K# I tried sim to yours but used 12BY7 (as I've alot of these)I config these as triodes but the gain is still too high proceeding split-load p/s and pull up the noise if prev stage is pentode as shown). I learnt my lesson from the Citation 2 which uses 12BY7's and piles of DC internal fb..despite the amp performance being reasonable with low thd & noise the stability and squarewave resp was questionable.
RF tubes can be flighty. In the circuit I sent I like to use local feedback from common mode driver cathode to each of the grids; this reduces IM thd, improves power headroom at modest thd and sorts out any misbalance. The ECL82 pentode strapped as triode is better in this aspect, the split loaded cathode resistor is a better value compared with that used with the 12BY7. The ECL82 has lower gain, poorer loadline linearity but responds better to the technique.
If there is a (cheapish) pentode which gives 15dB gain in triode mode at lower thd and better linearity than an ECL82 in triode mode then I want to hear about it.
Despite 7199 type signal pentodes brilliant for voltage gain, a small power pentode i.e ECL82 strapped as a triode has good power b/w at low output Z ideal for driving parallel p-p.
The troubles start with the feedback from anode to grid that the b/w high end rolloff even with 6dB feedback is vasty improved.
There ought to be a simpler configuration of getting the g2 of signal pentodes to behave better with my ascertion that really "g2 is an electrode that controls distortion".
richj
RF tubes can be flighty. In the circuit I sent I like to use local feedback from common mode driver cathode to each of the grids; this reduces IM thd, improves power headroom at modest thd and sorts out any misbalance. The ECL82 pentode strapped as triode is better in this aspect, the split loaded cathode resistor is a better value compared with that used with the 12BY7. The ECL82 has lower gain, poorer loadline linearity but responds better to the technique.
If there is a (cheapish) pentode which gives 15dB gain in triode mode at lower thd and better linearity than an ECL82 in triode mode then I want to hear about it.
Despite 7199 type signal pentodes brilliant for voltage gain, a small power pentode i.e ECL82 strapped as a triode has good power b/w at low output Z ideal for driving parallel p-p.
The troubles start with the feedback from anode to grid that the b/w high end rolloff even with 6dB feedback is vasty improved.
There ought to be a simpler configuration of getting the g2 of signal pentodes to behave better with my ascertion that really "g2 is an electrode that controls distortion".
richj
lt cdr data said:
Yes, indeed. High Fidelity Circuit Design had four chapters dedicated to stabilizing high feedback designs. However, Crowhurst certainly knew better than that. He never came right out and said so since that would have been the end of his career as a design consultant to the Big Box manufacturers, but he sure did between the lines of other publications. Driving NFB to such ridiculous levels is necessary only as a marketing gimmick so that one Big Box manufacturer can claim that their system is better than the other guy's simply because it has better looking distortion specs.
It's not necessary, and it's not desireable. Add excessive feedback and the result is poorer sound. That's why all those Big Box systems tend to sound alike, regardless of whether you're talking a hollow state or a solid state implementation. For a design I did, I used 6.9db(v) of local feedback from the finals, and no more than 4db(v) of gNFB. Anything over that, and the sound deteriorated badly. Giving some 20db(v) of gNFB made the thing sound as bad as any SS design. Given that the level of gNFB is not excessive, stability is not a problem, and fancy tricks are not needed.
BJTs need quite a bit of gNFB to sound good. MOSFETs can't stand nearly so much, and VTs require even less. If you have a good, fundamental, open loop implementation, then you sure don't need anywhere near 26db(v) of NFB. The only ones who need so much are the marketing dept.
I wired up some el86's in triode as a driver for the front end of my citV
with some very satisfying results. There not terrably expensive either.
I've been told the 6EJ7 is a good one.
http://www.tubes.mynetcologne.de/roehren/daten/ef184pentode_as_triode.pdf
Here's a couple cool links!
http://www.pmillett.com/pentodes.htm
http://www.tubes.mynetcologne.de/roehren/daten/as_triode_e.html
with some very satisfying results. There not terrably expensive either.
I've been told the 6EJ7 is a good one.
http://www.tubes.mynetcologne.de/roehren/daten/ef184pentode_as_triode.pdf
Here's a couple cool links!
http://www.pmillett.com/pentodes.htm
http://www.tubes.mynetcologne.de/roehren/daten/as_triode_e.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.