Calling Johan - Seas L26 + L18/L15 + 27TBFCG

[Dave Bullet]

Hi Rojah / Johan,

Thought I'd start a new thread for similar projects we are working on rather than threadjack this one:
Need help selecting a driver for my 3 way

I too am looking at an all metal 3 way. Using Seas drivers, I chose the L26 and 27TBFCG. I intend to go passive and have also got an L15 for the upper bass / midrange.

I know you are planning a L26/L18/27TBFCG, basing the mid/tweet crossover on Zaph's successful L18/27TBFCG.

I'm not sure I've picked a good mid for a true 3 way. The L15 is a bass/mid and the tweeter crossover point is limited to about 2KHz (or 2.5KHz depending on who you read) due to breakup (at about 8KHz).

I'll post some baffle simulations here and crossover ideas. Would be good if you don't mind sharing your approach. We might even have a couple of metal 3 ways for people to choose at the end of it (which seems to be lacking in the DIY area - at least ones that use a 10" woofer).

Cheers,
DAvid.

[Dave Bullet]

It's a slow moving project...

Here's a candidate crossover simulation and modeled on-axis (and reverse null) response.

Cavaets:
These are re-used driver measurements with simulated baffle diffraction added in
HBT generated phase using -1.018 and -7.677 inches for L15 and L26 offset respectively
SpeakerWorkshop used for crossover modeling

First the crossover:

[Dave Bullet]

I've aimed for 4th order acoustic slopes. Reason being - kill the resonances of the metal cone drivers - 4KHz and approx 8Khz for the L26 and L15 (respectively). Crossover points are 240Hz and 2400Hz.

I'd like to cross the L15 lower to the tweeter (say 2KHz) and eliminate as much 2nd / 3rd order distortion as possible.

[Dave Bullet]

Here is the modeled reverse null (L15 out reverse phase) - normally all 3 drivers are in phase (which seems odd to me).

[Dave Bullet]

And finally impedance. Minimums are around 3.1 ohms

[sreten]

Originally posted by Dave Bullet
- normally all 3 drivers are in phase (which seems odd to me).

Hi, perfectly normal, not odd, for 4th order acoustic slopes, /sreten.

[Dave Bullet]

Thanks sreten. Good to know. I just haven't seen a 3-way where all 3 drivers are wired with the same polarity so assumed I must have done something wrong (well, I probably have somewhere!).

Although 14 components aren't bad for a passive 3 way crossover (esp. considering metal drivers are used and cone breakups must be dealt with), it will be expensive. I will try and rationalise it down. I'll also include a padding option of the upper mid / treble response for bright rooms.

baffle simulation to be posted...

[simon5]

Let say alot of cheap 3 ways are probably 2nd order acoustic slope... so that means 2nd driver out of phase. Also alot of exotic designs are built with as few crossover components as possible, because for that kind of designers, less is more.

2nd order = 180° out of phase, so reverse polarity on the center driver and it's back in phase.

4th order = 360° out of phase, but 360° = 0°, it's a full turn, so you're back in phase, all drivers need to have same polarity.

[sreten]

Hi,

I suppose that for most 3 ways 4th order acoustic bass to mid
is not an option as it uses too many parts, therefore bass/mid
are usually out of phase.
Tweeter may be in / out with the mid depending on c/o.

/sreten.

[Dave Bullet]

I suppose that for most 3 ways 4th order acoustic bass to mid

I really struggled with getting good phase tracking with 2nd order acoustic between woofer / mid (a bit beyond my design capabilities... ), so resorted to 4th order.

I will try again since I would like to minimise the parts count.

I wasn't happy with a 2.4KHz xo point with the L15. The breakup is 8KHz and I wanted to be 2 octaves clear of it.

I've modified the crossover and got a smoother response that has even better phase integration and brought the tweeter level down. However - I feel it might be a little too low now (1725Hz). Although 4th order acoustic has been maintained on the tweeter and it is only 5dB higher through the mid stopband than the previous version.

Thanks,
David.