PDA

View Full Version : little bit of crossover design information please



req
12-04-2005, 07:58 PM
its my first time looking to build a passive xover for some speakers, right now - my stereo is a little lacking (i dont have a crossover atm). dont ask why, its a long story.

so - i was thinking of slapping some caps\inductors on the lines to get a little better sound out of them - seeing as caps\inductors are pretty cheap these days.

it looks like (from www.bcae1.com (http://www.bcae1.com/passxovr.htm)) that all ya gotta do is figure out where you want your points at, caps = highpass, inductors = lowpass, and use this formula=

1\2*pi*xover point*impedance = capicator in farads (multiply by 1,000,000 to convert to microfarads)

and inductors for highpass, formula=

impedance\2*pi*xover frequency = inductor in henries (multiply by 1000 to convert to millihenries)

then just slap them in series with the speaker (it didnt clarify weather they went in series with the POSATIVE or NEGATIVE lead)?

thats all a passive crossover network entails? aside from modeling how they will respond in the enclosure, and putting in +\- db slope configurations and all that fancy crap - just a cap and a coil in series with the loudspeaker and your done?

just clarifying this. any links to some good passive crossover how-to sites would be appriciated.

VOA_NIGHTMARE
12-04-2005, 08:02 PM
bump so i can hear it when its done

Bartak1
12-04-2005, 08:19 PM
Holy Crap. Req's askin a question. And this whole time I though he knew everything. lol

6 db first order xovers are really simple, just stick a cap on for the high's, and an inductor for the lows. They go on the pos. wire too, series wired in of course.

If you want 12 db/octave, its the same as above plus this-- on the highs' you stick an inductor between the pos. and neg. (if you get what Im saying when I say between-one lead on the pos. other lead on the neg).
And on the low side, you stick a cap between the pos. and neg.

Thats all there is to it.
Im not sure on the formula you need to get the xover point your looking for, I always just find a table that has the values I need for everything. Im pretty sure there is one on on partsexpress.com-but Im not positive (I cant remember if I found one there once, or if I was just looking there once).

I dont know of any how-to sites, sorry.

OlogyAudio
12-04-2005, 08:45 PM
its my first time looking to build a passive xover for some speakers, right now - my stereo is a little lacking (i dont have a crossover atm). dont ask why, its a long story.

so - i was thinking of slapping some caps\inductors on the lines to get a little better sound out of them - seeing as caps\inductors are pretty cheap these days.

it looks like (from www.bcae1.com (http://www.bcae1.com/passxovr.htm)) that all ya gotta do is figure out where you want your points at, caps = highpass, inductors = lowpass, and use this formula=

1\2*pi*xover point*impedance = capicator in farads (multiply by 1,000,000 to convert to microfarads)

and inductors for highpass, formula=

impedance\2*pi*xover frequency = inductor in henries (multiply by 1000 to convert to millihenries)

then just slap them in series with the speaker (it didnt clarify weather they went in series with the POSATIVE or NEGATIVE lead)?

thats all a passive crossover network entails? aside from modeling how they will respond in the enclosure, and putting in +\- db slope configurations and all that fancy crap - just a cap and a coil in series with the loudspeaker and your done?

just clarifying this. any links to some good passive crossover how-to sites would be appriciated.
This is all wrong... if you rly want to design a proper xover no 'textbook' calculator will work... they assume a flat frequency response and resistive impedance load... which isnít true...

www.speakerworkshop.com you will also need the frequency response combiner to extract minimum phase and well playing with things for a year atleast before your making anything better than the component set mfrs out there can in the under $1k/set bracket... If you are willing to invest 4000+ hours do it yourself and learn something :) -- if not go ahead and 1/2 *** it... thats what most mfrs do for everything short of their highest end lines...

BCAE1.com has a B in there for a reason :) It isn't all encompassing and this is one area where it only covers like 0.5% of the passive design process...

req
12-04-2005, 08:58 PM
This is all wrong... if you rly want to design a proper xover no 'textbook' calculator will work... they assume a flat frequency response and resistive impedance load... which isn’t true...

www.speakerworkshop.com you will also need the frequency response combiner to extract minimum phase and well playing with things for a year atleast before your making anything better than the component set mfrs out there can in the under $1k/set bracket... If you are willing to invest 4000+ hours do it yourself and learn something :) -- if not go ahead and 1/2 *** it... thats what most mfrs do for everything short of their highest end lines...

BCAE1.com has a B in there for a reason :) It isn't all encompassing and this is one area where it only covers like 0.5% of the passive design process...


like i said in my first post. thats basically it. nothing facny - but it will cross over a speaker at around the desired frequency.

i do understand that it wont be exact, nor will it be close (impedance rise, ect) but it will work for somthing that you just want to build a crossover for so that it does not sound like COMPLETE ***.

to be blunt - this is an INEXPENSIVE step up from nothing - correct?

