A 1959-ish 10W PP Class AB with 6K6s

[GrayDigger]

I recently finished a fun little project amp I call the “Marshmellow”. It’s a two-channel 1959 Super Lead-inspired 10-watt push-pull class AB amplifier using 6K6 pentodes.

Last year, a good friend wanted me to build him a “10-watt Plexi”. He wanted something that sounded reminiscent of a Super Lead/Bass, with decent headroom, yet could definitely get dirty. He wanted around 10 watts mainly for studio and practice.

Sounded like a fun project. But getting 10 watts in Class AB at turned into a bit of a head scratcher. To my knowledge, most of the micro/mini’s (1-5 wart) employ triodes – or if using a pentode – operate in Class A. Dialing down an 18+ watt AB amp (like a 1974) wasn’t attractive – losing the full power IMD loses much of the magic door me. So onward with the quest for a “tweener”.

I’ve built 6G15s clones with 6K6s as reverb drivers and thought these might be interesting to explore. They are plentiful and cheap (you can get a lifetime supply of NOS RCA 6K6s for less than a pair of current production EL34s). They fit a standard octal socket, have “soft” grid curve knees (true pentode vs. a beam tetrode) with similar distribution to EL34s, and have a max plate dissipation of 8.5 watts.

On the downside, the published max plate voltage is 315V. For push-pull class AB power amp, their relatively low power required a high OT primary impedance to generate load lines I liked. Although they were used a lot for radio and TV, I wasn’t aware of (m)any guitar amps that specified them. So… instead of being able to rip-off an existing design, I ended up designing one from scratch.

As I doodled around, I was struck by the relative paucity of information on the 6K6. The RCA data sheets are relatively sparse compared to other power pentodes. A 6K6 Spice model is available from (the awesome) Ayumi Nakabayashi, but I didn’t know how good a large signal model it was. So, I ended up designing with pencil and paper. I used LTSpice as a reality check. Once I was happy with my power amp, I used LTSpice extensively to tweak and voice the 1959 pre-amp to work with my power section. My longhand calculations got me right around 10 Watts.

For the build, I decided to:
* Use a standard BF PR chassis and cabinet
* Do it on the (relatively) cheap

I designed around the following transformers/choke:
* Hammond 290AX (VibroChamp) PT
* Hammond 155H (5H/50mA) choke
* Hammond 125E 15-Watt Universal PP OT

The primary focus was on making a 6K6 version that sounded (at least to me) close to a late 60’s Plexi yet keep it in the neighborhood of 10 watts at full power. I wanted goodly amounts of power tube distortion to mimic the “no master volume” originals.

But I couldn’t resist adding some extra stuff to my prototype, which included:
* Incorporating switchable SS/Tube rectification
* Making the power amp switchable between 6K6/6V6
* Using the extra 9 pin socket to add a transformerless Reverb.

Even with all the extra features, the build came out to less than $1k, including tubes, cover and all.

The playing and sound clips are not pro - just close mic’d, flat EQ with a touch of the reverb I built into the amp, MV dimed, gain noon to 3 o’clock.

As far as its primary goals, I think it sounds pretty much like what we had in mind. The reverb is a nice touch and doesn’t add much complexity to the build. I’d recommend including it (I used the PI B+ to supply the 12DW7 plates and decreased the final voltage dropping resistor between CF and Pre to 10K to compensate for the additional current draw).

If anybody is remotely interested in building it:

* I’d recommend keeping it simple and sticking with 6K6s. I’d also go with a tube rectifier. You can safely use anything from 5Y3 up to a 5AR4 and stay under 15W (My sweet spot is a 5R4 potato masher). The Hammond 290AX supplies up to 3 amps AC to the rectifier filaments, so you can run pretty much whatever rectifier you want. For more punch and less sag, you can upsize the reservoir caps to 30uF/30uF, or even go a bit higher. It’s been plenty loud (dimed with boost around 13W) for rehearsals and gigging in small clubs (about like a 5E3); stage volume is good even with a loud drummer. It is quite pedal friendly, the sustain is unreal.

* I was unable to directly upload my LTSpice models but would be happy to share.

* I also drew up an embarrassingly lame wiring diagram on the generic BF eyelet board (no reverb, 6K6, tube rectified version) for another friend (novice builder), and thought I’d include it in the attachments. (It looks like I did it in crayon. But that’s my level of drawing. There may be errors, so use at your own risk. I originally placed the PPIMV in the back of the amp where the ground switch cutout is, but the WD shows it on the front panel, which requires drilling an extra hole in the faceplate).

Thanks for the opportunity to share!

[brewdude]

I recently built a single channel inspired by the 6G3 deluxe using 6K6GT’s. I used Edcor iron.
http://www.ppwatt.com/node/18343
http://www.ppwatt.com/node/18344

6K6-GT GE Datasheet.pdf

I ended up rewiring the amp recently to utilize a 6G6B style tone stack. I love it now.
http://www.ppwatt.com/node/18407

6K6GT Blonde'ish 2.2.1.pdf

I just realized that one probably has to be a member of the Watkins forums to see the linked threads…

[romberg]

Cool project! Thanks for sharing. I bet it still gets plenty loud. No doubt the master volume on the back really helps make the thing sing at low volume.

Mike

[GrayDigger]

Thanks Mike.

Yes, I guess most of the time it’s loud enough even with our drummer

I’m always amazed at the depth of knowledge and wisdom so freely shared by members on this forum. So grateful for the opportunity to spend the wee hours (OK, too many of them) exploring a seemingly endless stream of incredibly cool rabbit holes, seeing the cool things folks have done, great techniques for tackling thorny issues, and references to great resources.

I think with a little of it sinking in, it’s really neat when projects turn out the the way you envision it.

Thanks forum!

[pdf64]

I’d never noticed that the 6K6 was a suppressor grid, rather than a beam, pentode.
And it’s the only info where I’ve noticed forward biased control grid current charted. I wouldn’t think that the control grid on such a small output valve could cope with 1.5W.

[martin manning]

And it’s the only info where I’ve noticed forward biased control grid current charted.

GE 6L6GC data sheet shows g1 current too.

[bepone]

2E22, SRS551,RS1003, many others too..

all possible to utilise in AB2 class, to squeeze out more power,

for transmitter tube RS1003 single ended, in A1 triode mode , it is possible to extract 12W and in A2 class 20w!! almost double, plus increased linearity

[nworbetan]

I wouldn’t think that the control grid on such a small output valve could cope with 1.5W.

The same control grid would probably have a harder time with 1.5W if it were inside a plate structure that was dissipating 25 watts.

[pdf64]

And it’s the only info where I’ve noticed forward biased control grid current charted.

GE 6L6GC data sheet shows g1 current too.

Thanks, I must’ve looked at those charts a zillion times but never noticed that, just assumed they were g2 current traces
https://tubedata.altanatubes.com.br/she ... /6L6GC.pdf