Interesting NFB control idea, Fender Silverface

[GlideOn-Designs]

Working on 79' Twin Reverb UL and the popular mod that others claim to do is to have the NFB on a variable or switched control.

Okay.

Ground switch has been freed up and there's maybe room for a petite pot with a very small knob on the rear faceplate as the hole is higher up near the lip of the chassis.

I see it in the circuit as the 820ohm before the speaker jack, but there's also a .01 cap in parallel seemingly to cut some highs late in the circuit which got me thinking...

I'm thinking a 10k‐25k pot with a series 470ohm resistor to set the total useful NFB control range.

However, in regards to this 'icepick cap' also in play which likely would be detrimental to altogether eliminate, could this cap value scaled larger corellating to the decreasing NFB resistance for a combined NFB/High Cut?

Basically a "Clean/Dirty" control?

 

[schmee]

I dont know if that little cap is for icepick or for noise. But the earlier amps didn't have it and it's probably not necessary.

 

[GlideOn-Designs]

I dont know if that little cap is for icepick or for noise. But the earlier amps didn't have it and it's probably not necessary.

It's certainly not a 'little' cap. That's 10,000 picofarad, 10 nanofarad, .01 microfarad which would absolutely neuter any highs used similarly elsewhere in the circuit.

Oddly it is still a 'musically bright' amp but I would not say 'ear-piercing bright' overall like I've heard some Fenders. But there are also no snubber/stability caps on the 47k/47k PI in this amp either, common feature I see on LTP designs which may explaining a bit more in regards the brightness levels.

The schematic is dead-on too.

Still plenty of highs to tame I would say.

 

[schmee]

It's a 'musically bright' amp but I would not say 'ear-piercing bright' overall like ive heard some Fenders. But there are also no snubber/stability caps on the 47k/47k PI in this amp either, common feature I see on LTP designs which may explaining a bit more in regards the brightness levels.

The schematic is dead-on too.

Still plenty of highs to tame I would say.

Does it have the bright switch? Options there. A change in the PI input cap (to pin 2) can tame brightness well also.

 

[GlideOn-Designs]

Yes, it has bright switches on both channels (technically activates 150k 'bass cut' to ground too) but those obviously are defeated once volume control is on full. PI Cap I have reduced to a .0047 from the stock .01 for tighter, less boomy bass. Not sure how that is relevant to the NFB function though...

 

[Ten Over]

It's certainly not a 'little' cap. That's 10,000 picofarad, 10 nanofarad, .01 microfarad which would absolutely neuter any highs used similarly elsewhere in the circuit.

It always depends on what is being used in conjunction with the cap. Here we have an 820R resistor in parallel with the .01uF cap. That combination has a corner frequency of 19.4kHz with an impedance of 580R at that frequency.

 

[GlideOn-Designs]

It always depends on what is being used in conjunction with the cap. Here we have an 820R resistor in parallel with the .01uF cap. That combination has a corner frequency of 19.4kHz with an impedance of 580R at that frequency.

That's true, the sum of the parts dictate the RC function, but it is still physically a large cap (one of the infamous "blue blobs") as far as passive caps are concerned and space is still limited.

The whole idea is to use a physically larger Cap/value as the NFB decreases with the turn of the knob, getting harrier, middier, brighter. Since it is already an arguably BRIGHT sound and less NFB would make it brighter still, I would like to maintain the same or greater degree of Treble Cut in the process.

So it would be more of a voicing switch all said and done. Shifts the focus towards a ruder upper mids rather than clean, pristine treble. Like a turn from Fender into something more Vox-like.

I don't know of any amps that come to mind that have this feature, certainly not pre 1980s. I'm sure someone has thought and implemented it by now.

Question is: What values?

 

[Wally]

I'm thinking a 10k‐25k pot with a series 470ohm resistor to set the total useful NFB control range.

The NFB signal is controlled by that 820ohm shunt resistor in relation to the 100ohm ‘tail’ resistor. Imho/ime, that equation might not need the 25k resistance to develop a significant decrease in the cancellation factor. I find a 5K pot to be a significant resistance to place in series with that 820 ohm resistor. You might find Randall Aiken’s article on designing NFB loops interesting…AIken Amps ‘Tech Info’ section.

 

[2L man]

I'm thinking a 10k‐25k pot with a series 470ohm resistor to set the total useful NFB control range.

470 ohm does not protect the potentiometer "slide" when current can come too high when you play loud.

 

[GlideOn-Designs]

470 ohm does not protect the potentiometer "slide" when current can come too high when you play loud.

Well, okay not necessarily 470ohm but some nominal minimum value. I was thinking slightly tighter but perhaps 820 is as low as it goes.

As far as current/power handling a 2w pot is found in 5k, 15k, 25k values easily enough but I haven't had any trouble with normal CTS pots in other amps I have variable NFB in.

Is 5k as high as it goes? - Seems to be the remaining question

 

[Wally]

Like I noted….that one resistance is not the story, is it? It is the relationship of the two resistances…. shunt and tail…that determines the feedback factor. When the loop has higher numbers in those two resistances, then the pot that is necessary to have a good sweep over the desired range will necessarily be a larger value. Ime, when the pot is too large, then the sweep gets affected in a manner that I don’t like.
Gotta find the spot!
When dealing with values such as 27k with a 5K tail, one would be wanting a much larger value pot than one would use when dealing with 820/100. It is interesting also to realize that the 5F6A Bassman has more cancelation effect than does the Bf circuit in this ULTR.

