re-designing a speaker crossover

mantis

Active Member
Hey guys,
I'm interested in trying my hand at designing a crossover for a 2-way speaker. Start with something simple right?

What kind of data must I have before me, and what are the best things to attempt to do:

-what slope crossover? -6,-12, -24? Which is best price no object?
-what about other tricks, are zobel circuits worth it?

Any help appreciated, thanks!
 
Register to hide this ad
Hey guys,
I'm interested in trying my hand at designing a crossover for a 2-way speaker. Start with something simple right?

What kind of data must I have before me, and what are the best things to attempt to do:

-what slope crossover? -6,-12, -24? Which is best price no object?
-what about other tricks, are zobel circuits worth it?

Any help appreciated, thanks!

http://www.amazon.com/Loudspeaker-Design-Cookbook-Vance-Dickason/dp/1882580109

http://www.parts-express.com/pe/showdetl.cfm?Partnumber=390-803

http://www.parts-express.com/pe/showdetl.cfm?Partnumber=248-625

http://pro-audio.musiciansfriend.co...241810&src=3WFRWXX&ZYXSEM=0&CAWELAID=26020141

http://www.fesb.hr/~mateljan/arta/

http://www.apicsllc.com/apics/Misc/filter2.html

It's not as easy as it looks...but starting with only a two way design is a good idea.

As for slopes, I like to use the minimal amount of components necessary, so I prefer to use 1st or 2nd order. 1st order is very tricky though and 2nd order has it's own issues. 3rd and 4th have less issues, but use much more(and sometimes larger) components.
 
All good advice so far.
For slopes that selection of what type of drivers you are using determines what direction to go.
Smoother speakers (usually more expensive and better quality)= simpler crossover.

Speakers with ragged slopes are tough to tame.
 
These are ar-4x speakers.

I've got preamp covered, I have a mic... its not extra flat or omni but its a fair condensor... its like a u87/4047 but with a slight spike at 12k... definitely usable from 500->10K. I also have sound absorbers... think small gobos... so I can surround the unit for testing if need be and get pretty few reflections.

With that sort of gear, is it possible for me to attempt to do anything useful?

I have a feeling that the safest way is clearly just to replace the original parts to spec... but that the alternative safe way is to construct a 4th order xover at the same original spec states... not too expensive, and should improve the original right?

Thoughts?
 
not a simple project

Passive crossover design is much more complicated than you would think.

Yes the parts (capacitors, coils, resistors) are readily available. And software to do calculations is, I assume available. 15-20 years ago when I first got into speaker building I bought a $10 DOS program to do calculations. And there was a program called LEAP for about $400 that was the top of the heap ( I never owned LEAP and not sure if it is still on the market) and several others in between. Today there are probably lots of options for crossover design calculations including software and you can even probably find formulae to do calculations manually (not recommended). But that is where the easy stuff ends.

You also have to do the following:

1) Determine the best crossover point. Set it too low and your tweeter will be vulnerable to damage. Set it too high and you lose dispersion for a band of frequencies because the woofer tends to beam frequencies at the higher end of its range. How will you choose the crossover frequency?

2) Determine accurate impedance values for your drivers. Just because the speaker is stamped "8 ohm" or "4 ohm" it doesn't mean you can use these values for design work. These are nominal values and likely not even close to the actual impedances that would be used for calculations. For example, a nominal 8 ohm driver can have actual impedance of around 6 ohms. But you don't know until you test it. It is essential to have the actual impedance to make a crossover that works correctly...otherwise your speakers will sound horrible (believe me, I have learned this the hard way, lol). To complicate matters further, for a 3-way system (which I understand yours is not) the mid-range driver will very likely have a significantly different impedance at its lower and upper crossover points.

3) What about level matching? Your tweeter may be louder than your woofer, or vice versa. Speakers manufacturers typically incorporate resistors in the crossover to properly match the outputs of the drivers in use. How will you match the level of your two drivers? Trial and error? Every adjustment you make will require a change in the other components of your new crossover.

4) Choose a crossover slope. There are trade-offs here. As the order of the crossover increases you get better protection for your tweeter and a smaller band of frequencies being produced by both drivers. But you pay for that in loss of efficiency. A 4th order crossover (generally) has 2 capacitors and 2 coils - with 2 elements in series with the driver and 2 elements in parallel. These things suck energy that would otherwise be used to produce music in the speaker.

I am only vaugely familiar with AR speakers. Are they not known for being efficient speakers? (please someone correct me if I am wrong about that) Anyway, efficiency is a very good quality for a speaker to have. It makes it possible to get good sound from lower powered amps and receivers. Now if you have a 200 watt amp maybe this is not an issue. But at 50 watts or less I would say that any speakers with a 4th order crossover may not be a wise choice.


I am not trying to discourage you or anyone from learning about or designing passive crossovers. But to do it right requires lots of testing, lots of equipment, lots of knowledge, lots of time. It is very easy to mess up a good pair of speakers. A proper crossover is critical to good sound. And the AR guys have already done this task for you.

There are lots of posts (either in DIY or the Speakers sections of AK) that discuss replacing capacitors in speaker crossovers. That operation is relatively safe and may even yield improvement in sound. May be a good place to start.
 
Last edited:
thanks for your post roger- heres the rundown to address the concerns your brought up:

-AR are very inefficient, but efficiency isn't important here... These things are 15W recommended, I'm using a 150 on them at the moment so theres lots of room to waste energy.
-I have a multimeter and can measure real nominal impedance values... I know they are way off from what they say.
-These speakers have a build in rheostat on the back for adjusting tweeter level, so thats covered.

with that info, where must I then go?
 
mantis,

If the AR's are "15w rated" that is a recommendation on how much power they can handle. Power rating is not the same as efficiency. Efficiency would be stated as, for example 90 db/w for a fairly efficient speaker or 86 db/w for a fairly inefficient speaker.

Ok Mantis, before I go further let me say that I am a bit fuzzy on details. It has been at least 15 years since I worked on any passive crossover stuff.

Measuring impedance is not as simple as using a DMM on the speaker terminals. For crossover design you need to plot a series of impedances through the frequency range where the crossover will be inplay. What is the impedance of your tweeter at 3k? Or your woofer at that frequency?

There is a procedure for doing that and honestly I don't recall it. But it requires at least a signal generator and I seem to remember that precision resistors were involved. It is not simply using a DMM because, ( I believe) the DMM is measuring resistance, not impedance. And there will be a difference that matters.

Again I ask, how will you decide the best crossover frequency? The AR designers knew all the parameters of the drivers. If you choose too low and blow your tweeter, will it be possible to find an exact replacement?

AS I said before, I am not trying to discourage you from learning. But reading beyond what I post here will be necessary to do this right. The others posted links for you. Check out books by David Weems or Vance Dickason.

If you proceed, by all means do not dismantle the original crossovers. In fact you could leave one speaker original and use it as a baseline for comparing to your new design. If you goal is learning about passive crossover design that would be the way to go.

On the other hand, if you goal is to improve the sound of a pair of speakers you like, simply re-capping (with better quality poly caps) is the simple, safe, and sure way.
 
heres the deal- the present "crossover" is a piece of junk, its just a 20uF cap for the tweeter and a 1.2mH inductor. My take is that ANYTHING will be better than that, so why not.
 
Simple does not mean it is bad.

What you have there is a 1st order crossover.

1st order/6db per octave = one element per driver
2nd order/12db per octave = 2 elements per driver
3rd order/18 db per octave = 3 elements per driver
4th order / 24 db per octave = 4 elements per driver

A 1st order/6 db per octave 2-way crossover is a capacitor in series with the tweeter and a coil in series with the woofer. Simple.

As mentioned before, there are trade-offs among the various designs. 1st order is the most elegant in many ways and is preferred by some purists. Of course there will be debate on this issue as with anything. The fact that your AR's use 1st order (assuming they are original) implies that the drivers are well behaved and lacking peaks that would otherwise necessitate the use of a 3rd or 4th order design.

Again assuming that those are the orignal AR crossovers, it would be a huge and disappointing waste of time to attempt to actually improve on them. The AR guys knew what they were doing or else their speakers from 30 years ago or whenever would not appear in AK'ers equipment lists as frequently as they do.

Just because the AR crossover appears simple does not mean that it would be easy to improve it. On the contrary, you could spend a lot of time and money (for testing equipment) and very likely not succeed in making any improvement.
 
heres the deal- the present "crossover" is a piece of junk, its just a 20uF cap for the tweeter and a 1.2mH inductor. My take is that ANYTHING will be better than that, so why not.

I think you're being presumptuous. Winslow Burhoe used a single cap in his EPI 100s and that module is legendary for it's sound. The fact that you're even considering rebuilding one on a set of 40+ year old speakers would indicate to me that the designers weren't totally inept and seemed to know enough to build a system that stood the test of time.

I'd just start with upgrading the cap. You may be surprised that single Dayton polypropylene cap may prove to be all they need...
 
I am not familiar with AR model numbers. Are these vintage?

If so, and if those are the original crossovers and if the capacitor is electolytic... you may consider replacing the electrolytic with a high quality polypropylene cap.

This is a common procedure and, as mentioned before, there are many posts here about it. Use the forum's search function, you might try "recapping" in the speakers or DIY sections.

It is possible that the original capacitors ARE NOT electrolytic though...I don't know the age of your AR's and whether or not they may possibly already have a better grade of cap.
 
hey guys,

I don't' think im being presumptuous, I know these are their lowest tier speakers and they needed to save cost by putting in a first order xover... why not go nuts if its so inexpensive to go the complex route... but if its not worth it to try to improve it then no need. I've replaced the caps with solens on my 2ax and 3a speakers, and they work, but these are the only ar models that I could comfortably "improve" since its 2 way and it wouldn't be crazy to attempt. Oh well.

Original cap is oil if I recall, I'm pretty sure all of them used those giant oil caps in the big boxes strapped to the backboard.
 
heres the deal- the present "crossover" is a piece of junk, its just a 20uF cap for the tweeter and a 1.2mH inductor. My take is that ANYTHING will be better than that, so why not.


I would agree with some of the others saying that a first order crossover may not be bad at all, in fact, it will probably be better than what you end up with without spending a couple of hundred on gear and books.

Low order crossovers are not something you should avoid because of a perceived lack of quality. Some of the best speakers made use first order crossovers.
 
I would agree with some of the others saying that a first order crossover may not be bad at all, in fact, it will probably be better than what you end up with without spending a couple of hundred on gear and books.

Low order crossovers are not something you should avoid because of a perceived lack of quality. Some of the best speakers made use first order crossovers.

I'm in agreement with Dave.
You shouldn't automatically assume that a first order crossover is an attempt to cut corners and save manufacturing costs. Nearly all the speakerbuilders I know will use a first order crossover if they can, due to the sonic merits that are available, that is, if the drivers have a smooth frequency response as well as roll-off. There are actually a few high-end designs that don't even use an inductor on the woofer and have a single high quality cap on the tweeter. Series crossovers are also said to retain some of the sonic benefits of a first order conventional crossover, but that's a totally different subject.

Now, one thing that hasn't been mentioned, is the ability of a first order crossover to pass a square wave. Some feel, and I'm one of them, that this ability translates into signal preservation when music is being played.

Some of the best sound available for playing music is through a single driver speaker, at least in the all-important midrange. No crossover at all and they can sound very nice, although they're usually compromised somewhere on the edges of the frequency bandwidth.

My advice is to replace the caps with decent Poly caps and then listen for awhile, you may be surprised how well your speakers sound.

Best Regards,
TerryO
 
Passive crossover design is much more complicated than you would think.

Yes the parts (capacitors, coils, resistors) are readily available. And software to do calculations is, I assume available. 15-20 years ago when I first got into speaker building I bought a $10 DOS program to do calculations. And there was a program called LEAP for about $400 that was the top of the heap ( I never owned LEAP and not sure if it is still on the market) and several others in between. Today there are probably lots of options for crossover design calculations including software and you can even probably find formulae to do calculations manually (not recommended). But that is where the easy stuff ends.

You also have to do the following:

1) Determine the best crossover point. Set it too low and your tweeter will be vulnerable to damage. Set it too high and you lose dispersion for a band of frequencies because the woofer tends to beam frequencies at the higher end of its range. How will you choose the crossover frequency?

2) Determine accurate impedance values for your drivers. Just because the speaker is stamped "8 ohm" or "4 ohm" it doesn't mean you can use these values for design work. These are nominal values and likely not even close to the actual impedances that would be used for calculations. For example, a nominal 8 ohm driver can have actual impedance of around 6 ohms. But you don't know until you test it. It is essential to have the actual impedance to make a crossover that works correctly...otherwise your speakers will sound horrible (believe me, I have learned this the hard way, lol). To complicate matters further, for a 3-way system (which I understand yours is not) the mid-range driver will very likely have a significantly different impedance at its lower and upper crossover points.

3) What about level matching? Your tweeter may be louder than your woofer, or vice versa. Speakers manufacturers typically incorporate resistors in the crossover to properly match the outputs of the drivers in use. How will you match the level of your two drivers? Trial and error? Every adjustment you make will require a change in the other components of your new crossover.

4) Choose a crossover slope. There are trade-offs here. As the order of the crossover increases you get better protection for your tweeter and a smaller band of frequencies being produced by both drivers. But you pay for that in loss of efficiency. A 4th order crossover (generally) has 2 capacitors and 2 coils - with 2 elements in series with the driver and 2 elements in parallel. These things suck energy that would otherwise be used to produce music in the speaker.

I am only vaugely familiar with AR speakers. Are they not known for being efficient speakers? (please someone correct me if I am wrong about that) Anyway, efficiency is a very good quality for a speaker to have. It makes it possible to get good sound from lower powered amps and receivers. Now if you have a 200 watt amp maybe this is not an issue. But at 50 watts or less I would say that any speakers with a 4th order crossover may not be a wise choice.


I am not trying to discourage you or anyone from learning about or designing passive crossovers. But to do it right requires lots of testing, lots of equipment, lots of knowledge, lots of time. It is very easy to mess up a good pair of speakers. A proper crossover is critical to good sound. And the AR guys have already done this task for you.

There are lots of posts (either in DIY or the Speakers sections of AK) that discuss replacing capacitors in speaker crossovers. That operation is relatively safe and may even yield improvement in sound. May be a good place to start.
Excellent post!
 
heres the deal- the present "crossover" is a piece of junk, its just a 20uF cap for the tweeter and a 1.2mH inductor. My take is that ANYTHING will be better than that, so why not.

- Those AR's are very decent speakers, and the crossover was carefully designed. I wouldn't change it without a clear definition of what I was was trying to accomplish. It is very doubtful that you will simply get "better sound" than AR's engineers were able to achieve, although it is certainly possible to change the sound to be more to your liking.

- If the crossover has both an inductor and a capacitor, it is 2nd-order, not 1st-order.

- You cannot measure driver impedance with an Ohmmeter. All that will give you is the driver's DC resistance.

- I've tried making measurements with the kind of mic you have. It won't work very well, because the polar response of the mic is neither even, nor flat.

- "Gobo" type absorbers are meant to reduce sound transmission. They will do very little to improve sound reflection. If it was this easy, people wouldn't bother building anechoic chambers.

-k
 
2nd order (12db per octave) requires 2 passive components per driver. For a woofer the coil is in series and the capacitor is parallel. For the tweeter the capacitor is in series and the coil in parallel.

A crossover such as the one described by the OP is 1st order because there is only one passive component per driver - a coil in series with the woofer and a cap in series with the tweeter.

It is not the total number of passive components (coils and caps) that determines a crossover's order (and slope) it is the number of passive components per driver. It is possible to have mixed designs for example 2nd or 3rd order on the tweeter and 1st order on the woofer. Or, as another poster referred to, a design with 1st order on the tweeter and nothing at all on the woofer.
 
It is possible to design a new crossover at home. Weather it will improve over the factory design is another question. The big issue with these AR's is the pots. If the pots are corroded you are not getting anywhere near the sound that the speakers are capable of. If I was going to mod these I'd start with bypassing the pots and possibly adding a fixed l pad instead if needed. Probably throw in a new polly cap too. Maybe a low resistance 12 gauge air core inductor. I think that the simplest crossover with the least resistace and number of components is the best route.
 
Back
Top Bottom