running tubes in push-pull at less than max output

[stormbringer]

I've been thinking about low-powered amps as of late. Here are three different approaches on my mind, with some example amps for each approach:

1. Single-ended (or parallel SE). Examples include the Fargen Mini Plex and the ThD Univalve and Bivalve.
2. Small-signal dual triode (e.g., 12BH7) in push-pull. Examples include the Snott Watt from Baron Custom Amps
3. standard PP designed for less-than-maximum output from the tubes. I don't know of any examples of amps of this type, but London Power sells power amp kits with octal tubes limited to 10 or 25 watts.

Has anyone here made a low-powered amp with octal power tubes running in PP? What would be the disadvantage of a 10-20 watt amp with 2xEL34s running in PP?

Cheers, John

[Gingertube]

Stormbringer,
Google ECC99 push pull, you will see heaps of examples, even a youtube video. ECC99 is a dual triode in an 9 pin socket, similar to 12BH7. Also covered at AX84

If you want octal then 6SN7 dual triode is the way to go.

For a bit more power (approx 12 Watts use 6V6)

London Power uses Power Scaling - an adjustable High Voltage Regulator with a tracking bias supply regulator which allows the amp power to be dialed from full power down to milli watts. I built one of these, I used a 50W Marshall Output Tranny driven by a quad of 6V6 for 35 Watts maximum and about 0.2 Watts minimum, a pair of EL34 could easily be substituted for the 4 off 6V6.

EL34 sound stunning in cathode bias at 375V for 35 Watts maximum output (much nicer than being hammered at 470V and fixed bias for 55 Watts). You could try that with a HV regulator to dial down power, with cathode bias you don't need a tracking bias supply regulator.

Simplest HV regulator is a MOSFET between the rectifier and filter caps, put a pot between the rectifier out and ground with the wiper going (via a 1k gate stop resistor) to the MOSFET gate. Series Diode between MOSFET and Filter cap to stop the cap back feeding at switch off. Also don't forget protection zener from MOSFET gate to source. With about 3 more components you can build in an adjustable SAG for Tube rectifier like compression.

Cheers,
Ian

[stormbringer]

Thanks Ian.

I've installed KOC's SB-2 power scaling kit (or whatever the budget kit for cathode-biased amps was called) in a project amp, and I'll put it in my next build along with the voltage clamping kit. That next amp is a 10W SE amp with one octal power tube. Now I'm interested more in the possibility of a 10-ish watt amp with 2xEL34s (with or without power scaling). Basically I see a lot of low-powered SE amps, but I wonder if there could be some tonal advantages to running the usual power tubes in PP. I know I can set up those tubes to run at 50Watts and then use power scaling to bring down the power and volume, but there are a lot of people out there who simply don't need 50W and would prefer a dedicated 0-10W offering.

Your recommendation of the EL34s in cathode bias at 375V for 35 Watts interests me quite a bit; although the possibility of even less power is appealing too in this new world of apartment and bedroom players (like me).

The London power kits I referred to are the new-ish PA66-10 and PA66-25 kits. Here's a link: Tube Power Amp Kits - - PA66-10 - Octal Push-Pull Power Amp Kit: PCB+Components+10W TX. Other than these kits, I don't know of any amps with these types of 'low-powered' output sections.

BTW - I'm a relatively new member at powerscaling.com (jmcd)

Cheers, John

[Gingertube]

John,
It seems that the max power of 10 Watts is dictated by the rating of the output tranny supplied with that kit. It is unclear whether the board includes the Power Scaling components to "restrict" the biggger octal tubes to 10 Watts out but I think not, those components should be on the recommended power supply. Those are questions to post over there. Having the ability to swap 6L6 to 6V6 to 6CA7 (EL34) etc. should allow you to find something to keep you happy. While I've built lots of amps of various powers for other folk The amp I buit for myself is push pull vintage 6V6G (the ST "Coke Bottle Shape) for around 12 watts. Even when playing with other guys in someones back yard that has allways been plenty of power.

The day of the "Stadium Crusher" 100+ Watt Amp is dead, thankfully we are returning to the quality not quantity style of amp. The attitude of the 70's and 80's players of "play it loud enough and they won't know if you are any good or not" is thankfully also dead.

Cheers,
Ian

[J M Fahey]

EL34 sound stunning in cathode bias at 375V for 35 Watts maximum output (much nicer than being hammered at 470V and fixed bias for 55 Watts).

