801a amplifier

[Paul Barker]

Meanwhile

[Paul Barker]

Roxy giving body language to say “I own this” or is she saying “ooh that’s nice heat”

So gm70 no good in gk71 base. Either need 813 base (which I have) : carefully drill out one ceramic hole.

Or scout eBay for gm70 bases, These look good.



I presume you solder to the silver sockets’ anyone?

Had to buy with graphite gm70. £108 total delivered and taxed take off £68 for cheapest pair of gm70 elsewhere = £20/base.

Yes, worth it.

[Paul Barker]

You may recall the concern I had at wearing out my few 801a valves. Down to 4! At same time I mentioned this I bemoaned the fact no suitable alternative valves.

So I was idling and looking once more at 1619. It fits the bill.

Using 6v6 triode curves. As Ive said before there is a misdirection about this valve on audio forums. It is not a directly heated 6L6, it is a directly heated valve closest to 6v6. Not intended as an equivalent, but should operate closely to 6v6, whereas its not on the same page as the 6L6. I think that whole thing was a mistype by one person once and that syndrome of repeating the error by people who dont verify things for themselves.

That corrected, the 6V6 triode connected into 9k operated at 14 watts dissipation, 350v quiescent in A1 provides 2.22 watts, @+10v A2 3.47 watts, @+20v A2 5 watts. (All calculated from measuring the characteristic curves) ((ve - va) * ( ve - va)) / 8*9000.

I cant calculate 2nd harmonic distortion graphically because the characteristic curves Ive found dont go beyond 500v and anyway the curves 2nd distortion pinch as expected. If A2 is encroached into as pointed out with 801a analysis above odd order distortion would sky rocket. But anyway, with this less power valve I cant ignore the A2 likelyhood. Which means fixed output valve bias, and in my case cf drive, since Im not interested in mosfet power drive. ACFII screams at me, but I may not add that task, rather pop a cf single stage in. Or probably nothing built.

Valve costs 1619 (which I dont need to buy, but replacement cost £15 each) 801a ( £70 to £150 each).

I need to drop the b+ from 630 volts to 350 volts. For breadboarding simplest option variac after seperating filament supply for rectifiers. Long term, alternate transformer.

The figures given are for one valve into 9k, but as with the 801a vesrion the 9k os what you drawer into the characteristic curves. The two output valves per channel are pse connected and the opt is 4k5. So teh actual power output doubles to 4.44 watts, 7 watts 10 watts.

What am I expecting ? a sweet sound rich in second harmonic, enough to play my music at home, much more distortion than the 801a version. But I may like that! Not a waste of effort anyway: plentiful, affordable replacement valves.

[ACPEN]

Couldn't agree more! recently put together a PSE 1619 SE amp (295V ht - 3.5k load - driven by 12gn7 a'la Pete Millett) very sweet sound + 1619's £8.00 from Langrex!) and ac htrs are fine!

[Paul Barker]

I null the hum anyway connecting ac filaments each if a pair out of phase. With push pull you connect the pair in phase. Seperate windings, same core. A different core for other channel.

The 801a connected so has no filament hum..

The last 1619 amp I built was push pull. It was mint. I lost the phase splitting transformers! Still smarting from that loss

[ACPEN]

Me too

[Paul Barker]

OTOH, I have 11 10y’s so as they use same base and heater as 801a one less thing to change. At 350v 25mA into +30v A2 7.4 watts x 2 per channel into full distortion. I was going to add a cf dc driver and negative supply anyway. Worth trying if only as phase one. But might surprise, one way or the other.