Starting an Active Open Baffle design. - Page 13 - 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 7th July 2013, 05:53 PM   #121
Rudolf is offline Rudolf  Germany
diyAudio Member
 
Rudolf's Avatar
 
Join Date: Mar 2003
Location: Germany
Quote:
Originally Posted by DrDyna View Post
I drug one of them outside this morning so I could get the MiniDSP set properly and the xml file saved so I could load it into the second one when it gets here.
Here's the result and what I did to get it. .... What do you guys think?
If this http://www.diyaudio.com/forums/attachments/multi-way/358275d1373039991-starting-active-open-baffle-design-minidsp-screen1.jpg
are your electrical correction filters, you basically got it wrong IMHO.

DrDyna, you just have become the victim of my sadistic urge to make things perfect. It’s not your fault – you just stood in the way.
Please let me explain why your approach is basically wrong. I will start with an EDGE sim of a 100x60 cm OB and a 15” driver centered at 27 cm from the bottom:
Dipole EQ01.gif
You can easily see how all dipole baffles start with a 6 dB acoustic high pass at low frequencies. Our example keeps this high pass to its dipole peak at 500 Hz (which is recognizable because the dipole dip at 1 kHz is clearly visible). It is very clear that the acoustic high pass will influence everything below 500 Hz.

Next simulation is for a real world Visaton BG 40 in this baffle:
Dipole EQ02.gif
It shows the 0, 30 and 60° response. You will notice how the real driver changes the response to a degree, but basically the baffle response rules. Note how the proportions of both diagrams are kept equal.

Next I have stretched the 0° response a bit along the x-axis to give room for the next filter variations:
Dipole EQ03.gif
I now consecutively applied a 6 dB electrical low pass at 50 Hz (black line) and an additional 12 dB electrical low pass at 100 Hz (red line). The non-EQ response is dashed:
Dipole EQ04.gif

Note how the acoustic high pass below 50 Hz is almost unchanged – it is almost 12 dB!
Note how the 6 dB electrical low pass not only straightens out the response (you can easily imagine to equalize it with some parametric EQs into the green range). It applies a falling acoustic response curve at the upper end of the woofer range too. Same for the 12 dB electrical low pass. Its influence at the lower end is small, but it works full tilt as a response leveler around 100 Hz and a cut-off above 100 Hz. Effectively nobody applied a real electrical low pass at 200 Hz, but we have got a working acoustical 20 dB low pass above 200 Hz.

Things get even worse when you install a passive EQ:
Dipole EQ05.gif
You remember the black dashed line? The continuous black line is the passive response with the 12 dB electrical low pass shown in the picture. You achieve an amazing 24 dB acoustic high pass below 50 Hz and a comparatively decent acoustic 12 dB low pass above 300 Hz.

What do I want you to understand in all this mess? It is the wildly different approach to filter design you have to take with dipoles. Symmetrical electrical crossovers are wrong in almost every case. You need to shape the acoustical response for every driver range to the intended form before you should even think about the application of over-all filters.

In my own case I have implied parametric EQs for the individual drivers only – there is not a single EQ for the summed response!

Rudolf
__________________
www.dipolplus.de
  Reply With Quote
Old 7th July 2013, 06:48 PM   #122
DrDyna is offline DrDyna  United States
diyAudio Member
 
DrDyna's Avatar
 
Join Date: May 2007
Blog Entries: 3
Quote:
Originally Posted by Rudolf View Post
If this diyAudioare your electrical correction filters, you basically got it wrong IMHO.

DrDyna, you just have become the victim of my sadistic urge to make things perfect. It’s not your fault – you just stood in the way.
Please let me explain why your approach is basically wrong. I will start with an EDGE sim of a 100x60 cm OB and a 15” driver centered at 27 cm from the bottom:
Attachment 358637
You can easily see how all dipole baffles start with a 6 dB acoustic high pass at low frequencies. Our example keeps this high pass to its dipole peak at 500 Hz (which is recognizable because the dipole dip at 1 kHz is clearly visible). It is very clear that the acoustic high pass will influence everything below 500 Hz.

Next simulation is for a real world Visaton BG 40 in this baffle:
Attachment 358638
It shows the 0, 30 and 60° response. You will notice how the real driver changes the response to a degree, but basically the baffle response rules. Note how the proportions of both diagrams are kept equal.