+1

[pdf64]

'EL34 sound stunning in cathode bias at 375V for 35 Watts'
What OT primary impedance have you found works well for these conditions?
3k5, as per the Mullard data sheet? http://www.mif.pg.gda.pl/homepages/f...129/e/EL34.pdf
Pete

[MWJB]

Run those EL34 in cathode-fed, self-split and you'll get around 10W from them....a PI can be switched in/out easily with a SPST grounding out the 2nd power tube's grid, to toggle between 10W & 30W-ish in cathode bias (things get a bit more complicated if you are intending to use fixed bias...which you wouldn't if aiming for low volume).

[Bruce / Mission Amps]

'EL34 sound stunning in cathode bias at 375V for 35 Watts'
What OT primary impedance have you found works well for these conditions?
3k5, as per the Mullard data sheet? http://www.mif.pg.gda.pl/homepages/f...129/e/EL34.pdf
Pete

Cathode biased amps sound great.
I've built lots of cathode biased amps... but, I've never been able to get a clean 35 watts out of any one of them when using a pair of 6L6s, EL34s, KT66s.. etc etc... in cathode bias.
I'd like to see a schem of that amp and some data... maybe an old dog can learn some new tricks too.

[Tage]

Here is an amp close to Ginger tubes description. Not sure if the schematic is 100% accurate but I've heard one of these and I like the sound.

http://www.webphix.com/schematic%20h..._hotcat_30.pdf

I have a build on the shelf right now that should run very similar to this, although with a more vintage gained preamp. Once its done I'm on to a self split 12au7 amp or a push pull fixed bias 12BH7, based on some of the ax84 stuff.

[Gingertube]

The schematic of what I built is attached. So for your possible interest the "Junkbox Trainwreck".
Bruce I measure 32 Watts RMS at just below clipping.

Cheers,
Ian

[Tage]

Hey Ian did you use a 4k output transformer? or something a bit higher like a 5-6.6K to go along with the lower, relative to a Fixed Bias, voltage?

[Gingertube]

Tage,
Output Tranny I used was wound in Sydney Australia, I was told that it was an exact copy (reverse engineered from a burned out unit out of a vintage JTM45) of a 50W Marshall with Raa = 3400 Ohms, I bought 2, the other one is in another of my builds (a London Power Standard from TUT5) being driven by a quad of vintage 6V6G.
Cheers,
Ian

[jpfamps]

We use EL34s in fixed bias with around 350V on the plate for about 26W before clipping into 5.8k, biased at 35 mA per side.

Biasing upto 50mA per side is safe using these voltages, however we felt the amp sounded better at 35 mA.

We also use 1.5k screen grid resitors, so the valves are running very conservatively, and should last longer.

We arrived at 5.8K by measuring the reflected impedance that gave the most power.

Incidentally, we also thought that the amp sounded better at this impedance.

Reducing the screen voltage would suggest using a higher impedance for optimum power.

Interestingly using a resistor instead of a choke and larger screen grid resistors means that the screen grids sag quite a lot. Thus the impedance needed for maximum power would be higher than in an amp with a very stiff screen supply, say with a choke and smaller (or even no) screen grid resistors.

[Dave H]

Cathode biased amps sound great.
I've built lots of cathode biased amps... but, I've never been able to get a clean 35 watts out of any one of them when using a pair of 6L6s, EL34s, KT66s.. etc etc... in cathode bias.
I'd like to see a schem of that amp and some data... maybe an old dog can learn some new tricks too.

Bruce, here's mine. B+ is 365V, output 30W, 500Hz @ 3.8% distortion. Primary is 4k.

J30 scm.pdf
J30 scope.pdf
J30 spectrum.pdf

[R.G.]

Just as a bit of rule-of-thumb understanding of tubes:
- The temperature and oxide coating of the cathode determine how many electrons can come out of each unit of area of the cathode per second.
- The area of the cathode times the electrons-per-area determine the current the cathode can provide and therefore the maximum current the tube can conduct.
- The area of the cathode and the temperature determine how much heat radiates out of it, and therefore how much heater power has to be provided to keep it hot.
- The current through the cathode (and plate) determines how much area the plate has to have to radiate away the heat generated by the current falling through the cathode-to-plate voltage.
- The nearness of G1 to the cathode and the fineness of the mesh openings in G1 determine the effective gain of the tube from G1. Fine openings near the cathode can turn off electron flow easily.
- The finer the mesh and closer to the cathode for G1, the more electrons hit it and the more grid leakage that happens, and therefore the lower the grid leak resistor required.

Electron emission and control from a heated oxide cathode for longest practical life and lowest reasonable manufacturing cost long ago hit a near-optimum. I suspect that the electrons-per-cathode-area is remarkably the same for all tubes, and the variables from 12AX7 to 6L6 and other power tubes is largely in the area of plate and cathode and the spacing of the grids for gain. Easier to get hundreds of times the cathode area in one big bottle than hundreds of little ones.