Well, I'm back after spending a few days doing
listening tests. I think I'm closing in on my
desired results. After playing around with the
crossover a little, things started to sound really
good, but there's still some room for improvement.
I had some Lpads laying around, so I used one to
attenuate the tweeter slightly. After reading
some other threads here, it sounds like the
easiest way to make a fixed attenuator would
be to measusre the impedance of the two halves
of the Lpad, and replace them with fixed
resistors.
The origional crossover I was using was a cheap
off the shelf 12dB/octave unit at 3.5k. Once
the tweeter was adjusted, the high end was
properly balanced, but the uppermidrange sounded
a little hollow. When I was digging around in my
parts boxes for the Lpad, I discovered that I
had a Radio Shack adjustable two way crossover.
I tried it out at the three different
frequencies, 2000, 2500, and 4000, 6dB/octave.
I like the sound of this crossover much better,
and things sound particularly good at 2.5k.
There is a barely detectable amount of
distortion in the lower range of the tweeter
however.
One other observation was that with the first
order crossover, the listening position became
less critical. With the origional steeper
crossover, in order to hear an even balance in
the frequencies around the xover point, I had to
position my speaker so that the midpoint between
the tweeter dome and woofer edge was level with
my ears. Deviating by a few inches from this
point resulted in a percieved phase cancelation
at certain frequenccies. With the first order
crossover, the sweet spot is much wider, and
the effects of moving out of this area is much
less drastic. I want to stick with a first order
crossover for my design, because it just seems
to work better at closer listening positions.
I think my most logical course of action would
be to proceed with the 6dB/octave crossover,
but go a little higher in frequency, to see if
I can get rid of the wisp of tweeter raspiness
that only trained ears will be able to detect.
The midrange started to sound hollow again at
the 4k. point, so I'm hoping that 3 or 3.5k.
will do the trick. The other possibility would
be a higher order crossover on the tweeter around
2.5k.
Before I go any further, I think it's time
for me to take some electrical measurements.
Before I start spending money on more crossover
components, I want to know what my driver
impedances actually are around my potential
crossover points. I've never made a Zobel
network before, and I'm also curious to see what
that will do. Fortunately the recording software
on my computer provides a very nice tone
generator function, and I have a good multimeter,
so I won't have to invest in any test equipment,
other than a load resistor.
So far this project has been really fun, and the
knowledge I have gained by searching the threads
and following the links from this forum has been
absolutely incredible. I wish I had found this
site a few weeks earlier before I ordered my
components. The prototype speaker I have now
sounds delightful. I spent all of
last night listening to my reference CDs, and
it's making most of the other speakers in my
house sound like garbage.
Building these speakers has also renewed an
interest in classic rock music for me as well.
Normally, I hate classic rock radio, because it's
all the same songs that I've heard 1000 times
before on 100 different stereos. It sounds pretty
tired after a while, but while trying to
tune my monitor speaker, playing my local classic
rock station was great, because I know how all
of those tunes should sound no matter where I
hear them. I'm looking forward to spending a good
part of my holiday weekend testing and tweaking.
listening tests. I think I'm closing in on my
desired results. After playing around with the
crossover a little, things started to sound really
good, but there's still some room for improvement.
I had some Lpads laying around, so I used one to
attenuate the tweeter slightly. After reading
some other threads here, it sounds like the
easiest way to make a fixed attenuator would
be to measusre the impedance of the two halves
of the Lpad, and replace them with fixed
resistors.
The origional crossover I was using was a cheap
off the shelf 12dB/octave unit at 3.5k. Once
the tweeter was adjusted, the high end was
properly balanced, but the uppermidrange sounded
a little hollow. When I was digging around in my
parts boxes for the Lpad, I discovered that I
had a Radio Shack adjustable two way crossover.
I tried it out at the three different
frequencies, 2000, 2500, and 4000, 6dB/octave.
I like the sound of this crossover much better,
and things sound particularly good at 2.5k.
There is a barely detectable amount of
distortion in the lower range of the tweeter
however.
One other observation was that with the first
order crossover, the listening position became
less critical. With the origional steeper
crossover, in order to hear an even balance in
the frequencies around the xover point, I had to
position my speaker so that the midpoint between
the tweeter dome and woofer edge was level with
my ears. Deviating by a few inches from this
point resulted in a percieved phase cancelation
at certain frequenccies. With the first order
crossover, the sweet spot is much wider, and
the effects of moving out of this area is much
less drastic. I want to stick with a first order
crossover for my design, because it just seems
to work better at closer listening positions.
I think my most logical course of action would
be to proceed with the 6dB/octave crossover,
but go a little higher in frequency, to see if
I can get rid of the wisp of tweeter raspiness
that only trained ears will be able to detect.
The midrange started to sound hollow again at
the 4k. point, so I'm hoping that 3 or 3.5k.
will do the trick. The other possibility would
be a higher order crossover on the tweeter around
2.5k.
Before I go any further, I think it's time
for me to take some electrical measurements.
Before I start spending money on more crossover
components, I want to know what my driver
impedances actually are around my potential
crossover points. I've never made a Zobel
network before, and I'm also curious to see what
that will do. Fortunately the recording software
on my computer provides a very nice tone
generator function, and I have a good multimeter,
so I won't have to invest in any test equipment,
other than a load resistor.
So far this project has been really fun, and the
knowledge I have gained by searching the threads
and following the links from this forum has been
absolutely incredible. I wish I had found this
site a few weeks earlier before I ordered my
components. The prototype speaker I have now
sounds delightful. I spent all of
last night listening to my reference CDs, and
it's making most of the other speakers in my
house sound like garbage.
Building these speakers has also renewed an
interest in classic rock music for me as well.
Normally, I hate classic rock radio, because it's
all the same songs that I've heard 1000 times
before on 100 different stereos. It sounds pretty
tired after a while, but while trying to
tune my monitor speaker, playing my local classic
rock station was great, because I know how all
of those tunes should sound no matter where I
hear them. I'm looking forward to spending a good
part of my holiday weekend testing and tweaking.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.