multiple bandpass subs

[DougSmith]

I am planning to build a trio of bandpass subs for my system using Earl Geddes' recommendation of having two broadband subs that overlap the mains in the 50-150 Hz region and one "ULF" sub to cover the 20-50 Hz region. I also plan to use the same high efficiency 12" and 15" B&C drivers as Earl does in his Abbey and Summa waveguide speakers. My question has to do with the use of electronic filters to optimize the response of the different subs. I am planning to drive the subs with a outboard amps and a DCX2496 to set levels, phase and filters, so I will have very good flexibility for doing that. Using WinISD to model the response and playing around with the box and port sizes for a while, I came up with the following - with and without filters.

Transfer function without filters

Transfer function with filters

The filters level out the response significantly and should allow me to blend the two into a flat curve overall, although I'm not sure exactly how to simulate the latter. If I look at the SPL curves, I get the following, using the filters.

SPL with filters

It looks like the maximum SPL with 200 watts (which is as high as I plan to go) would be about 107 db. I can't imagine wanting things any louder that that, especially since there will be 3 subs and some room gain as well.

So, my question is: Is it legitimate to use electronic filters and EQ in this way to optimize the response, or should I spend more time trying to optimize the bandpass parameters? Are there any drawbacks to this approach for bandpass subs? (I think I understand the pluses and minuses of bandpass vs vented plus sealed pretty well and I don't think I am ready to venture into tapped horns yet).

-Doug

 

[breez]

The efficiency of the ULF one doesn't look good. It may not be a suitable driver for the job, at least in the modeled box.

 

[DougSmith]

I have been debating that. The B&C 15" has a senitivity rating of 96db and high power handling of 1000W nominal with a 4" voice coil - so it has good thermal properties with low distortion, as I understand it (I am just a novice with this stuff, however). The Fs is 35 Hz and xmax is 9mm, so both of those could be more favorable for a ULF sub. The box could be bigger, but it is pretty large already. The question is: Is what I modeled good enough for my needs (music and HT, but not at room-shaking levels), or should I continue exploring different designs?

-Doug

 

[breez]

I would try modeling the woofer in a vented box and seeing if you can get better efficiency.It won't offer the benefit of acoustic lowpass (reduced distortion), but I think it is not so important for the 20-50 Hz sub anyway. Distortion products (2nd and 3rd harmonic) will stay pretty low in frequency and not cause localization problems. For the higher frequency subs the effect is more important as the distortion products appear quite high in the spectrum where they may be localized.

Also try different woofers for the ULF, maybe one with a lower Fs, so you can get better efficiency down there.

 

[SpeakerBob]

Why not design the subs to have flatter response before you use filters ?

Also why not use 6th order BP's ? They seem to give better output from what I've modelled.

Try modelling a conventional sub driver for the low box. Car subs model well in BP's too.

Rob.

 

[DougSmith]

Thanks Rob... I guess I'm missing something because when I put in your 4th order parameters for the BB sub I get a curve with a peaked top, not flat. My 6th order curves look about like yours (I changed the box sizes somewhat), but the output of the ULF looks lower. The parameters in the driver files I created look ok (some of the calculated values are a slightly off but changing them in the project file to match the B&C pdfs exactly doesn't affect things much).

6th order boxes certainly provide more control of the cutoff frequencies and look like they could provide a flatter response overall - and the output is indeed higher (both subs now reach 110 db). Are there any drawbacks realtive to 4th order I should be aware of, apart from the greater complexity of design and construction?

My results

- Doug