• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Info for 6b4g PP

Status
Not open for further replies.
Jump to Latest
#3
sparkx said:
Hello. This is my first post. 🙂
I have four 6b4g Svetlana and I want to build a PP amp.
I need schematics with:
- good sound;
- simple to build;
- not expensive (no interstage transformers);

For example, is this scheme good?
JE Labs PP 2A3 Amp Project

Best regards,

sparkx
Click to expand...

Welcome to the forum!

Let me start by pointing out the foreign object floating in the punchbowl- a decent DHT tube amp will NOT be cheap, even without an interstage. The schematic you linked to is a very unimpressive design (assuming you want an amp whose output replicates the input as opposed to an effects box)- it's basically a Mullard 5-20 with the feedback removed and a lower gain tube used in the first slot. The phase splitter stage is unbalanced, and the HF bandwidth is limited. The distortion will be high.

You can do better.

edit: I noticed that the writer began the linked article by saying that he really doesn't know how to design amplifiers. When someone tells you that, it's best to believe him.
 
#5
Sparkx,

How well matched are your 6B4G pairs? The J.E. Labs design is a Mullard circuit sans GNFB loop. It's easy enough to change from the self bias shown to "fixed" bias, which will deal with cathode current differences. However, unless the gm matching is reasonably close, you will have issues.

UF5408s in the PSU will be much quieter than the 1N5408s shown. Replacing the 12 KOhm LTP tail resistor with a 10M45S CCS is a worthwhile improvement. A separate 6.3 VAC winding and 50 Ω balance pot., for each 6B4 filament, will allow good control over hum.

Have you given thought to the O/P "iron".
 
#8
Yvesm said:
Hi Stuart !

Had an hard night ?
Click to expand...

Yeah, Rhone wine. You Frenchies are diabolical! 😀

Sparkx, HF bandwidth will be limited by two sets of significant Miller rolloffs. Distortion will be high because of the unbalanced drive from the phase splitter, the 120Vpp (or so) that the phase splitter will have to swing, and the reactive load of the phase splitter on the input stage. Again, some might like these effects (it's like a treble control turned down and an Aphex box turned up), and I can't argue that- thus my caveat about the amp as a pure gain device versus a device meant to change the sound.

For a really detailed analysis of the 5-20 input stage as well as detailed examination of the long-tail pair phase splitter, I'd highly recommend "Valve Amplifier," 3rd edition. It will cost about the same as a pair of 6550s and provide an excellent background for you to work from.
 
#10
Sparkx,

Let's deal with SY's legitimate objections, in a cost effective manner.

The 1st CMiller induced low pass pole is at the circuit's I/P. The grid to ground resistance of 270 KOhms is too large. Reduce that value to 100 KOhms. Also, the 6SN7 is not highly prone to parasitic oscillation. The 10 KOhm value for the grid stopper is complete overkill. Make that part 510 Ω/Carbon composition.

I previously stated that installing a CCS, instead of a resistor, in the LTP's tail improves performance. A CCS in that position forces symmetry between the 2 halves.

The 2nd CMiller induced low pass pole is at the 2A3 grids. Once again, the 270 KOhm grid to ground resistance is too big. In addition, switching to "fixed" bias (for balance reasons) requires the net grid to ground resistance to be quite low. Otherwise, the DHT can run away. Good as it is, the 6SN7 is not up to the challenge of simultaneously working into a low net impedance, driving the 2A3's considerable CMiller, and swinging 120 or so V. peak to peak. A change in type is definitely in order, for the LTP. I suggest the ECC99. While the μ of the ECC99 and the 6SN7 are roughly =, the '99 exhibits both a higher gm and lower RP. Those electrical characteristics are very important, when driving 2A3s.

BTW, in his write up, J.E. mentioned a high/low PSU. That's a direction this project will have to move towards.
 
#11
I'll stand by the high distortion and high output impedance criticisms.
Click to expand...

Euphony and mediocre damping factor are part and parcel of no loop NFB DHT amps. Some folks love that stuff. As Harvey Rosenberg stated, you're getting a show.

Proper speaker selection deals with the damping factor issue, but many good speaker designs are poor mates.

To be fair, the 2A3 family is more linear than the 300B and PP even order cancellation will also help, in the HD dept. Still, one has to wonder.
 
