EL84/6BQ5 triode pp with no cathode biasing

[DrewP]

el84/6bq5 triode pp with no cathode biasing

Folks, I've had the good fortune to inherit a second hand amp that was dead but with good output trafos and psu tranny.

My grand plan is to make a push pull triode-connected el84 amp using these trafos.

Input stage is pretty much negotiable.

My aims are to avoid cathode biasing resistors and caps if at all possible. I'd prefer to hook up an additional tranny on the chassis and take a bias votage to the grids with the cathodes connected straight to the centre tap. Additionally, if I can reduce the number of caps in the signal path that'd be nice, but I don't really want to go to the extreme of interstage trannies.

Planning on running minimal or no output stage or overall feedback. This amp will be hooked up to single driver Fostex FE168 Sigma speakers with a pretty benign impedance load.

I'm still a relative newbie, just want to get the best results I can out of the psu, el84's (4 of) and output trannys I have at my disposal.

Won't be filling a room any bigger than 12ft x 18 ft by 8ft (w/l/h) and no desire to bang head. 1.5 watts should do me.

Any suggestions gratefully accepted.

TIA

Drew

 

[ErikdeBest]

Hi DrewP

You are a lucky guy! Push pull amplifiers with EL84 are the center of discussion at diyaudio.

Jane has just posted a thread with DC coupled stages
http://www.diyaudio.com/forums/showthread.php?s=&threadid=77839
This one uses cathode resistors and capacitors...

gingertube is also active with his baby huey
http://www.diyaudio.com/forums/showthread.php?s=&threadid=72536
auto bias circuit - or just fit in SY string of leds: both have no resistor + capacitor to bias the EL84.

And then there is this last thread, with some more information on EL84 PP
http://www.diyaudio.com/forums/showthread.php?s=&threadid=79310

At glass-ware you will find some articles on biasing with the lm317 or lm337.
http://www.glass-ware.com/tubecircuits/Tube_Auto_Biasing.html
http://www.glass-ware.com/tubecircuits/Tube_Auto_Biasing_4.html

Good luck reading them all!

Erik

 

[DrewP]

True but all of these examples are using CCS, diodes or some similar arrangement between the cathodes of the el84's rather than an external negative bias voltage being supplied to the grids.

All those threads came up when I did a search.

Not to say that they're not good designs, I'd just prefer not to have anything under the cathodes if i could get away with it.

sorry, should have specified a desire to stay away from diodes etc as well. That said, the LM317 CCS could be interesting to play with.

many thanks,

Drew

 

[ErikdeBest]

Hi DrewP

Yes, you are right, none of the schematics use fixed bias and of course this is possible: take any RC coupled design, tie the cathodes to ground and put 10V negative on the gird, after the capacitor of the RC coupling. EC8010 defended the use of fixed bias and a interstage trafo in the design of an EL84 to prevent problems with recovery - this is also buried in some recent thread.

Erik

 

[Eli Duttman]

Drew,

IMO, the solution is pretty obvious. Build a variant of Poindexter's "Musical Machine" that uses EL84s as the "finals" instead of 6V6s.

Poindexter's Site

If you want vacuum rectification in the B- supply, emulate "El Cheapo". Sub. a 6AL5 for the 12AL5 and power its heater with the little Allied trafo's filament winding.

"El Cheapo" PSU Schematic

 

[SY]

Erik, the LED string and the voltage regulator feeding the cathode are actually fixed bias, not cathode bias (in the sense of the classic "automatic bias"). They just apply that fixed bias to the cathodes rather than to the grids.

 

[ErikdeBest]

Hi SY

Thanks for correcting me! So I learn a little more every day - and can say I am close to build my first fixed bias amplifier!

Erik

 

[rdf]

Re: el84/6bq5 triode pp with no cathode biasing

Planning on running minimal or no output stage or overall feedback.

Hi Drew. Mind if I ask, have you had a bad experience using the OPT secondary to apply cathode feedback? My experience, in SE at any case, is it works very well with EL84s.

