Series Crossover - Page 15 - diyAudio
Go Back   Home > Forums > Loudspeakers > Multi-Way

Multi-Way Conventional loudspeakers with crossovers

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 22nd March 2013, 04:51 AM   #141
rabbitz is offline rabbitz  Australia
diyAudio Member
 
rabbitz's Avatar
 
Join Date: Jul 2003
Location: Newcastle, Australia
Default AR Series components

To demystify the AR series xo, here's what the components do.

L1 sets the crossover roll off for the tweeter, lower the value to increase xo point and visa versa (very low DCR air core required).
R1 sets the tweeter level and must be above 3R so the speaker impedance does not drop. Tweeter SPL choice is important and must be higher than the mid woofer by a couple of dB to achieve this. Impedance of the tweeter can come into play (such as the XT25). Use a 10W ceramic or Mox resistor or equiv.
C1 sets the crossover roll off for the woofer in conjunction with L2 as well as the slope of the tweeter roll off. It effects both the woofer and the tweeter so a change effects both. Higher value increases slopes. Use a good quality MKP here.
L2 sets the crossover roll off for the woofer as well as the BSC. Larger value will drop the xo point, increase BSC possibilities and visa versa. It can be placed on the -ve line between C1 and C2. Use a good quality air core but DCR has to be noted in the modelling as does effect the result.
C2 and R2 forms the CR (zobel) and helps with final tweaking of the phase and roll off. Rarely is it a textbook value.
C2 value can be altered to change the woofer roll off slope etc (larger will make it steeper). Use a MKP.
R2 can be tweaked to help with phase at the crossover point so you end up with a deep reverse null of 30dB+. Use a 10W ceramic or MOX.

Optional components:
A cap can be included in series after L1 and before the tweeter but usually not required and can cause phase issues which means a complete redesign.
The tweeter can have a damping resistor across the tweeter to lower the impedance peak and 15R-22R is a good place to start. However, R1 would have to be decreased in value as you get added attenuation of the tweeter.
CR can be placed across the tweeter to help shaping of tweeter response and a good example of this is the 810921 tweeter.
LCR networks can be added if desired but every component in series xo effects everything else. Generally not required if driver choice is good.

Find Jeff Bagby's SCD, create some frd and zma files and have a play.
Attached Images
File Type: jpg AR_series_xo.jpg (13.0 KB, 268 views)
__________________
No longer DIY active
  Reply With Quote
Old 22nd March 2013, 05:06 AM   #142
diyAudio Member
 
wolf_teeth's Avatar
 
Join Date: Mar 2006
Location: Indiana
Blog Entries: 1
Quote:
Originally Posted by rabbitz View Post
C1 sets the crossover roll off for the woofer in conjunction with L2 as well as the slope of the tweeter roll off. It effects both the woofer and the tweeter so a change effects both. Higher value increases slopes. Use a good quality MKP here.

C2 value can be altered to change the woofer roll off slope etc (larger will make it steeper). Use a MKP.

Optional components:
A cap can be included in series after L1 and before the tweeter but usually not required and can cause phase issues which means a complete redesign.

Find Jeff Bagby's SCD, create some frd and zma files and have a play.
The electrical slope from the SXO cannot increase just by increasing the value of a capacitor, it will still transition to the inherent slope further down the rolloff. It can however change the shape of the knee at the xover which will add or remove damping to the knee depending on value. This is the same with parallel networks as well.

As to the AR-SXO design as a whole, yes- the added cap on the tweeter would alter how close the summation gets using this circuit and voicing by ear. It is likely that on a flat baffle you can't have the second order SXO and have it work well anyway. This is due to the acoustic offsets, and a typical 12/18 arrangement of acoustic slope is what makes them align better in general. The 1st on tweeter and 12dB on the woofer is why the alignment works as it is most of the time.

PCD now has SXO modelling in the same spreadsheet, so you only have to DL PCD for both. Be sure to get Response Modeler so you adapt your files to your boxes before using it though.

