Adding NFB to single ended marshall amp

[jordan86]

I have a single ended Marshall (AX84 SEL, modded to more SuperLead preamp spec) that I am considering adding NFB to. I don't know exactly if it will be a tonal upgrade but I'd like to experiment if I can without totally tearing the amp apart (adding NFB and possibly presence control). Since I am without a typical Marshall PI it gets a little more strange. I've read a ton and looked at 5F1, AB764 Champ, and some other schematics. My understanding is that the only two practical places I could insert it would be:

1) Preamp Stage 3 (just before CF/Tonestack)
2) At the KT88 cathode

I believe option 2 would be the better route, since it won't interact with the CF/Tonestack, but not sure. I've searched other threads here. Saw one comment from Sluckey about just removing the bypass cap from the power tube cathode (C6 for me). Is it really that simple? Also saw a comment from a guy over at chasing tone connecting a 68K straight from the OT to the power tube cathode (context was the similar SEplex build). Only other curveball is this build does have the variable cathode bias option and test point that AX84 has offered, so not sure how that complicates matters. I also don't totally understand what C5 is accomplishing on the screen grid.

I would love advice on how to implement the NFB and any potential values to test. I'm open try a pot or just alligator clip some values in to see what I get. Just need to know where to start.

Attached my schematic with my mods. The first two stages look weird, but they are just paralleled at V1 and controlled by a dual contrentric pot. Turned the 220K voltage dividers into mixing resistors. Only mod not shown is a virtual CT. 100R resistors running from Pin 8 to Pins 7 and 2 of the KT88.

[jordan86]

Side note: I understand its a fool's errand trying to make a 20W single ended amp ACTUALLY sound like a true PP 100W Super Lead, but I'd like to at least try to get as close as I can. I am hoping NFB might help the small amp hold together a little better. It's in a really good place already. I'd say I am content with it, but trying to eeek out that last 10% of potential. Guitar player problems :)

[shaun]

If your NFB is global - going from the OT to whichever point you decide to insert it - then in my experience, what you'll get is a taming of the overall amplifier - cleaner headroom, less distortion, and a tone you may or may not like. I thought the idea of Marshalls, generally, was balls to the wall (French term for lots of distortion), so it may be that you might prefer "local feedback," where you are basically tweaking a smaller section of the amp (a single amplification stage almost anywhere in the amp) by taking the signal out of that stage and returning it to the input side of the stage, possibly also running it through a filter capacitor or resistor combination. Rob Robinette has posted some interesting local feedback ideas on his site (should show up in a google search), and that's a good starting point. I'm sure builders on this forum have experimented with this endlessly.
What you'll probably have to do, after researching, is pick a couple of ideas that seem likely, then get out the gator clips and try stuff on the workbench until you get the sound you like. It's pretty difficult to determine exactly what sound you'll get from looking at a mod schematic, so you'll probably just have to try them until you get what you want.

[shooter]

Designing for Global Negative Feedback (aikenamps.com)


nice light read

[jordan86]

Thanks. I've read both the Aiken page and Rob's local FB ideas. Unfortunately, Aiken doesn't provide any Single Ended examples. 

And yes, I am great with balls to the wall....as the French say :) To clarify what I am going for:

-Less ratty distortion, but not less distortion. Maybe smoother though.
- More polished (fatter) high end and more punchy low end.

I typically run the bright channel volume at like 80-90% (sometimes dimed) and the Normal channel at like 20-30% for some fullness. Even then its not super high gain, so I'll boost it with a pedal for more saturation. Gets me somewhere between Angus and Eddie.

I guess its possible NFB is just a bad idea and not going to accomplish what I am hoping for. I'd more so just like to try it to see from experience what it would accomplish. I guess I could just do the 5F1 approach into the third stage before the CF. But I'm still curious if something at the power tube would be a better approach. 

I just don't know what I don't know :)

[jordan86]

If your NFB is global - going from the OT to whichever point you decide to insert it - then in my experience, what you'll get is a taming of the overall amplifier - cleaner headroom, less distortion, and a tone you may or may not like.

I think GLOBAL NFB is what I am after, maybe not much of it, but at least try to see how the amp responds.

[shooter]

An SE amp by design has "build in" NFB, self biasing.  You only have a 10uF for Ck, remove that and the tube will "try" a smidge better.
Adding UL sorta helps, Me, I add a big Ck, something between 47-100uf allowing more of the "good stuff" to pass through to the iron and speaker, at the end of day though, it's a Class A, A1 PA.  "force them to deep" and the growl back

[jordan86]

You only have a 10uF for Ck, remove that and the tube will "try" a smidge better.
Me, I add a big Ck, something between 47-100uf allowing more of the "good stuff" to pass through to the iron and speaker

Shooter, so you are saying I can remove C5 from grid to ground, or try replacing it with with a larger 47-100uf cap? My Ck (if I correctly understand that as C6, the cathode bypass cap) is indeed 100uf.

