Gain / supply rails query

[Khron]

I'm looking for opinions on a somewhat odd pairing of variables.

From what i've read since i got hooked on NatSemi's line of amp chips, the LM1875 and the LM3886 sound a certain way at certain voltages and/or at certain gain settings.

I'm set on making a pa-150-like amp for the mid-bass driver for an active 2-way (3xLM3886, paralleled). This one's gonna be "fueled" with about +-30V, although i'm not entirely decided on the gain i'll use (20 or less - 16-18).

Each of the tweeters will be powered by an LM1875. I *might* power these with less than the +-30V for the mid-basses (possibly by regulation), and i'm pretty much set on lowering the gain on these (ZeroOne on the Gainclone forum seems to have experienced improvements on the high-end of the frequency spectrum this way).


Anyone else have confirming / conflicting opinions or advice???

[AndrewT]

Hi,
compare the actual sensitivity of the treble and mid/bass drivers in their finished cabinet. Use this info to determine if you need a different voltage drive from the treble amp and mid/bass amp.
eg. treble 92db/2.8V & bass/mid 86db/2.8V the bass/mid needs double the voltage (-6db relative sensitivity) to give equivalent output.

I would try for equal gain in all amps and use attenuators to adjust volume into each channel. Many crossovers arrange for active/passive attenuation on their outputs.

When you reduce the gain across a chipamp (or opamp) you increase the risk of instability. Take care to follow the manufacturer's or experienced builder's advice if you go for unusually low gain through your chipamps.

[Khron]

The actual level/sensitivity matching will be done in the active crossover. My question actually was something along the lines of "Which configuration is known to sound better?"

[Khron]

PS: As far as i know, big (-ger) supply caps are recommended for optimising low-end response of the LM series, and small (-er) caps for better midrange/treble.