Needed: 2 way, phase linear crossover design references

[J.R.Freeman]

Hi there,

What resources are available (free on the web would be ideal!) for calculating and approximating crossover designs? The rough requirements would be a fairly steep (12 dB / oct) 2 way design, that can correct for phase distortions to some extent.

In addition to the crossover, how does one go about calculating a baffle-step compensation network?

Thank you,

Jim

[sreten]

http://www.rjbaudio.com/Audiofiles/FRDtools.html
http://www.geocities.com/woove99/Spk...esigningXO.htm

/sreten.


http://www.zaphaudio.com/
http://www.rjbaudio.com/projects.html
http://www.troelsgravesen.dk/Diy_Lou...r_Projects.htm

[djarchow]

Check out John Kreskovsky's site. Lots of transient perfect papers.

Here is a caveat though. I have designed a transient perfect 3 way speaker using John's work as it's basis. Designing a transient perfect speaker is challenging. It usually requires fairly extensive XO design skills. You will also need the tools and knowledge to be able to take accurate measurements of your drivers in your cabinets for frequency response/phase and impedance. Not to mention drivers which lend themselves to this type of design.

Your question about how to calculate a baffle step in a XO imples that maybe you are new to speaker design. If so a transient perfect or even semi transient perfect speaker will be a tall order.

Regards,

Dennis

[PeteMcK]

Rod's page has a couple of relevant articles under 'crossovers':
http://sound.westhost.com/articles.htm

[J.R.Freeman]

Thank you all for the references and suggestions.

It is true, I am rather new to serious design, and I have much to learn. Djarchow, a transient perfect design sounds interesting, but is likely beyond my means. Also, for my purposes, I can live with some phase shift and less than ideal transfer function. This is fun, I'm learning already! I thought 12 dB / octave was rather steep, but it seems this is only a 2nd order acoustic slope.

My goal is a moderately accurate 2 way, for modest SPL and near-field, smallish room duty. Zaph has a nice looking design in his SR71 (Lockheed Martin might have something to say about that name!). Hmm, that would fit the bill nicely.

However trying something on my own might make for an interesting failure at the very least. Though new to advanced loudspeaker concepts, I have a background in electronics, namely analog circuit analysis. I have enough test equipment and knowledge for crude frequency response measurements, and impedance measurements, but not distortion.

I have been eye-balling the Seas P18RNXP. Hmm, it is pretty! Zaph's impression of this driver seems favorable - and he makes mention of distortion characteristics in a mostly qualitative way. Its poly cone might not be as articulate as some, but it looks to be free of any serious breakup, which would simplify my XO (and thus reduce my opportunities for screw-ups!). 30 degrees off-axis still looks nice beyond 2K, suggesting a reasonable XO point, but Zaph mentions that HD levels increase beyond 800 Hz.

Anyway, this is mostly thinking out loud. It's all very interesting - thanks again for the links and such. I'm itching to start playing with some XO design!

Jim

[Andersonix]

Since your background is electronics, it would seem natural to go active crossover? Presumably you've seen Linkwitz' website, linkwitzlab.com?

[Dr. Ram]

Originally posted by J.R.Freeman
for my purposes, I can live with some phase shift and less than ideal transfer function.

I thought 12 dB / octave was rather steep, but it seems this is only a 2nd order acoustic slope.

I have enough test equipment and knowledge for crude frequency response measurements, and impedance measurements, but not distortion.
Jim

(i) Given your context that you have laid out so well, I would say phase distortion should be low on your list of priorities.
(ii) 12dB/oct would be a worry if you were concerned about intermodulation distortion. In your context (smallish room duty) you are looking at using barely a couple of watts; at those power levels, intermodulation distortion is not a concern and 12dB/oct is ample.
(iii) If you can take frequency response and impedance measurements you are blessed! Make sure you design your crossover using the *in-enclosure* frequency response and impedance measurements for your drivers.

-Ram