NFB insertion from o/p xfmr sec? Last stage is cath follower. PI is cathodine.

[rob_h]

To whom it may concern, or make curious, or who can bring forth enlightenment...

The configuration I'm working on, design only so far, is a cathode follower into bassman tone stack with master volume, into a cathodine PI, driving a push-pull output, no big deal, I know.  But I want to use negative feedback  I have found schematics with pentode and triode pre-PI stages using NFB where insertion is at the cathode of the last pentode or triode, both with no tone stack after, but all amps using cathode follower to a tone stack, to a cathodine PI have no NFB as far as I have seen.  

If I were to do it, I assume that inserting it in the cathode of the last triode is out of the question because of its voltage, so would it be inserted at the cathode of the first triode of the cathode follower pair?  Is there an example out there that I overlooked, or is there a problem with NFB in this confoguration?  I can always do a LTP PI.

Thanks for your consideration,
RH

[terminalgs]

The configuration I'm working on, design only so far, is a cathode follower into bassman tone stack with master volume, into a cathodine PI, driving a push-pull output, no big deal, I know.  But I want to use negative feedback  I have found schematics with pentode and triode pre-PI stages using NFB where insertion is at the cathode of the last pentode or triode, both with no tone stack after, but all amps using cathode follower to a tone stack, to a cathodine PI have no NFB as far as I have seen.  

If I were to do it, I assume that inserting it in the cathode of the last triode is out of the question because of its voltage, so would it be inserted at the cathode of the first triode of the cathode follower pair?  Is there an example out there that I overlooked, or is there a problem with NFB in this confoguration?

a schematic would be useful, even if its a rough draft...   I think I understand you to say:

" gain stage(s)->CF->tone stack->cathodyne phase splitter. "

Are you basically thinking of a 5F6-A with a cathodyne phase splitter instead of a LTPI?

Think in terms of the 5F6-A's LTPI providing some gain, and the cathodyne providing no gain.  If you substitute the cathodyne for the LTPI (in a 5F6-A), you loss that gain, so you should compensate for that gain somehow, and an extra gain stage directly in front of the splitter is a convenient way to do so.  If a 12AX7 makes too much overall gain, use a 12AT7 or 12AU7 (or voltage divider out front somewhere). Adding a gain stage there also provides a convenient place to insert the NFB at the cathode (Like a 5E5A Pro).

there are a lot of details to consider, but first maybe a schematic or more info would be the best next step....

I can always do a LTP PI.

what is driving the design elements of the amp?  why cathodyne instead of LTPI?

[sluckey]

I assume that inserting it in the cathode of the last triode is out of the question because of its voltage

Use a coupling cap and insert it wherever you like.

[HotBluePlates]

...  I have found schematics with pentode and triode pre-PI stages using NFB where insertion is at the cathode of the last pentode or triode, both with no tone stack after, but all amps using cathode follower to a tone stack, to a cathodine PI have no NFB as far as I have seen.  ...

I don't recall ever seeing a split-load/cathodyne inverter without a tube stage immediately in front of it. Meaning I've not seen a amp manufacturer place a tone stack right before a naked split-load inverter.

Are you sure you want negative feedback? Why can't you include the triode ahead of the inverter? Are you sure the cathodyne/split-load is the right inverter for the output stage you've selected (and the B+, screen voltage, bias selected)?

[rob_h]

To terminalgs: I went to a cathodine to save a triode to use as a second reverb recovery tube originally, but I didn’t consider the fact that it would have around unity gain.  If I have to use another stage to get back to the LTP PI gain, there’s no point; no savings on parts. Here are schematics of both. 

Also, as you can see, I can’t decide what to do with the reverb either.  Insert before and after a 3.3M and cap, or from the front end to the back end.  I think I saw some discussion on the subject in the forum.  The 6CM8 works very well, on paper, with a high gain triode recovery stage.  I guess I’ll have to get it up and runnning before I know if I need more wet signal or not.

To sluckey: I was concerned about trying it there, but the cap will isolate all, right, thanks.

To HotBluePlates: I can’t find an example either but I don't know why it isn't done. Yes, I think I do need NFB, but I think I’ll go with the LTP instead.  The reason for trying it was to save a triode for the reverb, but there’s no savings.

Also, as you guys can see by the switches, I couldn’t decide on a low gain, or high gain, so I decided to try channel switching a bit.  SW4 is to bring in more of a tweed sound in Clean modes.
 

