4 way speakers

I was working on this project during last summer. It was an experiment, where I tried to make a speaker with very simple crossover. Lower frequencies were filtered out with capacitors and higher frequencies were filtered with foam of varied thickness. The benefits of this design are low cost(no expensive inductors) and flat impedance (easy load for an amplifier).
This speaker still needs a separate powered subwoofer, because lower woofer is in bandpass enclosure and it's frequency response drops sharply at around 100hz.
I don't have photos of finished speakers yet.

.

Car Audio

I've been playing with different car audio setups for a few years. Here is where I am at now.
For tweeters I use modified  Aurasound NSW2 drivers (cone is slit and coated with damping spray and rear chamber is staffed with foam)  attached to small horns. they cover ~1khz-15khz
For midbass I use modified Dayton Audio PA165-8.
Subwoofer is 15in driver from diycable.com (very nice sub)

Foam acts as lowpass filter and provides smooth high frequencies roll off.

I use a single inductor for woofer's crossover. if you use metal drivers, notch filter has to be implemented, foam won't filter out metallic resonances.

This is frequency response(red) at driver's sitting position. It looks like high frequencies are recessed too much, in reality tonal balance is just right.

nearfield measurements of woofer and tweeter (system is playing pink noise)

frequency response on tweeter axis, mic is 8in away

So, my conclusion is that car system needs to be calibrated on tweeter axis, no matter where tweeter is installed. If I tried to achieve flat frequency response at drivers position, sound would be very bright and unnatural.

Of course there are still some problems in my install, particularly that hump at ~300hz

boombox, ghetto blaster

Here is my new project with drivers mentioned in previous post.
I used amp6 for woofers and t-amp for tweeters. Crossovers are passive (I didn't have time to design and build active ones). Crossover point is at around 1khz. Woofers' 7khz resonance is about 30db down, which is apparently not low enough, because at high volume sound becomes "crisp" It should be tested with higher power amp, because that sound could be caused by clipping amplifier. Anyway, it plays loud and sounds quite good for a boombox.
For power supply I use 10 rechargeable AA batteries.

Sunflower is not part of the boombox, it's just sitting behind.
Back of the boombox looks exactly the same as front(tweeter and two woofers)
Calculated frequency response.

Update, August 2013
I've tried to listen to this speaker through my UCD180 amplifier for a few days.
And I came to conclusion that notch filter must be used with metal woofers.
That spike you see around 7khz, 35db down is a big problem. It is not immediately apparent, but over time it develops into an unpleasant listening experience. If you are building speakers with metal drivers based on someone else's designs, I suggest you choose only those crossovers  that use notch filters.

aluminum woofers' cone treatment experiment

These modifications were done a few years ago. I'm just reusing these dayton DA175-8 woofers for a new boombox project. Cones were slit an covered with automotive damping spray. Small copper caps are glued on top of pole pieces.

Blue line in the graph below shows modified woofer's FR, measured on axis(in box).

Yellow line is FR measured with thick foam placed in front of the woofer.

Manufacturers' supplied FR

I was doing these modifications to reduce resonance above 5khz. As you can see resonance is still there. Unfortunately aluminum cones do not benefit from this kind of modifications.