[JoshK]

Agreed doug, that I would lean towards the non variable points.  But slopes are not variable AFAIK.  I think getting a cheap pro xo that is very flexible to experiment and then commit to a high end solution like marchand would be a better idea.  Maybe an even better idea is to design the passive line level xo we were talking about, but that is not without it's downsides as well.

[JoshK]

This camp is intriguing... hadn't heard of this before. Are any of the better known crossover guru's working in this arena? Are there any successful products on the market? Does it need to be tailored specifically for the speakers (like typical passives are)? Again, this is pretty interesting...

No idea bout the first question.  The later is of course, yes it needs to be tailored to each speaker in question, and for that matter each set of amps you intend to use.  You can see why this is an often overlooked solution.  It may in fact be one of the better solutions but it needs a lot of thought, care and planning to build the system up to work correctly in this case.  This is not the sort of solution an audio dealer wants to promote because it won't allow for amp swapping every month when the new and greatest product comes out.   Same with speakers etc.  So not for the flavor of the month club at all.  However, if you have committed to a pair of speakers indefinitely and want to take them to their best this may be a great way of doing so, I don't know for sure but it seems logical.

[doug s.]

Agreed doug, that I would lean towards the non variable points.  But slopes are not variable AFAIK.  I think getting a cheap pro xo that is very flexible to experiment and then commit to a high end solution like marchand would be a better idea.  Maybe an even better idea is to design the passive line level xo we were talking about, but that is not without it's downsides as well.

personally, unless the speaker drivers in question have weird response characteristics that would require special notch filters, etc, that wouldn't be addressed w/standard x-overs, i would just stick with 24db/octave slopes.  sharp cut-off & keeps the drivers in phase.

marchand also offers an interesting line-level passive x-over, but it looks like ya really would wanna know the slope before ordering; plug-n-play frequency changes don't seem to be part of the deal here.  perhaps phil could install sockets for swapping out the caps-n-resistors...

http://marchandelec.com/xm46.html



doug s.

[JoshK]

Well lets be fair, the 24db/octave slope keeps the signal in phase *only* near the xo point and not wideband.  Meaning the 24db/octave slope has, to the best of my knowledge, a larger impact on overall wideband phase then say a 6db/octave slope.  There are more schools of thought out there than just BC of VMPS that believe that anything above 12db/octave slope causes phase errors too great to provide seemless driver integration (outside of the xo point) and coherent sound.  I am in no camp here just stating what I have learned.   I would personally like to play it myself to form an opinion rather than just state what I have read somewhere.

Of course, if you use a digital xo then you don't have to worry about phase!

[doug s.]

Well lets be fair, the 24db/octave slope keeps the signal in phase *only* near the xo point and not wideband.  Meaning the 24db/octave slope has, to the best of my knowledge, a larger impact on overall wideband phase then say a 6db/octave slope.  There are more schools of thought out there than just BC of VMPS that believe that anything above 12db/octave slope causes phase errors too great to provide seemless driver integration (outside of the xo point) and coherent sound.  I am in no camp here just stating ...

yes, i know - a *lot* of speaker designers (thiel among them), believe 6db slope x-overs are the only way to go.  i believe that there's good stuff from *both* schools.  i like thiel speakers a *lot*...  

but, if *i* am gonna get into any diy-type stuff, then it's gonna be 24db/octave at the line level.  anything else is way too much above my knowledge & abilities.  and, i believe all the phase errors that brian (& jim thiel?) talk about are related to x-overs at the *passive* speaker level.  i do not believe these are of as much concern when at the line level.  even brian cheney uses active 24db/octave x-over in his top model, to cross his woofers over to his midrange panels.  
http://www.vmpseurope.com/e/p-st3.htm
and, this is what he recommended to me, when i queried him about going active & bypassing the passive low-to-midrange x-over in the rm-40's...

doug s.

[JoshK]

You are probably right doug, that active xo probably is far less harmful to phase at steep slopes but AFAIK I think it still does have the impact still.  I think I am going to try playing with a digital xo setup at some point in the near future, maybe not with my RM40's but probably with my own contraption.

[doug s.]

