crossover design for midrange in a three way (1 Viewer)

[David Giesbrecht]

When designing a bandpass xover for a midrange can you just have a standard high pass and low pass filter connected back to back or is ther someting I am missing? I am planning to use a peerless HDS 205 woofer a peerless HDS 134 as a mid and a Focal TC 120 TD5 tweeter. Xover points will be 500 hz and 3khz.

THX!

 

[Patrick Sun]

Yup, that's the idea.

 

[Dan Wesnor]

Depending on what you mean by "standard high-pass and low-pass", it may work, but it will sound like $#!+.

First of all, textbook filters don't work all that well to begin with. Second, the high-pass and low-pass will interact. The least problem this causes is bandpass gain, the worst irregularities caused by the impedance of one filter driving another.

But, if you're designing a 3-way using crossover simulation software and not relying on textbook formulas, yeah, you just put in the 2 filters back-to-back and optimize as usual. You may need to use a conjugate notch filter to knock down the mid's and/or woofer's impedance spike.

 

[David Giesbrecht]

Is there any software I can download for doing this? I am using PE box now and it's not any good for designing xovers for speakers better than 2 ways.

 

[Patrick Sun]

http://www.speakerworkshop.com

If you need to create the FR (Frequency Response) and ZMA (Impedance) graphs from existing manufacturer plots, you should look for a program called SPL Trace. This is in case you have no way of measuring your drivers while mouted in the front baffle of the speaker enclosure (this is the preferred way to get driver measurements, and then use this data in modelling the response while using speaker crossover design software).

So if you have no means of measuring the drivers, use the graphs on speaker manufacturers' websites (or on the PE website).

The learning curve for Speaker Workshop can be steep depending on your familiarity with all of the variables.

I did my Center channel speaker XO design by using Speaker Workshop just as an experiment. But I was lucky in being able to have my drivers measured correctly (installed in the front baffle).

 

[DavidES]

Thanks Patrick for sharing that link. I was looking for something exactly like that. Appreciate it.

 

[Dan Wesnor]

Be careful, as the path you choose at this point can lead to a lifetime of addiction.

 

[David Giesbrecht]

I'm already addicted since age 14 until now at 17 and beyond!

(I think that made sense?)

 

[David Giesbrecht]

If you wanted to lower the impedence of a speaker to increase its output could you just connect a resistor in parralel before the driver?

 

[Patrick Sun]

You're just burning power across the resistor.

 

[Dan Wesnor]

Did you mean parallel *to* the driver or parallel *before* the driver.

Impedance is not that simple. If you lower the impedance of the driver by connecting a resistor parallel to it, the changes you need to make to the XO to keep the transfer function the same might just knock the impedance of the overall circuit right back up again.

And why would you want to decrease the impedance?

 

[David Giesbrecht]

If you put the resistor in before the XO you wouldn't have to change the XO at all would you?

And why would you want to decrease the impedance?

To correct for different impedences in a speaker that uses 2 woofers and only 1 mid or tweeter.

 

[Patrick Sun]

You should only use resistors to attenuate the output, or use them in a zobel (impedance compensation), or for resonance traps and notch filters.

You can't use resistors to increase the output from a driver (though you can increase the amount of current the speaker network draws from the amp by creating a low impedance speaker by introducing resistors in parallel with the crossover filters), but the resistor(s) simply turns that added current into heat.

 

[David Giesbrecht]

Is there any type of circut you can build to increase the output of a driver not using resistors, there must be because I've seen lots of MTM kits wich use two 8 ohm midwoofs in parralel but only one 8 ohm tweeter and that have a final impedence of 4 ohms. Someone tell me if I sound like a total moron right now.

 

[Patrick Sun]

The nominal impedance of a MTM speaker with two 8 ohm midwoofers connected in parallel will almost always be around 4 ohms (if the crossover is done competently). Don't let the 8 ohm tweeter confuse the issue. The midwoofers will be drawing more power from the amp, so careful attention be be paid to make sure the impedance profile of the overall speaker in the low frequency region doesn't dip down dangerously under, say, 3 ohms, which would produce some clipping of the amp if played loud enough (if the amp isn't designed for low impedance loads). This is another reason why you will see designers use 16 ohm midwoofers in parallel in an effort to produce a MTM speaker that will present a more benign nominal load in the 8 ohm range, if the amp being used isn't designed for low impedance loads.

You get a 6dB increase in output by using 2 midwoofers wired in parallel, but some of that gets lost in the baffle step (where the speaker's output transistions from radiating all around the speaker in 4pi space in low frequencies up to around 500Hz to only radiating in 2 pi space, which is mainly in front of the speaker). To get a smooth sounding speaker, this baffle step issue needs to be dealt with, or you wind up with a "thin" sounding speaker. This is why MTM designs became popular, though they do present challenges in getting the lobing patterns correct.

BTW, you should probably check out "The Loudspeaker Design Cookbook" by Vance Dickason (Amazon, Barnes & Nobles should have it in stock) because it'll clear up a lot of these types of questions, but go into greater depth as well, and provide nice graphs as well.

 

[DavidES]

For some visualization help on those resistor questions above. Look at both L/R and center crossover schematics here. They do explain why in the text.

Is there any type of circut you can build to increase the output of a driver not using resistors

Not meaning to sound flippant if it comes across that way, David, but an amplifier uses resistor networks that include transistors to develop the larger signal that speakers need. Transistors are "resistor networks" that are manipulating, by design, voltage drops in order to manipulate current flows.
If I'm going over your head, I apologize. It hard to explain correctly in layman's terms.

A passive bandpass filter can increase output of midrange driver because of the interaction with the other 2 crossovers over a relatively narrow octave band, hence bandpass gain. That's why higher order passive 3-ways are very difficult to design accurately for summing and minumum phase shift. I think active crossovers like Linkwitz-Riley can do the above mentioned easily but I'm not familiar with them yet.