EL84 "Tweedle Dee" Modded Amp

[markjwgraham]

Evening folks

I've been lurking for a few weeks, soaking up the info for both amp mods and a cabinet build. Just completed a first foray into one of my amp setups thanks to all the resources here, so thought I'd share.

The Amp

Photo 31-05-2024, 17 27 26.jpg

This was originally a small 18W EL84 head from a UK builder. They were based on the WEM Dominator circuit, which he called the Bantam. I got one of the early ones, and enjoyed it so much I asked him to make a 5E3 variant, same power section as the Bantam but with a 5E3 type preamp.

Just the bright channel high input, with a toggle for triode/pentode mode on V1. It has a post-PI volume control on the front, and power scaling on the back. So it's a little different from a stock 5E3 to begin with.

How it started

Photo 31-05-2024, 17 17 58.jpg

I love the Tweed woodiness to the tone, but generally gravitate to cleaner sounds for vocal accompaniment.

So I started with some basic Robinette mods, e.g. clipping the V2a bypass cap to get it a bit cleaner, then as I got braver I started some of the simpler Tweedle Dee mods. I eventually went the whole hog and swapped out all the preamp changes for the Dumble schematic, and upped the filter caps (from 22uF to 30uF).

I used 3 x 1M Ohm resistors for the phase inverter LNFB (they are what I had to hand), so the tiny bit of electrical tape is to stop the 'in the air' PTP circuit contacting the ground in the event it gets knocked. I've got some 3M Ohm resistors arriving soon to swap in.

Where we got to
I also put the bypass cap on a SPDT switch and placed it just above the triode/pentode switch. I had it working but then OCD made me swap the leads round so the 'up' position for both gives lower gain, and 'down' position for both gives higher gain. This gives 4 different configurations of gain from the front. Sounds incredible.

Photo 31-05-2024, 17 27 37.jpg

Left to do/tinker... I need to swap out those 3 x 1M Ohm resistors for a single 3M Ohm, and I may make the bypass cap footswitchable, but those are for another day.

Thanks again for all the resources, pleasure to be able to share this with you.

Mark

[Stephen1966]

Welcome Mark.

I am curious. As I understand it, the WEM Dominator had a LTPI, and when you say you "swapped out all the preamp changes for the Dumble schematic" did you also include the cathodyne PI?

What about a schematic??? I'd be curious to see how it stacks up against the WEM Dominator and the Tweedle Dee.

http://wem-owners.com/schematics/

dommk3.jpeg

[markjwgraham]

Thanks Stephen

I believe you are correct the original WEM was a LTPI, and the first Bantam I had was wired this one.

But with this Tweed variant, the builder had indeed used a cathodyne PI, per the 5E3 circuit. So I didn't have to change this over. However the output stage uses EL84 tubes like the WEM, rather than 6V6. It also has a solid state rectifier.

I'm happy to try and sort out a schematic to highlight the cumulative differences. I'll need to find a way to draw it out for sharing, let me know if you have any software/website recommendations (or I may have to look at pen and paper!)

[Stephen1966]

Thanks Stephen

I believe you are correct the original WEM was a LTPI, and the first Bantam I had was wired this one.

But with this Tweed variant, the builder had indeed used a cathodyne PI, per the 5E3 circuit. So I didn't have to change this over. However the output stage uses EL84 tubes like the WEM, rather than 6V6. It also has a solid state rectifier.

I'm happy to try and sort out a schematic to highlight the cumulative differences. I'll need to find a way to draw it out for sharing, let me know if you have any software/website recommendations (or I may have to look at pen and paper!)

Interesting. Unless you are going to use it a lot the learning curve for schematic software can be too much effort - so cutting and pasting parts of different (existing) schematics, then pencilling in anything left over is as good a record of your work as any. KiCad is free software and used by a few around here with great results. Some prefer Visio for making layouts (instead of schematics) but in point to point that can be a bit harder to figure out what the circuit actually is meant to be doing. Others should chime in here though because there are a lot of options out there and none of the applications are distinctly better or worse than their counterparts though some may be easier to get your head around than others. I use DipTrace.

In the Photo, the pair of MOSFETS coming off what appears to be the Variable Voltage regulator looks like the Hall VVR for fixed bias. Curious to see if this is still operational and how you got around it for the cathodyne PI. Do you ever use the Variable Voltage?

[markjwgraham]

Thanks Stephen, I'll see what I can do.

Regarding the power scaling, that came installed when I ordered it. I had it on the original model, so I asked for it on this.

