GU81 Triode Connected Push Pull

[Paul Barker]

While the r81’s are at the lockup I,ll put them in the 833 drawer.

[Paul Barker]

So there are unique advantages to the r81.

Low ri anything from 300 ohm to 600. I came up with 500.

So when you want to parafeed so you can use a wet opt with poor L it may not be that bad parafeed. Also the 250 mA 30 henry humungus chokes a guy like me buys are the babies for the dc load.
The el38 amp you already made and enjoy with drive the r81 as no way are we going A2 no need either.

So much less to build to get 20 watts a walk in the park. No direct cf driver stage to build.

Otoh do we really want 264 watts of heat. Well actually yes in my Man Cave top floor loft room, unheated un insulated. If Piere Tribute lurks here. The room Pieter slept in.

It actually makes sense in that room. What brought me round was the low ri. The bases are in, this might be a project I do first.

Meanwhile my stroke recovery is coming on leaps and bounds.

[Paul Barker]

Same opt gives 20 watts into 8 ohm and also 20 watts into 4 ohm. Roughly speaking thd in both cases <2 thd se. The 4 ohm is lower only because it was indanger of terrible distortion so I made the current flow 250 mA per channel. The 8 ohm version only needs 200mA dc to get similar distortion.

So to power supply 600v plus autobias 130 to 140 v at 500mA is going to require out of my stock the big cube 3kv transformer on my big blue variac. Such is life. If you dont operate the r81 at least 250 mA you may aswell forget it. So you must use such monsters and you must parafeed or push pull. Obvs im parafeed because I like the harmonic result and dont like what pp gives which id 3rd alone even thought thats only 1%. Also I dont have to buy two more valves bases and top caps.

[Paul Barker]

Oh well, having both the bases and the top caps, the elephant in the room is to get sufficiently linear distortion numbers you have to hold this powerful valve as high as possible in the current. If you are operating it at the current flow of a gk71,gm70, 211, it has considerably more distortion than those mentioned.

So then you look more intently at the weight and cost of this direction you are lead. My 30 henry 250mA chokes are enough for the direct current. But its suddnly a monster. My power transformers are the weak point. Unless I bring home from the lockup the huge cube of 3kv power transformers that is so heaving it has lifting eyes for over head crane gantries. And the variac is heavy enough also.

On reflecting on these issues I’ve decided while sat listening to my triode connected el38 amplifier on the ‘57’s and I dont even need a dht.
So how do I get power from this idea?

Well I did varios load lines with the Sowter 8k into 8 ohm PP. if I put them on 4 ohm speakers they are 4k pp so doing the maths that provides 10.9 watts as a class A. You could call the 57’s 4 ohm, and the celestions are 4 ohm. Maybe 11 watts is sufficient? The ri is 1200 so 4k is 3.3 multiple so reasonable load.

I know that 4.5 watts se satifies me on the Quads, so 11 watts ought to sort them out permanently.

Unknown factor is the Celestions. But worth a try.

Beyond that se for the celestions 22watts se 211 . Yes I can easily get 30 watts with my big valves and especially those r81’s as triodes easily achieve 50 watts. But only if I need to go that extreme.

Sorry point of order, correcting my maths at 4k max volts 570-175 =395 x= 156025 / 32000 = 4.87 w edit 6 watts doubled for class A pp = 9.75w edit 12 watts. Forgot about the 8 watts dissipation of the screen. Got there in the end.

8k load 620v - 150v = 470 v x= 220900/64000=3.452 w x 2 for pp = 6.9 watts on 8 ohm speakers, but may be enough for the KEF Correlli’s.

So worth building and see .

Much simpler to build b+ 420 already available. Would fit the small chassis with in and out on the back, power plug, step atten.. holes punched, space inside for torroid 30 v for four pl38’s heaters.

I think the maths is correct now. Of course the 3rd h is less on the 8k load but less than .5% on the 4k load so should be ok. Some 2nd from the d3a triode volt amp which has cathode positive feedback from a 50 ohm unbypassed resistor but still achieves all the voltage gain required. should blend even distortion in with the 3rd of the output stage. The el38 only needs 20v peak to drive it to full power.

[Paul Barker]

Still got it wrong. Forgot to include the 8watts contribution of the screen. Power out on 4k loadline se is 6 watts, pp is 12 watts. That is better!