 |
Driver-Baffle Placement
What did I learn from the prestige? Well I don't like how the
horizontal dispersion is affect by the offset technique I used. In fact,
I think I hate it! I never have been able to feel the stereo effect from
the prestige and I think it has to do with the offset used. From now on
I will not sacrifice horizontal radiation symmetry for edge diffraction
reduction. In this design I chose to round the enclosures edges and flush
mount the drivers.
Enclosure Design
Resonance.. resonance.. resonance! I just can't stand it. I
like the appearance of the WS802 enclosure, but they sing a little to loud
for my taste. I guess I could have ripped them appart and added bracing,
but really it would be easier to engineer my own cabinet. Now I have a
rule of thumb. Take the longest dimension of a enclosure. For every
six inches of this length, there should be a well-engineered brace.
Well-engineered means a one-piece windowed brace across the cabinet interior
and routed into the walls. Now they shouldn't be exactly every six inches,
instead place them irregular to break up the smaller resonances which will
develop.
Study Phase at the crossover point! (or the 2nd/3rd rule)
This is very important and is covered in some and (surprisingly)
not all loudspeaker design books. Let me throw an example at you. You wish
to use a tweeter and woofer, crossing them over at 3,000 hz with a first
order crossover. How should you wire them in phase, or out of phase? If
you took the simplistic model you would wire them in phase and a huge response
dip could/would occur at 3,000 hz. Why? Let's do a quick and dirty analysis
of the woofer's phase at the crossover freqency. First, we know that the
inductor from the filter is acting with a -45 degree effect. Now, the woofer
itself is around -45 degree phase angle to begin with. In addition, the
woofer is about 15mm behind the tweeter (from acoustic center). This adds
an additional -35 degrees to the sum. So the woofer is -45 + -45 + -35
= -125 degrees. Much different from the proposed -45 degrees of the crossover,
huh? Now we look at the tweeter. It's placement has already be compensated
for in the woofer's phase, and it is in phase by default. This gives it
a total phase rotation of 45 degrees. Now, if we difference these phases
angles, We get 170 degrees out of phase. This is almost 180 degrees and
total cancelation! So you say, will just reverse the tweeter. Now
they are back in phase at the crossover point, Right? Yes they are, but
since they meet at a -3 db point from the crossover and get a +6db effect
from perfect phase, you now have a boost of +3db at the crossover. We'll
this is a bit of a pickle isn't it? Now, say we move to a second order
crossover. That gives us -170 degrees for the woofer and 90 for the tweeter.
Let's see that a difference of 100 degrees and they meet at a point -6db
below reference. So if we add the +3db since they are about 90 degrees
out of phase we get -3db below the reference line. We'll since that didn't
work, let's try second for the woofer, and third for the tweeter. Now we
have -170 and 135 degrees for a differnce of 55 degrees. So, we get about
+5 db from the phase sum and they meet at - 6 and -3 at the point. So this
is about +1db above reference, much nicer!. However, we still haven't worried
about roll-offs and I feel I've rambled so that's it for this theory crap.
So in ending, phase is very important at the crossover point.
|