Suncalc wrote:
I seem to have recovered my health and as such have tackled the issue of setting our overall frequency response for our 6SN7/300B amplifier. However, before I begin with the details, I believe that a few comments on frequency response in general are appropriate.
The required frequency response of an amplifier is a careful trade off between the necessary bandwidth to reproduce the signal to be amplified and the signal distortion and degradation which results from an excessive bandpass characteristic. If the bandwidth is too low (or misplaced), critical content will be stripped from the signal. If the bandwidth is too large, the amplifier can suffer from bias excursions, intermodulation products, and high frequency oscillation.
That's the first mention I've heard of bias excursions due to overly large coupling caps and it's an excellent point most people overlook. This website has an article and a calculator I found useful:
http://www.pentodepress.com/home/amplif ... excursion/I have one minor disagreement with your design philosophy. Although you can't hear 10 or 20Hz, you can feel it. That's why IMHO the low frequency response of the ears should not be used as a basis for determining the low frequency response of an audio amplifier. I can plug my ears and still "hear" (actually feel) the low frequencies through my body. That's one reason I don't care for headphones. Without real speakers there is simply no bass to feel. I would maintain that an audio system should be capable of reproducing frequencies down to 20Hz
without significant attenuation to be regarded as truly hi-fi. EVERYONE can feel deep bass.
High frequency response is a different story - not everyone has the ability to hear out to 20kHz. However, I found that the 300B, being designed for audio frequencies, to be incapable of extended frequency response sufficient to cause oscillation or high frequency noise amplification problems. My amp is down about 2 or 3 dB at 20kHz without any grid stopper resistor (my hearing is only good to about 16kHz and I can't feel the treble so I don't care if the response is down by about only 1dB at my limit of hearing of 16kHz). I could be the trannys or stray capacitance creating NFB at high frequencies and not the tubes - don't know for sure - that's speculation. Or maybe my proprietary 3D wiring.
Just my 2 cents.
After reading all the work that has gone into the design of this amp I feel like I cheated on mine by just using what was an off the shelf circuit, but one that was super easy to understand, draw load lines for, and build. I used very inexpensive PIO Russian coupling caps, Hammond OTs, simple (and cheap) isolation trannys for power, SS rectification and EH brand 300Bs and it sounds SUPER!
Can't wait to see the final design (and the final cost!). One question. Would you care to share with us the results of the testing that was done on caps that you mentioned? You've got me very curious. I've been doing a lot of listening tests on caps (and have formed some opinions about the differences in sound being related to the impedance of the source and load in the case of different coupling caps), but as an engineer I also believe in test data (when used appropriately).