OK…so I’m on wi’ summat

[Morgan Jones]

Nice neat construction with proper heater wiring.

[Cressy Snr]

Nice neat construction with proper heater wiring.

Cheers Morgan,
I learned how to do heater wiring from your “Building Valve Amplifiers” book.

[Morgan Jones]

Excellent; have a gold star.

What output transformers did you use? A friend with a pair of Lowther cabinets fitted with Fostex FE206 would quite like to make his first amplifier, and KT66 plus summat might do very nicely.

[Cressy Snr]

OPTs are James 6123HS.

[Cressy Snr]

Configuring the James OPTs for 3.5K primaries means one can use EL34s properly:

EL34SEUL.jpeg (89.87 KiB) Viewed 1392 times

I might install a double pole changeover switch so that the primaries can be switched between 5K and 3.5K (not on the fly!) to make it a nice versatile piece of gear.

It now puts out 10.8Wpc in single ended ultralinear mode. The Mission speakers specify amps of 10-100W capability, so it just about scrapes into the right area. Mind you I wouldn't want it anywhere near 10W, not with 10% THD at full chat without global feedback.
A small amount of cathode feedback via the OPT secondary might reduce the THD somewhat, but TBH I like the way it performs at the moment.

My mental health has come a long way since I started this thread a couple of months ago. At that time I couldn't think straight about anything much, let alone design something myself, so had to start with a published schematic. It gave me a good sounding device as a reset until I got my head in some sort of order. And whaddya know, I'm back with EL34s and KT66s. Can't get away from them!

Both valves sound magical in single-ended ultralinear mode, the EL34 especially.

[Cressy Snr]

Again, for the sake of completeness:

EL34SEUL.png (55.19 KiB) Viewed 1272 times

KT66 or EL34? Both are good. I prefer the latter, but others might prefer the former. It's all subjective in the end.

[Paul Barker]

Ive struggled to find this data, but in Vade Mecum it says single ended triode EL34 ri is 3k hold that thought, reading about 43% ultra linear in Wikipedia I think I found this “it can half the ri in Triode mode.” But I just know I read it.

I couldnt find the mu for UL but Triode single ended mu is 11. So I ran the cathode bypass cap in my spread sheet for that mu and that ri. Whether either of those are incorrect I dont know.

But I was struck by the small value cathode bypass cap and with those two assumptions its minus 3db at 23hz.

[Nick]

I have not tried it with a output stage, but certainly with interstages, small cathode caps can work a little different than you would from those calc's. The fact that you have a inductive load instead of a resistive one can give you a small (though helpful) inductive peak at LF via the LC circuit created that can mean that you have a LF peak and the other half of that is a 2nd order HP filter below that peak. The output transformer inductance should be considered as well.

[Paul Barker]

Probably Dave or Paul Joppa referred to the fact that protecting the se transformer from too much bass results in better bass. I think the tank effect referred more to the parafeed cap resonating with the two inductors in parallel (choke load and parafeed transformer) so I suppose you see a resonance between se transformer and the cap completing the ac path, in se also even if its the last power supply cap or the other ways to return the ac such as ultraapath and the hybrid partial ultrapath partial cathode bypass.

Weve mentioned the hybrid recently. I just dont recall if it ever went main stream andd was giben a name. Meither do I recall the folmula. But perhaps starting point was cap to b+ 1/3rd value of cathode resistor bypass but that would depend on mu which was probably 3 for the valves we used at the tome.

I used pio at the time and liked it.. 1/3rd was the ratio in my case robably with 6b4g or px4.

So Ive opened a cann of worms sorry Steve.

Im off pio atm its damaged my valves.

I suppose Mike could remind us what Mullard did with ultrapath bypass cap sizing.

[Nick]

Probably Dave or Paul Joppa referred to the fact that protecting the se transformer from too much bass results in better bass. I think the tank effect referred more to the parafeed cap resonating with the two inductors in parallel (choke load and parafeed transformer) so I suppose you see a resonance between se transformer and the cap completing the ac path, in se also even if its the last power supply cap or the other ways to return the ac such as ultraapath and the hybrid partial ultrapath partial cathode bypass.

No idea about most of that, but remember that in Steves case the cathode cap and the output transformer are in series WRT AC as the last cap in the power supply can be considered a short circuit (at least at 30 Hz)

Edit, well, he is using only 30uf, so not that much of a short. In this case the cathode cap and the last power supply cap could be considered as a 15uf cap at AC. I would expect that at the 30Hz point all sorts of interesting AC things are going on.

[Cressy Snr]

Bit of food for thought there.

[Cressy Snr]

Forgot about the caps being in series making 15uF in total, which puts the 3dB point at 27hz.

[Nick]

I have to be careful here as I have gone over to the dark side now, but I think that the preference for small caps in valve amps has two attractions. First, as Paul mentioned, taking the LF down too far just highlights the lack of flux in the output transformer, Secondly, if you have solved the flux problem either parafeed or push pull or something else you remove one of the things that make the amp sound different so makes it harder to distinguish from other amps.

[Paul Barker]

James transformers probably have good enough bass.

[Cressy Snr]

I've taken on board a few things, put in a bit of work and done some alterations to the power supply.

With a 100uF EL34 cathode bypass cap in series with the last cap in the power supply (changed to 50uF + 50uF, F&T dual electrolytic) I now have 33.33uF instead of the 15uF, giving a 3dB point of 12Hz or so instead of 27Hz.

I have a collection of quite valuable old rectifiers courtesy of a friend of our Ant's so I have gone belt and braces with the correct series resistors (when the winding resistance of each side to the centre tap is added in) followed by in-line solid state diodes, in order to protect the rectifier from flashover and destruction of the power transformer. Overkill it may be but there you go:

EL34PSU2.0.png (91.32 KiB) Viewed 981 times

EL34SEUL.png (54.79 KiB) Viewed 981 times

EL34PSUmods.jpeg (150.17 KiB) Viewed 981 times