https://sites.google.com/site/undefinition/classix-ii
Hi, You could do a lot worse than the above, rgds, sreten.
Hi, You could do a lot worse than the above, rgds, sreten.
Erik Squires has this correct. I would add one more requirement: that all 3 measurements performed without mic or speaker physically moved more than 1mm (untouched). All wiring changes done externally to speaker.
You can design a xo for a speaker on paper and it will capture the basic flavor, but unfortunately with XO's the bad/good/great lies in tiny details. Like is the acoustic offset 17mm or 28mm. Like does the placement of my tweeter at position X,Y on baffle width height W H combined with woofer location sum up after baffle step and diffraction to form a dip or a bump around the XO frequency.
When using actual FRD files from your driver on your speaker, you are assured a perfect simulation of the result.
Also note that ES is recommending LR4 - they are good but all XO's have advantages and disadvantages. LR4 may be steep and allow good separation but the phase is not ideal. If phase is important to you (it should be) try lower order first and LR2 is a good compromise. Sometimes asymmetrical slepes work well too. For first project I would go with a 2nd order LR. These slopes are "textbook" slopes. Meaning you want your final acoustical response to follow it. Electrically it may not be a LR2 filter - this is where Jeff Bagby's PCD is very handy. It generates the textbook curve say LR2 at 2kHz and Max 87dB as the target. You overlay your raw FRD and start applying circuit blocks and tweak values until the predicted individual response of the driver is a close match for the target. Do this for both drivers and only then will the XO exhibit that nice properties of the phase and ripple free response promised by its proponents.
It's fine to have a non textbook XO acoustically, just don't expect it to do what was promised by the textbook one. Some of the best XO's are not textbook and asymmetric and eclectic - but that's for advanced designers who know their craft.
So how important is the acoustic offset? Design your XO to your satisfaction without compensating for it. Imagine buying all the coils caps resistors. You are out $75 easy in parts. Build it and assemble and measure it. It may not look like your XO sim. if you simply go in and add or subtract 25mm from the driver offset you get a sense of this effect. Which way seems time consuming? To forgo the offset measurement and now have to redesign and re order new XO parts? It may be salvageable but either way still requires that offset to properly re design.
Last edited:
This was based on the last I knew of what Joseph D'Appolito was recommending for the D'Appolito (WTW) array. I personally haven't made a D'Appolito array or built an LR4 alignment in a crossover but in simulations I've attempted it seems they can be quite problematic to do correctly.
There's an article by Gravesen where he mentions attempting it between woofer and midrange but giving up on it for his center channel. If Gravesen gives up on them so easily it's a good sign it's going to be complicated for most of us.
Best,
Erik
Last edited:
Hrm.... I knew a double-woofer design would require nearly a double sized box, but CAD sims are showing that a WTW setup would be a bit bigger than I was shooting for. I was hoping a WTW setup would be small enough to do since I figured I'd run out of woofer steam before tweeter, but I'll stick to my WT setup for now and can do a WTW setup later if I really want to. I'm still waiting to get them in the mail though before I can make some proper progress... Stay tuned!
My speakers came the other day and look delicious! Although, work and such has prevented me from making any real progress lately. :/ I'm still bouncing back and forth between a WTW tower or WT bookshelf though. It bounces back and forth depending on the day. My main concern with a WT setup is that it will run out of steam on the low end at 'moderate' listening levels in a basement workshop or garage.
Last edited:
- Status
- Not open for further replies.