Adding some switchable NFB

[Tyrannocaster]

I am enjoying this tweed Fender but for the hell of it I would like to add some negative feedback (switchable). The donor amp has a switch that I haven't used yet and this might be a good way to put it to work. Looking at the various Fender schematics that use NFB I see considerable differences in where the FB is applied as well as how much resistance is used. Since the 5E5 didn't use NFB and the 5E5A did, I thought that might be a good choice, but while the amps are similar they aren't quite the same. Plus, and this really complicates things, the tremolo that I want to add would be attached at the same place the NFB is and I don't know if that would make any difference.

The 5C3 Deluxe uses NFB that goes back to a power tube input, not to the PI input, and it has a whopping 1 meg resistor. I'm thinking this will be a process of trial and error but if anybody has a suggestion on what to do I'm all ears.

[sluckey]

Look at this...

http://www.el34world.com/charts/Schematics/files/fender/champ_vibro_aa764_schem.pdf

[Tyrannocaster]

See - yet ANOTHER place to apply it!  They are everywhere! I think I should just add a pot temporarily and try different settings and even different places to insert the feedback - trying to use discrete resistors seems like a huge waste of time via trial and error.

Thanks for that alternative idea.

[sluckey]

I figured as long as you're using the champ trem you may as well use the champ NFB too. It's a proven combo. Be prepared to swap OT primary plate leads in case you end up with PFB. It's a 50/50 coin toss.

[jeff]

Looking at the various Fender schematics that use NFB I see considerable differences in where the FB is applied as well as how much resistance is used.

That to me makes sense. Using different power tubes with the same value NFB resistor will result in different amounts of NFB so different amps will have different resistors. Also using different taps will give you different amounts of NFB. If you use a schematic that uses 6V6s and use 6L6s but the same NFB resistor you wont have the same amount of NFB. Similarlly if you build the circuit according to a schematic that uses a 4 ohm OT and you use a 8 ohm OT, but the same resistor, you wont have the same amount of NFB. If you compare two Fender circuits also notice the number of speakers it uses. The two circuits are similar but they use different NFB resistors because they also use different OTs so it doesn't necessesary mean the one with the lower resistor has more NFB.
 
 Here's what I see.

If you do use NFB, how it's normally applied, it is injected into the stage before the PI and you have added a master volume, So by turning down the MV you're also turning down the NFB. I don't know if that's a good or a bad thing but I did something similar using a MV before the PI(although I put a pot wired normally and another cap between the cap and the grid/1M of the phase inverter instead of using the 1M as a variable resistor).
I don't know how this effects the NFB. I too wanted to add NFB but didn't because of this and didn't know where to inject it.

Is NFB before the MV a bad idea?

BTW do you find that wiring the 1M as a variable resistor works well as a MV?

[Tyrannocaster]

Well, in this case I took out the master volume because it just didn't work worth a damn. This amp isn't so loud that it really needs one anyway.

I used the 1 meg pot to try out the 5C3 NFB though - that's the one that uses a 1 meg resistor and goes to the power tube. I can't hear any difference, although if I turn the pot down it eventually starts to oscillate. I don't know if that means I have the OT leads reversed or the idea is just wrong in the first place for this amp.

I also tried the Vibro Champ's circuit using a 5k pot (switchable so I could verify the difference easily) but it adds distortion and lowers the gain at the same time at all its settings - not at all what I expected. Maybe this wasn't such a great idea after all.

Going to try other configurations but right now I'm too burned out on the amp and need to wait until tomorrow.

[HotBluePlates]

I also tried the Vibro Champ's circuit using a 5k pot (switchable so I could verify the difference easily) but it adds distortion and lowers the gain at the same time at all its settings - not at all what I expected.

Did you add the 47Ω resistor under the driver tube's 1.5kΩ resistor? If not, you just added an extra load to the OT secondary, rather than make a feedback loop. The 47Ω resistor is the shunt feedback resistor, mentioned below.

Looking at the various Fender schematics that use NFB I see considerable differences in where the FB is applied as well as how much resistance is used.

There's 3 big components to how the resistance values are selected:
1. Where in the preamp is the feedback going to be returned?
2. Where is the feedback coming from?
3. How much feedback do you want?

