DIYstompboxes.com

I'm in the middle of building a submini tube amp based on FrequencyCentrals [http://www.diystompboxes.com/smfforum/index.php?topic=78302.0]Superfly[/url].

I am a total noob with tubes, but I'm wondering if it would useful to do the same conversion with other amps that people like, similar to the tube-to-JFET conversion at Runoffgroove.

What sort of changes would need to be made to something like a Deluxe Reverb to use subminis instead of full-size tubes? Would copying the topology make for something pretty close in sound, or do the subminis inherently sound different from standard tubes?

I've seen the Submini Wreck (http://ampgarage.com/forum/viewtopic.php?t=6159), but am wondering about other classic amps.

I think that this is an awesome idea!  I don't have any experience with sub mini tubes, but I do have some decent experience with regular tube amps.  I think that it might be a good idea for me to do a sub mini copy of one of the amps that I own and compare the sound.  I have a Traynor Bass Mate and a Kalamazoo Model One that would both be fairly easy to copy, I been looking for a reason to try sub mini tubes and I think this is it!

http://www.dmitrynizh.com/submini-vibro-reverb.htm

Quote from: wavley on October 09, 2009, 03:48:19 PM
I think that this is an awesome idea!  I don't have any experience with sub mini tubes, but I do have some decent experience with regular tube amps.  I think that it might be a good idea for me to do a sub mini copy of one of the amps that I own and compare the sound.  I have a Traynor Bass Mate and a Kalamazoo Model One that would both be fairly easy to copy, I been looking for a reason to try sub mini tubes and I think this is it!

That would be awesome; I'd love to find out how these subminis compare to their big brothers in a fair comparison.

Quote from: frequencycentral on October 09, 2009, 03:48:47 PM
http://www.dmitrynizh.com/submini-vibro-reverb.htm

Whale Oil Beef Hooked! Awesome, thanks.

Puretube shooting at small tubes... (http://img.photobucket.com/albums/v437/latronax/tubeshoot174.jpg) :

(http://img.photobucket.com/albums/v437/latronax/tubeshoot175.jpg)



German-lesson: "Röhre" = tube;
                        "Röhrchen" = small tube;


:icon_razz:

I love Dmitry's page and it's wealth of information... but it'd be absolutely beautiful if the man had put in the specifics of the transformers he has used as well as the power supply requirements.

I recently learned something that is basic ohms law which states that the anode resistor to ground will give you the current required and that has made it easy to get the approx current needs for each stage, given the voltage is at the top of the anode resistor. That simple equation solves a bunch of problems with schematics like this but then we still need to be able to figure out what kind of transformers he is using for output and reverb duties. But then if we look at the bottom of the page I suppose we could guesstimate about 15mA at 230V which is about 15k but then by my reckoning we'ed want to bias the cathode at 5V at 5mA and that equals a resistor of 1k.

So, indeed, a very useful page. I'll be book marking that one.

And if one looks at merlins page (the valve wizard) hes got a bunch of reverb drivers that can be used with higher impedance tanks so one can bypass the reverb tranny using a single dual triode as a driver.

But, for myself, the beauty of full size tubes, is the amount of room available and the whole point to point thing which takes me away from PCB's and into a realm altogether different in space and scope. But who knows!

Here is a mini Train Wreck (using sub mini tubes) built by Dana (UR12 at 18watt and SE and PP watt) it sounds amazing.
http://www.ppwatt.com/node/16013
It was this thread that got me interested in sub mini designs then google found Ricks over here at DIY stompboxes.
He (Dana) also has a thread at amp garage on the micro wreck ,both places require registering to view,(no biggie if your interested in tube design)
I have seen several others basically scaled down versions of classic amps using sub minis.
When and as I remember them I will post links here.

