For those of you who have been able to experiment with different crossover slopes (12, 24, 48 db/octave, etc.), and types (L-R, Butterworth, Bessel, etc.), which have you found to perform the best in your systems? I currently am using 48 db/octave Linkwitz-Reilly, but I have been experimenting with others. They all sound different, and seem to have pluses and minuses. Anyone have any thoughts on this?
Hi,
- A simple enough question .
- FWIW one should separate electrical from acoustic slopes .
Complex Answer :
- I use 24 db LR electrical slopes on my twin woofers, which effectively translate into @ 17 or 18 db acoustic slopes , for my MTM woofers.
- I use an approx 18 db electrical slope on the horns ( call it a quasi Bessel arrangement ). This electrical slope when superimposed over the horn/driver combo translates into about a 12 or 13 db "acoustic slope" in the 1.5 octave area around the crossover point ( 800 hz ). ( 1.5 octave = .5 octave up & 1 octave down )
- The combination of these two acoustic slopes with their approx. 90° phase difference, work very well for my MTM setups. These differentials ( both phase and slope wise ) happen to conform pretty closely to D'Appolitos own theory (for best convention ) on what make MTM setups work . These choices seem to give the most coherent & seamless sonic glue throughout the crossover region .
Polarities ??
- In my setups : a positive signal "pulse" gives a compression wave for the woofers and horn . ( I don't like mixing together rarefraction and compression waves )
<> Cheers
A good explanation here:
http://sound.westhost.com/index2.html
Try this
Assuming a vertical speaker layout and a 2 way, hold your head horizontally a few inches from the speaker equally distant from the 2 drivers with voice playing.
Do you hear what sounds like 1 speaker or 2 speakers?
If it sounds like 1 you got it right.
Passive 24 dB/octave are a little difficult as the higher the order the more critical the accuracy of the component values become and the more critical it is that the correct impedance is seen by the crossover.
There is also the cost issue due to increased complexity and number of components.
JBL’s flagship K2’s use 24dB, not without reason.
When I build a 24 dB passive crossover I trim all components on a bridge and remove windings from the inductors to trim for the correct value.
Remember that most commercial designs are made to a budget and the crossover is 1 area where you can save money to meet a price point.
The lower the order the cheaper and easier it is to build a crossover.
Why build passive xovers when you can buy VERY high quality, fully adjustable active xovers (i.e. Ashly XR- series) on the cheap and bi- or triamp? Is it for the challenge of building such a device?
Passive crossovers are by their very nature not phase coherent and also they present a difficult reactive/capacitive load to the amplifier. Why do it?
Seems to me that if you are going to build a project speaker, you might as well get rid of the major fault of a commercially produced, passive x'd speaker, the crossover design.
Maybe I'm just nuts...
Steve,
Sounds like you are jsut over the river from Robert.
Robert is a dab hand with passive crossovers, perhaps help might be able to hep you out (see his diy 4 way XP system)
Forgive me but what sort of system were you thinking of organising?...I seem to recall you mentioning this in another thread.
I agree that active crossovers offer what appears to be a technically superior solution.
However if you look at many 24dB/octave commercial crossovers most are not that great, the main reason being that they do not follow the correct alignment due do the difficulty in adjusting the required number of components.
A 24dB/ocvtave Likwitz-Riley requires the adjustment of 6 resistors per section to change frequencies, difficult to implement.
The main benefit of a 24dB/ocvtave Likwitz-Riley being phase coherency and some other filter alignment designed for convenience of adjustment sounds wrong.
When you design a filter its phase coherency is defined by the filter type and it makes no difference weather or not the filter is implemented as an active or passive filter as to phase coherency or response.
A Passive filter that presents a difficult reactive load to an amplifier is also incorrectly designed.
The other day I had a commercial active crossover running on my speakers and wife took one listen and said that sounds awful.
On the other hand she like my home made 24dB/ocvtave Likwitz-Riley crossovers that run my left and right.
I struggle sometimes to decide which sounds better, my active or passive filters.
They both have the same alignment but sound different.
As I see it you get what you pay for, if you extend the argument to digital filters the cheap ones sound awful.
The same applies to analogue filters.
This is a very complex subject with many subjective impressions.
Absolutely...trust your ears..
Usually, you get what you pay for, or at least your supposed too! I like 18db butterworth filters. They sound good to me, and gives me a bit of overlap between the drivers allowing a more seamless image than 24db slopes! I always find 24 db filters to make speakers sound like individual bands, rather than the independent ranges gelling together as one.