Next I have stretched the 0° response a bit along the x-axis to give room for the next filter variations:
Attachment 358639
I now consecutively applied a 6 dB electrical low pass at 50 Hz (black line) and an additional 12 dB electrical low pass at 100 Hz (red line). The non-EQ response is dashed:
Attachment 358640

Note how the acoustic high pass below 50 Hz is almost unchanged – it is almost 12 dB!
Note how the 6 dB electrical low pass not only straightens out the response (you can easily imagine to equalize it with some parametric EQs into the green range). It applies a falling acoustic response curve at the upper end of the woofer range too. Same for the 12 dB electrical low pass. Its influence at the lower end is small, but it works full tilt as a response leveler around 100 Hz and a cut-off above 100 Hz. Effectively nobody applied a real electrical low pass at 200 Hz, but we have got a working acoustical 20 dB low pass above 200 Hz.

Things get even worse when you install a passive EQ:
Attachment 358641
You remember the black dashed line? The continuous black line is the passive response with the 12 dB electrical low pass shown in the picture. You achieve an amazing 24 dB acoustic high pass below 50 Hz and a comparatively decent acoustic 12 dB low pass above 300 Hz.

What do I want you to understand in all this mess? It is the wildly different approach to filter design you have to take with dipoles. Symmetrical electrical crossovers are wrong in almost every case. You need to shape the acoustical response for every driver range to the intended form before you should even think about the application of over-all filters.

In my own case I have implied parametric EQs for the individual drivers only – there is not a single EQ for the summed response!

Rudolf
I'm really making an honest attempt to read and digest all of this, but it's probably going to take me some time to figure out what I should change.

What I did when I measured was to measure each driver (with the others turned off) with the crossover engaged, and EQ a flat response within each band that I was trying to get each driver to cover.

Right now, if I play only one driver (the woofers together, or the midrage or the tweeter) they play relatively flat in the range that I've chosen.

Is the idea that the ranges I've picked for each driver should be different so that they are more "symbiotic" with naturally occurring dipole peaks and dips, or is the idea that what I've done is completely incorrect?

I really want to get this right, it's a learning experience for me, and I appreciate the time you spend talking about this with me

Edit: attaching measurements of the individual drivers. (not level matched)
Attached Images
File Type: jpg woofer.jpg (53.1 KB, 111 views)
File Type: jpg midrange.jpg (54.2 KB, 109 views)
File Type: jpg tweeter.jpg (47.5 KB, 109 views)

Last edited by DrDyna; 7th July 2013 at 06:57 PM.
  Reply With Quote
Old 7th July 2013, 08:10 PM   #123
Rudolf is offline Rudolf  Germany
diyAudio Member
 
Rudolf's Avatar
 
Join Date: Mar 2003
Location: Germany
Hi,
the individual curves of your drivers look decent enough - except a detail in the Neo response, which I will approach later. And the regions for each driver(s) are well choosen. I have no argument with them at all.

My curiosity is about the crossover regions, filter orders and phase relations mainly. Did you measure every two neighbouring regions together - in phase and one out of phase? It seems you are going with LR filters. Did you get the LR-typical 6 dB difference at the Xover frequency between single driver curves and summing curves? Deep notches for the inverted case?
Have you done off-axis measurements which you would show? What happened to the high-pass of the midrange driver? It doesn't work lower than -20 dB?

For the Neo3 I don't see any significant response dip at 7 kHz. This would be indicative for too much baffle area seen by the Neo3. Or did you apply some EQ there already? Which should not be done without parallel off-axis response control.

I hope that I don't come across too demanding or smart-alecky. Hopefully I'm only squeezing the best out of you.

Rudolf
__________________
www.dipolplus.de
  Reply With Quote
Old 7th July 2013, 08:52 PM   #124
DrDyna is offline DrDyna  United States
diyAudio Member
 
DrDyna's Avatar
 
Join Date: May 2007
Blog Entries: 3
Quote:
Originally Posted by Rudolf View Post
Hi,
the individual curves of your drivers look decent enough - except a detail in the Neo response, which I will approach later. And the regions for each driver(s) are well choosen. I have no argument with them at all.