 

[GlideOn-Designs]

Like I noted….that one resistance is not the story, is it? It is the relationship of the two resistances…. shunt and tail…that determines the feedback factor. When the loop has higher numbers in those two resistances, then the pot that is necessary to have a good sweep over the desired range will necessarily be a larger value. Ime, when the pot is too large, then the sweep gets affected in a manner that I don’t like.
Gotta find the spot!
When dealing with values such as 27k with a 5K tail, one would be wanting a much larger value pot than one would use when dealing with 820/100. It is interesting also to realize that the 5F6A Bassman has more cancelation effect than does the Bf circuit in this ULTR.

I understand the relationship of the two components and their functions as a whole, but for the purpose of this particular discussion we have to have a goal of being pragmatic and choose one of them to be adjustable and one of them to be fixed.

As you say, above 5k may not have the desired effect. That's as good a ceiling value as I've heard and read so far.

 

[2L man]

Well, okay not necessarily 470ohm but some nominal minimum value. I was thinking slightly tighter but perhaps 820 is as low as it goes.

As far as current/power handling a 2w pot is found in 5k, 15k, 25k values easily enough but I haven't had any trouble with normal CTS pots in other amps I have variable NFB in.

Is 5k as high as it goes? - Seems to be the remaining question

Yes with 2W potentiometer uou can use less series resistance but using 1/4W or 1/8W potentiometer I recall it was min 10k resistor needed for safe operation for 6V6 push pull 8 ohm output. There I used typical 0,1uF resonance capacitor, components 1kHz which amp can output full power.

 

[GlideOn-Designs]

Yes with 2W potentiometer uou can use less series resistance but using 1/4W or 1/8W potentiometer I recall it was min 10k resistor needed for safe operation for 6V6 push pull 8 ohm output. There I used typical 0,1uF resonance capacitor, components 1kHz which amp can output full power.

"Resonance Cap?" As in between the OT/jack and NFB control? Could you clarify as to the placement of the 10k (2w?) and 0.1 (350v +?) cap?

It already has a 22k 2w tail in the PI...

Also it is running 4x 6L6 @ 4ohms, the useful NFB range is going to be scaled down considerably lower than 8 ohm or 16ohm. I don't anticipate nor care about plugging into 8ohm extra speaker jack I suppose lowering the installed NFB knob would fix any side effects

 

[Wally]

"Resonance Cap?" As in between the OT/jack and NFB control? Could you clarify as to the placement of the 10k (2w?) and 0.1 (350v +?) cap?

It already has a 22k 2w tail in the PI...

Also it is running 4x 6L6 @ 4ohms, the useful NFB range is going to be scaled down considerably lower than 8 ohm or 16ohm. I don't anticipate nor care about plugging into 8ohm extra speaker jack I suppose lowering the installed NFB knob would fix any side effects

RE: that 22K tail resistor?? Has this aspect of this amp already been modified? If not, then the two resistors of interest in setting the NFB loop factor are a 820 ohm series and a 100 ohm shunt. (note: edited for correction of terms.)
Do you have a schematic of this section of the circuit? The only 22k resistor I see anywhere near the NFB loop is the 22k that
is one of the two resistors that establish the Line Out.

 

[GlideOn-Designs]

"When changing the feedback source from one output transformer secondary to another you change the feedback resistance by a factor of 1.41 for one step change (for example going from a 2 ohm speaker tap to 4 ohm), a factor of 2 for 2 steps (example 2 ohm to 8 ohm), or a factor of 2.83 for 3 steps (example 2 ohm to 16 ohm)."

- Robinette Pages.

About a 150% scale in NFB values when you increase the ohm/load rating.

@Wally was right in the money as 5k being near the upper limit of stable minimal NFB @ 4ohm load

So it could realistically be 5k, 10k, 25k pot but perhaps the larger pot would allow a better taper for the Capacitor in question.

Again, my desire is to have noticable high to mid-treble cut effect while increasing the grit and mids so taper and scaling are perhaps more important for the cap than pot. NFB resistance arguably can just as easily be set with fixed resistance in parallel with wiper and lug to set max NFB.

With an audio tapered pot for best results, which pot value seems to make the most sense?

 

[GlideOn-Designs]

This is what I'm going off of, the values are dead-on.

Ignore my doodle, that was an earlier mod (actually ended up on SPST value 1m pot)

 

[Wally]

Okay, a diagram will do. The NFB resistor values are 820 ohms and 100 ohms. What/where is the 22kohm resistor you mentioned?

 

[Wally]

Designing for Global Negative Feedback

Guitar amplifiers

www.aikenamps.com

 

[telemnemonics]

I was putting a pot on the NFB in all my Fender amps in the 1990s based on whatever info was available at the time.
The old well established values worked great!

As for a treble/ tone control, I dont see a reason to gang one onto the NFB?

Personally I just got rid of my UL Fender amps but if you have it and like it then have fun.
But ganging two functions onto one control is about the worst idea designers come up with, unless there is no space to put two controls so you can use them independently.

IIRC cutting NFB cuts treble by gaining up and fattening up the signal, resulting in less of the thin shrill tone a mid scooped Fender can have.