A couple more sub-mini amp ideas.

http://www.jjs.at/electronic/index.html

dave

This page has a nice list of sub mini projects from around th eweb including the ones we already mentioned.
WHAT SUX is that sub mini Princton PDF is missing all the pics and schematics  :icon_evil:
I have the Powerman bookmarked I had forgot about it.
But ALL of the Harmonic Appliances projects are top shelf stuff !! Listen to his song High Tide.
What should be a big plus to you guys is all the schematics and data sheet links  :icon_mrgreen:
http://amps.zugster.net/projects/subminiature

I guess to pull the focus of this question back a little, I really am just interested in how subminis differ from full-size tubes. I'm wondering: what changes need to be made in general to adapt a circuit designed for other tubes to subminis?

Some submini tubes are just lower voltage versions of the more common high voltage amplifier tubes.
Make a lower power supply and the rest of the preamp/signal chain values would be the same I would think.

John

Depending on the type of tube, there are really very few things that matter.  For a triode, mu is the most important, then there are things like transconductance, etc. that are secondary concerns.  For a power tube, you have transconductance, load resistance, output power and maybe a few other factors.  If you can match the values between submini and regular tubes, then there shouldn't be much of a difference sonically, if any.  I am going to do a submini JCM800 after I get a little more settled, though I also plan to do a submini Mark IV preamp and a submini deluxe reverb preamp at some point.  Tubes are tubes, they operate on the same physical principles regardless of package.  It is similar (though not identical to) SMD and through hole components.  Of course, the differing plate voltages and such will change, so you might want to play around so that you get plate voltages that lie at the same point on the characteristics graphs as those for a regular tube (i.e., try to find a similar response characteristic and the plate voltage that produces it as opposed to trying to straight up match voltages or simply maxing voltages).  Hope that made sense.

Thanks guys, that's the sort of info I'm interested in.

Hi, I see this topic is dead, but I just finished my submini JCM800 with a Max1771 SMPS.

I used the 2203 preamp schematic and 6N17B russian tubes,
and the 5-tube push-pull Fender derivative power amp  from http://www.dmitrynizh.com/submini-vibro-reverb.htm (http://www.dmitrynizh.com/submini-vibro-reverb.htm) with the 6N16B russian tubes.

The tubes are running at 200V obtained with the Max1771 SMPS developed by Frequency Central, i just made another layout to fit between both computer speakers that I'm using.

OT  has a 22.5k primary.

I made a small PTP layout on a PCB. I'm using a BMP tonestack after the cathode follower and a master volume. I also made another output to uso a bigger cab.

Power supply is 12V 2A.

A made some videos with the cellphone camera:







Hope you enjoy! :icon_biggrin:

Thomas

i think that sounds pretty darn good!!

Sub-mini Princeton 5F2-A

http://courses.physics.illinois.edu/phys406/Student_Projects/Spring01/JWicks/5F2-A_Micro_Princeton_Schematic.pdf (http://courses.physics.illinois.edu/phys406/Student_Projects/Spring01/JWicks/5F2-A_Micro_Princeton_Schematic.pdf)

regards, Jack

Well, I'm still trying to make this JCM in a 1590b,
so I will add some material...feel free to make suggestions
Schematic without SMPS
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/SubminiJCM_zps580c84d5.jpg)
SMPS: http://www.desmith.net/NMdS/Electronics/NixiePSU.html (http://www.desmith.net/NMdS/Electronics/NixiePSU.html)

Layout with SMPS with 73.4x44mm size
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/JCM800112_zps308bc00b.png)

Tone circuit soldered direct to the potentiometers, there is space to use board mounted jack's, but I prefer this way...
This layout is still NOT TESTED, but the circuit is the same from the videos, I just need to verify how noise his layout will be.

Cheers
Thomas

Oh no... Just what I needed to see. My submini Class AB amp (6112, 6021, 5902, 5902) project is nearing the smoke test point. I'll share the OSHPark board if it works out ok... So now I have a couple of new projects depending on how long this thread gets!

