possibe stupid question

[iknowjohnny]

but here goes. Has anyone here or any commercial designs used NFB that is sent further upstream than the PI, and if not why can't it be done?

[rp]

I think some hifi amps can use local feedback at various stages, but my technical knowledge is limited. There was a rebellion against this in the mid '80s probably cause the amps were getting too cold and sterile and simplicity came more into vogue. Others here will know more.

[iknowjohnny]

right, i've seen that in preamps, (i think trainwreck or dumble uses that) but i was curious if anyone had and if it was possible to take feedback fro the output just as is normal but bring it back to the preamp instead of PI.

[tubeswell]

AFAICT you can put NFB in wherever you want, but if its earlier in the pre-amp, you risk more oscillation (because more stages come into play in the loop), and you loose some of the early raw colouring of the critical 1st stage or two.

Either way you have to make sure the feedback signal phase is negative w.r.t. the respective signal phase at the insertion point.

Also remember that you will need to change the voltage divider in the NFB if inserting it earlier in the pre-amp, where the signal voltages will be lower. If you don't get an appropriate amount of NFB, the amp will go into oscillation more easily.

In addition, you can either feed it back to a grid, or to a cathode-biased cathode. But it its at a cathode, then the cathode will need to be un-bypassed either totally, or to at least some extent, for the NFB to have any effect.

[passfan]

HiFi amps have done it for years. Some even employ what is known as local Feedback for each stage taken from the ouput of each stage. In HiFi applications it is used more to control linearity than anything else. Eliminating distortion is the goal there and that might conflict with what one is trying to achieve with a guitar amp. Then again, maybe not. Look here for an in depth answer to this question and hold on because these guys are good with the theory and friendly too http://www.diyaudio.com

[rdjones]

It's not a bad question.

Yes, I'm aware of several Fender designs that feed the NFB back to the stage previous to the inverter. The push-pull version of the ('50s) Harvard, several of the Princetons, the ('80s) SuperChamp and my master volume Champ II all do this. Interestingly, I think these all have cathodyne splitters rather than the LTP. Even more curious is that the Champs have their master volume after the stage being fed the NFB.

There's also a few designs by Univox that use this configuration, one which has a pentode driver being fed NFB to it's cathode.

rd

[Cliff Schecht]

Those SuperChamps are hella cool amps and (I believe) were the last non-custom shop point to point amp Fender made. I've heard two and they both sounded phenomenal and put out a lot of sound for a small package.

[rp]

Those SuperChamps are hella cool amps a

The late Robert Quine, a great great player, swore by the original super champs. He had a great sound. It might just be the last collectible amp.

[Cliff Schecht]

Yeah unfortunately they cost a small fortune now. Last time I looked they going for $1000 minimum and I doubt that will change anytime soon.

[Jana]

If memory serves me, the Marshall Major fed the NFB into a stage earlier.

[iknowjohnny]

Thanks. I never noticed these before, but i may never have looked closely t any of those. Is there a reason i wonder why they all feed the cathodes instead of the grid?

[rdjones]

Thanks. I never noticed these before, but i may never have looked closely t any of those. Is there a reason i wonder why they all feed the cathodes instead of the grid?

It has to do with the impedances involved.
Remember, we are taking the NFB feed from the lowest impedance point in the amp, the speaker output tap of the output transformer.
The grid of a typical 12AX7 stage is a megohm or higher. The cathode is more like 1.5K or thereabouts.
Then there's the polarity of the insertion, it must be negative _not_ positive.

rd