Figuring out negative feedback

[rgorke]

So, I am trying to tighten up my low end. I had 100k/4 ohm. Lots of gain but flubby. I went to 47k/16 ohm and started getting this howling noise that changed pitch with the presence pot.

So, I went to 100k/16 ohm that tightened things up but maybe not enough. Now at 47K/8 ohm and things seem ok.

So, does it go 100k/4ohm > 100k/8ohm> 100k/16ohm > 47k/4 ohm > 47K/8 ohm > 47K /16 ohm?

Or is it not that simple?

Thanks

Yes, I have read the Aiken page but I need it dumbed down a bit.

[vh junkie]

Basically, all other things being equal, the tap and FB resistor determine the amount of feedback.

The amp puts out a larger signal as you increase the tap used, so FB increases.
4 ohm tap = least FB
8 ohm tap = more FB
16 ohm tap = most FB

A larger FB resistor retards this signal, while a smaller one lets more through
larger FB resistor = less FB
smaller FB resistor = more FB

So you can use the tap and resistor to determine the amount of FB. You should also be able to just use one tap (e.g. 8 ohm), and create the same amount of FB just by varying the resistor

Starting with the ubiquitous 100k/4ohm tap, this yields -7.5db if NFB, then extending to different values to achieve the same result:
-7.5db can be 100k/4ohm tap
-7.5db can be 150k/8ohm tap
-7.5db can be 212k/16ohm tap
These are all equivalent... and very loose

Tighter (more FB):
-10db can be 64k/4ohm tap
-10db can be 92k/8ohm tap
-10db can be 134k/16ohm tap

Tighter still (yet more FB):
-12db can be 45k/4ohm tap
-12db can be 66k/8ohm tap
-12db can be 95k/16ohm tap

Tight! (tons of FB):
-14db can be 32k/4ohm tap
-14db can be 47k/8ohm tap
-14db can be 70k/16ohm tap

So you could just pick one tap (probably 4 or 8 ohms) and just adjust the resistor to achieve the required FB.
A lot of other factors can come into play here. These example are for EL34s with a OT primary of 1700 ohms.

[rgorke]

Thanks!! That should be a sticky somewhere.

That is perfect. So, I am at 100k/16 ohm which is roughly equivalent to 47K/4 ohm. If I were to go to 47K/8ohm, it is too much NFB for my amp. I will stick with 100k/16, seems to be tightening up my looseness.

[stef]

why don't you try 27k-16ohm (that NFB setup was most likely in Ed's amp back in 1977) or 47k -8ohm

[rgorke]

why don't you try 27k-16ohm (that NFB setup was most likely in Ed's amp back in 1977) or 47k -8ohm

Because I get a howling when I dime the amp. If I get howling at 47k/16, I am bound to get it at 27k/16, no?

Seems like too much feedback is causing a bad reaction somehow. Not sure the cause but at 100k/16 is about as much as I can go.

[ampSnob]

If your getting any squeal or howl in any config you may want to consider moving some wires like the purple NFB, or B+ wires with a chopstick or something non-conductive and see if there is a magic position that cures the squeal. read on for why..

.... If I get howling at 47k/16, I am bound to get it at 27k/16, no?

It's the other way around. Negative feedback injects the 'opposite' signal back into the circuit, the idea being any irregularities in the audio caused by the amp after the NFB input point (the phase inverter) will be themselves be inverted and subtracted from the audio when it comes back in making the output more 'true' to the original signal without unwanted coloring from the amp. A design engineers magic wet dream cure-all, but of course we actually want the sound of that stuff (otherwise why bother with a tube amp in the first place) so using too much NFB can suck the life out of your tone. One of the side benefits of NFB is it can tame squeal because as soon as it happens the NFB circuit adds the exact opposite of it (subtracts it) from itself eliminating it. The better way to fix that problem though is through wire layout.

One more piece to the puzzle: a cranked presence control shorts some of the high end of the NFB out to ground before it can make it back into the phase inverter circuit. That is to say, with the presence all the way up, you are only negatively feeding back lower frequencies, therefore cleaning up the bass (tightening) and making the presence seem like it's just another treble control. With less high frequencies making it back through the NFB signal to cancel themselves out, it becomes unable to cancel out its own squeal. Therefore again, fixing the wiring is the best option as many people like the sound of the presence up if they can do it.

A lower value resistor lets more signal back in (more negative feedback comes back) which would usually mean less squeal/howling so you should be even safer from squeal/howl with 27k/16 if 47k/16 works.

[vh junkie]

While, It seems unlikely... because you probably would have stumbled on it before now, there is one thing that could cause the howling WITH the NFB in place:
Reversed OT primaries at pin3 of v5 and v6

If the primaries are reversed the NFB does not provide cancellation of problems in the main signal, but rather reinforces it.

[rgorke]

I've never had this problem before. Here is what it sounds like. The first change in noise it turning my PPIMV from 10 to about 8.5 or 9 back to 10. The change in pitch is turning the presence control. It seems to happen less when the NFB is less (more toward 100k/4ohm).

[stef]

more NFB (27K-16) = less feedbacking (howling, sqeeling noise etc)
move the pruple wire or do what vh junkie posted, because there's something wrong IMO

[dirtycooter]

TOTALLY A STICKY THREAD.......SHABAAAM!

[rgorke]

So, I have taken a major big boy step and removed the PPIMV that was in my amp since I got it. I now have a killer attenuator and the PPIMV has been dimed for the last month or more. When I took the MV out and put in the 47k NFB resistor, howling is gone. Not sure why but don't really care.

Also, I may start another thread on this but the amp does sound better without the MV in it. Maybe I am hearing the upped NFB but there is more clarity. The MV was in the indicator light hole and the indicator had a new home to the left of the switches. I am thinking of just putting a dummy switch in there to pretend to be a polarity switch.

If anyone has a busted on/off/standby switch they don't want, let me know.

I haven't had much time to play it but I am liking how it sounds now.

Thanks for the input and the NFB info that vhjunkie posted must be a sticky!!

[Ralle]

Just like to say why don't you try the 27k/16 or 8 ohm combo?
I allways tend to get a kind of thin spiky tone with anything higher than 27k... it ALLWAYS ends with me putting the 27k back in... If I want more gain or fatness and still keep the whole tone, I work with the taps instead... if that's not enough, the problem, or gain issue lays somewhere else...
Increasing the tap value isn't the same as increasing the resistor value... eaven if the two are basiclly in the same possition in the circut... I think it has something to do with the sound getting better if there's as little resistance as possible in the NFB circut... it's better if it's controlled with the OT's taps... With a small resistor the OT detirmines how much is feed back into the PI circut... the resistors value detirmines how much the OT will feed back...

Please correct me if I'm wrong...

[ezs]

Ive seemed to have a similar situation with the PPIMV in a build with 2 masters. Seems like theres really no reason for it being there if you have the 2203/04 type. I suspected instability and squeal having it in there, I used a closely matched Alpha too, clean install. Maybe in a 2 stage amp it works well but not right type a 3 or 4 stager imo. That being said, Ive never had any trouble with the NFB wire lying in any position or in any combo of res/taps, seems like a stabil part of the circuit. Glad you found it.

[rgorke]

bumping this back up to the top, I still think what vh junkie wrote should be a sticky somewhere.

[rgorke]

Bumparooski!!! Can we get a sticky for this? I keep referring to this thread. I am down to 27k on the 16 ohm tap with my selector on 8 into a 16 ohm cab.