Later,
Wolf
__________________
Photobucket picture pages: http://photobucket.com/Wolf-Speakers_and_more
Writeups/thoughts/blogs: http://techtalk.parts-express.com/blog.php?u=4102
  Reply With Quote
Old 22nd March 2013, 08:48 AM   #143
diyAudio Member
 
chefsinc's Avatar
 
Join Date: Jun 2006
Location: Perth Suburbs. Western Australia
Clearly I admit that I am a hack, with no software or measuring equipment used in my designs but as i have been building by ear for 7 years and analysing Tony Gee, Troels G etc. and great info From great people in this forum I am happy with my results. I am yet to burn a voice coil on any drivers with my guestimate designs nor send any amps into meltdown. So with that i will continue to build em like this.
of course what Rabbitz says is correct about selecting the correct drivers and on occasions i have ignored that but hey its my money, my time and my fun.

cheers friends.
just remember its just a bit of fun.
  Reply With Quote
Old 22nd March 2013, 01:03 PM   #144
diyAudio Member
 
chefsinc's Avatar
 
Join Date: Jun 2006
Location: Perth Suburbs. Western Australia
Quote:
Originally Posted by system7 View Post
I do hope this series discussion manages to stay on-topic and actually INFORM, because most of them don't!

Since Jonathan (chefsinc) started his SB 8" plus 1" project, I've spent a lot of time at Tony Gee's marvellous Humble Homemade Hifi website. The Extremon (found in components/downloads) is a very nice flat baffle 6" Paper plus 1" (SEAS 22TAF/G) design using a series network.

It's quite easy to get as good as you like using Parallel or Series networks, and you can throw some nice notches in with both. The interesting thing happens when you TIME ALIGN it by setting the tweeter back and flipping the polarity.

Suddenly you find all the notches are disastrous to phase, and you must dump them!

So you have to live with higher Fs resonance and 5kHz cone breakup. But the good news is 1) You get a simple fast circuit, and 2) All the resonances are PHASE ALIGNED, so may be less offensive. Series circuits become very appealing when time-aligned, but you CAN do similar stuff with parallel. Like metal cones, they are a strict discipline, but you will learn lots.

It's a matter of technique to adjust the circuits when dumping the notches, and I won't dwell on it, except to say that losing the 15uF DC-blocking capacitor was a very double-edged sword. But these are low-order circuits that make demands on drivers, and series filters tend to let a lot of low frequencies into the tweeter. So some drivers will work better than others, and SMOOTH is good. For instance, this polycone SEAS U18RNX/P looks fun.

After such an epic (IMO) post, I hope for some interesting replies...but I'm not getting my hopes up.

actually S7 I did dump the 15uf/or22uf in series with the tweeter in my schematic and now run it with just a 12uf after the .390mh inductor. no resistor either end, just the Lpad to dampen the tweeter output. so far 3 days and no distortion or burnout. I still think I could do a bit more to dampen the 3.5khz woofer breakup, but is only noticeable on CD's such as Adele's 21 with her high frequency voice.

Wolf, S7, Rabbits any suggestions? I tried to roll off the woofer with a bigger cap parallel to the zobel but it sounded a bit compressed. Go for a bigger L2 inductor? or just leave it as it is not that noticeable as I said only on certain CD's.

cheers guys
  Reply With Quote
Old 23rd March 2013, 01:56 AM   #145
rabbitz is offline rabbitz  Australia
diyAudio Member
 
rabbitz's Avatar
 
Join Date: Jul 2003
Location: Newcastle, Australia
Jonathan

I'm not familiar with your speaker but generally I tend to reduce woofer peaks at higher frequencies with a larger L2 and a larger zobel cap for additional shaping. If I'm understanding this correctly, the larger cap parallel to the zobel will not alter the slope but move the woofer xo point down so the effect on the peak would be minimal.
__________________
No longer DIY active
  Reply With Quote
Old 23rd March 2013, 05:23 AM   #146
diyAudio Member
 
wolf_teeth's Avatar
 
Join Date: Mar 2006
Location: Indiana
Blog Entries: 1
Quote:
Originally Posted by chefsinc View Post
actually S7 I did dump the 15uf/or22uf in series with the tweeter in my schematic and now run it with just a 12uf after the .390mh inductor. no resistor either end, just the Lpad to dampen the tweeter output. so far 3 days and no distortion or burnout. I still think I could do a bit more to dampen the 3.5khz woofer breakup, but is only noticeable on CD's such as Adele's 21 with her high frequency voice.

Wolf, S7, Rabbits any suggestions? I tried to roll off the woofer with a bigger cap parallel to the zobel but it sounded a bit compressed. Go for a bigger L2 inductor? or just leave it as it is not that noticeable as I said only on certain CD's.

cheers guys
You should place a 0.1-0.22uF cap across the L2, and then a 1.5uF cap across the woofer, and you may no longer require the zobel.

It'll look like this (1.5uF cap not pictured):
Click the image to open in full size.

