[bunnyma357]

I am starting to think about setting up a DIY active crossover tri-amping setup, using software to provide the crossover in the digital domain. I am interested in learning how others are dealing with active crossovers, particularly -

Are you using DIY speakers, or did you modify existing speakers to remove passive crossover?

What are you doing to provide speaker protection in case a full range signal accidentally makes it to a driver?

Are you using matched amps, or tailoring each amp to the specific frequency band it is driving?

How significant was the change from passive to active crossover?

Any problems that you didn't anticipate, that might be and issue for me to avoid?

Thanks for any help you can provide.


Jim C

[mjosef]

No software in use here, just hardware.
Modified my VMPS RM1 by removing the builtin passive crossover(for the bass drivers/mid panels) and am using a Marchand crossover to split the frequency @320Hz/ 24dB slope, which is higher than the original stock 166Hz.
No speaker protection in my case...no way an accident of that nature(full range freq. to a driver) can happen with a hardware approach...unless I am stoned out of my mind and then proceed to rearranging my system in that state. 

I use a tube amp for the md/hi freq. and a SS amp for the bass. Works well since I have pots(mono) on the outputs of the Electronic crossover...and I level match each channel with a SPL meter and pink noise.

Improvement was significant enuff to never will I go back to passive state. Hard to quantify such a subjective quality to another individual. YMWV

[NagysAudio]

1. - DIY, or commercial (without passive crossovers) speakers can be used, makes no difference.

2. - No speaker protection.

3. - I'm assuming you'll be using high quality amps? If so, then it really doesn't matter that much if you mix and match.

4. - If done correctly, active crossovers can sound SIGNIFICANTLY better than passive!!! But here's the catch... If you already have a properly manufactured speaker with a passive crossover, that was designed in an anechoic chamber (Thiele and Small parameters measured using added mass, input impedance, velocity, time/phase aligned, baffled step, edge diffraction, SPL, correct crossover slopes, etc.), by using an active crossover you will only make it perform/sound worse. Unless... You have an anechoic chamber and can do all of the above (plus another hundred or so) measurements yourself and tailor the active crossover to that particular speaker. Otherwise it's a complete waste of time and money.

5. - Stay away from dbx, Behringer, Electro Voice, etc. crossovers, they are garbage. Mixing pro equipment with consumer equipment can also cause a ton of headaches due to noise/hum issues (electronically balanced inputs, etc.). A lot of balanced consumer and pro equipment are manufactured incorrectly as far as grounding is concerned and sometimes they're not compatible. You might have to play around with different configurations of balanced cables (hanging shield, capacitor, etc.). Going from pro balanced to consumer single ended is another potential noise/hum problem.

[bunnyma357]

This is still just an idea, but hopefully I will end up doing it. My system is very bargain basement, so I'm not looking for absolute quality. I know there are many people here with DEQX processors or modded Behringers, and active crossovers also seem quite popular with the VMPS crowd. So I am hoping to translate that high end information into my low end system.

My speakers are AV123 Strata Mini's which visually look like they share some drivers with the VMPS 626jr, but probably have worse crossover components & cabinetry, so they should, in theory see a similar improvement as VMPS speakers. Right now I'm feeling that the speakers are the weak link in my system, and there are 3 areas that I can improve them - Diffraction (at some point I'll be contacting Jim Goulding again), Resonance (the knuckle rap test indicates these are more akin to a guitar than an inert enclosure - No Rez is in the future as well) and Crossover (after pricing a Skiing Ninja crossover upgrade, it occurs to me I can achieve more in the digital domain, and use the money saved on additional amps for an active crossover).

A software based system would be more prone (one wrong mouseclick) to sending a full range signal to a tweeter and potentially blowing it, so I was hoping that someone might have some experience of how to go about protecting a driver with a simple filter that wouldn't be in the active audio region.

My amps are likely to be in the price range of one or two 2 meter Nagy's power cords, so that may give some perspective on my expectations - I am seeking a high audio ROI - not the absolute best performance possible. Right now I'm using an Onix SP3, and my first instinct is to purchase 2 more of them to tri-amp the speakers. Most likely I will experiment with the amps I currently have on hand - Onix, Yaqin & Yamaha first.

As far as speaker design, mine are definitely designed to a price point, and one of the likely areas where costs were cut is in the crossover components - which is why upgraded crossovers are a popular mod for this speaker. I find speakers to be the most subjective component, and am not overly concerned with performance in an anechoic chamber, since I basically listen in a reverberant cube. After years of listening to various speakers I've determined the attributes that matter to me, and that sometimes seems to be things that don't test as well on paper.