I don't know too much about how it had been configured, but when you turn it down, you also have to pull the MV down to avoid fizziness. Perhaps an artifact of it being better suited to fixed bias vs cathode bias?

I used to use the scaling a lot, but like my headroom these days. I'm considering a 5E5 as a possible build experiment for this reason. I would consider upgrading the OT/PT on this one and switching to 2 x 6L6/EL34 output tubes, but I know nothing about the current transformers specs in order to make a fair assessment. If you have any advice please do share, but I wonder whether it may be simpler just to start a fresh build.

[Stephen1966]

Thanks Stephen, I'll see what I can do.

Regarding the power scaling, that came installed when I ordered it. I had it on the original model, so I asked for it on this.

I don't know too much about how it had been configured, but when you turn it down, you also have to pull the MV down to avoid fizziness. Perhaps an artifact of it being better suited to fixed bias vs cathode bias?

I used to use the scaling a lot, but like my headroom these days. I'm considering a 5E5 as a possible build experiment for this reason. I would consider upgrading the OT/PT on this one and switching to 2 x 6L6/EL34 output tubes, but I know nothing about the current transformers specs in order to make a fair assessment. If you have any advice please do share, but I wonder whether it may be simpler just to start a fresh build.

This is where a schematic would be helpful. The question is whether you need both a voltage regulator and a master volume for which, I'm assuming the latter is after the phase splitter - a dual gang pot maybe? I don't think it would make a difference whether the VVR is configured for fixed or cathode bias - as long as it's configured correctly - but it might affect a fizz if it is configured only for the power section, rather than the whole amp; a product of the different gain structures. More likely, the the fizz is because the minimum power is set too low. Even with cathode bias in the power section, the tone will start to fizz if the resistor on the ground side of the attenuating pot is too small. You might experiment with larger value resistors. You won't get the same amount of (or less) attenuation since the resistor sets the lower limit but this is not too noticeable anyway. VVR, Master volume, and pentode/triode switching all together though - do you need all those?

What kind of amp are you looking to transform this into? Is it headroom or grit, power or scalability? The Tweedle Dee is very pronounced in the mids and upper mids, is it more of the same or something a bit flatter, or something more scooped? You will find there are answers to all your questions when you start to drill down to what you really expect from it.

[markjwgraham]

Evening sir,

Current features
The VVR, PPIMV and triode/pentode switch were features on the original Bantam model. The idea behind them was the player could shift overdrive to pre- and/or power-amp with any of those controls. I'm also working on the schematic, but not sure I'm competent enough to include the VVR into it for you.

Amp Mods
Regarding amp mods, I'm not necessarily looking to transform the amp into anything specific. This is partly a learning experience (I'm comfortable with circuits but never worked on valve amps before), but also a process of slowly dialling in an amp to bring out more of the character I like about it.

For example, I like the woody tonal character of the 5E3, as it's a nice complement to my other amp which is a Fender 135W UL Twin. I don't want the 5E3 to be super sparkly, scooped or flat, etc, I've already got that covered. However the amount of dirt and breakup on the 5E3 was too much for my tastes, and I wanted more of the string sound to be retained even when cooking.

I found the Tweedle Dee mods (with 12AY7 in V1) really sharpened up the string sound and upper mids. More definition and clarity, even at gain. Putting the V2A bypass cap on a switch also helped nudge the response away from full compression and slightly stiffer feeling. Helpful that I can put that onto a footswitch too.

Power Section
My assumption with changing the power section to have more power would be that it would increase headroom and clarity in the tone and feel. It would also be a good learning experience.

Hopefully that's clear. I'd welcome your thoughts and input if you think anything would be a good avenue to pursue.

[Stephen1966]

Current features
The VVR, PPIMV and triode/pentode switch were features on the original Bantam model. The idea behind them was the player could shift overdrive to pre- and/or power-amp with any of those controls. I'm also working on the schematic, but not sure I'm competent enough to include the VVR into it for you.

Understood. As I understand it though, you want less dirt, more touch sensitivity but less compression.

Power Section
My assumption with changing the power section to have more power would be that it would increase headroom and clarity in the tone and feel. It would also be a good learning experience.

Hopefully that's clear. I'd welcome your thoughts and input if you think anything would be a good avenue to pursue.

More power would be my personal inclination here. Tubes with a lower Gm (6V6 or 6L6) should allow greater plate voltage swing. [EL34 and EL84 - higher Gm - so require less grid voltage swing to drive the plate voltage swing.] Also though, I would consider a reconfiguration of the whole 5E3 preamp with more negative feedback to give you the cleaner headroom.

