Parallel 2nd preamp stage - pentode/triode - possible?

[SixStringing]

I'm designing my next build, and it's probably going to be a Matchless Clubman inspired preamp with some tweaks:

V1 (triode) -> gain control -> V2 (pentode) -> Cathode follower -> Tone/fx loop -> PI -> output

I may have an extra triode (and room for one extra knob) and was wondering if it would be possible to split the signal after the gain control, run 2 parallel stage 2 (one pentode and one triode) mix them back with a single 'balance' control before going into the CF:

Code: Select all

                               /- V2a (pentode)-\ 
V1 (triode) -> gain control -|                     mix knob -> Cathode follower -> Tone/fx loop -> PI -> output
                               \- V2b (triode) -/

I expect I could hard-wire each V2 to give reasonable gain/distortion range, and compensate each V2's output pre-mix to account for different output gain voltages so that the balance control effectively blends (or chooses between) each V2.

Would this be doable? Is is too dumb to attempt?

Thanks.

[TUBEDUDE]

Nothing is too dumb to attempt. Even if the concept doesnt work out, something of value is usually gained in the process . Experimentation fosters creativity and flexible thinking.

[cdemike]

This could be a really cool project! Kind of makes me think of a Bassman or early Marshall in terms of the gain structure. I definitely think this is possible, though your preferences in terms of coupling the cathode follower will dictate how difficult this might be to implement. If you're looking just to lower the output impedance to drive the tone stack and aren't looking for some DC cathode follower compression, an AC coupled cathode follower would be fine and the coupling caps coming to and from the mixing control make this a lot easier. Something like the Echoplex mix control comes to mind for the control, and you actually wouldn't really need a tube cathode follower if the goal is just to lower output impedance, since a MOSFET source follower would lower output impedance while being more linear and more rugged/reliable.

DC coupling makes this a lot more complicated since you'd more than likely be changing the bias point of the cathode follower by changing the mix of pentode vs triode. My suggestions if you want have a DC cathode follower summing the signals would be to implement Merlin's "morph" control or to have some type of active mixer which in turn could be DC coupled to a cathode follower. This page from Merlin's site shows the general topology for the morph control: https://www.valvewizard.co.uk/ef86pcb.html
There are a number threads here, on the Hoffman amps forum, and on the Telecaster forum showing different values people have tried with 5879's and EF86's, so you wouldn't have to start from scratch figuring out values that work well to make the circuit work.

As far as active mixers go, I haven't personally tried or seen a report describing successfully using this circuit, but I came across this schematic on the Hoffman amps forum and thought it was interesting:

Zero gain mixer.jpg

One option that I think would probably be easiest if you wanted to have it perform similarly to a DC cathode follower and didn't want to use a morph control would be to AC couple the outputs of the pentode and triode stages and use a resistive mixer (either an Echoplex-style mixer or just a dual-ganged pot wired as two volume controls but with lugs 1 and 3 swapped for one gang such that turning one direction makes one side louder while making the other side quieter; the second option would still require mixing resistors off of each gang's wiper). From there you could send the mixed signal into the input of an AC cathode follower whose cathode is DC coupled to the grid of a DC cathode follower. That would provide a lot of current to drive the grid of the DC cathode follower with an added benefit of providing a softer startup condition for the DC cathode follower. Another option would be to use an anode follower, but at the expense of more noise thanks to the very large resistors required.

I don't think there's a clean solution that lets you "have your cake and eat it" strictly coupling the plates of the pentode and triode gain stages to the a DC cathode follower's grid. The two most common pentodes I know of for guitar amps -- EF86 and 5879 -- will probably not idle with similar voltage drop magnitudes across their plate resistors when compared against a 12AX7. Most schematics I've seen describe voltage drops considerably greater across the plate resistors in both cases as compared with a 12AX7 with the same value plate resistor and similar voltage. Additionally, pentodes generally have much higher output impedance than triodes, connecting connecting their plates as a kind of common plate load mixer would almost certainly involve a very heavy load for the pentode. A solution to would be to add resistance between it and the triode's plate, amounting to a kind of resistive mixer, but any kind of resistive mixer DC coupled to the cathode follower's grid would limit the amount of driving current available. That would limit the amount of current available to drive a DC cathode follower's grid current, which in this setup kind of defeats the purpose in my opinion since grid current is what causes the asymmetric wave form compression.