Hi,
some updates:
I read some articles about SMPS layouts very interesting to reduce noise issues:
http://ecee.colorado.edu/~ecen5797/course_material/layout.pdf (http://ecee.colorado.edu/~ecen5797/course_material/layout.pdf)
http://cds.linear.com/docs/en/application-note/an136f.pdf (http://cds.linear.com/docs/en/application-note/an136f.pdf)
and made some changes in my layout, as the separated ground, smaller terminals for the sensitive points:
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/JCM1590a_zps93f84620.png)

Another change came from some reading of PP output stages. I wanted to calculate the load line for the 6N16B tube and Matec, from another forum, plotted the load lines for various voltages and impedances:
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/6n16b150V_zpsde1bcdad.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/6N16b180V_zps8072f9b6.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/6n16B200V_zps95b91de8.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/6n16B230V_zps88fa193c.jpg)

As you can see running the amp at 180V and the 22.5k transformer could be dangerous, a better configuration is the 230V with a 44k transformer.
It's easy to use the 22.5k transformer with the 4 ohms tap to obtain that impedance.
The new cathode resistor now is 2.2k instead of 1.2k as 2*2.1mA*2.2k=9.24V, very close to the 9.1V specified.

Before I start this new build I will recalculate each stage to verify the operation points. Next stage will be the PI!
Cheers
Thomas

are those 'sub-mini' tubes similar to the ones used on the z-vex nano head? if so, they can be a real pain in thes ass to change as most were designed to be soldered in. the thin leads have to be perfectly straight in order to fit properly into the tube socket & it usually takes several attempts, one of the primary reasons I opted for the rock block...it uses standard pre-amp tubes (i.e. 12ax7, ecc99 etc.) & has far more clean headroom than the z-vex which is little more than a glorified fuzz pedal running on russian satelite tubes.

Hi,
yes, they similar, but Z-vex uses the american versions. The socket is a problem indeed, I'm using IC's sockets so far.
The clean headroom in the nanohead is very reduced, as it only uses one triode in SE, it will only output 1/4W or so.
These one has two triodes in PP and a real Pi, not self-split like the Superfly. It's louder, something like 1W, like the load line shows. I'm redesigning the amp to obtain more headroom, and max. power from these tubes and the SMPS
The 12ax7/12au7/ecc99 amplifier is better, and the JTM1 clone will be one of my next projects, but i'm kind of a miniaturizing maniac, addicted to sub-miniature tubes.  ;D

@thomasha

Hello, any updates on your JCM800 project? Or even any news about JTM1 project?  :) Thanks in advance!

Hi,
after some soldering I finished the board, take a look:

(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/DSC_0444_zps0108f823.jpg)

After some testing, there was a lot of noise in the preamp, using only the PI and power stage it was nice and clean.

I'm drilling an old enclosure to test how the circuit behaves.
If it works like the first build, the proximity with the enclosure removes the SMPS noises.
Cheers,
Thomas

Wow Thomas! That's a lot on a small PCB...hope it works out.

Rick

Hi!
unfortunately that board was very noise... back to project....

The SMPS near the first stage was a horrible layout idea. I'm thinking of reducing it to 3 tubes, same as the little 1W marshalls and use a cathodyne PI. (more space, less heat)

And make a separate board for the SMPS, with a cooler, or heatsink the fet. The best fet I could find was the IRF644, but it still gets really hot after one hour playing...a better low ESR capacitor and inductor are in to buy list.

Just thinking of it's worth dropping the electrolytic capacitor for 2 film capacitors...easier to find at least.

Well, I made another video of the first build:

I salvage this same kind of speakers yesterday, I think this amp needs a brother, I just need to find a nice wood piece for it...

Thomas

Hi all,
new video, this time with a real speaker!


Cheers
Thomas

Hi,
I made another prototype, this time using 3 tubes. The PI is now a cathodyne, and the B+ is 234V.



