[sdiy] Re: crystal clear

[Magnus Danielson]

From: "Forbes, William - EE - UK/Leamington" <william.forbes at luk-asg.com>
Subject: RE: [sdiy] Re: crystal clear
Date: Thu, 8 May 2003 15:04:07 +0100 

> But aren't inductors and capacitors reactive components and therefore will
> only have imaginary impedances.

Only in theory, only in theory. Also, I've seen several passive cross-overs
that include resistors as well, so there you are. Inductors can have quite
significant resistanses and when you enter the electrolytic caps field the
losses also seems to come along naturally.

> Losses will be due to resistive components (copper losses in inductors
> etc.).

Yes, in whatever form is in there.

> I agree that there will be some loss in a passive network but what sort of
> magnitude are we talking about?
> Remember loudness is measured in dB. Therefore 3dB means 2x the power!

I'm not saying that it is the only drawback of passive crossover filters, but
rather that it is one which we better do without when we can, and my point is
that it is feasable to do active crossover filtering.

Remember that a 1 dB loss becomes quite alot of watts when you toss a medium
power signal (150-300W) signal onto things.

PA systems tends to be used at and beyond their limits in power all the time.
The problem is how do you get the maximum power out of the system, avoids
burnups, avoid distorsion while getting the maximum clarity in the actual sound
at the same time? Active filtering and separate amps tends to make the equation
come together MUCH quicker, so much better that I left the passive filtering on
the exhibition shelf. I can't see any theoretical or practical evidence that
the passive crossover design have any major benefits to contribute over what
the alternative offers, it does have a number of drawbacks and we can avoid
those fairly easilly, so why keep suffering under the old burden?

Now, the benefits of passive filters is that when the amp clips, the passive
filter keeps some of the energy of the elements in frequency ranges they are
not able to handle power well. This seems to be a huge benefit, but it isn't.
Remember the mantra - limiter, limiter, limiter  - you have a limiter per band
to make sure the amp doesn't clip and doesn't push out more power than the
element can handle. If the amp doesn't clipp, there is no reason for any filter
to divert the energy from clipps.

A simple argument for limiters is this: Consider a gig, consider the guesting
house mix technician (or maybe just forget the last word there), consider the
influence of alcohol, consider that his hearing becomes a little more muffled
as the gig goes, what happends? Levels go up, up, up, up... limiters are there
to protect the system from misshapps and morons. When the limiter kicks in (you
*SHOULD* have LEDs visible showing limiter status) you don't get more, so then
you can calmly say to the guy to back off, since he won't get more. If he backs
of he gets more dynamic so it "feels" stronger. So you want limiters anyway.
A single amp with passive filtering can naturally have limiters, but it is a
cruder approximation and the limiter level is set so that you don't burn up the
most sensible element. Per-band limiters allows for much better control and
also less influences in sound. Oh, limiters and all that is dimensioned for
maximum gain setting throughout the post-crossover system, so there is no knob
to tweak by mooron to fool the system to get more. Limiters are locked by
design. All mooron available controls is pre-crossover filter.

Cheers,
Magnus