audio-talk

The Super Rocky I built during this lockdown is so good with the KT120s that I’ve commissioned our Ant to build me an
18 x 14in chassis out of afrormosia, so that everything can be spaced out more.

Because of the multiple LCLC power supplies, the existing chassis is pretty cramped, and the KT120s fit with only one eighth inch spare between them and the front faces of the output transformers. This is not good, as the close proximity of the KT120 glass to the OPTs is heating the end bells and from there, the heat is transferring to the laminations. In other words, the output transformers are acting as heat sinks, so that after a couple of hours they are getting a bit too warm for my liking.

The deeper chassis will solve the space problem and will introduce the possibility of maybe using KT150s further down the line, when I’ve saved enough pennies for a pair. I will most probably be looking to a different driver arrangement and am open to suggestions.

I have 12GN7 and EF80 for a two stager, with pentode drive, or 6SN7 cascade three stager might be good, or 6SL7, DC coupled to a 6H30 cathode follower.
Bearing in mind that the pentodes have B9A bases and the 6SN7 are octals, I need to get this right as I don’t want to be hacking the top plate about after the event, so some advice on pre/driver stages would be most welcome.

Thanks.

How about a 6SL7 / 6H30 Aikido driving the beastie? of course that has double the number of initial valves to 4 so might not be possible for you...

or 6BR7 pentode aikido front end with 6H30 aikido back end?

12GN7 is rather good in its own right as a driver. Pete Millett found it as good in pentode mode as it in triode mode with a CCS load... http://www.pmillett.com/pentodes.htm

He also found the 6SN7 cascade as pretty limited see bottom of the same page... and I have to say that I found the same in the old WAD 2A3PSE so replaced them with a 3Cg pentode...

There are nine pin version of 6SN7 and 6SL7 too so maybe just go for nine pin holes and be done with it?

Looking forward to this evolution of the amp

JamesD wrote: ↑Thu Jun 11, 2020 8:49 pm How about a 6SL7 / 6H30 Aikido driving the beastie? of course that has double the number of initial valves to 4 so might not be possible for you...

or 6BR7 pentode aikido front end with 6H30 aikido back end?

12GN7 is rather good in its own right as a driver. Pete Millett found it as good in pentode mode as it in triode mode with a CCS load... http://www.pmillett.com/pentodes.htm

He also found the 6SN7 cascade as pretty limited see bottom of the same page... and I have to say that I found the same in the old WAD 2A3PSE so replaced them with a 3Cg pentode...

There are nine pin version of 6SN7 and 6SL7 too so maybe just go for nine pin holes and be done with it?

Looking forward to this evolution of the amp

Cheers James.
Plenty of food for thought there.

Interesting link there James..
Steve, my way would not have excessive gain or stages
The kt120 is quite a sensitive valve and, I'd guess will be set up around - 15v so a choke loaded capacitor coupled a voltage gain of around 25 is where I would be aiming as a start.
I doubt the kt120 will need much driving, (but don't hold me to that) but your unlikley going to be feeding it more than 10 volts with normal listening.. A single triode would my choice and least stages the better.. But it always depends what your driving of course
Pete millet highlights that pentodes have more distortion at the voltages you're using (10v column ), for me that rules pentodes out.. I know they can sound good as drivers.. But I always feel the're second best.

It really depends on what you want from the amp.

Brave thing, asking us lot...

Now that is interesting. I look at the triode curves for KT120 and KT150 and I see the op point at around Ia=100mA, Va=400-450V, Pa=40-45W and Vg=-50 ish volts to yield around 8W with a 5K load. SO it requires up to 50V of signal. I guess if we assume CD levels input we have 2.8V max of signal and that need a gain of, say, 20 for the max 50V signal. Conventionally if we assume 1V input we need a gain of 50 - so that sets the two edge cases for gain. Certainly the first is possible with a single triode but the second case requires a trioded pentode such as D3a or el802 at least in a conventional design - I guess we could make a triode based Gm amp to get the gain...

ALternatively we could look at Steve S's suggested op point of Vg=-15V... Then we have Va = 200V at Ia=100mA which might give us 1W with a 2.5K load.
Which would sort of work with the Fane's but not nearly as well as the conventional case.

I think we need a three stage amp with older nice british triodes or maybe an early pentode as front end of an aikido...