You are probably right doug, that active xo probably is far less harmful to phase at steep slopes but AFAIK I think it still does have the impact still.  I think I am going to try playing with a digital xo setup at some point in the near future, maybe not with my RM40's but probably with my own contraption.

i'm not so sure the negative impact has to do w/active vs passive so much, as doing it at the line level vs doing it at the speaker level.  (ie: before vs after the amps).  which is why marchand has a special *passive* x-over to work at the line level, *before* the amps.  

in any event, i would certainly be interested in your results w/your "contraption".       someday i hope to do a line-source "contraption" of my own!  

doug s.

[JoshK]

I just don't see how doing it line level instead of speaker level is going to get around the problem is all.   You could be right but it doesn't seem obvious to me.

[doug s.]

I just don't see how doing it line level instead of speaker level is going to get around the problem is all.   You could be right but it doesn't seem obvious to me.

to be honest, i am not sure myself, on a technical level.  intuitively, it makes sense, as on the line level, it's only the amp that is seeing the change from a line-level x-over.  distortion/efficiency-reduction levels are typically far lower here - look at typical frequency response, distortion levels & signal-to-noise ratios of active x-overs.  this way, the speaker is only seeing a limited frequency from the amp, & its drive ability is not effected by capacitors, chokes, resistors, etc...  it would be nice to get some feedback about this from experienced diy speaker builders or professionals...

doug s.

[woodsyi]

I certainly am no expert.  Rather consider me a blind monkey fiddling with crossovers to see if monkey music comes out of speakers.  I can howerver report what I hear from an active triamping set up.  I am using a Marchand XM 44 with RM 40s.  You can check my 2-channel system for associated equipment.  Sticking strictly with 4th order slope, I got the Larger sub to 80 H, two woofers between 50Hz and 600 Hz, and the ribbons (using the existing passive xover network between the panels and the FS tweeters) down to 200 Hz.  What I get is a very clear sound.  I just about have the tonal balance where i want it.  I have tried all sorts of passive setups-- sigle wire, bi-wire and passive biamping with attenuator on the tubes.  I get the best sound with active x-over with  these speakers.  With the pots on the speakers and line level control on the x-over I can balance the all important (for vocal representation) 200-700 Hz balance just right.   FYI, Mega woofer rolls off at 100Hz and midbass woofer rolls off at 600Hz.  Both see the same signal but there is a filter that limits its range for the megawoofer.

[JoshK]

so you allow significant overlap b/w the panels and the upper woofer?  interesting. Is there any perceived cancelation?

[woodsyi]

I tried the lowpass for the woofers at 200Hz as Brian recommended, but it sounded too bright and thin.  The serial first order coil Brian uses slopes up to 600 Hz. So, I tried 200, 300, 400, 500, 600 and 700 (all 24dB) and settled on 600.  I don't hear any cancelation.  Rather the woofers add that woomp to the voices that makes Johnny Cash sound like Johnny Cash. It also provides that huskiness to Natalie Merchant and warmth to Norah Jones that makes them special.   On operatic music, I hear the voices through the speakers just like what I hear in opera houses. All in all, the sound is clearer with active x-over and you have more options to get the right tonal balance.

[doug s.]

I tried the lowpass for the woofers at 200Hz as Brian recommended, but it sounded too bright and thin.  The serial first order coil Brian uses slopes up to 600 Hz. So, I tried 200, 300, 400, 500, 600 and 700 (all 24dB) and settled on 600.  I don't hear any cancelation.  Rather the woofers add that woomp to the voices that makes Johnny Cash sound like Johnny Cash. It also provides that huskiness to Natalie Merchant and warmth to Norah Jones that makes them special.   On operatic music, I hear the voices thro ...

is the passive between woofs & midranges disconnected, or still in the circuit?  when i discussed going active w/brian for these speakers, he recommended disconnecting the passive x-over...

doug s.

[woodsyi]

It's disconnected.  Take the megawoofer out and the stuffing and you can see the coil.  You just have to desolder one wire and resolder bypassing the coil.  If you don't, the active crossover just won't kick in as it parellels the passive x-over.