This should fit in a 1590b box.

Cheers,
Thomas

I'm enjoying your submini tube odyssey! Cool build. I started this thread but never got my Superfly working back then. Maybe I'll try to resurrect it one of these days, as my building skills have improved somewhat in 6 years.   :icon_lol:

Hi,
this was already posted elsewhere, but for completeness, here is the final version of that amplifier in the video above!

(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/DSC_0383_zpsn1zpl0zd.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/DSC_0378_zpsfyxqnkn8.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/schematic%20submini_zps7ji4jw29.png)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/DSC_0934_zpskxhqa3uw.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/DSC_0941_zpsg7zlfrzd.jpg)

and I made a small doc. if someone is interested.
https://dl.dropboxusercontent.com/u/10420127/RussianJTM.pdf (https://dl.dropboxusercontent.com/u/10420127/RussianJTM.pdf)

cheers,
Thomas

That is one of the coolest things I have ever seen! :icon_biggrin:

Quote from: thomasha on July 22, 2015, 06:45:53 PM
Hi,
this was already posted elsewhere, but for completeness, here is the final version of that amplifier in the video above!



cheers,
Thomas

That is very cool, sounds nice too!
Wonderful and inspiring build!

cute-ist amp ever...sounds cool too.  8)

Hi,
made a new video of this amplifier:



and this:


cheers,
Thomas

ha ha i just saw that on youtube.. ;)

cool.. 8)

Fantastic build!

Another one to the list:

(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/Photo%2010-12-2016%2020%2016%2050_zps2mgmh3es.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/Photo%2010-12-2016%2020%2015%2048_zps4mobb8fj.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/Photo%2014-12-2016%2021%2017%2006_zps8jw02odl.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/Photo%2010-12-2016%2020%2017%2023_zpsa6i74owb.jpg)
(http://i1303.photobucket.com/albums/ag159/Thomas_Eduardt_Hafemann/Photo%2010-12-2016%2020%2018%2038_zpsyh7wz2ke.jpg)

same schematic as the previous one, but using the 6111 and 6112 tubes, that required some changes.
I used a prototyping board this time.
Cheers

Did you change the bias on the tubes or use the same as the Russian ones?

changed the cathode resistor to 1k.
Current for both output triodes is 8mA at 220v, roughly the 1.8W max. dissipation.

Wow, that is one great looking sub-mini tube build Thomas!  8)

Awesome!!!


Ham fisting my way through one build to the next since 2003. 

I haven't even come close to finishing my submini blackface style amp  :icon_lol: 

Lookin' great Thomas!

finally made a video for this one, is there someone still interested in these small amps?

That's great Thomas.
Yes I'm still intrested, I will eventually get around to doing one.
Don't give up the good work mate.

Alright, here it is in its open, unboxed form. Only took 3 years to make.



It isn't quite a classic amp topology but it is based on a Top Hat Club Deluxe which is based on an AC30 Top Boost channel. I used 6n17b for the preamp, 6n16b for the PI and 6p30b for the output stage. The whole thing runs off of a 12v DC adapter which then goes through 2 SMPS: one MAX1771 for the HV and one TPS564201 for the heaters. I'll post a full schematic when I get around to drawing one; the one I worked off of to build it has some mistakes and I changed quite a few values in order to get it working.

When I first plugged it in, it sounded grainy and was barely audible. So I went back in and replaced the PI circuit with the one Thomasha used for his JCM800, which didn't solve the problem. Then I noticed I forgot to ground the OT secondary, which made the whole amp stop making sound. Then I figured out I had flipped the secondary common and one of the primary leads  :icon_redface:. It's a wonder the thing output any sound at all, but somehow I didn't kill the tubes or the transformer so it's fine I guess. Now all that's left is to take the whole thing apart and put it in the box.

I'm definitely still interested in this subject. Metaldestroyer, that looks and sounds awesome! I'd love to glance at a schematic if/when you have time to draw one.