:)


PS the speakerworkshop program, sweet link and great download. but what do i got to do to find a tutorial on it? does not seem like there are any *how-to's* on that website.

and Bartak1-

I have never claimed to know everything, I meerly choose to talk about the things I have knowledge about.

thch
12-04-2005, 09:04 PM
its my first time looking to build a passive xover for some speakers, right now - my stereo is a little lacking (i dont have a crossover atm). dont ask why, its a long story.

so - i was thinking of slapping some caps\inductors on the lines to get a little better sound out of them - seeing as caps\inductors are pretty cheap these days.

it looks like (from www.bcae1.com (http://www.bcae1.com/passxovr.htm)) that all ya gotta do is figure out where you want your points at, caps = highpass, inductors = lowpass, and use this formula=

1\2*pi*xover point*impedance = capicator in farads (multiply by 1,000,000 to convert to microfarads)

and inductors for highpass, formula=

impedance\2*pi*xover frequency = inductor in henries (multiply by 1000 to convert to millihenries)

then just slap them in series with the speaker (it didnt clarify weather they went in series with the POSATIVE or NEGATIVE lead)?

doesn't matter. components will be in series. at least for a simple filter.



thats all a passive crossover network entails? aside from modeling how they will respond in the enclosure, and putting in +\- db slope configurations and all that fancy crap - just a cap and a coil in series with the loudspeaker and your done?

just clarifying this. any links to some good passive crossover how-to sites would be appriciated.
the filter you have described is 6db/oct optimally. The speaker may not be a resistive load which further alters the response. this may or may not matter. for instance if the speaker is no signifigantly reactive near the crossover point there shouldn't be issues. higher order filters (higher slopes) will require more components. it can get fairly expensive when you need 8 inductors and 8 capacitors. (stereo, 24db/oct slopes).

there are a lot of design issues that are interesting to look at if you are interested. most people aren't and go for simple filters like 6db/oct and 12db/oct then use an EQ to fix any errors. high order filters require higher Q stages which require more precise values for capacitors and high "Quality" inductors.

try sound.westhost.com 's articles section for some information.

req
12-04-2005, 09:10 PM
thanks for the help so far guys, i will most definatly build a simple 3db\6db crossover to get started with a little desktop speaker project for my computer. seems easy\inexpensive enough :)

i know it is not a great way to do it, nor is it the most optimal (not even close) but again, its a step up from nothing, and i figure its better than two tiny little aura 2" fullrange speakers running mono off a healthy 2w :)

JimJ
12-04-2005, 09:14 PM
2W is all you need :D

With slightly different speakers than 2" full ranges, though ;) Maybe 8" full ranges...

Bartak1
12-04-2005, 09:15 PM
and Bartak1-

I have never claimed to know everything, I meerly choose to talk about the things I have knowledge about.

I know, and I didnt mean for it to sound that way. I simply said that because I see you making a lot of good informative post in which you really seem to know what you are talking about.:) It wasnt meant as an insult if you were thinking that, I was just pretty much saying I think you have a pretty great deal of knowledge.

req
12-04-2005, 09:23 PM
I know, and I didnt mean for it to sound that way.

lol, didnt take it offensivly bud :laugh:

but thats why i seem to know what im talking about, i try to help people on topics that i know more about, and direct them to information\people that know about the topics that i know less about, or i just dont reply at all. thats the way it should work for everyone :(

thats why i dont like the lounge that much.


2W is all you need :D

With slightly different speakers than 2" full ranges, though ;) Maybe 8" full ranges...

:laugh: dont talk to me you jerk :up2somet:

so really, thats all i need to know about this? sweet. ill try to learn more as i go - but for now this should work :veryhapp:

thch
12-05-2005, 11:36 AM
yep. Fc = 1/(2*pi*T). T = RC = L/R.

so L = (1/2pi) (R/F)
and C = (1/2pi)(1/RF)

Inductors (L) for lowpass, Capacitors (C) for highpass. its in series, so wire it to either side of the speakers. if you want to get fancy, there are "series tuned filters."

http://sound.westhost.com/parallel-series.htm

and if your interested, here is an example filter with "stopband zeros". typically called "elliptic" but realistically ellptic refers only to filters with passband and stopband ripple. elliptics have the sharpest slope. http://sound.westhost.com/articles/ntm-xover.htm

edit -- stopband zeros...

ngsm13
12-05-2005, 03:54 PM
Do it right n00bx.

NG

req
12-05-2005, 05:49 PM
wow ng, you really never have any input to my threads eh? lol

thanks thch, but alot of the crossover lingo i dont understand :)