 

[DrewP]

No, no bad experience, just prefer to avoid if possible.

I'll be treating the amp to a pretty benign load and there's something about sending a correction from before to be processed through the amp now that my mind doesn't really like the idea of, no matter what the lovely sine waves tell me.

 

[rdf]

Trust me Drew, I t'aint no sine wave guy though I don't object when they look better either. Cathode feedback stays within the stage, except arguably for those who consider the OPT another stage. I highly recommend giving it a spin while breadboarding, you might be surprised. I was.

 

[DrewP]

Can anyone explain to a dumb newbie the relative sonic merits/downgrades in using a silicon CSS such as the LM317 etc in place of the regular cathode resistor/cap network?

I'm guessing that the CSS is controlling the 10 or 20ma quiescent current through the tube rather than having the current level determined by bias.

(Well obviously, as there's no where else to current to go to or from)

Some folks I know have said that they preferred the sound of nothing under the cathodes and the negative bias being applied to the grids from external source, some say that a resistor/cap network sounds fine, some swear by diodes.

certainly the LM337 option looks simple, but how does it sound?

 

[ErikdeBest]

Hi DrewP

the advantage of a 317 in the cathode is that it will always draw the same current through the tube. This can be very benefical for the OPT, mainly when you are using a (mains) toroidal one.

About sonic advantages I can't say much - yet.

 

[arnoldc]

I've been using LM317 (w/ Black Gate bypass) at the cathode of my 12B4 preamp and it made a positive difference compared to the RC combination. I posted it on the 12B4 thread.

For the amp, this is also something I have been contemplating. I'm even looking at replacing my 45 RC cathode bias with a CCS.

 

[kathodyne]

hi,

I've heard a few times now, that an lm317 is not a good choice for using,,,,,,,, because it operates badly >1kHz.....is that something to worry about???

 

[arnoldc]

Hi kathodyne, where will the anomaly show? The scope visuals on my 12B4 with LM317 on the cathode has excellent square wave response up to 30kHz, and perfect 1kHz.

 

[rdf]

Have you considered the IXYS 10M45S class of current regulators? 450 volt capable, one or two resistors to set (the second as a gate-stopper in case of oscillation) and plenty good performance. Dirt cheap too, via Digikey.

 

[arnoldc]

Hi rdf, I have been WANTING those. But unfortunately, not available from either Farnell or RS Components

I could try DigiKey ultimately.

 

[kathodyne]

When DC voltage is used, power supplies mostly are regulated. Voltage regulators like the LM317, LM337, LT1086, and the 78-series have very limited bandwidth. Most already loose their good properties above 1kHz, in fact there is no regulation anymore at 100 kHz, just where the problems start. They should be used to get rid of hum alone. Hf is almost passed completely, both when they are used as voltage or as current sources. These passing Hf harmonics influence the sound a lot. They make sound harsh, uneasy and reduce image and make the background less dark.

from: http://www.machmat.com/sales/kits/index.htm
look under miscellanous (ccs/vccs)

 

[arnoldc]

Hi kathodyne, I would have no doubt that a "proper" CCS like that of the diyAudio CCS will do a better job. But for whatever its worth, the LM317 improvements were very audible for me, and none of those described in the link you posted above manifested on my 12B4 preamp, both on the scope and to this ears

The LM317 is my starting point in venturing to this CCS stuff, and I must admit I had a negative impression about this when I previously heard a Bottlehead Seduction and Foreplay "upgraded" with their CCS. To me they affect the tube sound so much that the sound becomes clean *and* lean.

Thanks to this forum I've opened my eyes and experienced many things I was rejecting in my "religion"

 

[DrewP]

Hmm, all this is rapidly encouraging me to go with the original plan of "bollox to anything in the cathode circuit, ground em both and apply the bias to the grids from an external source instead."


Doing it that way avoids getting bogged down in all the arguments about what sounds better/worse in the cathode cct. I'm guessing that nothing probably sounds better than most of the other options.

That said, being able to use an LM337 to get perfect current ballance in the output trafo instead of a few mA this way or that is attractive too.