The problem I am having is mainly due to the fact that I'm new to building tube amps. Other than specific transformers that are named, I don't understand what to look for. I do understand the risks and have worked with high voltage circuits before.

I'm guessing the SMPS circuits negate the need for a power transformer, but what about the output transformer? What do I look for in the primary and secondary coils? Am I limited to only a few options or are there myriad choices out there? I'm keen to learn more about this component in particular, but Google falls a little short on the detail. Thanks in advance.

That TPS564201 looks like a great chip. Looking forward to seeing the schem. Great work!

Nice build, hoping to see some pictures too, it already looks interesting with those chicken head knobs.

You could also tell us more, if there were any problems with your build

I wanted for some time, to make something fail proof, so it would be easier to other members to build, but the SMPS is always the most problematic part and introduces a lot of noise if not tamed adequately.
If you managed to get around this would love to see your layout and which inductor  and mosfet you're using with the MAX1771.
I found some good inductors some time ago, but during me testing there was also a lot of them that just vibrated too much, and produced this coil squeal sound.
Another member showed me a guy at ebay selling smd SMPS that could deliver up to 20W at 200v. Maybe that could make the whole process easier.

You had luck with your build problems, the tubes and OT are more forgiving than the SMPS chip, I think I lost the track of how much I already burned just because one filter cap had a bad solder joint.

keep the good work going!

I'll just drop this off here  :icon_biggrin:

(http://i67.tinypic.com/2j15qol.jpg)

I'll do a full write up with schematics, stories, pictures, etc. when exams stop

wow that looks awesome, need to make some pictures like that.

Some time ago I finished another build, that was more like a remake of an older amplifier.
The superfly using russian tubes never had enough gain to me, (at lower volume settings...).
So I changed it to something in the line of the jjs amplifier with only one triode in SE.
The circuit and the looks where both night train inspired, and I named it the "Tiny Nightingale"

Here some pictures:
(https://i.imgur.com/oy5dwiel.png)
(https://i.imgur.com/9kmmoXal.png)
(https://i.imgur.com/A4G8t4wl.png)
(https://i.imgur.com/3aDcYXtl.jpg)
(https://i.imgur.com/tsxCUIxl.jpg)

Cheers,
Thomas

Can you make your pictures larger please?

Quote from: tonyharker on September 13, 2017, 08:33:06 AM
Can you make your pictures larger please?

If you take out the "m" at the end of his links, you can see the larger version.

Maybe 5 years ago I made a 2W plexi/JCM800.  It was based around a 1959 super lead, but when I saw how similar it was to the JCM800, I worked up a relay switching system that would allow me to be in plexi mode with either input 1 or inputs 1 and 2 jumped and then I could flick a switch (or use my foot switch) and go into a JCM800.  It sounds quite good, but I don't use it a ton.  I'm thinking of building something else into that chassis.

Hi, just adding another build!

This time using the 5672 direct heated tube in push-pull at only 87v.
The preamp still uses the 6N16B and 6N17B russian tubes. In the future I plan to use direct heated tubes for all the amp to go low filament current and all battery if possible....

For now I tried a AC30/liverpool mixed schematic for the preamp and I loved the sound of it!
It can blast 0.5W on a 8 ohm load when fully distorted!

The schematic follows>
(https://i.imgur.com/7pHWLQ5l.png)

As can be seen the last stage is biased using the 44 ohms and the 62 ohms resistors, which gives a grid bias of about -7.5v.
Using the 22.5k transformer from Hammond it only goes to class AB in with the maximum voltage swing.
Although there is a layout for it, to make it faster I used a perf board:
(https://i.imgur.com/PxwIvWEl.jpg)
(https://i.imgur.com/eUkHQyfl.jpg)

Some measurements:
(https://i.imgur.com/MtttwdZl.jpg?1)
Here is a measurement of the signal at the speaker terminals (8 ohms)