[rob_h]

I was thinking about what terminalgs said about the “added triode before a split-load/cathodyne inverter”.  If a LT PI has a gain of around 30, then using the “added triode before a split-load/cathodyne inverter” approach would give a higher gain for the two tubes compared to a LT PI.  If the added triode gain was at, say 60, the result would be double the gain from two triodes of the LT PI.  I’ve read pros and cons on the two PI types, but don’t know from experience. What matters to me is the tone.  My amps are LT and SE.  Do any of you guys have an opinion one way or the other?
 

[HotBluePlates]

I don't know where to begin...

It seems like there is no one amp we can point to and say, "This is a ____, but with these changes to the preamp." And because it seems like you've tied to include every feature and design element you think you might like, it appears to have the kitchen sink in there. I'm thinking that starting with 90% of an existing amp design, but leaving room or control positions for the planned changes, would give you a solid foundation to get working before adding all the unusual features.

If you amp were a blank sheet of paper, you'd have to start with a desired output power level, a knowledge of what power and output transformers can be had, what supply voltage that means you'll likely work with, and the output tube type that will get your desired power given all these other considerations. You complete a design of the output stage, and the bias of the output tubes tells you how much signal your phase inverter has to throw. That signal size and supply voltage might dictate to you whether a split-load or a long-tail (or paraphase) inverter is the best choice.

You design the amp from output backwards to the input. But I don't see anything definite on the power supply voltage or output tube type to tell us anything useful about what the requirements are for phase inverter.

Other items:

• It's not clear to me the signal will split off properly to the reverb circuit, as the 3.3MΩ that makes that happen is only in the circuit part of the time. Additionally, there is a block "REV" at that resistor which correctly indicates the tapping-off and return point for the reverb, but you have the reverb circuit proper connected to different places in the preamp.
• The caps at the reverb input don't seem right, nor does the 16kΩ grid reference for the 6CM8. The 470pF across 16kΩ rolls off highs at/above ~21kHz (which isn't needed as your guitar speaker doesn't go nearly that high), while the 0.001uF working against 16kΩ is a high-pass filter that's -3dB at ~9.9kHz (above all guitar notes and most harmonics, and also at the extreme top end of your speaker's range).
• If you keep SW4, you'll want to implement it with a DPST (instead of the double-throw not being used), with each section shorting out a ~100kΩ resistor between the cap you're adding and the cathode resistor.
• The 6BR8 pentode at the tail end of the reverb circuit seems like unnecessary gain, while the two volume control which follow it are unnecessary loss (especially the 50kΩ pot which will kill the pentode's gain). Without really breaking it down & analyzing, it seems like the pentode will mostly add noise.
• A "poor choice" of setting for SW1 & SW2 can leav you with a single triode trying to drive the tone stack ahead of the phase inverter, which will almost certainly not leave enough signal input driving the phase inverter, no matter if you choose a long-tail or split-load. It seems unwise to give yourself an option of a non-workable switch setting. So you probably need 2 triodes in your clean preamp, not counting anything you do with the reverb.
• SW3 and the volume trim for the tone stack bypass seems unnecessarily complex. If you want to bypass that tone stack, a SPST switch can disconnect the Mid pot from ground... and tone stack bypassed. If you want to insure that switch won't pop when thrown, maybe put a 250kΩ-470kΩ resistor across its contacts (i.e., between the Mid pot's ground lug and ground, but shorted by the switch when you want to engage the tone stack).
• C1 and R2 seem to be sized to offset each other. The tube already will have ~100pF of input capacitance, so a 100pF to ground seems like an unneeded expense. And if you want to roll-off treble with C1, why make R2 smaller-than-typical, which reduces its interaction with the tube input capacitance to roll off treble?
• There are a handful of other resistors that don't seem like the right value for their position, but picking them out might not be fruitful if you're going to make broad changes or updates.

[PRR]

> I don't know where to begin...

Agree.

Without the extensive look-thru you have done, I see:

* level at the send/return jacks is power tube grid level, 15V-45V (can't tell). This is FAR more than can be handled by most things you'd put in the loop.

* Taking NFB *around* both tone and FX tends to *negate* whatever the tone/FX do. (Just as it reduces flaws in the power amp when you take NFB around a power amp; the NFB can't know that a bass-cut or echo-box is "wanted", it just knows it does not compare to the original signal so it tries to cancel-out the difference caused by the tone/FX circuits.) (And taking NFB around any kind of delay, such as you might stick in an insert loop, is SURE to cause horrid instability, because the NFB is now attemping to "correct flaws" from 90 milliSeconds ago.)