[shooter]

my bad...eyes, yes C6, not sure what C5 is doing so can't comment on that


my understanding of Ck, the bigger it is the less "natural" NFB for an SE amp
but the bigger it is, the more frequencies, especially the lows, get passed on
I believe gain is also a factor, less cap, less gain, but i'm not gonna go back n re-read 

[jordan86]

I'll try lifting C6 to see how that goes. Also, will check with the circuit designer on C5.

[jordan86]

To clarify the KT88 cathode section....

There is a 1 ohm resistor and two test points coming off Pin 8. Next it goes through a 1K bias adjustment pot, then through a 100 ohm resistor to ground.  I'm thinking that 100 ohm could possibly function as a NFB tail if I inserted global NFB there? I could run a resistor straight from the OT to the top side of that 100R resistor. Something in the 820R to 5K range?  820/100 would be Fender-esque ratio. Where as something upwards of 2K+ would be more Marshall-esque. Mind you none of those values ever happened on a power tube, so it's all uncharted territory.

[2deaf]

C5 is just a filter capacitor for a branch of the power supply.

[2deaf]

This one is real easy to try.  For more NFB, take it from the 8r tap.  Probably need to reverse the OT primary wires (Hammond doesn't give us the polarity info).

[Willabe]

This one is real easy to try.  For more NFB, take it from the 8r tap.  Probably need to reverse the OT primary wires (Hammond doesn't give us the polarity info).

In the schematic, what is connected to A from the KT88 K?

Can't see it in the drawing. 

[sluckey]

filaments

[jordan86]

filaments

Yes, filaments. Raised heaters.

[sluckey]

If you decide to remove C6 for some degenerative feedback, point A will no longer be a dc voltage. It will be guitar signal. Not sure what that will do to your heater elevation or your guitar signal. Let us know what happens.

[Willabe]

Where is the KT88 K grounded?

[jordan86]

Where is the KT88 K grounded?

The entire amp is grounded with a single ground bus bar.

[d95err]

The Marshall SL5 (according to a schematic I found) uses a gain stage after the tone stack where NFB is mixed in. It also has a presence control there.

The Marshall Class 5 also has gain stages after the tone stack (but no NFB, I think).

[shooter]

ground is at B
I got tangled for a minute with that

[PRR]

You can't get there from here.

You have artistically developed the amplifier gain staging to have just-enough gain to brush overload.

NFB requires a consistently large excess gain. That means more stages and loops free from gain-knobs like tone and volume controls.

[jordan86]

Man....PRR the dream killer. And on a Friday afternoon. Brutal! :)

JK. JK. That's the sort of wisdom and insight I needed to hear. Sluckey's insight on removing C6 and all its repercussions was HUGE too. I guess I could remove one of the paralleled stages up front and put that back in series. Maybe as a stage 4 after the CF tone stack. Or add another tube as a stage 4. Would definitely be an SEPlex at that point. https://chasingtone.com/the-seplex/

I think I will keep it as is though. It's a good sounding little amp. Just enough gain when pushed to get me into Angus type territory, with a dash of Tweed Champ DNA. And with a pedal it does the EVH fairly convincingly.

I'll post a sound sample with video here one of these days so you all can appreciate it, and not just be solving my problems. :)

Really appreciate it everyone. Hope you all have a great weekend.

[shooter]

I just passed on my "version" of your amp, it left on it's 4th iteration with a schematic taking it to the 5th version, if implemented.
 the topography is my favorite "all around" LoWatt SE
IIRC I was able to get 80+Vac to the 88, not sure if it was 2 AX or 1 AX 1 AU


enjoy the amp, mix it with an "equal" PP fixed bias, 2 2X12 cabs, faces will blow back 

[jordan86]

enjoy the amp, mix it with an "equal" PP fixed bias, 2 2X12 cabs, faces will blow back 

It does some mighty good in stereo with my 18 watt, EF86 based Dr Z.

[jordan86]

the topography is my favorite "all around" LoWatt SE
IIRC I was able to get 80+Vac to the 88,

Shooter, do those pairs of 100uf in series filter better than just a single 50uf filter cap? Or did you do that for voltage purposes?

I’d love to hear that amp. I’ve started running a Gold Lion KT66 in mine recently. I don’t necessarily need the extra volume the 88 provides, and the 66 warms and fatten things a touch too. 

[shooter]

volts, unloaded, or power up it's >500vdc for ~~30 seconds
The poor 88 was my breadboard tube, it's run at 150% max, it's been fired up without a load but WITH signal.
There's some odd IM that I never chased down, so the new owner was been advised to start with a new or different tube.


If you're anywhere around Indianapolis I can probably get you some jam time on it, although I kept the good speakers