EF86 driving an ODS style tonestack?

[Lynxtrap]

I kind of already know that the answer is "won't work well". I found a topic with a similar question from 2009, although with no conlusive solution.

I have loose plans for a pedal steel amp, and got inspired by the Dr Z Surgical Steel amp. It uses an EF86, TMB tonestack and 2xKT88 (UL), presumably an LPT PI. That's all I know about the circuit.

Having a small stash of NOS EF86s and experience with the Route 66 and Z28 preamp, I thought that topology might be a good idea (perhaps with KT66s and NFB with a presence control).

Also, the ODS type tonestack is IMO the best there is (of the ones I've built and experimented with), so I immediately thought that would be the one to go for.

The goal here is obviously maximum output (in the 50-100W range) with the least amound of distortion and a well balanced tone from a simple and straightforward circuit. I'm not ready to get into a SSS build, and also don't want a Twin Reverb type amp.

Now, having a stash of tubes is no good reason to choose tube types, so I'm prepared to change my mind on that

So I guess the question is, what is your take on this? How much does the ODS tonestack mess with impedances and drop gain? Grateful for any advice and ideas!

[FUCHSAUDIO]

Here's the issue: Pentodes can't drive much of a load, and a tone stack is a somewhat complex load. This is why most Pentode preamps have simple high cut pots or low frequency rolloff switches etc.

I've made a few pentode based custom builds for clients, and the easiest bet is to add a cathode follower (one half of a 12AT7 works well), drive the stack with that, and use the extra stage elsewhere.

If you want more gain, use a coupling cap from the CF into the stack, add a gain pot after the stack, and then add a post tone stack gain stage (with that extra tube section) with another volume control on the output, so you can even get some overdrive.

If you don't want nor need overdrive, direct couple the pentode plate to the CF grid, which eliminates a few parts. You can use a 12AX7 for more gain and overdrive potential.

Have fun.

[johnnyreece]

Here's the issue: Pentodes can't drive much of a load, and a tone stack is a somewhat complex load. This is why most Pentode preamps have simple high cut pots or low frequency rolloff switches etc.

I've made a few pentode based custom builds for clients, and the easiest bet is to add a cathode follower (one half of a 12AT7 works well), drive the stack with that, and use the extra stage elsewhere.

If you want more gain, use a coupling cap from the CF into the stack, add a gain pot after the stack, and then add a post tone stack gain stage (with that extra tube section) with another volume control on the output, so you can even get some overdrive.

If you don't want nor need overdrive, direct couple the pentode plate to the CF grid, which eliminates a few parts. You can use a 12AX7 for more gain and overdrive potential.

Have fun.

A one-tube option for this might be a 6U8. It's a triode/pentode combo. I believe the pentode is EF86-ish, and the triode is 12AT7-ish. Just a thought!

[roberto]

I've made a few pentode based custom builds for clients, and the easiest bet is to add a cathode follower (one half of a 12AT7 works well), drive the stack with that, and use the extra stage elsewhere.

Do you change the TS to match the eq response with the fact the input impedance is passed from something less than 40k to around 200 Ohm?

If you don't want nor need overdrive, direct couple the pentode plate to the CF grid, which eliminates a few parts. You can use a 12AX7 for more gain and overdrive potential.

To reduce parts even further, you can supply g2 of the EF86 from a voltage divider used as load of the CF: split the 100k into 33k and 68k and connect EF86's g2 in the mid point: this way you'll change the curves of the EF86 into something inbetween a pentode and a triode (you oblige it working in UL). The loadline becomes less dependant from g2 as well, because the knee of the pentode becomes wider and softer: within a certain range you can decide how the tube compresses and reacts when approaching g1=0 V.

[10thTx]

You can use a mosfet cathode follower after the EF86 and then the OD tone stack. I can NOT tell any difference in tone between a tube triode or a mosfet as a cathode follower.

Google "mosfet follies"

IF you are wanting to just experiment with a pentode in an OD section, I'll offer this idea and you can substitute an EF86. I've used a 5879 a number of times as an OD tube and like it. I've tried a 5879 tube alone, a 5879 tube with a cathode follow and 12A_7 triodes into 5879's. I've tried numerous tone stacks in the OD with mixed results and satisfaction. This James Tone stack can give a mid scooped tone or a psuedo mid boost tone. (See Duncan Tone Stack Calculator for example) So far, it has easily been my favorite in the OD topology of all I've tried. See attached schematic

With respect, 10thtx

[roberto]

A source follower is indeed another option, if designed correctly and far from clipping.
I'd avoid what you usually find online (IRF820 et similia), and I'd go to something like the STU9HN65M2 or newer.

[Lynxtrap]

Here's the issue: Pentodes can't drive much of a load, and a tone stack is a somewhat complex load. This is why most Pentode preamps have simple high cut pots or low frequency rolloff switches etc.

I've made a few pentode based custom builds for clients, and the easiest bet is to add a cathode follower (one half of a 12AT7 works well), drive the stack with that, and use the extra stage elsewhere.

If you want more gain, use a coupling cap from the CF into the stack, add a gain pot after the stack, and then add a post tone stack gain stage (with that extra tube section) with another volume control on the output, so you can even get some overdrive.

If you don't want nor need overdrive, direct couple the pentode plate to the CF grid, which eliminates a few parts. You can use a 12AX7 for more gain and overdrive potential.

Have fun.

Thanks! I don't need overdrive. But is there a possibility/risk that the CF will be overdriven when hit by the pentode coupled straight to the grid?

[roberto]

If the gate of the source follower is half the supply voltage, and you have the pentode as first tube, you can't normally overdrivde it.

[Lynxtrap]

If the gate of the source follower is half the supply voltage, and you have the pentode as first tube, you can't normally overdrivde it.

You mean something like this..?

[roberto]

I meant DC coupled, so working on previous stage working point to set the (source in that case, cathode in your case) follower working point.

[Lynxtrap]

Not sure I got this right, but do you all think something like this would work? Perhaps even an ECC82 for the triodes?
The component values are just aproximations.

Preamp draft.jpg

[roberto]

It's an idea taken from Rupert Neve early mixers' input of the '60s.

Rupert Neve.png

A possible evolution is chosing the required g2 feedback and using it as a recovery stage before the cathodine PI.

Rupert Neve evolution.jpg

[Lynxtrap]

It's an idea taken from Rupert Neve early mixers' input of the '60s.

Thanks! The Neve input looks interesting. But what exactly is happening there? Just an easy way to keep the screen grid at a lower voltage?
What happens to that voltage when current throug the triode varies? Will there be some kind of negative feedback effect?

[roberto]

In Neve configuration the 220n to ground is a lowpass at 6 Hz: it is a way to have a lower impedance supply for the preamp screen, so a better regulation.
On the other configuration I posted there's a-g2 feedback on the EF86 through the cathode follower.

[Lynxtrap]

I made a first draft, using the basic idea for the input stage and CF posted by Roberto. The basis of the tonestack is a version I've used previously for a bass preamp and the mid boost borrowed from an ODS clone I own. just to start somewhere. The triodes are a 12AU7.

Would I need to change the component values in the tonestack as it is driven by a CF here? I know Blencowe recommends altering the values depending on the source impedance, because of noise issues as I understand it, but would there be other consequences?

Steel TS.jpg