NFB Switch

[TIMBO]

Hi guys,Just wanting some ideas on having NFB resistors on a rotary sw so as to be able to dial in the feedback. I don't fully understand how the resistor is calculated for that circuit or is it a case of that resistor is a good amount of feedback for the required sound,if that makes any sense  If this is feasible what values would work?

[Platefire]

Why not just use a pot and dial it in that way?

[TIMBO]

Tone Junkie said the same, thought that this might be something that could be more precise to the required feedback 

[jojokeo]

I've seen this (a NFB pot) installed and it being called a "gain" control. Start w/ a low-end value resistor btwn the pi or speaker jack and when the pot is full counter clockwise (or wired the opposite way) it won't have zero resistance btwn the two. And if the pot is large enough then it will have high enough resistance to sound effectively out of circuit when fully turned the other way. Depending on the sound or range your looking for and values used you may want a 10k pot or 100k?

[LooseChange]

It will have DC on it and it will make noise. Experiment with a few values. Add a switch to remove.

[Tone Junkie]

I like a 27k resister with either a 100k or 220k pot depending on how high gain your going . yes it has DC on it but once you tweek it you pretty much leave it . i like the 27k to start with old fender values I think. This gives a nice cleaner tone lots of NFB all the way up to splawn quickrod at 220k which doesnt give so much nfb.
If you dont like pots.
you do paralleled resisters on a switch to get your differant NFb, set up your switch then use the 220k pot to dial in what you like then measure the pot to get your two favorite put the highest one strait then parellel your other resister to get your other favorite, if that makes sence.

For me I like the 220k pot gives you alot of freedom. but you have to be carefull to high a resister or pot and you find your instabilities quickly what I mean is oscilations that might be hidden with a 47k will show at 220k.
Having said all that put your stuff on a rotory switch try it before you drill a hole you probably will find 1 or 2 that you like the rest of the switch will be a waste. ultimatly my friend your ear will tell you.
A pots quicker to find the sweet spot though.
Thanks Bill

I must be getting tired Im starting to ramble 

[The_Gaz]

Tone Junkie, I think I understand you, but I would be more concerned about an oscillation arising with too much feedback, not too little. If one has an OSC going on with an open loop (no global NFB) amp, then I think there are other lead dress issues that should be attended to. On the other hand, too much feedback can cause incurable oscillations, especially motorboating, that can only be fixed by reducing the amount of NFB or slugging the dominant pole as Merlin (Valve Wizard) discusses in his preamp book.

You can get rid of the scratchiness on the pot by adding a large series blocking cap (1uf or higher) with the pot. I'd actually not recommend it though because it adds another phase shift in the feddback loop which can  make the amp more likely to osciallate - not always though.

[jeff]

The schematic shows all 47K resistors so all positions except for the first would sound the same.

 You could wire it that way with different values for each position, but for ease of wiring I suggest wiring resistors in series from lug to lug on the switch.
Use a "make before break" switch to avoid volume popping.

[jjasilli]

The usual NFB loop works by supplying voltage to the cathode of a small bottle tube.  (Usually the PI if there is one, or the last preamp tube driving SE power amp.)  Note that only voltage is important; there is almost no current.  Therefore the size of the series resistor to limit current is not a factor.  The NFB loop works as voltage divider.  Hence, in theory, the ratio of the series:shunt resistors is what matters; not the absolute value of those resistors.  E.g., for pure voltage dividing purposes, a 26K : 56K pair is as good as a 2K6 to 5K6 pair.  Except that in practice, the NFB shunt resistor does double duty as (part of) the cathode resistance in the tube accepting the feedback.  So the NFB series resistor must be sized to yield the proper divider ratio with the cathode resistance.

The series resistor can be made variable as set forth in the posts above.  Theoretically the shunt resistor could could also be variable.  But this is not done, because it would also change the bias of the tube!

What amp??? If the tube receiving the NFB has a cathode bypass cap, then split cathode resistors are needed to keep the NFB signal from bleeding to ground through the cap.  See VibroChamp:  http://www.el34world.com/charts/Schematics/files/fender/VIBRO_CHAMP_AA764.pdf

[TIMBO]

Thanks for the reply guys, The idea behind this post was just to be able dial in some nice to really full on feedback but this is only govern by what will sound good and that might only be a couple of values of resistors. I 've done a schem that might do the job and the rotary switch is set up to give between 2-6 positions so i can adjust the number of values that are a better sound. The values on the schem are what i have seen on other schems so i'll use these as a starting point.Thanks

[jeff]

The values on the schem are what i have seen on other schems so i'll use these as a starting point.Thanks

It's not just a matter of 27K VS 100K.

