Cabinet Passive to Active Crossover Conversion

[yrrapt]

Hi,

I'm going to be making my first speakers soon and will be making the Amigas from undefinition. I'm planning on running them actively with crossovers either in the digital domain with an FPGA or in the analogue domain with 4th order sallen-key filters.

Would I be correct in saying that I can just calculate the crossover frequencies from the passive crossovers on undefintions site and implement my filters to match the response and suitable tweeter attenuation? I'm an electronics engineering student so have no problems with filter design I just need to make sure I'm not making any incorrect assumptions before I spend money.

Thanks in advance for any help.

Tom

 

[sreten]

Hi,

No. For an active x/o you need the same transfer functions as passive.

Basically a typical active 2-way is a waste of money. An active 2 way
should be bass and mid/treble or a 3 way with passive mid/treble.

rgds, sreten.

Unless you know shedloads about speaker design you can't simply
change a design to 4th order active and what that implies acoustic.

http://sites.google.com/site/undefinition/diy
(see if nothing else, the excellent FAQs)
http://techtalk.parts-express.com/showthread.php?t=219617
http://www.zaphaudio.com
http://www.zaphaudio.com/ZA5/
http://audio.claub.net/Simple Loudspeaker Design ver2.pdf
http://www.rjbaudio.com/Audiofiles/FRDtools.html
http://web.archive.org/web/20090902124715/http://geocities.com/woove99/Spkrbldg/DesigningXO.htm
http://www.rjbaudio.com/
http://web.archive.org/web/20090902202231/http://geocities.com/woove99/Spkrbldg/
http://speakerdesignworks.com/
http://www.htguide.com/forum/showthread.php4?t=28655
http://www.deadwaxcafe.com/vzone/david/david.htm
http://www.troelsgravesen.dk/Diy_Loudspeaker_Projects.htm
http://www.humblehomemadehifi.com/download.html
http://www.quarter-wave.com/
http://www.frugal-horn.com/
http://www.linkwitzlab.com/
http://www.musicanddesign.com/

Great free SPICE Emulator : http://focus.ti.com/docs/toolsw/folders/print/tina-ti.html

 

[yrrapt]

Thanks for the response.

That makes sense, a bad oversight on my part.

So entering a theoretical world, if I mimicked the passive crossover response with linear phase FIR filters in an FPGA the principle would work? Further, assuming that I already have an FPGA at some point in my chain and the cost of implementing digital crossover's is nothing (let's exclude the added cost of extra DACs, amps etc.), how much of an improvement would you expect, if any?

Cheers,
Tom

 

[kalinowski]

I wouldn't say emulating analog crossover response using digital filters is a good idea. Thing is you can do so much more in digital domain than using simple passive elements. What's more, once you set up your hardware for a digital crossover making changes costs you nothing.
If I were you I'd check out crossover points for your construction and then design a filter with a linear phase for example. You're a step ahead analog stuff already and now fun begins. Try some room correction etc, you should have lots of fun and a good sounding speaker at the end.

 

[john k...]

Thanks for the response.

That makes sense, a bad oversight on my part.

So entering a theoretical world, if I mimicked the passive crossover response with linear phase FIR filters in an FPGA the principle would work? Further, assuming that I already have an FPGA at some point in my chain and the cost of implementing digital crossover's is nothing (let's exclude the added cost of extra DACs, amps etc.), how much of an improvement would you expect, if any?

Cheers,
Tom

The HP and LP filters of that design are minimum phase. You must digitally emulate the minimum phase response of those filters digitally. Itdoesn't matter whether you do this with FIR or IIR filters but if you use FIR filters they should not be linear phase. (FIR does not imply linear phase.) As Sreten said, you must emulate the transfer functions of the passive filters. That means both amplitude and phase.

Additionally, using linear phase filters doesn't really get you much since you will stillhave to deal with the driver's phase. So just applying linear phase filters to a speaker system doesn't mean you end up with a linear phase system.