#12
Eli, have to disagree with a couple of your points. First, the input rolloff is unlikely to be determined by the 270k- remember, that resistance is shunted by the source resistance of the preamp, probably one or two orders of magnitude lower. The first Miller pole is between the output of the voltage amplifier and the phase splitter. The second pole is between the phase splitter and the output stage. (BTW, one might want to check that the heater-to-cathode voltage isn't being abused...)

Indeed, a CCS in the long tail will clean up the CONSIDERABLE second order distortion from the unbalanced phase splitter. It will not reduce the high distortion from trying to drive the large swing needed by the low mu output tubes.

The 270k grid resistors for the 2A3 are within the max rating for self-bias, so I'm not sure I understand your objection- the source impedance is dominated by the phase splitter.

Like I said, it's an effects box, not an uncolored amp.
 
#13
The 270k grid resistors for the 2A3 are within the max rating for self-bias, so I'm not sure I understand your objection- the source impedance is dominated by the phase splitter.
Click to expand...

Getting the "finals" properly balanced requires "fixed" bias and individual "idle" current set trimmers. Betcha 270 KOhms, in the grid to ground position, is now way too "tall". You can go to the bank on Billington not matching the tubes as well as McShane does.

Indeed, a CCS in the long tail will clean up the CONSIDERABLE second order distortion from the unbalanced phase splitter. It will not reduce the high distortion from trying to drive the large swing needed by the low mu output tubes.
Click to expand...

That's 1 of the reasons I suggested switching to an ECC99, as the LTP. While its μ is, for all practical purposes, the same as that of the 6SN7, its much lower RP allows stage gain to be greater. Please look at the ECC99 data sheet. Up to a 400 V. anode to cathode potential is permitted. A condition set permitting the swinging of 120 V. peak to peak should be "in the cards". Obviously, the simple Greinacher doubler PSU J.E. shows will not do. MOSFET buffering of the LTP may be necessary, even with a '99 on the job, to keep stage gain up.

BTW, one might want to check that the heater-to-cathode voltage isn't being abused.
Click to expand...

Right on! "Rat Shack" offers low cost 12.6 VCT filament trafos. Dedicate 1 of those, whose current rating is appropriate, to energizing ECC99 heaters. Bias that dedicated heater supply off B+ to control noise and keep the heater to cathode potential well within established limits.
 
#14
How about another starting point?

I always thot this amp would be a good candidate for 6B4G instead on EL34s strapped in triode.

pp_1c_s.gif


It is the basis for VSE's 300B PP amp.

dave
 
Last edited:
#15
planet10 said:
How about another starting point?

I always thot this amp would be a good candidate for 6B4G instead on EL34s strapped in triode.

pp_1c_s.gif


It is the basis for VSE's 300B PP amp.

dave
Click to expand...


Buffer the differential cascode with MOSFET source followers and that will do for driving the small grid to ground resistance of "fixed" bias 6B4G "finals". Hmmm, 6922 cascode anode potential will be in the general neighborhood of 200 V. That, suggests that the little ZVN0545A, with its tiny capacitances, could be an excellent fit.
 
#16
Problem with that circuit will be low distortion at high swing- an ECC88 cascode can do very well at 40V or so, but a 2A3 or 6B4 needs much more. ECC99 might work, but I'd be inclined to stick with a 6SN7, but change the operating conditions- input couple via a step circuit that puts ~20-50V on the 6SN7 grid, CCS tail, 5mA or so current per side, 47k plate resistors, and a 450V supply. A pair of CFs or source followers would be very advantageous sandwiched between the phase splitter and the output stage.
 
#18
Guys,
The circuit posted by Planet 10 drives trioded EL34's perfectly without any need for ** MOSFET followers.

Changing the input tube to the 6H30 running at 15mA with 12k anode loads drives 300Bs PERFECTLY - as evidenced by the great reviews our dpa300B poweramps have received. Check out my website.

I do NOT agree with adding MOSFETs into the signal path of such an amp - unless you can't hear what they do.

And Sy, the 6SN7 is not a great tube for cascode - far too little gM. The 6H30 though is a killer at 15mA/V and lots of current.

I want speed and I want definition.

Regards, Allen
 
#20
even 6Н6П is good "enough"

yup - Allen's circ is good for start :clown: ;

then one CCS from each output tube cathode - to neg PSU , cathodes coupled with CRC T network , already known here ;

that way both DC and AC balance of output stage are perfect

in that way - even toroid OPT can be used
 
Status
Not open for further replies.
Top Bottom