Give that a shot,
Wolf
__________________
Photobucket picture pages: http://photobucket.com/Wolf-Speakers_and_more
Writeups/thoughts/blogs: http://techtalk.parts-express.com/blog.php?u=4102
  Reply With Quote
Old 23rd March 2013, 07:30 AM   #147
system7 is offline system7  United Kingdom
diyAudio Member
 
system7's Avatar
 
Join Date: Aug 2008
@chefsinc. I can't quite get a handle on the filter for your SB Acoustics 8" woofer yet, but I just have a feeling that a Zobel around 7.5R and 10uF might be as good as it gets after the baffelstep. That's what you do with mildly over lively woofers in general. It's not tidy or rolled-off enough to just run, almost unfiltered, on a mere coil like the famous polycone Vifa P13.

wolfie's tank is a nice idea, and quite Paul Carmody "Overnight Sensations", but sends the phase at 7.5kHz to hell in my opinion. I still don't think notches work with series circuits. What's wrong with a Zobel here?

Click the image to open in full size.

Remember samadhi's lovely Stellan speaker above? Uses a Visaton W130S 5" Woofer and a Visaton TW70 Cone Tweeter. I like the TW70, it just FITS with paper woofers very nicely. Quite how nicely, you are about to see!

I knocked up this 2nd order series crossover with the indispensable attenuator on the tweeter:

Click the image to open in full size.

Here's the bit that made me fall off my chair in amazement.

Click the image to open in full size.

THAT, my friends, is PERFECT phase alignment. I just can't do anything as good with a parallel circuit. It just doesn't get better than that!
Attached Images
File Type: jpg stellan_by_samadhi.jpg (63.1 KB, 12 views)
File Type: png Stellan_Crossover.PNG (5.9 KB, 37 views)
File Type: png Stellan_Phase.PNG (22.3 KB, 33 views)
File Type: png Stellan_Freq_Resp.PNG (19.1 KB, 43 views)
File Type: png Stellan_Transfer_Function.PNG (17.9 KB, 46 views)
__________________
Good Night, and Good Luck. Best regards from Steve in Portsmouth, UK.
  Reply With Quote
Old 23rd March 2013, 08:44 AM   #148
diyAudio Member
 
wolf_teeth's Avatar
 
Join Date: Mar 2006
Location: Indiana
Blog Entries: 1
The phase goes bouncy at 8k in my sim.

Click the image to open in full size.

You're already far off the xover point, so it doesn't matter as much. This is w/o the 1.uF to limit the top-end woofer ripple.

Depends on your compromise.

And to say it's Paul Carmody-esque is funny- I've been doing this a lot longer than he has.

You can notch an SXO.
Wolf
__________________
Photobucket picture pages: http://photobucket.com/Wolf-Speakers_and_more
Writeups/thoughts/blogs: http://techtalk.parts-express.com/blog.php?u=4102
  Reply With Quote
Old 23rd March 2013, 11:02 AM   #149
diyAudio Member
 
chefsinc's Avatar
 
Join Date: Jun 2006
Location: Perth Suburbs. Western Australia
cheers guys I will give it a shot once back from my Holiday.
  Reply With Quote
Old 31st March 2013, 08:36 PM   #150
system7 is offline system7  United Kingdom
diyAudio Member
 
system7's Avatar
 
Join Date: Aug 2008
This post is a DOOZY, though I say so myself!

Have a look at this well behaved bookshelf design, using Visaton Drivers:

Click the image to open in full size.

Do you notice something about the ratios of the components? Remember that crossover frequency is Root 1/LC, and impedance is Root L/C. Crossover at 3.5kHz, and phase and flatness is exemplary. It's a 2:1 ratio of course.

Now here is a well behaved 3.5 kHz Series crossover:

Click the image to open in full size.

This has a 3:1 ratio.

See, it's adding in some bafflestep inductance and making it into a standmounter that skews the values so you can't see the pattern. BTW, you can make most stuff fit just by adjusting the red resistors. Isn't that COOL!
Attached Images
File Type: png SC_Parallel_Bookshelf.PNG (8.4 KB, 14 views)
File Type: png SC_Series_Bookshelf.PNG (7.4 KB, 17 views)
__________________
Good Night, and Good Luck. Best regards from Steve in Portsmouth, UK.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
series components in crossover REALFLEO Multi-Way 6 28th December 2006 10:53 PM
AR series crossover pei Multi-Way 3 9th December 2006 02:00 PM
Series crossover problem F1 FAN Multi-Way 22 27th June 2006 11:09 PM
B&W 801 matrix series 2 crossover mod help musicalmaniac Multi-Way 0 10th January 2005 01:17 AM
Series crossover question Peter Daniel Multi-Way 9 29th May 2003 07:43 AM


New To Site? Need Help?

All times are GMT. The time now is 03:30 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2