Feedback on 1st Crossover Attempt - W130s-4/BC25TG15-4 TM

[Colonel7]

Hi all - This is my first ever crossover attempt so I'm looking for feedback and to learn (and not make any terrible errors). I bought 4 of the buyout W130s-4 drivers and a couple tweeters to try and learn by doing. I made a sealed 12.5 liter prototype cabinet based on the T/S parameters I measured after burning in the drivers (using WinISD). I made a couple attempts at a crossover using PCD because that's what I'm the most familiar with. .

The first is crossed at 2400 Hz. Not sure if the values are visible on the schematics. Woofer: L2: 0.6 mH; C5: 11 uF; R7: 5.6 Ohm. Tweeter: C9: 12 uF; L9: 0.2 mH; C10: 50 uF; R15 2.0 Ohm; R 16: 2.2R.

The second is crossed at 2800 Hz. Woofer: L2: 0.5 mH; C5: 11uF; R7: 5.6 Ohm. Tweeter: C9: 9 uF; L9 0.2 mH; C10: 27 uF; R15: 2.4 Ohm; R16: 3.7 Ohm.

Any thoughts would be great. Hopefully I can get far enough to order the crossover parts and they'll be in stock.

Once I finish the TM, I plan on tearing it down and making a floorstanding MTM with basically the same crossover but with the woofers wired in series. Of course that's probably another month or two out.

[rpb]

It looks like you used the factory measurements. They will not include the baffle step. Therefore, the output at 1k and above will be much louder than shown in your sim. I would suggest increasing L2 to about 1.3mH. That's a guess. Make the transfer function show a reduction at 1k of 6dB,.

Maybe add a resistor before the tweeter filter to increase the speaker impedance at higher frequencies.

[Colonel7]

Thanks for the help! I will give it a try. Newbie questions: How do I make the transfer function reduce 6 db? Adjust the tweeter target level an additional 6 db and then the component values accordingly for the tweeter?

You are right that I used the manufacturer specs for the frequency response. I did take frequency response measurements (nearfield and at 1m for each driver and a combined at 1m)) but I need to troubleshoot why PCD won't accept the FRD even though I thought I had re-formatted it correctly. I may need to learn XSim to use them - I'm not sure if they're corrupted or just due to the limitations of PCD being old and not updated anymore.

[rpb]

The transfer function just shows what the filter is doing. Change that coil value to 1.3mH, and then look at the transfer function. It may be close to -6dB at 1k, and that would indicate that you are on the right track.

On the measurements. Look at them in note pad, or maybe word, and look at the values. On my setup, the values are too low to show up on the screen in PCD. They may be in the 30s or 40s, when they need to be up around 90dB. You can correct this with Excel, if that's what is wrong.

[Chris Roemer]

You didn't share YOUR (measured) T/S parms w/us, but 1/2cu.ft. box for this driver SEEMs large enough to port.
Using mfr. specs, I only see 80Hz sealed, while a 1-1/2"id x 4" long (rear) port probably gets down closer to 50.

To see how "baffle-step" compensation looks, try a 2.0 series coil on your woofer in your sim (along w/your Zobel).
Then, sim the tweeter as 2nd order, using a 6uF series cap and 0.70mH shunt coil, and an L-pad with SR = 5ohms and PR = 1ohm. (I don't really go for a 1ohm parallel resistor on an "L", but try it (in this instance) to see what the overall FR curve looks like.) You should see a slow (but steady) decline rising from around 300Hz up to maybe 1.5kHz? Then the tweeter will look like it's running about -6dB below the woofer's level (near 2-300).
This is "full" (-6dB) BSC, but many shoot for less (like maybe only 3-4dB). Based on personal prefs. AND room position.

[rpb]

Another way to look at this..... Look at your 1 meter on axis measurement. (Hopefully a gated measurement.) You will notice that the woofer is louder at 1k, than 400hz. You want a coil big enough to make the woofer, ( and the completed speaker), SPL at 1k about the same as at 400hz.. If you have some xo parts, you can do things by trial and error. Takes longer sometimes, but it's fun to me. I have lots of parts.

[rpb]

You didn't share YOUR (measured) T/S parms w/us, but 1/2cu.ft. box for this driver SEEMs large enough to port.
Using mfr. specs, I only see 80Hz sealed, while a 1-1/2"id x 4" long (rear) port probably gets down closer to 50.

To see how "baffle-step" compensation looks, try a 2.0 series coil on your woofer in your sim (along w/your Zobel).
Then, sim the tweeter as 2nd order, using a 6uF series cap and 0.70mH shunt coil, and an L-pad with SR = 5ohms and PR = 1ohm. (I don't really go for a 1ohm parallel resistor on an "L", but try it (in this instance) to see what the overall FR curve looks like.) You should see a slow (but steady) decline rising from around 300Hz up to maybe 1.5kHz? Then the tweeter will look like it's running about -6dB below the woofer's level (near 2-300).
This is "full" (-6dB) BSC, but many shoot for less (like maybe only 3-4dB). Based on personal prefs. AND room position.

I believe he is using a single 4 ohm driver. I think 2mH might be too much, but maybe not. For a sim, no worries. Try it and see.

[Colonel7]

Below are the T/S parameters for the W130s-4 after burning-in the driver with music and sine waves below the FS. I chose sealed b/c I want to see how it sounds and what the measured performance will be like compared to how it simmed in the box. I can add a port before tearing it down. Later with the MTM I aim to choose a volume that I can have a port with the option to bung it to make it sealed. And thanks for all of the advice. I will try it all and report back!

W130s-4 Unit 1 Loudspeaker parameters:

Fs = 63.83 Hz
Re = 3.70 ohm[dc]
Le = 102.49 uH
L2 = 124.39 uH
R2 = 6.18 ohm
L3 = 320.73 uH
R3 = 1.72 ohm
Qt = 0.53
Qes = 0.62
Qms = 3.82
Mms = 7.23 grams
Rms = 0.749310 kg/s
Cms = 0.859898 mm/N
Vas = 6.33 liters
Sd= 72.38 cm^2
Bl = 4.170500 Tm
ETA = 0.26 %
Lp(2.83V/1m) = 89.54 dB

Added Mass - Constant Bl Method:
Driver unbaffled
Added mass = 6.25 grams
Membrane Diameter= 9.60 cm

[Colonel7]

This was all really helpful; thanks rbp and Chris. I'm now understanding how the different component values are affecting things - in the past I was too cautious and not altering the crossover component values enough to draw definite conclusions. Still playing with them and will probably try full and partial and see what sounds best (and how it measures). I don't know what I prefer as this is new to me. And the one gated measurement I saved as an image rather than frd/text does show the textbook 6 db loss.