Of course its possible to use any op point and just use 10V of drive and I agree that Steve is unlikely to use that much in practise but I think the Valve op point needs to be around 40W anode dissipation to work well.

The answer is to make two drive circuits and see which sounds best LP2 loaded with a Hammond 126C anyone???

My mistake James I had wrongly thought it was lower grid volts. Don't know where I got that figure from..
-50 is another ball park.
two stages needed... Like the idea of old British triodes. (my valves of choice). I quite like ml4 driving ml4 or eqvs

Alex Kitic visited here a couple of years ago and has had success with his RH series of amplifiers. I have been using a RH84, RH807A Super and a RH Universal since 2013 with no issues other than "great sound" and plenty of power (using a pentode as a single ended pentode).

http://rh-amps.blogspot.com/2013/05/rh- ... ersal.html

Thanks chaps,
At the moment, in my existing amp, the KT120s are operating at -30V and 110mA. HT is 383V at the anode, which is 353V across the valve, so I’m running at about 38W dissipation.

-30 is about where most of my amps run with a single stage, choke loaded mh4 or lesser gain valve and an interstage

Mh4.. is my choice I could let you have a pair of mesh plates.. ac/hl, 41mhl will all do it.. the higher gain valves of the era could be over driven with a CD player
Idh also makes things simple with no loss of fidelity in my experience.. but I don't mind being educated on that one.

that looks right in the sweet spot Could run it with a bit more Va and a bit less current for a bit more output but why bother when you don't need the gnats of extra power...

Alex Kitic's Schade feedback (or Partial Feedback as JRB styled it) amps are fine - not optimum designs by any means but perfectly fine - for a more optimised take on pentode shade feedback amps have a look at Gary Pimm's Tabor design. His website has gone but its still on the Wayback Machine https://web.archive.org/web/20150217180 ... abor3s.gif

MH4 mesh plates would be very nice - go for it Steve!

James

edited for spelling schade...

JamesD wrote: ↑Fri Jun 12, 2020 12:38 pm optimised take on pentode shade feedback amps have a look at Gary Pimm's Tabor design. His website has gone but its still on the Wayback Machine https://web.archive.org/web/20150217180 ... abor3s.gif

I saved that circuit recently, and bought some mint Siemens 6AU6's. I have the Purvine 6.6k outputs with Litz ... need Chris to send me some 1624's ...

steve s wrote: ↑Fri Jun 12, 2020 12:35 pm -30 is about where most of my amps run with a single stage, choke loaded mh4 or lesser gain valve and an interstage

Mh4.. is my choice I could let you have a pair of mesh plates.. ac/hl, 41mhl will all do it.. the higher gain valves of the era could be over driven with a CD player
Idh also makes things simple with no loss of fidelity in my experience.. but I don't mind being educated on that one.

Steve that’s fantastic! You have a PM

JamesD wrote: ↑Fri Jun 12, 2020 12:38 pm that looks right in the sweet spot Could run it with a bit more Va and a bit less current for a bit more output but why bother when you don't need the gnats of extra power...

....MH4 mesh plates would be very nice - go for it Steve!

James

Well that’s that settled then. I’m doing it.

Not wanting to throw a spanner in the works... You probably know the gain your power amp needs to splice into your system coz you have several. Can't you measure the gain, subtract output stage gain which leaves driver/input gain and choose accordingly? Therefore no need to theorise about standard level signals for full power etc.

Yes I could measure the gain and design accordingly as you say, but going from standard levels to a design is easier, and with my track record, going standard is the best move I’ve made. This back-to-basics amp is giving me more listening pleasure than the past 29 amps put together.

There are two phrases that have brought me to a standstill since I built a Shishido Loftin-White as my first ever scratch build 15 years ago.

1. For two stages you need a high gain first stage.
2. But you can’t use an ECC83.

Feck! Now What!

Trying to get around that problem has been a complete PITA, and my amplifier gain structures have been complete crap no matter how many I’ve built or how frustrated I’ve got. This is the key to why so many of my devices have sounded weedy and anaemic. Never enough gain. Now I’ve got paralleled 6SL7s the gain problem is no more, but we’re now sort of back with the ‘you can’t use an ECC83’ problem.

But I’d rather have the gain and a decent sound than the weedy wonders I’ve had in the past. It’s revelatory to go back to 2005 and start over.

ECL86 maybe. Triode for gain stage, directly coupled to pentode as cathode follower cap to kt120.