|
FIGURE 6 W70SF crossover
(modified)
First order to keep it simple
although not as simple as Jeff's W70, which had only the woofer low pass
filter and the tweeter high pass filter. The ten inch unit in that
system was running full range and had no level control.
It should be noted here that the
original crossover in Jeff's W70 units has the 4mh low pass filter in
the woofer section and the 4uf high pass filter in the tweeter section
along with a 50W pad. The ten inch unit
operates full range.
.
These potentiometers (pots/pads) are not
L-pads despite that they resemble such due to the schematic. They are a
high quality wire-wound pot about an inch in diameter and if they are as good as
those used in the wood boxed Wharfedale crossovers of the fifties, they'll last
for decades without having to be cleaned, after which they'll last for more
decades.
Jeff's W70 had such a pad in the tweeter
circuit and after removal and a good alcohol soaking, the device performed as
expected. While such a pot will upset the impedance more that a true
L-pad and cause a slight crossover point shift, this is not worthy of
concern. The response curves of fig 4 put that concern to rest.
If one is very concerned about impedance
variations due to L-pads, then perhaps they should use T-pads, or pi-pads or
even H-pads. Such pads are not readily available but can be made.
The L-pad requires two variable resistors, the T and pi pads will require 3 and
the H will require 5. They can also be made with fixed resistors.
An associate of mine from the nineties was
one such person with impedance variation due to L-pads so, after designing the
crossover, he tuned the mid and tweeter levels with fixed resistors, not as
fixed L=pads but as a single series resistor to the speaker. So much for
the concern about impedance variation.
In our flagship design in the 90's, we used
an MTM configuration with a second order crossover. This required
reversing the mid section polarity. While that may be true for a single
mid-range, the MTM configuration may be an exception. That system always
seemed a little bright, so to speak. Years later, when the LMS was
acquired, a response showed a nasty peak centered around 3khz and about an
octave wide. Upon pulling the midrange units out, it was noticed that
their polarity was reversed, so it was put back to conform the polarities of the
woofer and tweeter. The peak dropped.
In all the crossover design circuits I've
seen in books written by engineers, no mention was made of polarity exception
using two midrange units as per the MTM configuration. The transfer
functions used for filter design seem to assume a single point source for each
speaker and how their wavefronts interfere. It's similar to the assumption made
about pressure wave amplitude decay following the inverse square law. That
law assumes radiation into free space or an an anechoic chamber, there are
no boundary reflections. In a reverberant field, such as a room, that law
goes to hell in a handcart, PDQ. (Pretty Damned Quick)
|
