or Connect
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › what's the big deal about crossovers?
New Posts  All Forums:Forum Nav:

what's the big deal about crossovers? - Page 2

post #31 of 66
Marked for later.
post #32 of 66
Well I do have a suggestion that is currently being tossed around in AES (Audio Engineering Society) about offering a list of members expertise that they could provide to the general public.

All the suggestions above are great and after building speakers for more than 30 years the resources now versus then are incredible.

My suggestions is to find the local chapter of AES and found someone who could mentor you in the building and designing process. I can state that these members have the expertise and equipment and most of them love to share the knowledge that you wish to acquire.
post #33 of 66
Here is the link to the Pacific Northwest chapter.

http://www.aes.org/sections/view_section.cfm?section=42

good luck

The best resource I have used in the last 15 years has been LinearX company. Chris has a very good understanding of loudspeaker modeling.
post #34 of 66
I am back into speaker building after a 7 year hiatus, back then LEAP/LMS was the standard for the rich kids and Madisound and Solen were the providers of choice. When Dynaudio got out of providing drivers to DIYers, I lost interest. That plus I got a good job in the industry for a while.

There are many great posts on this thread that should help you.

As far as Ray's links page... I think we have the wrong people doing cancer research, Ray would find the way to have them all licked in a weekend.

The crosover point and slope between your mid and tweet should be more of a result of the tweeters capabilities ad the expectations for playback level. If this is a chamber music playback system for a tiny den, you can use lower x-over points and less steep slopes. A benefit to 6Db or 12Db slopes over more complicated designs for 18Db, 24Db and steeper slopes is that with fewer components, the signal path is shorter and the detrimental effects of each of these crossover components on the signal will be reduced.

For your first ever DIY, you may wish to calculate out a basic first order (6dB) or second order (12dB) and assemble it yourself point-to-point and be gentle on the final product.

If you are going to listen to "Live At Leeds" at the level of the original performance, or use the speakers for HT, then use a steeper slope design and maybe raise the x-over point a touch higher. Get an advanced helper and some modeling and measurement software ad prepare to experiment. Impedance compensation circuits (to keep the crossover point accurate by compensating for the varying impedance Mark spoke of) and bypass circuits are difficult to plan, and easy to screw up.

This is called "voicing" the system and it is an art form.
IMO, don't buy some off the shelf generic pre-fab x-over from the supplier, build it yourself or you may as well just have bought the entire speaker pre-built.

Heed the advice of Colofan and find the local speaker nerds, I used to rap with the PSACS builders here in IL, and long time audio reviewer Tom Nousaine was a regular. Once I tagged along for a private loft sale at Madisound just for the PSACS members, somewhere I still have 2 pairs of 8ohm Vifa cast basket 180mm mid-woofer drivers...
post #35 of 66
A crossover is a necessary evil for most loudspeakers, all passive networks de-couple the amp output from the speakers voice coil (bad thing) and waste a lot of power. The biggest problem is the more components you add the more damage you do. The best crossover would be no crossover, but unfortunately, no one has made a driver that can properly reproduce the entire spectrum properly. The next best thing is Active Crossovers where the frequency division happens before the amplification, this reduces the need for compensation networks and conjugant filters and with some (Linkwiz Riley 4th order) keeps phase intact. Properly designed Active networks introduce very little noise into the system and increase dampening (control of the drivers VC) by 10 fold, and don't waste any power.
post #36 of 66
Quote:
Originally Posted by Kingdaddy View Post

A crossover is a necessary evil for most loudspeakers, all passive networks de-couple the amp output from the speakers voice coil (bad thing) and waste a lot of power. The biggest problem is the more components you add the more damage you do. The best crossover would be no crossover, but unfortunately, no one has made a driver that can properly reproduce the entire spectrum properly. The next best thing is Active Crossovers where the frequency division happens before the amplification, this reduces the need for compensation networks and conjugant filters and with some (Linkwiz Riley 4th order) keeps phase intact. Properly designed Active networks introduce very little noise into the system and increase dampening (control of the drivers VC) by 10 fold, and don't waste any power.

Interesting, So is this what Thiel does in their CS speakers where they put integrate the tweeter and midrange onto the same voice coil? Although they still have a crossover between the bass and midrange...

Are you suggesting that if someone came out with a way to integrate all 3 drivers (hi, mid, bass) onto the same voice coil, this would be optimal? Or am I mis-interpreting you?
post #37 of 66
This is all physically sound (sic) and good. But it is very difficult to back up the analysis by double blind listening tests. This seems to suggest that the real-world gain is not really that high...

