Why triode Mk3 is louder than stock Mk3

[Wieslaw Lipowsk]

One of my Mk3 with a half 12AT7 and 6CG7 confugured as long tailed pair, and with KT88 output strapped in triode mode plays a lot more loudly than my other, unmodified and stock Mk3. Why is it so?

The only way to adjust the volume in the amp is through the negative feedback level. I am afraid to fiddle here, however, for the balance between the vocals and the instruments could be destroyed.

I'd be grateful for any advice and explanation.

W.L.

[SY]

Chances are that either a feedback connection is made improperly or that one of the feedback components has gone bye-bye. If there's a cathode bypass cap in the first stage, that can be a problem, too.

[Wieslaw Lipowsk]

Thanks, I thought about that too.

I modified my Mk3 as per Kara Chaffee project, Audioexpress 2001, where there is 1.8K resistor in the negative feedback loop, and then 20R to ground. No cathode bypass cap in the first stage. That 1.8K is said to be an optimum value sound-wise.

[Wieslaw Lipowsk]

Now that I thought about it again I came to the conclusion that the sensitivity of the new tuned Mk3 might be much higher than it was before.

So the output seems louder, too. But the difference is astonishing, the modified amp seems to be not 30 watter (in triode mode), but as much as 90 against the stock one.

[Dave Cigna]

I know what you mean. In the past I have converted PP 6V6 amps from pentode with considerable NFB to triode with little or no NFB and they seemed to play just as loud. The common wisdom is that the NFB causes harsh clipping that is audibly ugly sounding. Without the NFB (or with just a little) the clipping is more benign. That means you can play closer to max power; transients might clip but the distortion is not objectionable.

Besides that I can think of one other reason. The 6CG7 LTP is a fairly beefy driver compared to the concertina in the stock unit. It might be capable of driving the grid a few volts positive (AB2 operation) without choking. That means you might actually be getting more than 30 watts out of the amp. Depending on the operating point and load, it is sometimes impossible to get a pentode to operate in AB2. If the loadline crosses Vg=0 near the knee in the curve then driving the grid positive will not get you any more power; the amp just clips. OTOH, you can usually push the grid of a triode quite a bit positive and get more power, assuming the driver is up to it and doesn't distort itself. I'm not familiar with Kara Chaffee's circuit....

-- Dave

[Wieslaw Lipowsk]

One half of a 12AT7 (actually GE 6201) directly coupled to a 6CG7 (green Sylvania). It is a Schmitt type or long-tailed pair type of phase splitter.

Quite similar circuit to that you can see in an old Marantz 8 and Eico HF-60.

Here, there is no cap in the global nfb circuit, but "lower" 390pF mica cap was retained.

When replacing the 1.8K in the global nfb with 800R, the amp plays as loudly as the stock UL one. But my ears can hear that it is not an optimum setting.