Selecting Crossover Frequencies for 3-way build

[Passinwind]

Originally Posted by Nate74

The active crossover I'm familiar with was in a "big" PA system. It was fed with the +4dB signal then sent into the various power amps. Would a similar setup be needed here? With a separate amp for each speaker?

Yep. I use a speaker management system (the Driverack is one well known example) to do cab development, it has some extra useful features compared to more basic active crossovers.

One thing to note though is that passive crossovers have all kinds of interactions that can actually work in your favor sometimes (and quite the opposite many times too). You need a very solid understanding of this to model with good correlation using an active system.

You might want to download Passive Crossover Designer and try modeling something you already have heard/understand/have access to good specs for.

[Nate74]

On a side note, before work today I tried running the 12" as it originally had been with everything below 3.5kHz going to it and then used the only half of the second crossover with the 6.5", basically using it as a HPF at 800Hz. I think the cabinet sounds better so I'm now thinking I might try my hand at building a HPF. It's just once cap right?

[will33]

The passive crossover IS the speaker level, "after the poweramp" crossover. Usually built inside the speaker box, as the good ones are designed for the specific drivers being used in the specific box they're in.

A generic, "off the shelf" crossover geared for 8ohms for example, can work "alright", IF, the drivers impedance at the corner frequencies are at least close to 8ihms, which most are not, especiallythe filter that rolls the topend off the mid driver. Speakers impedance rises as frequency rises....a lot.

I use one of those in my PA speakers, and they sound "pretty decent", but not as smooth as they could with a specific, purpose built crossover for them. I've also spent some time swapping around drivers, finding ones that "play nice" with the existing crossovers. That sort of going about it backwards.


The active crossover as you say works at line level. Preamp---->crossover---->, then the various to different dedicated amplifiers driving each part of the system.



You mentioned earlier wanting to hear what your highs sound like at 1khz-4khz, if I understood correctly. Careful with that. Most small tweeters like that can't handle being passed lower than around 3khz, and that's with a fairly steep, 18db/octave slope.

[Passinwind]

Originally Posted by Nate74

On a side note, before work today I tried running the 12" as it originally had been with everything below 3.5kHz going to it and then used the only half of the second crossover with the 6.5", basically using it as a HPF at 800Hz. I think the cabinet sounds better so I'm now thinking I might try my hand at building a HPF. It's just one cap right?

One cap makes a first order filter -- 6dB per octave. Look at some better options for your purposes here. I generally use at least 3rd order for a 6.5" driver in a bass instrument system.

[will33]

Originally Posted by Nate74

On a side note, before work today I tried running the 12" as it originally had been with everything below 3.5kHz going to it and then used the only half of the second crossover with the 6.5", basically using it as a HPF at 800Hz. I think the cabinet sounds better so I'm now thinking I might try my hand at building a HPF. It's just once cap right?

There are some HF components that can handle being driven much lower, but those aren't used in most bass cabs. Search up Dukes "Thunderchild" line of cabs to see examples of those.


A single cap is a 1st order filter, meaning it only filters out the lows at a rate of 6 decibels per octave. That doesn't offer much protection, is an inexpensive way of doing it, should be set no lower than 5khz, and can still sound harsh and possibly blow under power.

Yours likely uses at least a 2nd order (12db/oct.) or even better, 3rd order (18db/oct.). That would be cap--->inductor shunt--->cap, to make up the filter.


Some basic, beginner reading in the matter here, although certainly not the whole enchilada. http://www.diyaudioandvideo.com/Tutorial/Crossover/

The different slopes will also change the phase relationship between the drivers, and must also be taken into account. Nothing beyond a simple, one part, first order filter is easy.

[will33]

That link I posted there is an alright, beginer level understand of things, bit doesn't go into specific enough detail as far as some driver specific uses and "tricks" you can use. Some of these pros here could probably point you to other places. Also some things that might be considered "no-no's" in lower power, hifi stereo type stuff can work to your advantage in high power, live sound reinforcement type stuff.

You basically never stop learning and I'm way at the beginning of the curve there, but I learn from these guys all the time. Just search up some of their writings in other relavent threads, etc.

[Nate74]

This his all been very informative. I'm really thinking this could get me hooked...

