago, when I put a 3.2 ohm resistor in series with my AR-1 loudspeaker, attempting to emulate a DF of 1. I did not like the power loss across the resistor, BUT it did sound better, so I kept it.
About 25 years ago, I put a 0.5 ohm resistor in series with my Otala and a WATT-Puppy combination. I had to add it, because the amp would 'sing' at 2KHz if I didn't. No real harm was done. In recent years,(almost 20) I added 0.3 ohm in series with a Parasound, as this was part of a Bybee Purifier and things sounded OK. Today, I don't use a series resistor with Sasha's and it still sounds OK. Let's be practical. Try it yourself.
About 25 years ago, I put a 0.5 ohm resistor in series with my Otala and a WATT-Puppy combination. I had to add it, because the amp would 'sing' at 2KHz if I didn't. No real harm was done. In recent years,(almost 20) I added 0.3 ohm in series with a Parasound, as this was part of a Bybee Purifier and things sounded OK. Today, I don't use a series resistor with Sasha's and it still sounds OK. Let's be practical. Try it yourself.
Achieving very low output impedance in an amp is composed of 2 parts. Of course, negative feedback lowers (apparent) output resistance, but that is not all. First, you have the OPEN LOOP output impedance which is a function of the CIRCUIT TOPOLOGY. Voltage followers are best, and bipolar transistors are champion. However, IF you use a bipolar output stage, you MUST drive the final stage with a VOLTAGE DRIVE, and this is not so easy. Many will use a triple Darlington and this can work well enough. I personally use a complementary power mos fet driver that drives the bipolar output stage. This achieves pretty low output impedance without feedback, so added feedback is just a 'bonus'. Now, does the combination that I use make the best amp possible? It will give you great specs, but in truth a constant resistance over frequency appears to do just as well.
Last edited:
The trouble with series resistors is not power, simply use MPC type of metal oxyde 5 or 7 Watt types, if necessary with two or three in parallel, but finding the just right value for your speakers takes time and a reasonable stock of relatively high power resistors at hand.
And once you do get it right, that's good just for you, elsewhere it may be all wrong.
And once you do get it right, that's good just for you, elsewhere it may be all wrong.
Gee, Dejan, you're implying I'm argumentative - me, the nicest guy in the room!
Thanks for that clarification - this sounds like digital line level behaviour is not 100%, not enough low level detail is being reproduced well enough for the threads of the sound to be clearly separated. "Good" CDs sound good in this situation because the levels are carefully engineered so that all the happenin' stuff stays above this lower level, one is never aware that that there may be issues in the quieter regions.
Even though you may feel otherwise, my experience has always been that the information is there, except that there is too much blurring in the replay chain, making, for example, the voices hard to understand. That blurring is a defect of the audio playback system, I have proven this over and over again to myself, when a "thick" recording unravels and myriad strands of low level sounds emerge from the soup, and all make sense - I can choose to "listen" to just one of those sounds and follow it through the recording; like having a comprehensive multi-track machine, and just punching the Solo button to hear the contribution of one element.
- Status
- Not open for further replies.