DIY speakers modification

[Tube Radio]

I built my DIY speakers about two years ago and at the time I used a 6db/octave passive crossover with them. I recently built a small pair of bookshelf speakers (cabinets from a pair of auvio 6.5" two way bookshelf speakers with better speakers + crossover installed) and used a 12db/octave crossover with them. They sound much better than my DIY home stereo speakers so I'm gonna re-do the crossovers to be 12 db/octave which should get me better sound. I am keeping the l-pads in place as I have no clue what the efficiency of the midrange compression driver is with the horn lens I have installed on it. Now I could figure out an L-pad circuit using two resistors for the tweeter, but I like to be able to adjust the midrange drivers and tweeters to the amp I am using which right now is a Scott LK-72A. Now OTOH if I had something to measure the midrange driver's efficiency I would do away with the variable L-pads altogether and go with a fixed L-pad circuit. Maybe y'all can help.

I am using these components
Aurum Cantus G2Si Ribbon Tweeter SPL 96.0 DB 1W/1m with 2 ohm resistor in series to provide for 8 ohms impedance.

Eminence PSD:2002S-8 1" Titanium Driver 8 Ohm Sensitivity: 105 dB (1W @ 1m on axis horn) connected to a Dayton H08RW 8" Round Waveguide.

Eminence Delta Pro-12A 12" driver SPL: 99.2 dB 1W/1m

Looks like I don't need an L-pad circuit on the tweeters and apparently would appear I need to get tweeters with a higher output.

 

[RayW]

Why do you think 12dB/octave filters will sound better than 6? You can't base the decision on the fact that you heard speakers with 2nd order speakers and they sounded better than what you have. Listen to some Dunlavy SC-VI. They have 6dB crossovers and sound pretty darn good.

Arbitrarily changing filter slopes and crossover points rarely results in an optimum speaker. The slope and frequency has to be tailored to the drivers. For instance, looking at the published response for the Delta 12A if you put a 12dB/octave filter at 2kHz then it will counter peak that occurs around 2500Hz and you will end up with something that has an actual -3dB point somewhere around 2800Hz.

Looking at the response on the 2002 you could do 12dB at about 3kHz and probably end up with something reasonable.

As for the slope/point for the mid to tweeter crossover you have a lot more leeway. There is a lot of overlap and you should be able to pick a point pretty much anywhere you want north of about 5kHz.

If you run the tweeter wide open then you can pull the pad out and maybe use it on the woofer if you need it. Just watch for power ratings as it'll see a lot more on the woofer than on the tweeter.

 

[Tube Radio]

I forgot to mention my crossover frequencies.
Currently at 6db/octave
Midbass 1,414 Hz
Midrange 2,545-12,738 Hz
Tweeter 12,738 Hz

Will be at 12db/octave
Midbass 2,545 Hz
Midrange 2,545-12,732 Hz
Tweeter 12,732 Hz

When I first built these speakers I per GordonW's instructions crossed over the midbass a little lower to counter the peak which works. I think my main problem is that I am adjusting the L-pads to balance the sound at the crossover frequencies which would seem louder as both speakers would be reproducing near those frequencies whereas with a 12db/octave Linkwitz-Riley crossover would make the output more balanced at the crossover frequency.
I use a couple subs run on a separate amp with a low pass crossover of 150 Hz.

The main reason for me going to a 12db/octave crossover is because when I have the L-pads set to put the tweeters and midrange at the proper levels the tweeters and midrange drivers sound a little too loud at the crossover frequencies.

EDIT: I just thought of something. I see why a lot of people do not like to use L-pads. It is because they can add unwanted inductance due to the way they are wound.

 

[RayW]

Ah, now I understand your concerns. If it's just around the crossovers that you're having trouble you could try a broad, shallow notch filter centered around the XO. It could be outboard and easily reversible. Just something else to mess with.

As I stated above, I wouldn't go with a symmetric XO from the woofer to the mid because of that 2500Hz bump in the woofer response. I'd back the woofer filter down to around 1800-2kHz or so and feel around there for what works.

 

[Tube Radio]

I'll see what I can come up with. All I gotta do is plug the 1,414 Hz crossover frequency for the midbass driver into a formula and then make adjustments to the frequency to get a standard coil value.

Also a 12 db/octave crossover on the speakers will make them in phase with the subs instead of being 6 db out of phase like they are now.

I plugged in the numbers into an L-pad calculator and got these results:
Resistors
R1 = 3.9 Ohms 19.49 Watts
R2 = 8.42 Ohms 9.99 Watts

I adjusted the L-pads for the midrange horns to 8.42 ohms across the driver (driver disconnected) and they sound good like that so I'll order those two resistors along with the crossover components and see how they sound once I make the crossovers 12 db/octave.

I can get the 3.9 ohm resistor easy, but the 8.42 ohm resistor even if I get an 8 ohm resistor and series it with a .42 ohm resistor will be near impossible to do. Now would it be possible to just use a .4 ohm resistor in series with an 8 ohm resistor without affecting the circuit any?

Just looked and I can get a .43 ohm resistor which has a 5% tolerance which will be close enough.

I also have a Zobel network on the midbass driver.

 

[Tube Radio]

Ok I got the fixed L-pad network installed on the midrange horns and they do sound better. Now I just gotta wait for the crossover components I ordered yesterday to come in then I can see how the 12db/octave crossover works. I will need bigger boxes for the crossover so I can keep my 6db/octave crossovers together except for the 1/4" jacks just in case the 12 db/octave crossovers don't sound good.