[Passinwind]

Originally Posted by will33

That link I posted there is an alright, beginer level understand of things, bit doesn't go into specific enough detail as far as some driver specific uses and "tricks" you can use. Some of these pros here could probably point you to other places.

The links on greenboy's website go very deep if you take the time to follow them.

[will33]

Originally Posted by Passinwind

The links on greenboy's website go very deep if you take the time to follow them.

I've read through those, great stuff. And I keep going back to them to look stuff up, it's just too much information to retain all at once.

Setting up a different computer here with those modelers and looking for a decent, affordable, measurement mic/RTA system, through links you've posted before, as well as Bill F., John K, fdeck, etc. There's some measurement mic system Bill linked to in some thread by the guy who was doing transmission line/hybrid plans. Pretty sure I saved it in here somewhere. Basically a measurement mic driven by a usb adaptor.

[will33]

It's the stuff linked to in post #121 of this thread. Transmission line bass cabs

Thoughts?

Looking for affordable but effective here.

[Nate74]

Ok, next question. If I want to calculate a frequency of say a second order butterworth crossover using the L and C values, what's the equation?

[Passinwind]

Originally Posted by will33

It's the stuff linked to in post #121 of this thread. Transmission line bass cabs

Thoughts?

Looking for affordable but effective here.

Holme Impulse is great IME, and that hardware looks fine for your purposes.

[will33]

Originally Posted by Passinwind

Holme Impulse is great IME, and that hardware looks fine for your purposes.

Cool, thanks. Can get going the right way here without having to concieve a firstborn so I can sell them off.

[Passinwind]

Originally Posted by Nate74

Ok, next question. If I want to calculate a frequency of say a second order butterworth crossover using the L and C values, what's the equation?

You can find the math on Rod Elliot's site, IIRC.

Sadly, it's not that simple. The easiest thing to do is probably to use the calculators I linked to, plug in frequencies, and see when L and C get close to what you have. But since driver impedance changes with frequency, things get complex really fast. Passive Crossover Designer is freeware that goes several layers deep and is worth learning if you are getting the bug.

I use LTSpice a lot, which is also freeware. If I am confronted with an existing crossover design I just plug in the values and get a ballpark electrical response curve. I have some acoustic response Spice sims too, but that is the deep end of the pool for most of us!

You'll quickly find out that filter alignments outside of the classical ones tend to work best in many cases. Greenboy's Cheap But Good designs are worth studying to see how that works.

[Nate74]

I think just for my understanding, leaving the speaker impedance fixed will suffice. I'm just noting that capacitors and inductors come in nice round number ratings, so I'm curious as to what frequency they'd calculate at a fixed impedance.

[DukeLeJeune]

Originally Posted by Nate74

Do you keep a bunch of inductors and capacitors on hand?

I have two boxes of inductors, two boxes of capacitors, and three boxes of resistors, in assorted values. And I can combine them (or wind them down in the case of inductors) to get just about any value I need to try.

I normally build a "breadboard" when working on a crossover design, using bus-bar type connectors at each junction, all on a big piece of plywood, so that I can shuffle components in and out without soldering them or over-twisting the wires. You can just use jumper wires with alligator clips for simple crossovers, but for anything complex that approach quickly turns into a bowl of spaghetti.

Originally Posted by Nate74

On a side note, before work today I tried running the 12" as it originally had been with everything below 3.5kHz going to it and then used the only half of the second crossover with the 6.5", basically using it as a HPF at 800Hz. I think the cabinet sounds better so I'm now thinking I might try my hand at building a HPF. It's just once cap right?

Two possible issues. First, you may be getting an undesirable impedance dip by not low-pass-filtering the 12" woofer and then running the midrange in parallel with it from 800 Hz up. Second, depending on what the impedance curve of your midrange looks like in its dedicated sub-enclosure, a simple series capacitor may not adequately protect it.

Originally Posted by Nate74

Ok, next question. If I want to calculate a frequency of say a second order butterworth crossover using the L and C values, what's the equation?

You can find the equations and an online calculator here:

http://www.ajdesigner.com/crossover/crossoversecond.php

You may find that the inductor value matches one frequency, and the capacitor value matches another. In that case, try changing the impedance assumed and see if you can get them closer. Or just assume that the actual electrical crossover frequency is in the middle.