Passive Crossover Network
•   Custom   Designed   by   Mark   Schifter   and   Danny   Richie
•   Air   Core   Inductors
•   Polypropylene   Capacitors
•   Non-Inductive   Wirewound   Resistors   
• Fourth Order Slopes on High Frequencies
• Second Order Slopes on Low Frequencies
   
Crossover Points
• 8” Woofer = 27Hz to User Definable
•   5.25”   Mid   Bass   Coupler   =   80Hz   to   650Hz
•   8”   Planar   Magnetic   Midrange   =   650Hz   to   4100Hz
•   1”   Planar   Magnetic   Tweeter   =   4100Hz   to   35KHz

These are the current crossover specs for the speakers - as a total layman, as I understand it the higher the order, the steeper the slope, and the worse the phase error - hopefully, applying the crossover in the digital domain can keep steep rolloffs without having an adverse effect on phase, and also allow for time alignment of the 4 different drivers.

I appreciate the warnings about pro gear, but at my price point pro gear is and will be a part of the system. The performance of my Behringer Feedback Destroyer Pro has greatly improved the quality of my system - compromised bass adjusted to eliminate room modes sounds way better than pristine bass with huge peaks. I've also been happy with the performance of my Focusrite Saffire Pro 26, which would be at the heart of the system - providing the 4 stereo outputs for the software active crossover.

I realize that this isn't the cleanest, most ideal approach to achieve the best sound quality, but it seems like the possibility for good sound at a very reasonable price may be attainable. Thanks for any help in figuring this all out.


Jim C

[Noseyears]

The only disadvantage with active crossovers is that every speaker needs its own amplifier channel. And the system gets more complex that way.

[macrojack]

Eliminating passive crossover increases efficiency and allows for use of smaller and more specialized amplification. I find that trade off attractive.

[pjchappy]

Instead of the Behringer crossover, look into Ashly.  I've heard from more than one person these are much better quality than Behringer, with better parts to begin with.  No mods necessary.  It is more expensive, of course, but much cheaper than many modded Behringers.

http://pro-audio.musiciansfriend.com/product/Ashly-Audio-Protea-3.24CL-3In-3Out-Speaker-Processor?sku=181034



Paul

[jtwrace]

Instead of the Behringer crossover, look into Ashly.  I've heard from more than one person these are much better quality than Behringer, with better parts to begin with.  No mods necessary.  It is more expensive, of course, but much cheaper than many modded Behringers.

http://pro-audio.musiciansfriend.com/product/Ashly-Audio-Protea-3.24CL-3In-3Out-Speaker-Processor?sku=181034



Paul

+1.  I love my Ashly.  It's very user intuitive and when you call Ashly direct, a english speaking person answers the phone and is more then willing to help.  I asked if there was a way to dim the screen and he (Jim) sent me the diagram with which resistor to remove.  Yes, they're that great!

 http://www.ashly.com/protea324cl-cl-d.html

[bunnyma357]

Well part of my plan is to actually eliminate as much hardware as possible. So Crossovers and Parametric EQ/Room Correction will all be done in software.

I'll probably be able to get rid of a Belari Phono Pre, a Zero DAC, a Squeezebox 3 and a Behringer Feedback Destroyer Pro, and if I can get the Apple remote to conveniently and transparently control volume, I might even eliminate my Meridian Preamp.


Jim C

[pjchappy]

So, you plan on using a soundcard w/ a multi-out, then?  That's a possibility for me for my future active set-up.  Haven't decided yet.


Paul

[bunnyma357]

So, you plan on using a soundcard w/ a multi-out, then?  That's a possibility for me for my future active set-up.  Haven't decided yet.


Paul

I have Focusrite Saffire Pro 26 i/o firewire interface, which I bought as the interface to run Pure Vinyl and use as a DAC and Phono Pre Amp. I have been very happy with it, after eliminating some initial hum issues. It has 8 analog channels for output, so I started thinking that it might be a cheaper and more effective mod to buy 2 new amps, rather than redo the passive crossovers in my speakers.

http://www.focusrite.com/products/saffire/saffire_pro_26_io/#Specifications




Jim C

[Kevin Haskins]

I love active.... but you do have to design it right.   The single most important key to any loudspeaker is the design and that isn't something that is trivial to reproduce.    The best parts put together poorly will sound like crap.    Average components designed correctly can sound incredible.   

You don't need a chamber to design a loudspeaker but the skills, equipment and learning curve are not trivial so I'd recommend copying something that is designed by someone who knows what they are doing.   Linkwitz or someone of his background I'd feel comfortable recommending.   I'd have reservations about 90% of what you see online...... stick with someone like Linkwitz or someone who at least has some experience designing loudspeakers and electronics.   

[bunnyma357]

I love active.... but you do have to design it right.   The single most important key to any loudspeaker is the design and that isn't something that is trivial to reproduce.    The best parts put together poorly will sound like crap.    Average components designed correctly can sound incredible.   

You don't need a chamber to design a loudspeaker but the skills, equipment and learning curve are not trivial so I'd recommend copying something that is designed by someone who knows what they are doing.   Linkwitz or someone of his background I'd feel comfortable recommending.   I'd have reservations about 90% of what you see online...... stick with someone like Linkwitz or someone who at least has some experience designing loudspeakers and electronics.   

Kevin,

It seems like an active crossover would be pretty easy to implement, with just adjusting the crossover frequencies and slopes and then testing the results and making small adjustments. Starting with the published specs and going from there. What aspects do you feel would be hard to deal with?

It seems that many people are doing this with hardware active crossovers, what specific areas do I need to think about before going down this path?  Is it likely my current passive crossovers are doing more than just splitting the signal between drivers?


Jim C

[jtwrace]

What aspects do you feel would be hard to deal with?

Jim C

I know this was directed at Kevin but I feel the need to answer this from an non-industry professional side.

The time and patience. 

Unless you have someone at your disposal via email and phone you will be very surprised on how many hours, days....you can spend on this.  It's not just put in the current speaker specs and whalla! it's done.  If you strive for what a typical audiophile strives for you'll be listening / measuring many times over.  In the end you will have learned more then you might of wished but after many iterations you'll have a unique setup and have a much better understanding of what actually happens.  Last, you'll more then likely have a better ear from it.

This was exactly my experience and I'm very happy I did it.  I am amazed at how much time I spent doing it.  No, I'm not retired. 

One last thing.  As I was doing this whole process, I kept thinking this is not an active crossover.  Active in my world would actually do it automatically.  I would love to develop an active system that you put in the specs and it would run a full DOE on the full spectrum and at the same time auto save the results (output to say MatLab) and then you can have program numbers to go back to.  That IMO would be a true active bad-ass setup. 

[macrojack]

Jason is definitely onto something in terms of market needs if not in terms of approach. I certainly don't have an answer but I do have the questions.
As audio evolves more and more of us are looking for high performance without high prices and speakers seem to be the area where the pricing curve runs away from the performance curve most dramatically. There are many great driver options and lots of very doable cabinet designs but the crossover, delay, impedance, Eq, and gain matching parameters, and their interdependence, baffle most of us. It is easy enough to buy world class speaker management (prices are reasonable by audiophile standards) but the gear is meant for pros and doesn't cater to my ability level. I have located a pro sound man in my area who is supposed to come by the house next week and set up my XTA so I can benefit from it.

I agree with Nagy that Behringer, DBX, etc. is not built to the standard of the other equipment in our system and it seems counterproductive to introduce something so crude in a chain that you are trying to improve upon.

I remember reading that Tyson had great success with DEQX but endured a strenuous learning curve assisted by test instruments to get the job done. Wouldn't it seem that someone could design a device for a couple of thousand dollars that does it all, does it well, and does it for you? If not, why not? Certainly the need is growing but the fact that this is for 2 channel people and only those who are outside the box, so to speak, it may not represent a large enough demand to justify the R&D. Certainly, it isn't anything to turn Sony's head. If it is to come, it will have to be from the pro side.

My XTA accepts PC cards, whatever they are. I wonder if someone could create program cards for various DIY combos and just sell the card to the end user to put in his machine.
Wasn't there a subwoofer manufacturer years ago who made a sub that could be mated to various speaker systems by the insertion of a custom XO card?

If that idea would work, I already have the hardware. All we need is someone to create and market the software. Maybe XTA?

[jtwrace]

mj-

I have no doubt that something like this could be developed.  The issue of course is cost.  The audiophile world is so small that unless you were selling many of these systems I don't think you could get the cost low enough.  I have a rough idea of what it would probably cost just to get going..it's a lot!  Just the code is thousands of hours. 

In the Pro Audio world I don't think it's an issue as that's what they do and the enviroment / audience is much different. 

[Kevin Haskins]

Kevin,

It seems like an active crossover would be pretty easy to implement, with just adjusting the crossover frequencies and slopes and then testing the results and making small adjustments. Starting with the published specs and going from there. What aspects do you feel would be hard to deal with?

It seems that many people are doing this with hardware active crossovers, what specific areas do I need to think about before going down this path?  Is it likely my current passive crossovers are doing more than just splitting the signal between drivers?


Jim C

It is pretty easy if you have the design already done from an engineering perspective.   Hell... designing a speaker is pretty easy but one person's easy is another enigma.   

Let me outline what is necessary to design a loudspeaker.   It isn't rocket science but it does take some understanding of what is important and just as importantly what is not.   It takes some basic knowledge of acoustics and it helps to have a physics background.   Having said that I know people who run high-end audio companies who have no math/science background and they have managed to design something that sounds good enough to keep them in business...... but I digress.

#1.   Have a goal.   You have to know what you are aiming for and it has to be something that can be measured.    You have to understand what is most important and how to achieve it.    I'd say start by reading Toole's book and you at least have a good general concept of what the goal should be.   Toole's book isn't a book on HOW to achieve the goal.   It is more about what the goal should be.   

#2.   Know how the arrangement of various transducers in various mechanical/acoustical environments translates to the goal.

#3.   Know what to do with the electrical signal that can shape what we measure in the mechanical/acoustical sphere into our final objective.

#4.   Probably should have pointed this one out first..... know how to measure acoustical properties of loudspeakers.    Learning how to measure is critical because garbage in garbage out certainly applies.


You then have to buy the equipment to do the measurements.... prototype enclosures... and go through the trial/error of testing new ideas.   Import the measurements into modeling software to determine the electrical filters needed and then finally design and build the electrical filter.    At that point you can listen and start to tweak it for subjective preferences and perhaps go back through the measurement/model/build process again for as many iterations as you find necessary to achieve something you are happy with.


If you start with a design.... and you just want to convert it into a different type of electrical filter that is obviously much easier.   But... most manufactures are not going to give out electrical transfer functions and when they do.... they probably are not going to support the endless questions you will naturally have concerning reproducing their passive network with a DSP.   

As an example.... this is the measurement of my old EX-6.5 mid-woofer in a 0.5 ft^3 ported cabinet. 





Now.... tell me which filter and at what frequency should I crossover?   Keep in mind this is one measurement on one axis.... so you ideally wouldn't design around just this one measurement.    Also... some of the data in that measurement is not good data.   Do you know what is good and what is bad?    You have to know enough to translate the measurements you get and have a general idea of how they are going to look on other axis of measurement and of course... you have to have some certainty that those measurements you are chasing have subjective consequences.

Here is a tweeter FR.   What filter should I use and at what frequency to make this look good?   Note... this one changes significantly off-axis also so you need more than one measurement to do good design but I figured I'd start with something to illustrate my point.   

[macrojack]

It appears to me that Emerald Physics has crossed part way into this realm by providing a pre-programmed DSP specific to his speakers as part of the finished package. He could upgrade his product by simply offering better DSP devices as options. Other companies could conceivably do something similar using better pro gear like XTA or Ashley.

If they became a dealer for a given line, they could just program the unit for their speaker and sell it at retail along with the speaker as a package. Wouldn't that ultimately provide better performance, brand loyalty, and profitability while eliminating the need to build passive crossovers?

[TomS]

Clayton Shaw has already been doing some of this with Spatial and EP, at least the notion of room interaction, measuring and tuning remotely.  Not so much crossover alignment, but it is interesting nonetheless.

http://www.spatialcomputer.com/page10/page10.html

[Kevin Haskins]

It appears to me that Emerald Physics has crossed part way into this realm by providing a pre-programmed DSP specific to his speakers as part of the finished package. He could upgrade his product by simply offering better DSP devices as options. Other companies could conceivably do something similar using better pro gear like XTA or Ashley.

If they became a dealer for a given line, they could just program the unit for their speaker and sell it at retail along with the speaker as a package. Wouldn't that ultimately provide better performance, brand loyalty, and profitability while eliminating the need to build passive crossovers?

You could..... it is not always as easy as using an off-the-shelf DSP.    I wouldn't do it for the very simple reason that you can get into an endless loop of technical support questions.    If you sell the pro-gear DSP you have to support it and it would open up endless technical support questions that would never end.    The more you sell the worse it would get.   

Also... once you design the electrical transfer function ideally it should never need to change.   Opening up access to the filter design is like giving a monkey a gun.    It is essentially giving the user a very complex tone control but one that has the potential to totally screw up the design and possibly damage transducers.   

The only equalization that is useful for tuning systems is one that provides a couple bands of PEQ under about 200Hz or so.    That would allow a user to measure and equalize some room issues but that still requires knowledge and skills and equipment to perform right.