Selecting sub crossover frequency

[bfg9k]

I am thinking about building a subwoofer - I'm thinking about on the Shiva setups. I have an Adcom power amp I can use to drive it, so I'll just need an outboard crossover. I have links to the PLLXO x-over design page and ESP's active Linkwitz-Riley crossover and would probably try the passive one at 12 db/oct first for simplicity's sake.

The main speakers are a pair of Energy C-1 speakers, which are good to about 55-60 Hz although I'd like to crossover a little higher than that, say around 70-80 Hz. The amp is an NAD 304.

How should I go about selecting a good frequency to start with?

 

[roddyama]

Bfg,

The best x-over is no x-over. Currently your C1’s have an f3 of 53Hz as a ported design. In this configuration, the port output is dominating the lowest half octave to the overall FR. It may be possible to eliminate the port output (block the port) and get a closed box roll-off of (ultimately) 12db/oct that is 3db down somewhere in the range from 70Hz to 90Hz (depending on the actual alignment). That would make up your hi-pass section.

 

[Ken L]

How should I go about selecting a good frequency to start with?

Generally, for best integration you want a full octave of flat overlap above and below the crossover point.

Said another way - where you have two full octaves of flat overlap of the drivers (in their enclosure) , the center point is the best initial target for your crossover point.

It has been my unscientific observation that 4th order L-R is frequently used successfully on subs -

FWIW, I tried numerous slopes and points with a digital crossover and found L-R 4th order to be the best for my configuration.

In my case, cone breakup distortion sometimes would occur with the mid (actually full-range) until I filtered the lows with the crossover.

Many ways of doing this and looking at it - but in my case the same drivers/cabs with the former plate amps and a 12DB slope never did integrate well -

YMMV

Regards

Ken L

 

[bfg9k]

4th order L-R

If I were to pick a frequency as you suggest that's well within the flat response for the sub and the Energies, and that's above the f3 of the Energies, should I bother with blocking the ports? It's an easy enough experiment.

I guess what I'm wondering is how much like a ported speaker they'd act like if they're not seeing any input near where the output of the port is significant.

At 24 db/octave there shouldn't be much interaction between the drivers, if at all... I have some opamps kicking around so I'll have place a Digikey order for some caps and resistors to try to make a test circuit.

 

[roddyama]

Re: 4th order L-R

Hi Ken,

Generally, for best integration you want a full octave of flat overlap above and below the crossover point.

Said another way - where you have two full octaves of flat overlap of the drivers (in their enclosure) , the center point is the best initial target for your crossover point.

Generally, I agree with this approach. Although in the case of a mid-bass/sub x-over, it's not always the case that the mid-bass driver will have the needed extension, particularly at the higher outputs.

It has been my unscientific observation that 4th order L-R is frequently used successfully on subs -

FWIW, I tried numerous slopes and points with a digital crossover and found L-R 4th order to be the best for my configuration.

Even with the the approach I've outlined in my previous post, a Linkwitz 24db/oct hi-pass cut-off is possible. It will take a little rework of the Energies. They would have to be tuned in a closed box alignment with a Q = .707. Then all that is needed is a 12db/oct active or PLLXO hi-pass before the amp.

In my case, cone breakup distortion sometimes would occur with the mid (actually full-range) until I filtered the lows with the crossover.

I have found my system to respond quite well with the mid-bass aligned with a Q of .5, a f3 of about 65Hz and a PLLXO consisting of a single cap before the amp.

Bfg,

If I were to pick a frequency as you suggest that's well within the flat response for the sub and the Energies, and that's above the f3 of the Energies, should I bother with blocking the ports? It's an easy enough experiment.

I guess what I'm wondering is how much like a ported speaker they'd act like if they're not seeing any input near where the output of the port is significant.

At 24 db/octave there shouldn't be much interaction between the drivers, if at all... I have some opamps kicking around so I'll have place a Digikey order for some caps and resistors to try to make a test circuit.

It is simple enough to block the port and construct a PLLXO for the mid-bass amp. You can do this and everything is completely reversable. You're also going to add a sub (I like stereo subs) anyway. The only thing extra you would need to implement the approach I've described is a variable frequency lo-pass filter (1/2 the active x-over) with the cut-off slope you want.

Once you have everything together, you only need to adjust the lo-pss cut-off frequency to your liking (or flatest FR if you have an RTA).