A very stupid mistake!!!!!

[Joel]

Originally posted by Fuling
...No opinions about the cathode follower output idea?

I'm researching the idea of a 6L6GC as a triode cathode follower, with a choke cathode load, in a parafeed configuration, as we speak. What do you think about that?
I will use a Hammond 125BSE (very tiny) as the parafeed output transformer, with the 5k primary selected. 20H/100mA cathode choke.

Could be a nice n´grunty driver stage for a big triode amp...

Please don't say "grunty" - it's so Brett. Also, do a search on "slew rate distortion" and you'll see that the high current driver stage everyone thinks they need for the "big triodes" is actually a myth. The current required to neutralize their input capacitance is actually quite low.

Anyway, please post your CF output stage.

joel

[Fuling]

Wouldn´t a 125B (no SE) be a better choise for a parafeed amp?
No airgap should mean higher primary inductance, right?
By the way, I´m waiting for the parts for a 12B4 parafeed amp to arrive. I ordered them in february, but shipping from Canada takes time...
30H/40mA chokes and 125C OPTs. I also ordered two of those 150H chokes for the input stage, but those will probably end up in the driver stage in my 6B4G amp together with a pair of 6SN7:s replacing the current 6SL7:s.

Since there´s no airgap in my small surplus 2k/8ohms trannies I´ll
ditch the cathode follower idea for now. They will probably work alot better in a parafeed circuit, where there´s no DC through them.
Not that I like pentodes, but since I have a few EL86:s lying around why not put them in service? They seem ideal for this application.
Refer to the classical Philips SEPP OTL amp for 800 ohm speakers(You know the one, right?), but replace the lower pentode with the power section of a 6EW7 and add the other section of the 6EW7 section as an input stage. Should work, I think.

[7N7]

Originally posted by Fuling
The 6EW7 looks alot nicer with a horizontal load line, which I guess represents a CCS or plate choke load.
Could be a nice n´grunty driver stage for a big triode amp or maybe even a hybrid

316A who frequently posts here has done active load parafeed experiments; I have no experience with these.

I agree that the second section of 6EW7 (as I wrote earlier) would make an excellent driver for a big valve.

I disagree that a weedy valve will do; you need current to drive big valves (this was all thrashed out in a long previous thread). Additionally, beefier valves have lower anode resistance, very important when your driver is looking at the low value of grid leak traditionally used with big valves - important in terms of distortion.

I am currently building a new stereo mid-range amplifier; this (like my INSTOCK project - previous thread) will use beefy DC coupled cathode followers to drive the output valves and as with the INSTOCK project I hope to couple without grid leaks: if I got away with it with 6528 with its very high gm, I should be able to manage with the triode-strapped dual beam tetrodes that I am using now.

Incidentally, at my operating point, 6528 has ra of about 390 ohms per section. Careful measurement showed that best distortion was achieved with a load of 6k a-a - far more than would be used traditionally.

7N7

[Joel]

Originally posted by 7N7
...I disagree that a weedy valve will do; you need current to drive big valves (this was all thrashed out in a long previous thread).

Let's make a distinction here - you are saying that, subjectively, you think a high current driver sounds better, right? Surely you don't disagree with the statement that, mathematically high current is not required, do you?

[Joel]

Originally posted by Fuling
[B]Wouldn´t a 125B (no SE) be a better choise for a parafeed amp?
No airgap should mean higher primary inductance, right?

Yes, all things being equal - but in this case they are two different transformers, and the BSE has the better specs.

[7N7]

Originally posted by Joel


Let's make a distinction here - you are saying that, subjectively, you think a high current driver sounds better, right? Surely you don't disagree with the statement that, mathematically high current is not required, do you?

You have offered two areas in which I have to make a distinction:

First the observation that I think a high current driver "sounds" better and second a mathematical matter.

Neither applies I'm afraid; I doubt I would detect a difference aurally, and I certainly could not do the necessary mathematics.

However the arguments I referred to in my previous posting, made by those better qualified than I were very convincing.

Further, my point that one needs low output resistance: this I consider to be irrefutable since I have first read about this, second tried it, and third observed the results.

High current every time for me - it is worth bearing in mind also that high gm (usually a consequence of high current) does improve noise performance.

7N7

[Joel]

Originally posted by 7N7
...However the arguments I referred to in my previous posting, made by those better qualified than I were very convincing.

I see. Well, could you explain one or two of those for us? (or point me to which ones you are referring) Because I have not run across any that held up under analysis.
The High Current Myth is a derivation of the use of tubes with low plate resistance as driver stages. The low Rp is what was originally sought, and the high plate current was an (unfortunate) side effect. Those not aware of the necessary calculations have since focused on the current as the important factor, which simply is not the case. A tube like a 2A3 needs only 2 mA in a driver stage, and this can be proven both experimentally, and on paper, quite easily.

[7N7]

Originally posted by Joel


I see. Well, could you explain one or two of those for us? (or point me to which ones you are referring) Because I have not run across any that held up under analysis.
The High Current Myth is a derivation of the use of tubes with low plate resistance as driver stages. The low Rp is what was originally sought, and the high plate current was an (unfortunate) side effect. Those not aware of the necessary calculations have since focused on the current as the important factor, which simply is not the case. A tube like a 2A3 needs only 2 mA in a driver stage, and this can be proven both experimentally, and on paper, quite easily.

Unfortunately, I cannot remember the title of the thread - I certainly did not understand the technical arguments (way above my head) but since I opt wherever possible for the lowest anode resistance (or at least high gm in my cathode followers) I am obliged to be on the high current side - a glance at the anode characteristics is usually enough to convince me.

7N7

[fdegrove]

Hi,

Please don't say "grunty" - it's so Brett. Also, do a search on "slew rate distortion" and you'll see that the high current driver stage everyone thinks they need for the "big triodes" is actually a myth. The current required to neutralize their input capacitance is actually quite low.

There we go again...hardly a month has gone by and Mr. Whimpy Driver rears his uggly head again?

There's more to life than overcoming the Miller capacitance, you know.

Why some people, like yourself, insist on taking pride in designing for mediocracy is beyond me.

Brett's going to love you for it, Joel, and so do I to name a few ...we often wonder why we go through all the trouble explaining it all just for you, just to see you adhere to the same old logic again and have the nerve to insult all of us in the process.

Whatever you do at home is utterly your own business, but must you really spread it as the ultimate "Truth" ?

Be a man for a change and admit you can't read beyond an RCA datasheet or better still prove all of us wrong?

Cheers,

[EC8010]

Hello Joel,

I'll disagree that a high current driver is not mathematically required. Slew rates were discussed ad finitum on a previous thread, but just because a stage can swing the required current without hitting cut-off or grid current does not necessarily mean that it can do so linearly. Capacitive load are particularly unpleasant because they force vertical movement of the loadline (turning it into an ellipse), and gm is non-linearly dependent on Ia. Once you start looking at the loadlines and real valve characteristics, the rule of thumb 10x doesn't seem quite so daft after all.