Crossover Design

[maldecido33 10+ year member]

I've got decent knowledge of crossovers; what the components in them do, how they work, etc. I'm looking for some good resources to design my own fairly basic active crossovers. They're for a simple 2-way bookshelf louderspeaker.

What sort of resources are out there?

Click to read more...

 

[thylantyr 10+ year member]

Might be cheaper to buy an active crossover.

DIY link

http://sound.westhost.com/projects-3.htm

 

[maldecido33 10+ year member]

I'm sure it'd be cheaper, but I'm looking to actually learn the design process. I can design a 2nd-order LR, Butterworth, or Bessel crossover, but the only factor I'm taking into consideration is a given x-over point. I'd like to learn the processess associated with selecting crossover frequencies based upon various parameter's of the driver.

 

[thylantyr 10+ year member]

I'd like to learn the processess associated with selecting crossover frequencies based upon various parameter's of the driver.

I see this as two paths.

The hard path - Electrical engineering

The easy path - Variable crossover, common sense, listening test

You can use your pencil, paper, calculator to figure out the technical aspect

of crossover design factoring in tons of variables, this takes time. People

are moving away from this methodology because there is speaker software

on the market that does this, but you have to learn the software which takes

time.

Or

You can build a loudspeaker using a flexible active crossover, apply some common

sense when you make adjustments, do some listening tests using a test box

and zero in on what settings gave you the best sound. Then you have choices.

1. Continue to use that 'development' processor in your system

2. Take those parameters and make a passive crossover equivalent.

3. Take those parameters and make a active crossover equivalent.

Note, if you built an active crossover, if you choose your components well

you could make your active crossover have variable frequency selection by

using precision ganged pots instead of fixed resistors. Another method is

to use fixed resistors to select crossover frequencies, but you can use

sockets and headers to make resistor modules to change the frequency,

you see this in car audio.

Which path do you take? Depends on your goals. If you want to dazzle people

with a crossover engineering marvel to show off, then make a technically

correct design, but there is no guarantee that it offers better SQ.

If you want to zero in on best SQ, then do the simple approach and don't

worry about if it's technically perfect. All that matter is that you liked the sound.

Or do a combination of both.

For car audio because of the environment, the tweaking method is ideal. You want

to be able to change variables.

For home audio you don't have installation limitations so you can choose any method,

but I still prefer the 'listening to the speaker' method and turning knobs to get the best sound.

The simple method, IMO is superior because it tackles the root cause, what gives me the best perceived sound.

But the simple method also assumes that you used good drivers and you executed a good loudspeaker recipe,

nothing dumb.

 

[maldecido33 10+ year member]

I see this as two paths.
The hard path - Electrical engineering

The easy path - Variable crossover, common sense, listening test

You can use your pencil, paper, calculator to figure out the technical aspect

of crossover design factoring in tons of variables, this takes time. People

are moving away from this methodology because there is speaker software

on the market that does this, but you have to learn the software which takes

time.

Or

You can build a loudspeaker using a flexible active crossover, apply some common

sense when you make adjustments, do some listening tests using a test box

and zero in on what settings gave you the best sound. Then you have choices.

1. Continue to use that 'development' processor in your system

2. Take those parameters and make a passive crossover equivalent.

3. Take those parameters and make a active crossover equivalent.

Note, if you built an active crossover, if you choose your components well

you could make your active crossover have variable frequency selection by

using precision ganged pots instead of fixed resistors. Another method is

to use fixed resistors to select crossover frequencies, but you can use

sockets and headers to make resistor modules to change the frequency,

you see this in car audio.

Which path do you take? Depends on your goals. If you want to dazzle people

with a crossover engineering marvel to show off, then make a technically

correct design, but there is no guarantee that it offers better SQ.

If you want to zero in on best SQ, then do the simple approach and don't

worry about if it's technically perfect. All that matter is that you liked the sound.

Or do a combination of both.

For car audio because of the environment, the tweaking method is ideal. You want

to be able to change variables.

For home audio you don't have installation limitations so you can choose any method,

but I still prefer the 'listening to the speaker' method and turning knobs to get the best sound.

The simple method, IMO is superior because it tackles the root cause, what gives me the best perceived sound.

But the simple method also assumes that you used good drivers and you executed a good loudspeaker recipe,

nothing dumb.

I'm an electrical engineering major, but with only two semester's under my belt. Only taken an intro EE class, computer-science for EE, and EE Digital Design. Haven't gotten to the point where'd I'll learn this stuff in school, I just wanted to learn it on my own first. I've got some good budget drivers and a x-over that game with the partsexpress showcase for these drivers, and I'm messing around with tweaking stuff on the 2nd-order Linkwitz-Riley with Zobel Filter and L-Pad. After I get this down I'm gonna mess with 3rd and 4th order x-overs.