V1a with LNFB => Tone stack with preamp boost/bypass => V1b => Master => LTPI with global feedback (presence etc optional) => x2 6_6s/EL34s

Feedback networks are going to give the headroom and there are ways you might switch them in and out for effect. You would need different transformers for more power. A PT with a fixed bias tap and a 5V (rectifier) tap or a separate transformer if you want to drive relays from a footswitch.

The choice of tone stack is going to influence the overall tone more than perhaps the choice of power tubes but you can experiment with them quite easily. https://www.guitarscience.net/tsc/info.htm

As for the maths, if you aren't going to cut and paste from other designs but (shock) do it "properly," you might want to pick up a copy of Blencowe's, "Designing Tube Preamps" as a starter https://www.valvewizard.co.uk/ along with grabbing yourself a copy of the Radiotron Designer's Handbook 4th Ed. [You can find the RDH4 free to dowload, online.]

[markjwgraham]

Thanks for all your input Stephen.

I've managed to install a GNFB loop today. 56k resistance with presence pot (10k with 0.1uF on centre lug to ground). My word. What a difference! Also reduced the noise floor enormously. Had to swap the OT primaries around and undo some lead dressing to get them long enough (insufficient OT secondaries to reverse them on the speaker jacks), but all done now.

Picked up the parts to add a Bassman tonestack but will leave it for now. The interaction of the 5E3 tonestack with the feedback and presence control is plenty to be experimenting with right now.

I'll be looking into rejigging the V1A/B triodes to be better deployed, per your suggestions.

[Stephen1966]

Congratulations! Isn't it great when a plan comes together?

[markjwgraham]

MJW Modified 5E3.gif

There we go! I've omitted the VVR just for the sake of space. It only operates on the B+1 and B+2 voltage lines.

One thing I do notice is just how little voltage dropping is going on. 1500 resistor and 8200 resistor, assuming I'm reading the values right.

Any thoughts on this?

[Stephen1966]

MJW Modified 5E3.gif

There we go! I've omitted the VVR just for the sake of space. It only operates on the B+1 and B+2 voltage lines.

One thing I do notice is just how little voltage dropping is going on. 1500 resistor and 8200 resistor, assuming I'm reading the values right.

Any thoughts on this?

Sure, do you have your voltages? You don't have any screen grid resistors in the schematic so the 1k5 dropping resistor will drop your screens' voltage 10V or so, which is not far off normal. 470 Ohm to 1k are typical values for this resistor. It's a doing a few things, but essentially, it's interacting with the B+ cap to provide an RC filter, and it keeps the voltage to the screens just below that going to the plates. I recommend you take a look at the Valve Wizard pages on output stages, single-ended and push-pull.

The 8k2 resistor could be allowing a higher than usual voltage to hit the cathodyne but it depends on the HT supply of the transformer and how much current is being drawn by cathode biased EL84s, with the 180 Ohm cathode bias resistor.

I would not worry about it, everything you have mentioned so far indicates the amp is functioning without any major issues. Personally, I find the cathodyne doesn't really like too hot a voltage but voltages higher than the design centre values found on datasheets can be exceeded without valves melting down. You might be looking at shorter valve life in the long term if they are biased too hot. The maximum datasheet values are on the conservative side.

If you have duplicated the Tweedle Dee preamp, did you also include the extra filter stage?

[markjwgraham]

I didn’t add the extra filter stage. Went with Robinettes minimal mods. I also undid the LNFB on the PI, 4.5k PI cathode bias and the 500 ohm v1 bias. Kept the 100k plate and tail resistors though.

I don’t have my voltages. Next step is to measure this stuff. I’d like to ascertain what the PT is capable of delivering to see if I can make 6v6 or even 6L6 work.

[Stephen1966]

I didn’t add the extra filter stage. Went with Robinettes minimal mods. I also undid the LNFB on the PI, 4.5k PI cathode bias and the 500 ohm v1 bias. Kept the 100k plate and tail resistors though.

I don’t have my voltages. Next step is to measure this stuff. I’d like to ascertain what the PT is capable of delivering to see if I can make 6v6 or even 6L6 work.

To be fair, it probably works okay but it's looking a long way from a Tweedle Dee. It will be useful to see the voltages and as a reference you might use the modern schematic attached as a guide, to record a comprehensive dataset in your schematic before you make any further modifications. You might find that some of your ideas don't pan out the way you expect them to, so it's a good idea to have the record to return to. It's also time to start looking at the tube datasheets if you are serious about changing the power section.

Fender-57-Deluxe-Schematic.pdf