[cdemike]

This is the paired cathode follower suggestion I was describing. The input would need to be after some kind of resistive mixer.

Mixed inputs to drive DCCF.PNG

[soundmasterg]

For an active mixer, Fender did it with one of the Bassmans, AA864 I think. Most people didn't understand that circuit and modify that channel and get rid of it, but an active mixer can work very well and results in better bandwidth and less reduced gain when compared to a resistive mixer.

If you have an extra triode you could parallel the first stage in your proposed project instead of trying to mix a pentode and triode gain stage after the volume control. The pentode section would have more gain than the triode section and would have to be compensated for. You could hook each parallel input stage up to push pull pots and have different cathode resistors on each section for different gain levels, and switch one or the other out. I did that on a project and it works pretty well. I went with a parallel 12AX7, then a tone stack and volume control, then into a 5879 and then PI. The power tubes are 7868's so they don't need much gain in front of them to be driven. The amp goes from tweed clean levels to AC/DC territory when full up and is SUPER touch responsive, I think mainly due to the pentode being hit with the gain stage in front of it, and then pentode smashing the phase inverter, but only on one side of the waveform....so some asymmetrical distortion. I don't know though for sure...I never sat down and analyzed it on a scope after I got the sound I wanted.

Greg

[SixStringing]

Thanks everyone for the great ideas.

For this amp, I'll be using a 6SH7 as in the original Clubman circuit. From a quick glance at the data sheet, it seems to be able to handle slightly higher voltages than the 5879, so maybe they could be driven off the same supply? I'll be driving a pair of 6BM8s (using the triodes for LTP PI), so I don't need a huge amount of voltage going into the PI. B+ at the plates of the 6BM8s will be around 315V.

Yes, I was definitely hoping for a 'having my cake' moment - being able to blend/choose 2 different V2 circuits with one knob. A truly DC-coupled triode/pentode might just be too much to ask for one control knob - branching into 2 separate V2 circuits and mixing after coupling caps would probably make the most sense.

I was also planning on a parallel triode 1st stage as in the Clubman, and I like the idea of a push-pull to throw this stage off balance a bit.

I'll post up a proposed schematic once I sort things out.

[Stevem]

I In a way have played around with such a thing 35 years ago in ended up to be something like a 4-10" super reverb at least in terms of speakers .

Let me ask this of you that I think you need to come to to terms with first since you are putting so much effort and thought into just preamp sections.

Get to the bottom line first, make a list of what you like about a triode gain stage and a pentode gain stage, but leave the gain differences between both out of the picture, do this before you even build anything if you can.

Make your list consentrating on the harmonic differences they produce .

I built each preamp section first and drove headphones with each to leave the power amp section and the coloring of the speakers out of the picture and just focus on the harmonic content of each first.

I would suggest you do the same and then make another list of what you like and then see how what's in your head and in reality compare.

What I am getting at here is to me, I found that interms of a complete amp that each preamp section really wanted a different output section!

I will post back to you my findings and what I ended up doing when and if you reply back with what your bottom line likes are for each kind of preamp section .

[SixStringing]

Stevem - you make excellent points and I'm sure I'm getting ahead of myself on this concept I'll probably save this experiment for another build.

soundmasterg - even with the parallel triode in first stage, I'll have a spare triode. I do think I'll follow your build (and that of the Clubman) and put the tone stack between the paralleled first and pentode second stage, use the extra triode for a recovery stage for the effects loop, and use the one extra front knob for a pentode morph control (and maybe even push-pull for different bias/bypass of the parallel V1 stages). The send/return level pots will fit on the backside of the chassis.