What crossover order to choose? Need your experienced opinion.

[Misa A.]

I need experienced people’s opinion on the following question, please.

I am working on a crossover network for my speakers made of Scan Speak 18W/8545 x 2 and D2905/9500 in MTM configuration. So far I have a parallel filter with a crossover frequency of 2 kHz and with a 2nd order electrical and 4th order acoustical slope for both tweeter and woofers networks. Drivers are connected in phase. Everything is fine except that at high signal levels tweeter seems to start to distort in its lower region.
I don’t like the idea of moving up crossover frequency, so I need to increase steepness of tweeter filter to unload the tweeter at lower frequencies. I suppose guys at North Creek put a notch filter in a tweeter network below crossover frequency for this purpose.
My simulations show me that I can get theoretically good results if I change the filter order for tweeter and woofer networks to 3rd electrical order, which gives me 5th acoustical order respectively. Crossover frequency will remain 2 kHz and one of drivers will reverse polarity to radiate in phase.

My question is: out of general theory and personal experience of people here, will such 5th order crossover with reversed tweeter polarity sound worse comparing to 4th order in-phase filter? Should I try it or it is better to stick to 4th order and try to increase the crossover frequency or implement notch filter to unload the tweeter?
I ask this question because I know the general rule that says, that the higher the filter’s order, the worse the phase characteristics of a speaker and it's transient response.
Thanks to all who answer my question.

Misha

[pinkmouse]

Simulations? Have you actually measured anything?

You really need to find out what you have in reality before you can start tweaking!

But as general advice, as long as there are no nasties in the woofer response, i would just move the crossover up a bit.

[SY]

2K is just too low for that tweeter. Going to a higher-order network will not help much. If you can squeeze out just an extra half-octave or so, the tweeter will be much happier.

[Ken L]

Originally posted by Misa A.
Scan Speak 18W/8545 x 2 and D2905/9500 in MTM

http://www.madisound.com/solist.html

The Madisound Scan-Speak Solist is similar but uses an 8545K and probably a different box. I used to own these and can find the crossover schematic

Take a look at the differences and if you think you're pretty close I will either scan my copy of the LEAP designed crossover or post the values.

Got to run for now, taking my wife out for Birthday stuff and will check this thread later to see if you want me to gather this info.

Regards

Ken L

[Misa A.]

Thank you all for quick response. As I see, nobody recommends me to increase the filter order.

Yes, I measured both in-cabinet SPL and impedance curves including phase with Speaker Workshop and JustMLS, using Eric Wallin’s Jig II and Preamp with a Panasonic WM-62 microphone, though not calibrated. A friend of mine has LspCAD and he helped me with SPL measurements and filter simulations. Now it is already 4th version of filter working in my speakers and so far what I get exactly matches design curves.

North Creek people cross SS 9500 at 1.8 kHz and as they say, find it the best sounding solution after a lot of tests. http://www.northcreekmusic.com/NorthCreekKits.html (Just checked their web site and saw that they changed 9500 to 9900 in their Thor kit.
I would also like to have crossover frequency as low as possible in order to keep vertical directivity of speakers wide as my main listening position has to be 15 deg off axis in vertical plane.

Ken L , I would greatly appreciate if you help me to get Madisound’s design of crossover for their Solist-9500 with element values. I know, that 8545K is 0.5 dB less sensitive and has a different SPL curve, but I could use this schematic as a good starting point. Can you please E-mail it to me to ma35@mail.ru ?

Best regards,
Misha

[Misa A.]

Sorry, I meant of course North Creek Rhythm kit, not Thor.

[Ken L]

Originally posted by Misa A.


Ken L , I would greatly appreciate if you help me to get Madisound’s design of crossover for their Solist-9500 with element values. I know, that 8545K is 0.5 dB less sensitive and has a different SPL curve, but I could use this schematic as a good starting point.

I believe this is probably a very good starting point - I would like to point out that I communicated with Madisound and that they do not consider this to be a proprietary schematic and OK'd it for me to post in this forum. They consider it to be good public relations.


QUOTE]Originally posted by Misa A.
Can you please E-mail it to me to ma35@mail.ru ?

Best regards,
Misha [/QUOTE]

I would rather post it than e-mail it. Otherwise I will get e-mail from somebody years from now asking me to e-mail it to them _grin_

Regards

Ken L

[jewilson]

Ken,

The speaker diagram is interesting but what are the values, I guess I am confused.

[5th element]

I have used the 95 at 2k and it took this with no problems. 4th order acoustics, although this is active so power issues are not relavent. Clipping could obviously alter the ability to handle 2k and if this is at loud levels it wouldnt be suprising if clipping was occuring making it sound rough and harsher anyway.

Look at linkwitz site and see how he calculates excursion and see if your 95 will be exceeding xmax at the levels you are using.

[Ken L]

Originally posted by jewilson
Ken,

The speaker diagram is interesting but what are the values, I guess I am confused.

On the top of the schematic-

Choke (inductor) symbols are like C's on top of each other ( or Dolly Parton _grin_

.6 is a .6 mH 16 Ga choke, etc.
.7 is a .7mH 16 Gauge choke
.27 is a .27 mH 26 Gauge choke (inductor)

Capacitors are in microfarads - and the symbol looks like two opposing t's with one bolder than the other

8CP = an 8 micro farad Cap, etc.

so you have 3, 8, 15, 20 Microfarad capacitors, etc

The sawtooth looking things are resistors -

which are in watts and ohms. I don't remember what MO stands for in ohms.

I don't have the crossover to look at and have no information other than this.

Due to size restrictions on posted images at the forum, I did not scan the entire document - However, the only thing left off was the LEAP logo,etc. This appears to be a standard format of output from a program where apparantly they expect the operator to know AC/DC ratings of the caps and parts, etc.

This is not my strong suit and it's been like three years since I assembled this_grin_.

I'm just reading it as what I _think_ is correct.

Hopefully, someone else may chime in with further explanation.

Other than that, all I can think to add is that these are not small crossovers _grin_, nor are they cheap.

Regards

Ken L