Loudspeaker design really must be very difficult, as there are few absolute references. One thing may work good in a given setting, and less so in another. And the absolute verdict (talking about listeners ears here, not his Wife ;-)) is not easily predicted from a stack of measurements

regards
Knut Inge
post #38 of 66
I think it's pretty well known among designers (or agreed upon) that any frequency dividing network is a degradation to making natural organic music. The best crossover is NO crossover, but current speaker limitations lead most to a dividing network. If I have my pick of putting a wad of reactive components between the amp signal and driver (destroying impulse and control), OR a hand full of grungy op-amps, I'll take the grunge over the decoupling any day. Grunge can be overridden or tamed. However, the loss of dampening makes micro dynamics and phase alignment nearly improbable IMO.
post #39 of 66
Coaxial drivers can be effective,but they have their own flaws too. There is on magic answer.

Quote:
Originally Posted by Kingdaddy View Post

I think it's pretty well known among designers (or agreed upon) that any frequency dividing network is a degradation to making natural organic music. The best crossover is NO crossover, but current speaker limitations lead most to a dividing network. If I have my pick of putting a wad of reactive components between the amp signal and driver (destroying impulse and control), OR a hand full of grungy op-amps, I'll take the grunge over the decoupling any day. Grunge can be overridden or tamed. However, the loss of dampening makes micro dynamics and phase alignment nearly improbable IMO.

Ofcourse theres many ways to go,its like going out for a meal.You could go to mcdonalds and be happy for 5minutes and feel sick later,or you could go get some italian or whatever and enjoy the quality and flavour.

The driver itself is reactive ! Can a resistor or reactive component,add a distortion where as the driver itself doesnt? I mean the electrical part alone of the driver. I know that many people do think that resistors and capacitors can be heard. I think that those people are pretty much doomed in audio. Theres only so much you can achieve without electrical components

Its a design choice with its own pros and cons to go crossoverless. People often add a 'supertweeter'with a capacitor approximating a 1st order filter. But then the DI is mucked up since the single driver is beaming into a tight acoustic pattern while the tweeter is behaving in an omnidirectional way.

Many do enjoy single drivers,like fostex. Usualy- little- girl -with -guitar- music - no offence intended,they often state this themselves.
For dynamics people tend to go with more efficient/higher power handling systems.

Go to linkwitzlab
publications.htm#12%20-%20A%20Three-Enclosure%20Loudspeaker%20System%20with%20Active%20Delay%20a nd%20Crossover

Regards
Mike.e
post #40 of 66
How does one safely remove a passive crossover from a manufactured loudspeaker?
post #41 of 66
im not familiar with that brand also...i'll wait..
post #42 of 66
Quote:
Originally Posted by RxNinja View Post

How does one safely remove a passive crossover from a manufactured loudspeaker?

The only way to remove a passive crossover and safely drive your tweeters (especially) is to substitute an active crossover and use a crossover frequency fairly close to the original design and bi-amp or tri-amp. Active or 'electronic' crossovers along with bi-amping offers benefits in dynamic response, phase coherence, even power handling in some cases.
post #43 of 66
Awesome thread.

I agree with just buying a kit from GR Research. I'm a hands on guy myself, but I'd just build a test box and start screwing around with crossovers with cheap components to start out with. IMO you'll learn more working with low end drivers than expensive ones.

Until you get a lot of experience I'd stick to building good kits for speakers you actually want to listen to.
post #44 of 66
wow..thanks for all the links...
post #45 of 66
I've built a couple speakers designed by Dennis Murphy. His designs are very highly regarded in many forums. One of his designs includes the GR-Research M-130. Check out his website. He does all his work for pure enjoyment and shares his designs completely free of cost.

I've found working with very good proven designs helps in learning speaker and cross-over design.

murphyblastercom
post #46 of 66
This thread (Ray's Links) is beyond incredible. Thank you!
post #47 of 66
Thanks for the compliment erandmckay!
post #48 of 66
Another option which I've fooled around with and have achieved some outstanding results is to use a wide range driver with a piezo tweeter. I am using a wide range made by css, the WR125S and some piezo tweeters I had laying around, no crossovers.
post #49 of 66
I didn't see this on your list...



http://www.josephaudio.com/essay.islope.html
post #50 of 66
Thanks SleeperSupra. Included it at the bottom of my links thread. Also included the active DEQX processor/crossover which can go from 100-300 dB/octave. Here are the links added:

Infinite slope crossovers http://www.josephaudio.com/essay.islope.html
DEQX Digital Calibration Processor/Crossover http://www.deqx.com/pdc26.html
post #51 of 66
You're missing three great links Ray,

Passive Crossover Designer, free passive crossover design software - http://audio.claub.net/software/jbagby.html

FRD Consortium, more free design and modeling software - http://www.pvconsultants.com/audio/frdgroup.htm

ARTA - a powerful and free loudspeaker measurment suite - http://www.fesb.hr/~mateljan/arta/
post #52 of 66
Thanks TurboFC3S, added those links at the bottom...
post #53 of 66
Quote:
Originally Posted by M NEWMAN View Post

HOLY SH#@!!, Ray....nice work!

I thought I was going to end up in china! Too much time on your hands!

