feed back etc?

[steve s]

As a relative novice in amplifier design, and from a prespective of simple is best i have been reading some of my old books. one quite topical is the about the mullard high quality amp..
the book in question and that i refer to is is called high quality sound reproduction and the section is a reprint of a wireless world artical by w a ferguson
the reason i have posted is i am interested in why no feed back amps sound better and feed back amps can have better bass .. and why he differences in the sound of feed back and zero feed back

i think this artical explanes it..

the more stages in an amp the more phase shifts you get...
when you add feed back you get more phase shifts .. could it be these phase shifts that change the sound..you know what i mean by that...

it is also very difficult to measure phase shifts ( so are they not measured?)

it is very hard to apply feed back over the full frequencey range so most applications it is at the mid frequencies.. so I would deduct that the better bass discribed by some, is may be due to less/ no feed back at the low frequencies..?? think less mids turn it up a crack = more bass

so is it the 'difficult to measure' phase shifts that (spoil) change the sound.. and the easy to measure distortion, that is corrected, maybe through the phase shifts

remember that most amps where built for the masses.. thats not me..


the last point was about cathode by pass caps, i found on driver stages quite an improvement with out them.. so i was suprised to see part of the text and graph showing the difference between the two on the output valves

if there is an increase in output transformer impedance on a triode connected el34.

the by pass caps could be dispenced with, and there was a improvement in sound quality (and still achieve full triode power) , the graphs show this quite well...

an increase in output transformer impeadance would give a better quality and lower distortion in any case. but the curves are quite different beween the two on the graph

any comments. or more knowledge on the subject

I have not gone into all the detail.. but later in the week i will be removing the by pass caps on the pt15 .... and as i don't use feed back.. well

cheers steve

[vitalstates]

Hi Steve

I dont know anything about what the various amp designs do to phase exactly but I do know that its super critical when recording. From experience I know that if you record one sound source, e.g an acoustic guitar, with two mics(which is common practice) it is super critical to get the distances from instrument correct. Get it wrong and the worst case is that the two phases cancel each other out and no sound is actually reproduced. This is the most eerie phenomena when you hear it for the first time. The 'not so worst case' is that the focus disappears and its not possible to mix the instrument precisely into the sound stage.

I would imagine, although I'm not certain, that if an amp messes with phase the same blurring or attenuation will take place.

anybody else?

Ed

[petercom]

This is too simplistic a view I am afraid.

Phase shifts are part and parcel of amplifier design and any good designer takes them into account when working on the topology of a circuit including the feedback loop.

In the case of the forthcoming WD88VA, which has to use a modicum of feedback because of the anode impedance of the KT88s, the time constants of the complete circuit are first carefully considered in the Spice model and then 'tweaked' in the prototype to make sure they don't cause any undesirable effects such as instability and poor transient response.

[IslandPink]

Some good discussion of feedback in Lynn Olson's site in the design philosophy sections and the Aurora/Karna amp pages :

http://www.nutshellhifi.com/triode4.html

http://www.nutshellhifi.com/triode1.html

There are other effects considered here - extension of distortion products and intermodulation distortion . There a lot of questions that arise when the output transformer ( or any transformer ) is part of the feedback loop . On the other hand , local short (device) feedback loops can be more successful .

Mark

[dave dove]

hmmm

dave

[steve s]

Thanks for the input.. that 1st link is very interesting mark, the 2nd link, although valid seemed more of a justification for taking that design route.. again me being simplistic... very like the mullard booklets i have?


but .. why does feed back change the sound in the same amp... there are the time delays or is it the phase shifts, or something else, and if phase shifts are hard to measure,. but surely still measureable,

mullard seemed to know there stuff, but they discribe a 3 or 4 stage amp with feed back over the full freq range to be quite a formidable task.. and that it is usual to find feedback decreased at upper and lower frequencys... but have just enough to provide sufficent damping...

I'll stop rabbiting on...

cheers steve