|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
77.43.10.3
I have designed, built and tested several tube amplifiers (from 4W to 40W, OTL, PP, for hi-fi and guitar, all with global feedback applied over 2 or 3 stages).
In my experience I have always found that some global moderate feedback helps to get the best overall performance of a tube amplifier (I mean with the particular speakers that I use in my listening room) and to stabilize performance against components ageing.
However I see in the asylum that there are different opinions about the use of negative feedback.
Many inmates are quite strongly convinced that the quest for zero feedback is a must for every serious audiophile.
This is the reason why I propose this call for the survey of the opinions (and or the experimental findings) of all the interested inmates.
A speculative byproduct of this survey might also be the definition of some novel measurement technique able to better correlate laboratory measurements with the results of live listening sessions.
ecc230
Follow Ups:
This is a subject that has attracted many opinions, as you'd be aware. The extreme view that "all NFB is bad" is an unproven generalization to which I wouldn't pay any heed. Sometimes it helps to use NFB and I'm not afraid to do so in such cases.
While effective in reducing overall THD, loop feedback actually enhances the odd-ordered harmonics used by the human ear to detect the volume of a sound (the 5th, 7th and 9th). Our ears, IOW, are so sensitive to these harmonics that enhancement (distortion) of them by only 100ths or 1000ths of a percent is audible. This was proven by General Electric in the 1960s.
Additionally, there is a myth that negative feedback is supposed to stabilize amplifiers, but it is actually a destabilizing effect. In fact **positive** feedback has the stabilizing effect: once the amp is in oscillation, its behavior is considered stable (oscillating).
With negative feedback, the behavior of the amplifier is considered non-linear, when subjected to a constantly variable input signal and non-linear load (typical of most speakers). Chaotic behaviors can be observed: bifurcation of the input signal and other non-linear effects.
Additionally, each loudspeaker has its own critical damping value of output impedance in the amplifier that is not served by a set feedback value. In fact, there are no known examples of a speaker that benefits from damping factor values over 20:1, but there are speakers that benefit from damping values as little as 0.1:1. If the speaker is overdamped, it is much like trying to move the cone through fluid rather than air; the cone will see less excursion and bass response can be down by as much as 8 db as a result, in addition to loss of transients.
see:
"Critical Damping: Missing Link in Speaker Operation Parts 1 & 2"
at:
http://www.pearl-hifi.com/06_Lit_Archive/07_Misc_DownloadsMisc_Downloads.html
also:
If the speaker load isn't a perfect flat 8 ohms, without resonances and especially has areas of capacitance, then the amp has to fix it up with a band-aid to cover the weakness of these speakers. That would be NFB.
A perfect speaker would also be linear and need 1 watt of power for maximum needed volume to keep an amp linear without NFB. Well the speakers are usually not that good, yet some very costly horn systems do approach this ideal.
So what do you want to compromise on? An amp that's compromised and fits poorly designed 2 ohm minimum impedance compromised speakers and buy both compromised components, or an amp that's almost ideal with speakers that are almost ideal and buy both. It takes years of experimenting and spending vast dollars to understand this concept, and then only to find nothing is really perfect as something is always flawed. :-)
-Kurt
I enjoy committing audio heresy myself and have done so by applying Feedback to my DIY 2A3 Single Ended stereo amplifier.
A little bit of global feedback on this amp was able to tighten up an otherwise bloated and pretty poor Bass performance.
Just like too much of anything is bad for you, a little feedback properly applied can improve the measured performance of an audio circuit.
Yet some critical ears seem to get annoyed by the use of Feedback if it is really not necessary.
Like a spice, feedback is only a part of a larger whole. The end results seem to indicate that it's not the villain some claim it to be.
You should do a search for this. As a topic, this one has been thoroughly discussed in the Asylum.
After reading many of the past posts on the subject (most of them are more than 5 years old) I understand why no one will likely expose again himself by repeating his opinions on this matter.
Once you had get burned by hot water you fear cold water as well.
It clear to me now that the feedback/no-feedback clash will go on forever (unfortunately with no benefit to the open minded audiophile).
I have realized that the research for some better experimental technique for correlating the results of subjective listening to the quantity of feedback applied to a particular tube amplifier has been blocked by the vendors lobby who are not interested in the possible scientific results but rather to use the feedback argument as a sales tool only.
All audiophiles should oppose to this status of things and I will certainly do.
ecc230
"research for some better experimental technique for correlating the results of subjective listening to the quantity of feedback applied to a particular tube amplifier has been blocked by the vendors lobby"
That's a pretty strong statement. I wonder what the evidence is to substantiate this assertion.
Personally, I believe that with high efficiency speakers and a reasonable impedance curve, triode amps don't need feedback. Pentodes need feedback to lower the output impedance.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: