Going from my Atlas amp to the sub's high level (binding posts) does not mean the Atlas amp is driving the two subs? Right? The plate amp in the sub still powers the subwoofer and the Atlas amp is simply sending out the signal for the plate amp to amplify. This is the way I have always understood it?
Yes, again you are completely correct. The amp would see a load impedance equal to the impedance of the main speakers (8 or 4 ohms or whatever) in parallel with the input impedance of the high-level input of the sub, which for all practical purposes will be essentially identical to the impedance of the main speakers. For instance, if the high-level input impedance of the sub is 20,000 ohms, that in parallel with 8 ohms is:
(8 x 20000)/(8 + 20000) = 7.997 ohms.
My pre has 3.3uf coupling caps. No, my second output is not buffered separately. I use low capacitance Blue Jean cables that are 10 foot long on the subs.
Assuming that the length of the cable running to the main amp is also around 10 feet or less, and given the very low capacitance of the Blue Jeans cables (12 pf/ft if you are using the LC-1 cable), you should be ok in the present configuration with respect to the possible effects on the upper treble that I mentioned.
With respect to the possibility of deep bass rolloff, 3.3 uf represents a capacitive reactance (the capacitive form of impedance) of about 2413 ohms at 20 Hz. In terms of the magnitude of the overall output impedance of the preamp, that combines with the resistive component of the preamp's output impedance (which to a reasonable approximation is probably about equal to the 5K specified nominal output impedance) as the square root of the sum of the squares of the two numbers. So the 5000 ohm value at higher frequencies would increase to 5552 ohms at 20 Hz. That is not much of a difference in the context of a 20K load, but some amount of phase shift will also occur as a result of the capacitance (I won't take the time to calculate how much, but I believe it will be small), and the slight impedance rise together with the phase shift might, I suppose, be enough to be very slightly perceptible.
Also, keep in mind that if you connect the subs to the amplifier outputs instead of the preamp outputs the signals seen by the subs will reflect the sonic effects of the amplifier, which of course are not reflected in the signals seen by the sub in the present configuration. Very conceivably that could be the most significant factor in the sonic differences between the two configurations.
In any event, I would definitely give the speaker-level hookup a try, before investing in a buffer (which was an excellent suggestion, but as my math indicates might not be necessary in this case even with the preamp-to-sub configuration).
Regards,
-- Al