The point in the preamp where you'll be injecting the feedback already has circuit components around it. The value of these components will determine the size of the feedback element at this point. For example, if you're bringing the feedback to a point between a cathode resistor and ground, and that resistor is 1-2kΩ, then adding anything bigger than 100-200Ω starts to change the way the circuit works even before feedback is applied. But if the circuit components are more like 100-200kΩ, then even 15-20kΩ is a small percentage of the existing resistance.

So, where you intend to apply feedback changes your choice of feedback circuit components, generally the shunt feedback resistor running from the loop to ground.

Along these lines, sometimes a cathode resistor is chosen to do double-duty as a cathode resistor and shunt feedback resistor. In which case we move on to...

Where is the feedback coming from? Generally, we'll source it from the OT secondary, because we're using the feedback to normalize the response of the output stage and transformer. With voltage feedback (which is the kind we use from OT secondary to earlier stage), we care about how big the voltage is at our feedback source.

Best way to figure that: If you have only 1 OT secondary tap, assume the speaker is a resistor. Figure or assume your power output to the speaker. Calculate the voltage needed across your "speaker resistor" to dissipate the power the amp will deliver.

If you have multiple taps, you can derive the feedback voltage from any of them. Be aware that if you tap off the 16Ω tap instead of the 4Ω tap, you will have a larger feedback voltage. It doesn't matter which tap has a speaker connected, the calculated voltage will always be present at the tap you derive the feedback when your assumed power is delivered to the speaker.

(If you choose to measure to verify this, use should use a resistor load in place of the speaker. Otherwise, you and your meter will be confused by reactive power and read some voltage other than what you expect).

So now you know how big your source of feedback is, and what allowable resistance you can have at the injection point. You need to figure out how much feedback you should apply, and use that decision to determine the series feedback resistor (2.7kΩ in the VibroChamp). The series feedback resistor forms a voltage divider along with the shunt feedback resistor to reduce the voltage at the source of feedback to just the amount you decided you needed.

So if you copy from a given amp, you need to double-check that you're copying all aspects of it, or you add well-chosen adjustable parts to dial in the sound you want.

[Tyrannocaster]

Did you add the 47Ω resistor under the driver tube's 1.5kΩ resistor? If not, you just added an extra load to the OT secondary, rather than make a feedback loop. The 47Ω resistor is the shunt feedback resistor, mentioned below.

No, I didn't, because here ( http://www.el34world.com/Forum/index.php?topic=13867.msg130021#msg130021 ) in response to my question about whether to keep the 47 ohm resistor or not, you said :

What you have will work just fine.

The 47Ω resistor is the other half of the feedback loop; the loop requires a series resistor and the 47Ω shunt resistor to define a voltage divider determining how much feedback there is. You have no loop, so you don't need the 47Ω resistor.

I probably misunderstood what you were saying anyway. Maybe I should try it again with the resistor, but I'm putting the whole NFB thing on hold until I can get the tremolo to work.

 

[HotBluePlates]

 

You'd originally asked about whether the tremolo would work without the 47Ω resistor. It will, because that resistor is part of the feedback loop of the VibroChamp, which at that point you weren't discussing as a feature you wanted to incorporate into your amp.

So the trem will work without it, but the feedback loop won't.

Hopefully, your other thread has uncovered the problem in the trem, and adding the 47Ω resistor between the 1.5kΩ resistor will fix the feedback loop. The catch is the 1.5kΩ cannot be connected to ground; the 47Ω resistor goes between it and ground. That may necessitate an extra eyelet or moved ground point, depending on the particulars of your layout.

[Tyrannocaster]

You'd originally asked about whether the tremolo would work without the 47Ω resistor. It will, because that resistor is part of the feedback loop of the VibroChamp, which at that point you weren't discussing as a feature you wanted to incorporate into your amp.

So the trem will work without it, but the feedback loop won't.

Hopefully, your other thread has uncovered the problem in the trem, and adding the 47Ω resistor between the 1.5kΩ resistor will fix the feedback loop. The catch is the 1.5kΩ cannot be connected to ground; the 47Ω resistor goes between it and ground. That may necessitate an extra eyelet or moved ground point, depending on the particulars of your layout.

Okay, I got it now. And I realized my post might have sounded sort of whiny and "Well you said blah blah" and I didn't mean for it to.

Thanks for all your help; it's a lifesaver.