This build also required the traditional small combo, as used in my previous builds. This time I wen't with a semi finished box, cut it in half, remade the finger joints in the bottom and it ended up looking like this:
(https://i.imgur.com/nxSkGVkl.jpg)
(https://i.imgur.com/FuCtd4Nl.jpg)
(https://i.imgur.com/DOlfm8Ll.jpg)
(https://i.imgur.com/hwCUCW4l.jpg)
(https://i.imgur.com/hgfghgPl.jpg)

To keep it simple I just used some wax this time,
I liked it better than the shiny finish of my first builds (less finger prints).

I'm planning to make some videos soon, to show how it sounds.
Not as loud as the previous builds (what I was looking for! I haven't added a master volume in this one), but as good sounding.
The advantages are:
-less filament current than using 3 or 4 russian tubes. (500mA against 800mA)
-push-pull output stage which gives (IMO) a better tone.
-lower wattage (around 200mW ), TV output level in the small speaker or even higher if using proper 12" speakers.

As I said before, If I could do it only using direct heated tubes with series filaments this could have an even lower filament current, which would allow using some 3.7v Li-Ion batteries (3x to have almost 12v).

Cheers

That is so amazingly cool!  I'm working on a little headphone amp with integrated cab sim and overdrive that will get a dovetailed walnut head cab, but that little Vox is just so cool!

I made a quick video clean/dirty with the volume maxed,
to add the distortion I increase the guitar's volume.

Trebble was like 3 o'clock, while bass was at 9 o'clock.

I know this thread is semi-old, but I just happened upon a stash of 6N16, 6N17, and 6P30, and I'm commenting here to remind myself to do something cool with them :P

Quote from: vigilante397 on March 20, 2018, 01:19:54 PM
I know this thread is semi-old . . .

This thread is never too old for stuff like this:

Quote from: thomasha on September 12, 2017, 06:01:43 PM
(https://i.imgur.com/3aDcYXtl.jpg)

8)

Quote from: vigilante397 on March 20, 2018, 01:19:54 PM
. . . but I just happened upon a stash of 6N16, 6N17, and 6P30, and I'm commenting here to remind myself to do something cool with them :P

I wonder if any of Rick's old submini projects would suit these, perhaps with a bit of tweaking?  I've built a few and really like them.

The problem I seem to be having is that half the links in here are dead and the other half are blocked by the firewall at work :P

I'm thinking I want to make a little high-gain amp because that's the main tone I don't really have covered in my current amp collection. I was hoping Metaldestroyer would post details on his amp, but he seems to have vanished. Oh well, to the breadboard then ::)

Sorry I disappeared for a year. I decided the best place to put the write-up would be my own site and it took me until now to get around to it. Here it is everyone:

http://projectstudioproject.com/2018/05/30/subminiature-tube-amplifier/ (http://projectstudioproject.com/2018/05/30/subminiature-tube-amplifier/)

Hi guys, I already posted over madbeans, but maybe there is someone interested in this one here too.

I started another subminiature amplifier where the main idea was to use the pencil tubes I got over ebay. I have some old 5678 and 5672s that are already used, but still working. Someone probably took them from a military radio or something.

This time I aimed at a mesa boogie kind of distortion, similar to what was done with the DR Boogei pedal. Actually I started working on the preamp, using the 5678 in triode mode.

I started from this schematic, since it has some voltages.
(https://music-electronics-forum.com/attachment.php?attachmentid=36025&d=1444386792)
It was originally posted here: https://music-electronics-forum.com/showthread.php?t=40525 (https://music-electronics-forum.com/showthread.php?t=40525)

I checked the load lines just to find out that most of the stages are center biased, with exception of the cold clipper with the 39k resistor. My first attempt already sounded pretty nice:


But it required a proper design.