I use an active crossover the Behringer DCX2496 and I'm more than happy with the affordability and flexibility of it where I can chose any crossover filter or cross off slope.
post #54 of 66
It would seem to me, and I am certainly no expert on speaker design, that if I was going to build one, active crossovers are a must. At lease for testing. When we design a high end audio system for a club, there are NO passive crossovers. All the crossovers are done via dsp ( either in the amps or seperat dsp units ) I find this desierable as even with the manufacture's speaker tunings, they are only a starting point. Active crossovers also let you time align each driver.
post #55 of 66
I am new to the AVS world.This would take a few years for me to read it all. Considered this page bookmarked. Thanks!!!
post #56 of 66
Thank you dinoman22 & fellow members for all the compliments...
post #57 of 66
cool
post #58 of 66
Quote:
Originally Posted by Kingdaddy View Post

I think it's pretty well known among designers (or agreed upon) that any frequency dividing network is a degradation to making natural organic music. The best crossover is NO crossover, but current speaker limitations lead most to a dividing network. If I have my pick of putting a wad of reactive components between the amp signal and driver (destroying impulse and control), OR a hand full of grungy op-amps, I'll take the grunge over the decoupling any day. Grunge can be overridden or tamed. However, the loss of dampening makes micro dynamics and phase alignment nearly improbable IMO.

I just spoted this thread and note that what you said is truth. May be I am one who sacrificed to speaker filtering most of all. Knowing unpararell performance of one widerange driver I tried between 1976 and 1990 frankly anything to repeat it with multiband systems and the conclusion was undoubtfull - it won't go. Why? Because all frequencies in one widerange are radiated from the same diaphragm an have the same time to reach ear. Therei is in my opinion no other reasons (like lost of energy or "something eating sound" in passive elements ) If the problem is to be solved first we should not complicate it and take care for what is obvious. And for three drivers obvious is 1) not fixed one common distance to ear, 2) covering of performances bands between speakers 3) resonant (full of no-watt energy) character of switching beteween them. In 1990 I just kicked all known stuff and assembled something what was driven mostly by prediction in my head that it had to be possiblie. It was not easy way to make it just right - at the beginning I couldn't recognize even why it works. Even it was success only in one field and sound was still far from perfect this field was most important - First time in multiband system I noticed fluent operation of switching between speakers. If standard filters deliver non compatiblie parts this simple system delivered something what seemed to be prenivelated. The effect is that when one of the drivers is completely disconnected next to him takes over their common radiation . Even the slope is ranging between 12 and 18 dB/ oct the ranges are covered in some strange way - my conclusion was that they pass unwanted job to cooperating driver. Today it is already developed in nearly perfect way and You may check that circuitry on my page Michal Czabajski D.I.Y. Nonfatigue Sound System .. The details may vary depending on used speakers and it is painfull to hit the point but main advantage is that at last there is the point

Rgds
post #59 of 66
Quote:
Originally Posted by RxNinja View Post

How does one safely remove a passive crossover from a manufactured loudspeaker?

Fabricate a new termination for the crossover with the necessary external connections and wires to connect to the drivers.

Gain access to the interior of the speaker and disconnect all drivers from the existing passive crossover.

Connect the new termination to the drivers. You may wish to leave the passive crossover and its termination and controls in place.

Test new terminations, reassemble speaker, test again.

One issue is whether or not you wish to ever reuse or resell the speaker with its original passive crossover. If you do, then you want to minimize the impact on the speaker. I would favor creating a wiring harness with Speakon connectors that passed through a small hole in an inconspicuous place for the new termination.
post #60 of 66
Quote:
Originally Posted by Kingdaddy View Post

I think it's pretty well known among designers (or agreed upon) that any frequency dividing network is a degradation to making natural organic music.

That's why virtually every speaker has some kind of a crossover. The designers just put them in to make the speaker more expensive and sound bad.

No, its not well known that frequency dividing networks have to degrade speakers. The degrading design feature that everybody would prefer to avoid is multiple drivers. Trouble is, we just can't make good speaker drivers that are full range and practical.
Quote:
The best crossover is NO crossover, but current speaker limitations lead most to a dividing network.

No, the best tweeter is no tweeter, but rather we would prefer have a woofer that simply does it all. Trouble is that with current technology, the single driver full range speaker is mission impossible or at least impractical. Even planar speakers have multiple drivers and crossovers.
Quote:
If I have my pick of putting a wad of reactive components between the amp signal and driver (destroying impulse and control), OR a hand full of grungy op-amps, I'll take the grunge over the decoupling any day. Grunge can be overridden or tamed. However, the loss of dampening makes micro dynamics and phase alignment nearly improbable IMO.

I guess we can interpret that as a yes for active crossovers. On that we can agree.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Audio theory, Setup and Chat
AVS › AVS Forum › Audio › Audio theory, Setup and Chat › what's the big deal about crossovers?