Reckon we need a new forum....

[leok]

mdeneen,

Both speaker and resistor are wired to the output taps of the output transformer (all external connections).

Different configurations were fun and educational. They were done along with discussions on this BB with mobile, arco, & Conductor. Following is text from a note I wrote then:

-----------------------

Based on the above discussions, I tried a few experiments with my 2 Heath W5Ms.

The amp is normally configured as ultra-linear with KT66s biased dissipating just over 20 Watts each, at idle. One of my Sovteks died so I had only a set of Valve Art tubes which seemed to lack high frequency capability.

I had re-wired the g2 directly to 400V dc to operate in tetrode mode, which is designed to reduce Miller capacitance and improve high freq. response. This worked, but there was clearly some distortion at the high end.

Mobile strongly recommended the triode configuration, so I tried it; connecting g2 to the anode. I had expected high frequencies to vanish, but they didn't. Much of the high end distortion also vanished. What was left was some swirly sounding stuff on the "S"s.

Mobile even more strongly recommended single-ended triode so I disconnected the signal input to one of the KT66s (leaving it in to balance the bias into the output transformer). This was very good. In fact, this was much better than I would have thought. The sound was cleaner, more articulate, and the swirly stuff was gone. At this point, what I had was very much like a big operational amplifier, with a voltage gain of 10, and a single-ended output. This is a really cool thing and one would do well to stop there.

But Mobile also mentioned that single-ended triode with no global feedback is worth a try, and there are plenty of reviews that agree. Also, there is plenty of engineering that agrees. So I removed the first gain stage and phase splitter (which wasn't helping anyway). The results were very remarkable except that the KT66 had absolutely no local feedback and was scrambling the high end again. So I re-connected the ultra-linear connection to give it some negative feedback (why the triode connection doesn't do this is a mystery to me). This is a tradeoff between high end distortion and upper mid. detail. I'm sure this is the starting point for a real single ended triode amp. The point for me is I have never heard anything that comes close to this.

----------------

Since then I ran across an idea on the Plitron (transformer) webpage which featured push-pull output, triode connection, class A, no global feedback, and a little local feedback via 10 Ohm Cathode resistors. Initial gain and phase splitter were carbon copy of the 12AU7 circuit already in the Heaths. This, for me, has been the most successful arrangement. What I had wanted to address here was lack of dynamic capability of the set hookup. I just don't think the KT66 and Peerless transformer adapt that well to set operation. With the set connection, when sound became complex and loud it would loose definition. With the push-pull setup, full orchestra remains clean, with stable imaging.

The final change has been to swap ECC88s in for initial gain and phase splitting (there was some very minor bias adjustment for this) instead of the 12AU7s. his change was impressive. Everything sounds more real and the very high end is very clean. One reason I pushed for this last change was a disk I have of Andrew Manze and Richard Egarr (violin and harpsichord) playing Handel's violin sonatas. These period instruments deliver a load of upper harmonics that can get real screachy if distorted. The ECC88s realy cleaned up their harmonics. This is probably the first time I've heard my recordings of period baroque instruments sound right. I heard Manze and Egarr live a few weeks ago, and what I have now is reasonably close.

I use RF-7s and I think they're great. I have Forte IIs also. I think the RF-7s sound more natural, but I'll never give up the Fortes. Crossover components for both speaker sets have been upgraded (air core inductors and Musicaps). I'm aware that a pair of Cornwalls and a real s.e.t. might sound better, but I don't have those. not yet anyway.