For the 5678 the load lines at 45v or 60V look like this:
(https://i.imgur.com/lYwKGBhl.png)
Here the values
(https://i.imgur.com/XjcoMyDl.png)

Since the 5678 and the 5672 have directly heated cathodes, the cathode follower would be really difficult to implement without a proper transformer. Here I wanted something that could run from 9V and use as little current as possible. In a series string the filaments only require 50mA!

So for the cathode follower stage I used a lower gain/ higher current stage, with a 20k plate resistor. Not sure if this is the best approach. What would you do?

The important values at the table are the cathode voltages, since I will have to bias each stage at that value. In the meantime I tried to make a spice model for the 5678 tube to simulate it, and compare with the simulation of the original preamp. Therefore I used the paint_kit tool. My lines look like this:

(https://i.imgur.com/ZXf3GVo.png)

And the schematic:
(https://i.imgur.com/bi91WwN.png)

I was very confident as I compared both sine waves, since both have a similar shape. I could get it even closer by adjusting the tone.
The only difference is the voltage swing, since in the simulation I was testing a new voltage divider before the last 5678 tube, where the cathode follower, with unity gain, should be.
(https://i.imgur.com/PpdU92vl.png)
(https://i.imgur.com/8OVbbQWm.jpg)

For the output stage I tested some different configurations, but ended up using the 5672 tube. I even used the specified values at the datasheet, with 65V and a bias of -6.5V.

Since all the filaments are in series I decoupled the signal to ground at each filament and used fixed bias as the difference between the cathode voltage and the grid leak resistor of the previous stage. For example, the seconda stage has it's cathode at 2.5V (negative side of the filament), while the grid is at 1.4V approx., resulting in a bias of -1.1V. I Used a trimpot to adjust the voltages and also decoupled every trimpot, so that the signal sees the trimpot as ground. The extra current from the anode is really low in comparison to the filament current (0.1mA vs. 50mA), so that no parallel resistor is required.

For the output stage, on the other hand, there should be at least 3.3mA, so that I added a 1.5k resistor to ground (filament is at 7.5V). The final schematic looks like this:
(https://i.imgur.com/AbKv2Oc.png)

So the good news is that it sounds awesome, and is really quiet. The bad news is that the 5678 is quite microphonic. I did all my testing over the table, and as I put the board over my 1x12" cab it had some ringing. I could use the 5672 in place of the 5678, but the filaments must be reversed, or I build the amp inside a small cab where I could use some rubber feet to dampen the vibration. Actually it's not that bad, since it's not that loud, and I can always put it on the nearest unplugged amp.

I have some 5672 that I used in other projects, which I could remove just to test the circuit and make a proper amp, without the microphonics. What do you think?

> ...directly heated cathodes, the cathode follower would be really difficult to implement  ... So for the cathode follower stage I used a lower gain/ higher current stage, with a 20k plate resistor. Not sure if this is the best approach. What would you do?

I would do Anode Follower. Grounded cathode (filament battery), Plate-loaded, with ~~unity-gain NFB from plate (after coupling cap) to grid to source. Since real work happens, I would run it as Pentode (G2 to 45V), and looks like 47k plate load for best swing.

Thank you PRR! I just tested a NFB 500k pot in the existing circuit and it sounded better, less hairy.
The output dropped a little bit, but that's because I have the voltage divider before the 80k grid stopper.
The cathode still was at -1V, but bypassed with the 100uF capacitor.

Is DC coupling a better option than AC? I don't think I can use a DC coupling scheme, because my grid is elevated to be at the same voltage as the cathode in the filament string, which is at around 6.25V at the negative side.

I imagine I could bypass the upper 470k resistor and have the full blast of the previous stage at the grid stopper. This way I could decrease the gain of the anode follower and still have a voltage swing big enough to achieve the full output of the power stage (65mW),  with it's grid at -6.25V (12.5V p-p for full power).