Notice also on those schematics what OT impedance is being used.
Fender used 27K, 56K and 100K NFB resistors on different models. But check out some Fender schematics and notice the value of the resistor VS the number of speakers. I think Fender used different values on different amps not to change the amount of NFB but to keep it the same from amp to amp depending on how many speakers the amp used.

If you have an amp with 4,8,and 16 ohm taps the voltage will be different at each tap. So a 27K with a 4 ohm OT won't have the same amount of NFB as a 27K with a 16 ohm OT. Theoretically instead of changing resistors you could use the same resistor and change the amount of NFB depending on which tap you took it from.

I know you're experimenting with the amount of NFB but it's something to keep in mind if you're ever building a stock Bassman(2 ohm) but want to use a 8 ohm transformer. Change the OT, change the NFB resistor too to get the same sound.

[TIMBO]

Hi jeff, The ot is a standard 18w and i usually just use the 4 and 8 ohm taps with the NFB connected to the 8ohm tap, did not know that connecting the NFB to a different tap would give a different sound, what would be better? I like the idea of the resistors in series but it makes it harder to do the trial and error thing to find the right combo.Thanks

[jeff]

I'd use the 8 ohm tap. You don't get a better sound using the 8 ohm tap VS the 4 ohm tap. By using the right value resistor you can get the same voltage from either tap.

 I wasn't suggesting switching taps, just pointing out when you compare schematics, it's part of the equation. The important thing is the voltage you're sending to the PI no matter which tap you use to get it.

Fender used 27K with 2ohm and 56K with 4ohm. I think(have to check) 100K/5K(presence) is usual for an 8 ohm tap so center around that. Using the 27K on the 8 ohm tap will give you a lot of NFB. Good for experimenting.

I think you're on the right track experiment with a switch to find what you like. I just built an amp without NFB. It sounds great but you got me thinking.
I might try this out too.

[TIMBO]

Jeff, I'll think i'll start with these.

[jeff]

Looks good to me. Let me know what you find.
Only thing I'd change is the position of when it's disconnected from first to last. Only because the volume will increase as the resistance increases and when it's disconnected it will be the loudest. Instead of going quietest to loudest, right now it will go loudest to quietest to louder and louder.

[jeff]

I edited my last post because it was wordeded confusioningishly.

[TIMBO]

How's this, I don't think WORDEDED is A real Wheard.  

[jeff]

 
Another thing to consider is the presence pot is part of the NFB voltage divider. I've never seen a 10K presecne pot paralled with a 4.7K resistor. I not saying that can't be done but changing that changes the NFB. If we had a 56K with a 5K presence pot and used the 10K pot + 4.7K resistor for a smoother pres knob, to get the same NFB we'd have to change the 56K to a 36K.
     
          4.7k || 10K =  3.2K~
        56K/5K = 35.8K~/3.2K~

Again, doesn't really matter much because you're experimenting with values, but something to consider. If it were me, I'd try to play with the values to have a 36K position on the switch. I'm probally splitin' hairs, but that's just me.

By the way, I like what you've done with the preamp.

[jjasilli]

I'd use the 8 ohm tap. You don't get a better sound using the 8 ohm tap VS the 4 ohm tap. By using the right value resistor you can get the same voltage from either tap.

Yes.  Note again that the critical factor is the value of the final NFB voltage injected into the cathode of the receiving tube. Ea OT secondary tap puts out a different voltage.  The higher the Ohm rating of the tap, the higher the voltage.  If you double the rating of the tap, e.g. 4Ω > 8Ω, the voltage increases I think by 1.4.  Amp manufacturers tend to tap NFB off the highest Ohm tap, then divide NFB voltage down from there.  But you can source your NFB from any tap you want.  Then design the NFB circuit with the correct voltage division to get any particular, final, NFB voltage.  

So, if you have a multi-tap OT secondary, you can make NFB switchable by using the SW to select the different OT taps.  Then a pot, or alternate resistors, are not needed within the NFB circuit itself.  Some guys make the NFB source switchable with a Speaker Ohm selector SW. That way the NFB voltage automatically stays in sync with speaker voltage.

For objective purposes, NFB can cure the operational issues of oscillation, or of fizzy speakers.  Otherwise, the amount of NFB voltage to use, if any, is a tonal issue.  As such it is purely subjective.  There is no magic voltage or OT tap to use; except that one or more particular settings may sound magical to you.  

[TIMBO]

Thanks again for your input,it all sounds great but a little over my pay grade. I do tend to do the trial and error thing a lot and i try to do my builds to be able to tweak parts if needed.This project was presented by Tone Junkie awhile ago and it caught my eye as well. I'm not sure how this one will sound because it have changed a lot from the original.Thanks