• E3443.png

    WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.​

How much feedback is too much feedback

well, the subject says it all.
I'm just wondering how much feedback that it's normal to use in tube amplifiers.
I have no sense with how much feedback I should use in my poweramp, of course I could experiment a little bit and see what I find best but I would also like to hear what you have to say about it(I know that a lot of you prefer no feedback at all;) ).

The first amp I made I used variable feedback, and I could even turn it on and off, and for that amp I often preferred to have little or no feedback at all. I felt that the music became more alive with low amounts of feedback, but for larger amounts of feedback the amp became more "correct".

The only thing I can say I know about the amount of feedback applied is about opamps (and a bit for transistor stages, although not used for audio). Opamps have a large open-loop gain so when resistors is applied the amount of feedback will often be pretty large.
Yes, SS amps typically use more than 60dB (dBV I presume.. but no one ever specifies, damnit! :mad: ), but they have to cover their high distortion butts with it and get ungodly damping factors for why-I-don't-know...

Triode amps don't need much global because of the internal NFB in triodes. Pentodes need it to reduce distortion and increase D.F. The tradeoff, pentodes vs. triodes, is you get more power and less distortion from pentodes. People seem to prefer distortion for some reason. :clown:
In any case, more than 40dB should never be necessary for tubes. Typical is 6-10dB for a triode amp, and 20-30dB for pentode.


P.S. Frank: what's a penthode? ;)

A very old rule says 10 dB of “overall” feedback is the limit. And this rule still holds for tube power amps, whether it is a triode or a pent(h)ode ;) amp. Using more feedback does reduce distortion but blows out all the music of your amp, not to mention the risk of instability and peaking at the frequency ends (oscillations, ringing and/or “motorboating” ).

I myself stick without overall feedback for my UL pp amp.

In general I agree with you Peter, but as always it is largely a matter of taste. The local feedback of the Ultra Linear topology suits me best. It has some lower DF than real triodes, but impedance correction of the speakers with a zobel sweeps out much of the importance of DF. My speakers have a 16 ohms hump at 1.8 kHz, which makes them quite forward sounding with most tube amps. The zobel corrects for this.

I am planning to use the Plytron/Amplimo 4070-CFB OTF’s for a new amp. It uses cathode feedback for the output tubes. Let’s see what this will bring. Maybe you are right and do I gain nothing from it.



diyAudio Moderator Emeritus
2002-10-24 10:19 pm
Peter's answer is pretty close to mine. It depends on the amp and also on how you add up the feedback. For example, do you count cathode degeneration? I've got source followers driving my output tubes, do I count their degeneration? After all, that's feedback.

If you've ever had a chance to listen to some Futterman OTLs, preferably with Quads, you've listened to some of the highest feedback tube amps ever made. Yet I don't hear too many complaints about their sound.

What I personally would avoid is amps with moderately low feedback, say 6-10 dB; that's the point at which you've extended the harmonic spectrum (see, for example, Baxandall's series of articles in Wireless World), but not really knocked down the distortion to super-low levels.
I did actually think about global degenerative feedback, taken from the OPT secondary and fed to the first tube.

Well, I know that grounded anode stages have a lot of serie-shunt feedback, without it the grounded anode would be rather useless.
and the same with many OTLs, but I didn't express myself good enough.

The reason why I asked was because I doesn't have much sense about how much feedback tubeamps "normally" use.
Come on SY let make things not too complicated. As far as I understand Gunderz’s question it is about tube amps in general. So let us stick to general OT coupled amps and feedback in a sense of overall feedback (usually from output to the cathode of the first stage). In that case the “10 dB rule” is still a valid one.


Edit: Oops Gunderz, we did post at the same time :)


diyAudio Moderator Emeritus
2002-10-24 10:19 pm
Well, that leaves out classic amps from Marantz, Audio Research, and a few hundred others that run more like 20+ dB.

10 dB as a rule was more typical in the Williamson era, when iron wasn't as good as it later became. At that level, you're only knocking down distortion by a factor of about 3, while significantly increasing the order of the distortion. In reality, if you've got a good margin, there's no reason not to use more.
SY said:
10 dB as a rule was more typical in the Williamson era, when iron wasn't as good as it later became. At that level, you're only knocking down distortion by a factor of about 3, while significantly increasing the order of the distortion. In reality, if you've got a good margin, there's no reason not to use more.

Well there is a good reason to use no overall feedback at all, despite the lack of distortion reduction and lower DF and even with modern high performance output transformers. This is when it comes to overload behaviour, especially with modern high dynamic sources like CD/DVD-A. In another thread I already posted a link to this article:


Although it is written almost half a century ago, most is still worth noticing.


diyAudio Moderator Emeritus
2003-01-18 7:57 am
Near London. UK
It depends a lot on the loudspeaker. Loudspeakers are high-pass filters, and require a low source resistance to set their Q correctly and achieve the designer's intended bass response. This low resistance can only be achieved by feedback.

However, there are some exceptions. If you put a put a loudspeaker on an open baffle, a triode amplifier often has the correct output resistance without needing any feedback. Similarly, loudspeakers that are very efficient (horns) or have very light moving parts (ribbon tweeters) do not rely on electrical damping.

If the amplifier has an output transformer, you can't really apply more than 30dB of global feedback without risking (usually HF) instability. As for setting a number between 0dB and 30dB, I wouldn't like to set any hard and fast rules. I know that's not much help, but that's the way it is.

Sch3mat1c: Feedback is simply the ratio by which gain has been reduced, so it is just dB. dBV is a ratio relative to 1V, and therefore specifies an absolute voltage.


diyAudio Senior Member
2002-08-21 1:20 am

And I think the moderate amounts of NFB are responsible for a lot of the particular 'sound' of many current and older PP amp designs.

Wouldn't you agree that a well designed PP does NOT need any extra global FB?

All in all I'd much prefer to use local NFB than Global NFB, there a few handy tricks that allows us to avoid it provided we design with linear amplification blocks in the first place.

I'm aware of another thread running somewhere else on this forum covering a related topic and am still shocked by some statements I read there...engineers and music?

It confirms to me that (U.S) engineers are extremely field limited, but that's about all I'm going to say about it.

Cheers, ;)

P.S. The shock was laughter induced...I'm slowly recovering but my jaws still hurt.