Hi,
After some tests, the final result
(https://i.imgur.com/2YvHw6Z.png)
1 - I moved the tonestack one stage forward, so that there is still a driver stage for the 5672. The cathode follower has no gain anyway, and I didn't want to add another 5678 to the circuit, 6 tubes seems enough.

2 - I added a clean switch, to bypass some stages and get a nice clean or overdrive with the gain maxed.

3 - The SMPS I used was based on the 555, with some minor changes, my layout for the MAX1771 was too noisy.

4 - Some stages have a bias trimpot. The bias is calculated between the negative side of the filament and the grid voltage. Since the filament voltaeg is fixed by the series string, I adjust the grid voltage, as in a fixed bias amplifier. Third stage is biased at almost 2.25V, close to cut-off, a in the mesa. The 5672 is biased at the same time, and has a cold bias, as specified in the datasheet.

Here is the board
(https://i.imgur.com/8EpvP45l.jpg)
(https://i.imgur.com/sHJdiiRl.jpg)

I fixed the metal pin of the socket directly to the board. A slightly undersized hole keeps it in place.

Here, another board with an extra RC filter for the 70V.
(https://i.imgur.com/NcXNivcl.jpg)

I adjusted the trimpots before mounting it in the box. So I could also make some measurements.

The tonestack was monted in the box. Since I only had a PCB potentiometer for the mids I had to use a small board and solder it to the neighbours. I thought about doing a tonestack board, but there is no space for resistors or capacitors between the pots, so it makes no sence. This way I can still test different tonestacks
(https://i.imgur.com/2Z7NZxbl.jpg)

I had to sand the lid a little bit, so that it would fit
(https://i.imgur.com/Ndinmgbl.jpg)

The etch was inspider by Homer and Scylla, the six headed monster. (six tubes ...)
(https://i.imgur.com/A9GA1njl.jpg)
(https://i.imgur.com/NZQZU4ql.jpg)

The SMPS would not fit in the box, so I butchered it!
(https://i.imgur.com/0cNNRQBl.jpg)
(https://i.imgur.com/wJHlE1xl.jpg)
I had to cut it, and turn it 90 degrees, so that it would still fit.
(https://i.imgur.com/IsQsqwWl.jpg)
The tubes were soldered to the sockets to make things easier
(https://i.imgur.com/hSPoiCgl.jpg)

And the final result
(https://i.imgur.com/dRrFQxjl.jpg)
(https://i.imgur.com/45KoA1nl.jpg)
(https://i.imgur.com/u93308vl.jpg)

Here a not so short video, in case you're wondering how it sounds.

Hey Thomas, your amps are inspiring! I was looking at the Bassman Micro project recently (https://robrobinette.com/Bassman_Micro.htm#Bassman_Micro_LTP_EF80), as the idea of building a baby bassman/plexi type amp is very appealing. The designer Rob has recently added an EF80 output stage version of his design to emulate pentode overdriven sounds.

I was wondering if I might ask you if you thought combining the preamp of the Bassman Micro with the output section of your high gain/5672 output design might be possible? Essentially I would like to build a Bassman/Plexi with an output of way under 1/2W, but I'm not concerned about small packaging! Cheers.

If not concerned with small packaging why not use regular tubes? Just use a low power output tube like a 6AK6 and run it at lower voltage?

Yes I am open to any solution. 6AK6, EF80 or subminiature.

Part of my attraction to submini tubes was I suppose an expectation that there may be an option out there to run a pair of tubes in push pull closer to (or a bit past) their limits, and still have a nice low audio output. Rather than running larger tubes capable of a greater output colder to achieve similar audio output.

Ideally I'm looking to build a power amp section of 1/4W or less, perhaps 100mW is my ideal, and pair this with the Plexi style pre amp section from Rob Robinett's Bassman Micro.

If you really want a pentode power amp, you can give it a try on the EF184 tube. It's cheap, and I believe you'll have a low output power from a pair of them in PP.