diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Multi-Way (http://www.diyaudio.com/forums/multi-way/)
-   -   Crossover filter point spread (http://www.diyaudio.com/forums/multi-way/87015-crossover-filter-point-spread.html)

mcdoug88 23rd September 2006 07:59 PM

Crossover filter point spread
 
2 Attachment(s)
Hello,

I am building a pair of DIY speakers using the Dayton RS-180S Aluminum driver for the woofer and the Visaton KE 25 SC ceramic dome for the tweeter. The enclosure is the Dayton .75ft (21L) box which will be used in a vented configuration. I have used Bassbox Pro 3 and X-Over Pro 3 for crossover design. I was able to use provided test data from Dayton/PE for the woofer. I had to manually enter response data based on the curves shown at the Visaton site for the tweeter. I double checked the resultant curve for the tweeter in xover pro against the one at the Visaton site.

After trying many variations of filter orders and crossover points, the best response curves appear to come when I use a third order Butterworth for both the woofer and tweeter. The best "net response" comes when I set the LP filter at 1250 Hz and HP filter at 2250hz giving an effective xover point of about 1250hz according to the resulting graph. In a sense this is good because online research indicates that the RS-180S is best crossed low and keeping the tweeter above 2000 Hz maintains good power handling. BUT, is it OK that the crossover points in X-over Pro are set 1000 HZ apart between woofer and tweeter? Does it happen this way in other speaker designs? Have I possibly forgotten something? I know I have read about designs that mix filter orders but donít ever remember seeing anything about different crossover points.

Iv'e included the response and phase graphs below.

I have included an impedance resonace network for the woofer. The Visaton tweeter has a resonance network already built in that has no specs provided at their site. THe Visaton tweeter also has an essentially flat response down to 1250Hz:

Visaton KE 25 SC

Any advice or constructive reaction is appreciated. This is my first post in this forum so please, be kind. javascript:smilie(':xeye:')
xeye

Thanks, Doug

Salas 23rd September 2006 10:51 PM

Seems you have spot on 90deg difference between drivers. Because 3rd order sums power flat and loudspeaker drivers are voltage controlled devices you got a +3dBSPL bump on axis. So you spaced em apart. Its normal practice. Thing is both drivers gonna contribute weaker of axis in the cross point, and maybe you are going to experince less energy in a normal room. Better design for 20deg horizontal off axis at a 2m distance.

mcdoug88 27th September 2006 05:44 AM

Thanks for the response. Un fortunately. My software does not offer simulation of off axis response. I will design with a slight bump at the xover point so that i will hopefully cet more level off axis response.

Shaun 27th September 2006 07:07 AM

If your software does not model off-axis, you could opt to make a set of off-axis measurements and then substitute these FR curves once you have completed the on-axis design (about now would be appropriate time).

I want to point out that the theoretical summing of different roll-off curves usually assumes that the drivers are acoustically aligned. This can be achieved physically by moving the tweeter back wrt the woofer, or electrically by inserting a delay circuit in line with the tweeter.

Another way of dealing with non-coincident acoustic centres is to misalign the crossover slopes, which effectively adjusts the phase, enabling the system to sum flat at crossover. Slightly offsetting the f3 point can yield similar results. Now, judging by eye, it does not appear that the respective roll-off curves are symmetrical. The tweeter initially has a slower roll-off than the woofer, then at around +-1300Hz it suddenly falls much steeper. It really makes nonsense of the use of an f3 point to describe what the curve is doing.

I don't think this is anything to worry about at this point. I doubt that the horizontal off-axis response will be much affected, but vertical off-axis response is more sensitive to these variances. What remains now is for you to evaluate how good this filter is. The ultimate test of course is to listen. But if you're a tech-head, you can go ahead with the off-axis measurements (recommended; after all the effort you've doubtlesly put in up until now, it pays to know all the facts about your speaker, especially for future application).

Good luck


All times are GMT. The time now is 01:11 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2