Pentode(EF86) as asymmetric split load (gain +follower)?? Zinky tail!

[kilowhat]

Hi guys,
I was looking at some 6g15(standalone outboard reverb)-ish schematics and came across the fender dual professional Zinky designed (see attached schematic). At the input stage it uses a tail beneath the cathode resistor (like a sort of asymmetric cathodyne/concertina PI), serving double duty as both (modest) gain stage and cathode follower for mixing buffered dry with recovered wet. As the signal is miniscule i dont see any risk of the next stage clamping its grid (thus have the follower make its anode go full shitshow), plus the tiny signal in means the tail dont steal too much swing headroom from the anode. I think the solution is quite genius in its simplicity, effectively saving a triode.

This got me thinking: in my estimation the "achilles" of the 6g15-design is that a single triode give not enough gain(50), but two cascaded provide too much gain(2500), forcing us to dump a lot of it before the tank driver (without utilizing it for EQ-purposes). Did someone say pentode?? :)

My question: Would this magic Zinkytina-Zinkytail setup of an asymmetric split load be feasible with a pentode?
Say EF86 datasheet standard: 100n input, 1m grid leak to 1k cathode(full bypass), 100k anode, 390k screen+100n cap, but now tied to a 10k tail, serving as follower for dry mixing. What are the implications? I dont see us loosing too much swingroom, but do we loose substantial gain?

And how does the constant cathode current of a pentode come into play? I do remember having read in Merlins book that given the screen cap be tied to the top of the followers tail (i.e not to ground), a pentode could be used for follower duty. But i cant see in my head HOW, given that pentodes running class A should be (near) constant current (thus no tail wiggle?).

Ive been trying to get this straight in my head for days (to no avail). The combination of follower feedback, asymmetric anode, screen current and the constant cath-current nature of pentodes turn any cerebral simulation into proper mush. If you can, please help!

Many many thanks! /kw

[kilowhat]

Bonus question: if a small signal pentode is a (near) constant current device, do a partial bypass cap (e.g 2.7k/0.68u) have noticeable effect on EQ? If the current is constant the resistor shouldnt swing(!). So why does every EF86 schematic and datasheet bother with cathode bypass caps at all?

It feels like theres some critical angle im oblivious to here. /kw

[shooter]

So why does every EF86 schematic and datasheet bother with cathode bypass caps at all

blind stab in the dark
helps mitigate noise, hum n other unwanted junk in hi gain stage, possibly help stabilize tantrum prone tubes 

[kilowhat]

Hi shooter, thanks for your reply!
But doesnt the high internal resistance of pentodes shove almost all supply noise straight onto output signal from the anode? Maybe youre onto something with the soothing tube tantrum aspect.. Thank you! /kw

[kagliostro]

I'm out of the discussion, my Know How is too poor

May be you can find something of interesting here

http://www.r-type.org/articles/art-097.htm






Franco

[DummyLoad]

blind stab in the dark
helps mitigate noise, hum n other unwanted junk in hi gain stage, possibly help stabilize tantrum prone tubes 

wondering how so? - cathode bypass is positive feedback - so wouldn't PSRR be lower with bypass in place? disconnect cathode bypass and the AV drops to about 1/3 of with a bypass. IME, keeping the screen supply as clean as possible and plate load as low as needed for the output swing and AV required work best.


--pete

[shooter]

 

that's why I prefaced it 

[PRR]

> Ive been trying to get this straight in my head for days (to no avail)

May I suggest? You could BUILD it in one fevered morning and TRY it.

If you like it, you don't really have to understand it.

I don't know Zinky. My *guess* is that he designed with a soldering iron more than a tube-book.

[shooter]

My *guess* is that he designed with a soldering iron

If I understood my Kid, he started with a guitar and didn't like the noise his amp made, so went with your design assumption, and liked it 

[tubeswell]

The dry bypass in the Dual Professional input stage is designed to put out a low strength ('guitar pick-up' level) dry signal, because if the dry signal is too high, the wet signal from the reverb recovery stage won't be able to cut through in the mix.


The 6G15 dry bypass signal was unity gain for the same reason. In the Dual Professional there's no CF bypass for the reverb part of the circuit. Bruce Zinky was saving himself a triode.


There's no reason why you couldn't do the same trick with a preamp pentode in V1. Merlin Blencowe covers this in his 1st designing tube pre-amps book. You can shunt Cg2 to ground or to the cathode - it will work both ways. If Cg2 is shunted to ground, the CF part of the pentode would be effectively triode-strapped. If Cg2 is bypassed to the cathode, it will work in pentode mode. (Don't forget to put Rg2 in - otherwise the cathode will end up being shunted to HT.)