Be aware that textbook filters only work as advertised into a relativly smooth impedance curve (which is where Zobels and resonance peak filters come in handy, if you don't mind the added complexity). As a first approximation, look at the published impedance curve and use the impedance of the driver at the crossover frequency. Also, remember that this approach doesn't take into account the native frequency response of the driver itself (in both SPL and phase), and the net result will be the combination of the electrical filter and the driver's native response. This doesn't mean you can't get decent results with some trial and error, but don't forget to try reversing the polarity of the high-passed driver to see if that gives you a better blend in the crossover region, as there are so may real-world variables at play that it's a roll of the dice.

If you're looking for a relatively affordable yet highly capable measurement/modelling suite, this one looks good to me:

http://www.woofertester.com/stproduct.htm

At a minimum, I recommend the latest edition of The Loudspeaker Cookbook, by Vance Dickason.

If you get into using a modelling program, something you might want to do is this: Model the impedance curve of each filter (highpass, lowpass, bandpass) into an open circuit. This is what the crossover will see if the driver blows. With crossovers more complex than first order, in some cases you can end up with a highly counter-intuitive impedance dip that is virtually a dead short.

If you're still excited rather than discouraged, you're already addicted.

[Nate74]

Officially into the technical area known as "over my head." But I do really appreciate the info and suggestions.

[Bassmec]

Originally Posted by Nate74

Officially into the technical area known as "over my head." But I do really appreciate the info and suggestions.

There is a heck of a lot of info to take in vis a vis passive crossovers.
It's often wise to let others do the testing and only attempt to build stuff based on years of experience and testing by experts, it's like various old
passive studio monitors, there is all sorts of aftermarket hopped up crossovers if for nothing else than including modern poly caps in place of the old bipolar lytics etc.
Some classic designs like transmission lines have a huge range of aftermarket crossover attempts even without changing the cab or original speakers in any way.
Has anybody mentioned crossover phase and time alignment yet?.
Cos that's a whole extra subject that even some well established designs
haven't been able to perfectly sort out.
However sharp you decide the passive design slope should be in the crossover area, it's going to have some phase angle differences between drivers.
The best thing about modern active crossovers is the 24 dB per octave filters with micro delays to correct these phase problems.
Class d amps are cheap and light these days, if you want to have the best experience with multiple loudspeakers in different bands filling your frequency response, active is your very best bet.

[DukeLeJeune]

Originally Posted by Bassmec

Has anybody mentioned crossover phase and time alignment yet?.
Cos that's a whole extra subject that even some well established designs
haven't been able to perfectly sort out.

The research I've read on the subject indicates that the ear can hear the difference between two different crossovers that have different phase characteristics, but even if one is technically "correct" and the other is not, listeners generally don't have a strong preference and do not tend to consistently pick the "correct" version over the other one.

On the other hand, the ear is quite sensitive to the spectral balance (frequency response), so that is where most designers concentrate their efforts. In other words, frequency response matters far more than phase response.

While the eye can easily see the obvious distortion of the waveform if the system is non-minimum-phase, the ear/brain system does not hear waveforms a such, so it is much less sensitive to phase shift than our intuition would lead us to believe.

[Bassmec]

Originally Posted by DukeLeJeune

The research I've read on the subject indicates that the ear can hear the difference between two different crossovers that have different phase characteristics, but even if one is technically "correct" and the other is not, listeners generally don't have a strong preference and do not tend to consistently pick the "correct" version over the other one.

On the other hand, the ear is quite sensitive to the spectral balance (frequency response), so that is where most designers concentrate their efforts. In other words, frequency response matters far more than phase response.

While the eye can easily see the obvious distortion of the waveform if the system is non-minimum-phase, the ear/brain system does not hear waveforms a such, so it is much less sensitive to phase shift than our intuition would lead us to believe.

I work with phase relationships with extreme care having to time align
lots of different microphones all collecting basically the same sound stage every day.
Probably makes me very sensitive to any cancellations caused by any Phase angle anomalies in my passive monitors, they can have a serious effect by causing cancelation dips in the crossover area frequency response. Passive crossovers can be a bit of an impedance nightmare
for amplifiers and they are certainly not lossless.
Quite a lot of DIY passive crossovers I have heard are a bit like DIY sex, not quite the whole nine yards if you ask me.