Funky local NFB on EF80 single ended power tube?

[Porpoise Head]

Hi all,

I got some great help here before when building a dual preamp (dual rectifier/ matchless lightning). I'd like to build an all pentode pre + poweramp as my next project (EF86 -> attenuation -> EF86 -> EF80). The regular way to do NFB is obviously from the secondary of the OT to the cathode of the tube driving the power tube. However, I'd like to be able to bypass the cathode of the second EF86 so that option is out of the window.

Then I saw an interesting scheme on old Matamp schematics. I added my take on this scheme as an attachment.  I added an SLO-style depth control. This seems to work great around an 12ax7 in LTspice, but will it work (with different values) around an EF80 single ended power tube? The gain of the EF80 is roughly 60 (i think), which is similar enough to an 12ax7. I'm interested in your thoughts.

[tubeswell]

What is your goal with the feedback? (More HiFi?, lower output? Less clipping?, something else?).

[passaloutre]

Why can’t you bypass the cathode of the second EF86 receiving NFB?

[Porpoise Head]

My goal is to be able to get a somewhat scooped response from the poweramp if I like to do so. The two resistors with labels will be pots in the actual Amp. The frequency response is above. However, I think you are right and I'm being stupid. I can offcourse do nfb the fender Princeton way and Largely bypass the second ef86 cathode but not entirely, and insert the nfb there. Thanks! Any thought on whether the 'matamp' or Princeton way is better? Both will enable a scooped response.

Edit: I remember why I thought the Princeton way wouldn't work: the two ef86 of the preamp have a shared cathode with switchable bypass and switchable screen bypass. This would make the feedback both positive and negative at the same time? Unfortunately I could not attach a schematic of the preamp part. The filesize of my photo is too big.

[tubeswell]

The scooped response is from the filtering. You can get that either with filtering a feedback loop or filtering another part of the circuit.

If your goal is about HiFi (and HiFi helps set you up for a nice scoop* by flattening the response and increasing the audible bandwidth**), then by all means include global NFB.

*but the scoop comes from filtering, not from merely adding NFB. So you still need the filtering.


**from the audible perception of balancing the frequencies (whilst reducing the amplitude all round)

I wouldn’t be too concerned about reducing gain from having an unbypassed cathode resistor on one of the EF86s - bypassing the screen has more influence on gain in a pentode (than bypassing the cathode).

If you aren’t concerned about HiFi and just want a mid scoop, you can play around with filtering the cathode resistor bypass network (and/or the screen resistor bypass network).

Another way of getting a ‘cleaner’ signal (albeit with reduced gain) in a pentode, is not having a screen bypass capacitor. The screen will still act to boost the gain (above what it would be in triode mode), but the screen current feedback within the pentode will keep a lid on the gain. You will lose some (but not all) of the ‘bite’ of a pentode, but the added compression from an unbypassed screen can also be pleasant in a guitar amp (if you have enough gain elsewhere in the circuit).

[Porpoise Head]

Great suggestions tubewell.

I'd like to do the scooping hand in hand with nfb to get a bit more 'hi fi' at the same time. This way I can run:

a) a dirty fully bypassed preamp into a more hi fi poweramp, or by throwing some switches:

B) a mellower preamp into a poweramp that's blasting.

From a clipping point of perspective I'd like to do any scooping late in the amp. Changing the cathode bypass network of the ef80 is then also an option. However, I do not think one can get a scooped response from the cathode. How does that work? I only know flat, mid+treble boosted or treble boosted. Highly interested.

[kagliostro]

I don't know if you'll obtain what you are looking for but one thing you can easily try is this

Note the way the cathode resistor (and bypass cap) are connected to ground (via the positive of the secondary of the TU)






Obviously you can use an SPDT switch to connect the cathode resistor directly to ground or via the secondary of the TU

this way you insert and exclude the added feature

Franco


p.s.: Read here the reason for that way of connection

https://el34world.com/Forum/index.php?topic=31667.msg348700#msg348700

[Porpoise Head]

Thank you kagliostro. I'm going to check that link for how it works because I've never seen this before!

[Porpoise Head]

Kagliostro, I have one more question about the cathode feedback: how does one change the amount and/or frequency response of the feedback? With normal nfb you would escape the fb loop to ground. In this case that would mean a resistor and/or cap in parallel with the speaker. That seems a bit odd. Adding series resistance would change the bias point of the power tube, so that's not an option either

[HotBluePlates]

... The regular way to do NFB is obviously from the secondary of the OT to the cathode of the tube driving the power tube. However, I'd like to be able to bypass the cathode of the second EF86 so that option is out of the window. ...

Look at the Vibro Champ; they, too, bypassed the cathode resistor of the stage before the power tube.

A negative feedback loop is made of 2 resistors, though folks often focus only on one of those.  In the Vibro Champ, those resistors are the 2.7kΩ that connects to the speaker, and a 47Ω to ground (at the "ground-side" of the 1.5kΩ cathode resistor).  Those two resistors form a voltage divider to allow only part of the speaker signal back as negative feedback.

Just do what Fender did with the Vibro Champ: add a small resistor between the cathode resistor & ground.  "Small" to avoid changing the DC bias of the tube too much.  Now if you also size the series feedback resistor appropriately for the new "small resistor," you can inject negative feedback at the stage before the output tube.

[kagliostro]

If you don't have a Tone Control between the tube that precede the power tube and the power tube grid you can use the Vibro Champ solution

If instead you have a Tone control in the middle it isn't good practice to connect the NFB signal there and will be better to try if you have a benefit using the connection proposed on the schematic I posted

Franco