With the Rane AC23, I dunno, it just starts to sound compressed as you push the system. So Bryston 10B LR were tried, same thing, as you push, the sound compresses!
I have no real experience with passives so I have nothing to say in that area!
One of my favorite active electronic xovers is still the Urei 525.
I have a new BSS FDS 366T, so I will be trying out a digital unit once again, I will experiment with the different filter types!
I also would like to hear Ians all discrete, class A crossover, it looks good, what I have seen of it here on the forum!
scottyj
I experiment with phasing to voice the tweeter and midranges., reversing the phase shifts the localization of the sound around and shows me what needs to be changed to get the drivers to balance out. I use a VCB-100 virtual xover box to quickly A-B changes to get the "voicing" right. I start out with 12 db slopes and use the natural rolloff as much as possible. I shoot for a point of origin that does not seem to come from one driver, rather from the entire speaker, then I know I have the crossover matched. I seldom see the calculated values when I am done. Being able to A-B the changes is important since it is difficult to remember what was going on 20 minutes ago accurately.
As to electronic xovers, they cannot help too much in time alignments and other nuances like passive parts can. The abrupt rolloffs can do more harm than good sometimes.
Part of the art of good speaker design is in the crossover, and that is usually what separates good speakers from great ones.
Passive crossovers also do the balancing and the EQ.Originally Posted by Alan Fletcher
Some folks would rather have a wire than a rack full of dedicated gear to upgrade.
[O.K., two wires, then....
Actually, I had intended a discussion on the various merits of different slopes and types, not passive versus active, but what the heck! I'm also using a BSS FDS 366 (not the "t", but with the new software so I think they are the same). I really do like the sound of this unit.
What about using one slope for the high pass and a different one for the low pass? I haven't tried this, but under what conditions would you want to do this? I suppose it might help get a flatter response through the crossover region, but couldn't you do the same thing with a little EQ?
The issue of group delay is beyond my limited measurement capabilities, so I don't know what effect this is having in my system. I'll have to try experimenting like duaneage and jkc suggested and see if I can hear any differences.
This is done quite often. Usually the tweeter is crossed over steeper to increase power handling, especially when trying to match with a larger woofer. Volumes have been written on the pass charactoristics of crossover networks, not to mention the impedance and phase that results when you inject a crossover into a system.
As for using a little eq, sure why not. But if your developing a speaker for a market you cannot count on an EQ to be present. In your home you can do what you must to get the sound right. Most rooms need a LOT of eq anyway, few of us live in studios with properly designed walls.
Most important point is to make sure the drivers are protected and that the impedance does not go too low or deliver an unusual phase angle to an amplifier, that might get something hurt.
It depends as Earl said on the characteristics of your drivers and how you propose to use them..(baffle layout).
I would definately try the simple 6db slopes and then progress from there: there is no right answer.
Unlike many on this forum I like the low power 45 triode amp, at 1.8 wpc I have the flexibility to use the 1st order crossovers. 1st order crossovers are really quite forgiving but require knowing your drivers well.
With JBL 15" drivers like the 2226, they are a high inductance driver and coils do not always act like crossovers. In fact the "Pi speaker forum" uses them for duty up to 1.6khz with just a .7 mh(for the 2226H 8 ohm model) to tame a rising response from 800hz to 1.6khz. The 2226 actually is down 20db at 2 khz due to its natural behavior.
When used with a 3rd order at 1.6 khz hi pass crossover they operate in a very seamless manner. I am currently using a 1st order crossover on my compression driver on Edgar round wooden horns. For now, Altec 902/650hz Edgar tractix. I am awaiting my order from Bruce(Edgar) for my 500hz tractix and my 399's from Great Plains.
Reading from prior posts on this forum discussing crossovers with comments from "Earl K", I experimented some and I am using a 3 parallel cap combination. I have a 2 ufd oil cap/12 ufd Dayton MPT(not sure what the "T" stands for, maybe just a fancy way of saying MPP?)/.68 Dayton MPP cap. This is all fed directly to a 8 ohm L-pad.
Sound is good at just the right volumes, too low or too high and the sound is just off a bit, but to get it right and the 18" 2242 in Giskard cabinet with PE 250 watt sub plate amp makes music downright scary in a good sense.
I am using Mike Bakers old E130/2225 recone with a "perfect-lay 14 guage 1.2 mh coil" and they mesh very well together.
For higher power though I use a 3rd order on my compression drivers.
Just my thoughts for the fans of the "peanut powered amps"
Ron
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
Reply With Quote