Pete Milett 가 잡지에 발표할 앰푸의 회로랍디다. 잡지에 싣기전에 자세한 설명을 안하기로 계약되어 있어 회로만 공개 한것이고 만든 다음 왼전히 튜닝되어 진 것은 아님

14 와트에 output impedance 가 2옴이라니 조금더 negative feedback을 걸어 1 옴 이하로 되면 비교적 싼 가격으로 요지음 소리통을 울릴 수 있는 SE amp 를 만들 수 있는 회로로 보입니다

이 앰푸의 negative feedback은 출력단의 UL탶에서 드라이버단의 플레이트로 연결되어 있슴 -- 똑 같은 방법이 차덩앰푸란의 비스므리한 넘에서 구연되어 있고 이를 E-Linea 라고 별명을 붙인 것 같음

조금더 negative feedback 거는 방법은 cathod bypass 콘을 제거하면 어떨까? 생각해봄

Goal
1. high power
2. low output impedance
3. relatively low cost

을 충족 할수 있지 않을까 생각해봄

소리야 나중에 만들어 보야 알거고

영문은 Pete Milett가 쓴 글

It's good for ~14 watts at clipping (5% THD)... 1 watt is about 0.4% THD, but it's optimized for a nice harmonic profile (2nd harmonic highest, then 3rd, then 4th, ...) rather than mimimum THD. Zo is less than 2 ohms, mostly due to the local feedback from the UL tap to the driver plate circuit.

The D3a (also 7721) is the best driver tube I've ever used.
The power supply is conventional, 720VCT -> 5U4GB -> 40uF -> 5H -> 40uF, using PP-in-oil motor caps. Around 400V, I think, at B+.

---------------------------------------------------------------
 
I thought it would be interesting to try feeding the whole circuit directly from a choke (the 1.5H choke), with only the ultrapath cap - in other words, without the Solen. That's what I started with.
It worked pretty well, but I had a little hum on the output (around 3-4 mV). I remembered reading about schemes to use a small cathode bypass cap in addition to the ultrapath cap to null out any residual power supply ripple. The 20uF cap did a good job at that.

Of course, as Serus pointed out, the other way to kill the ripple-induced noise is to eliminate the ripple. Today I tried 2 ways to do this: changed the output cap in the PS from 50 to 200uF - the noise was about comparable to using the 20uF solen. Then I changed the PS output cap to 100uF and added a 40uF oil cap from the OPT B+ to ground, essentially making the 1.5H choke just another LC section of the PS. With this the noise is even lower, near the limits of my measurements at around 120uV (I read 140uV with the amp off and the flourescent lights on!) I'd expect that a proper layout on a metal chassis will drop it even more.
(BTW, the filament winding CT is grounded and doesn't appear to be an issue. Also, the power supply is for a stereo amp, so the other "channel" is currently a 3k dummy load).
There wre small differences in distortion performance between the three cases but probably not enough to worry about. And you could also make a valid arguement that any noise under 1mV or even 2mV doesn't matter, especially with my <~92dB speakers.
So maybe you guys are right, and the "conventional" connection (if one can call ultrapath "conventional"?) is best. Intuitively I agree that it's best to keep the AC current path as short and simple as possible, through the best quality caps.
I also think that the bass sound is usually better with more/bigger filter caps. Using a small choke as the 2nd LC stage forces any HF current demanded from the power supply to flow through the last cap (which should be a good quality cap), but still allows current to flow from the bigger PS caps (which aren;t so critical) for high current, low frequency pulses.
So, I'll probably build it without the Solen and with a little more capacitance... sometimes "novel" is not "best", eh?