SpeakerTalk

[KefHeir]

Dear Kef comrades

I'm in the midst of converting my DIY Kefs to an active configuration. Initial results are mixed to disappointing so I thought I'd share the experience here in case anyone may have some thoughts on where I could go with this experiment.

A bit of back story. I inherited a pair of DIY kef speakers a couple of years ago which sounded remarkably good for their 1980s vintage. I decided to use them as my main speakers so I set about giving the cabinets a makeover, replacing the crossovers, and refreshing all the drivers with a new set from Falcon.

While the upgrades gave clear and obvious improvements, there were a couple of problems: the bass levels were too high and prone to boominess, and something odd was happening in the mid range which seemed to lack detail and generally sounded a bit asphyxiated. I got support from a professional engineer to take some measurements which reassuringly confirmed what my ears were telling me: there was a huge suck out between 100 and 1KHz and a sharp peak in bass at 60Hz, falling off sharply to the left and right. The response curve was basically saw toothing between 60 and 90db up to 1kHz where it then became more level. Suspicion fell on the band pass phasing of the new crossover (not the Falcon one I should probably make clear), so the initial plan was to build new, bespoke crossovers. It was during these discussions I decided to go the whole way and try an active set up.

I asked the same engineer to build an analogue active crossover. For this he said it was important to use balanced connections due to the risk of noise from the AXO, but he hit a problem with my QUAD QSP power amps which use 15 pin d-sub connectors rather than the conventional XLRs. He tried to patch in to the balanced signals in the QUADLINK system but without success so we decided to drop the analogue approach and do the filtering in the digital domain using a miniDSP (4x10). I had my doubts about this, but persuaded myself the benefits of active amplification might outweigh any drawbacks of the miniDSP electronics.

And so it's all built and connected and I had my first listen last week. There were clear pros and one very big con. On the plus side, the much flatter response was self evident. And now a lot more detail is present in the mid range. The bass was also much more under control and flatter. The benefit of the active tri-amping was also clear, with the speakers now able to cope with difficult orchestral and piano music without getting overwhelmed and compressing the details. Despite all these good things there was one major downside: the sonic illusion was broken. Even with the apparent deficiencies of the old passive crossovers the speakers were always hugely listenable, and at times could make musical detail sound exquisitely real. With the DSP that authentic voice or sonic texture or whatever you might call it has all vanished. The sound now has what I would describe as a quite unpleasant, synthetic quality, and everything that was musical and involving about the old sound has gone. (There's also a severe white noise problem which may or may not be a fault with the DSP.)

So, quite disappointing but nevertheless I'm enjoying the experimentation. I've read very positive comments in this forum about DSPs so I'd be very interested to hear anyone's thoughts on this. My immediate question I suppose is whether the quality of the sound could be improved by alternate DSP programming, or is what I'm hearing just the fundamental limitation of the miniDSP electronics, which I assume just don't measure up to the QUAD which I've listened to for a long time?

I hope this hasn't deterred anyone who was thinking of a DSP / active set-up, but on the basis of what I've heard so far it hasn't worked for me. Would be very grateful if anyone has any suggestions or thoughts on further things I could try, otherwise it looks like I'll be reverting to passive, albeit hopefully with better crossovers.

[audiolabtower]

The miniDSP does have some comments about noise. The consensus seems to be that if using the non balanced output, set it as high as possible and then attenuate before the power amp to reduce the noise while still getting a useful range on the volume control. It is not the last word in digital quality (or analogue op amps inside) but I would have expected a reasonable result, if not really high end. Does Quad not give a pinout on the d-sub? If not they are not the company they once were.

[KefHeir]

Thanks Audiolabtower, much appreciated.

[exkefman]

[KefHeir]

Thanks Tim. Sorry for slow reply, I haven't logged in for a while.

Unless I'm not following, what you describe sounds like a conventional (analogue) active configuration but using tape out rather than the standard line outputs. What would be the advantage of that?

(Not sure if this is going too off topic for the Kef forum, but the moderator can advise

Darren

[exkefman]