My curiosity is about the crossover regions, filter orders and phase relations mainly. Did you measure every two neighbouring regions together - in phase and one out of phase? It seems you are going with LR filters. Did you get the LR-typical 6 dB difference at the Xover frequency between single driver curves and summing curves? Deep notches for the inverted case?
Have you done off-axis measurements which you would show? What happened to the high-pass of the midrange driver? It doesn't work lower than -20 dB?
I'm not sure about the high pass on the mid, really I just picked where I wanted it to work and went with it. As far as summing, I didn't have to do too much, except for the midrange, which I applied a small amount of delay (0.15ms) and watched the dip in the response rise until it was where it looked good, then I stopped adding delay.

I do probably need to do more work, and that's what I'm looking for. I'm at a spot right now where it seems to be "ok", but I'm really looking for hints and tweaks to make it better.

Quote:
Originally Posted by Rudolf View Post
For the Neo3 I don't see any significant response dip at 7 kHz. This would be indicative for too much baffle area seen by the Neo3. Or did you apply some EQ there already? Which should not be done without parallel off-axis response control.
The responses in the screenshots I posted are after EQ. They represent the chosen XO frequencies, (48db/oct, although I've been trying some 24). Some of it was quite peaky before EQ. I really try to do most of my EQ through subtraction, what I did was input the range of interest for each driver into REW and let it generate an EQ curve that gave me a reasonable response, then I added a little "flavor" or my own to the final EQ, such as a little dip around 6-7k to get rid of some sibilance. (that's not in the screenshots)

Quote:
Originally Posted by Rudolf View Post
I hope that I don't come across too demanding or smart-alecky. Hopefully I'm only squeezing the best out of you.

Rudolf
Oh, no no no, please...I can handle criticism, that's how these things get better! I will honestly try to absorb any information you give, and I will try as best I can to make a good attempt to make any corrections or implement any advice.

You won't hurt my feelings at all. <3 In fact, if I make major changes to anything, I keep all the xml files for the MiniDSP so that I can inspect them and try to learn through listening what each change means. I'm totally fine with having a minidsp-rudolf.xml So far, any changes that I've made that were a result of feedback have been mostly positive, so, by all means...

Give me sh*t!

  Reply With Quote
Old 7th July 2013, 10:14 PM   #125
DrDyna is offline DrDyna  United States
diyAudio Member
 
DrDyna's Avatar
 
Join Date: May 2007
Blog Entries: 3
Also, after reading over this a little more, I guess what I need to do is do some off-axis measurements. I'll try to do some tomorrow.
  Reply With Quote
Old 7th July 2013, 11:29 PM   #126
Rudolf is offline Rudolf  Germany
diyAudio Member
 
Rudolf's Avatar
 
Join Date: Mar 2003
Location: Germany
Quote:
Originally Posted by DrDyna View Post
then I added a little "flavor" or my own to the final EQ, such as a little dip around 6-7k to get rid of some sibilance. (that's not in the screenshots)
....
Give me sh*t!
Some observations of my Neo3. First a comparison of my equalized system (free hanging Neo3 on top from 2.5 kHz upward). On-axis measurement on tweeter height from 40 cm distance (top) and from listening position (bottom):
Dipole EQ06.gif
You see how the 7 kHz dip gets more shallow with distance/reflections and off-axis. At 30° the response is quite smooth without any EQ:
neo3 pur response 0-90.gif
Measurements are 0-90° at 10° intervals. 30° is pink.
If REW is trying to equalize that 6-8 kHz hole - don't allow it. This EQ will be a severe source of sibilance.
Listen to the response hole without thinking about the curve.
__________________
www.dipolplus.de
  Reply With Quote
Old 8th July 2013, 07:50 AM   #127
diyAudio Member
 
gainphile's Avatar
 
Join Date: Dec 2007
Location: Melbourne the sunny city!
Quote:
Originally Posted by DrDyna View Post
I had thought about raising it up high, I remember seeing the photo of when you measured yours and it was up on an ironing board

So, I guess it should be up for only the mid measurement? I remember Rudolf saying that the woofers should be near the ground, a concept I'm familiar with, which is why I left it where it was.

Out of curiosity, what sort of effect will measuring it like I did have on the mid?
Yes, the ironing board

Correct, only for midrange. Especially to find that pesky frequency, depth, and Q of that midrange dipole peak!

For tweeter, indoor with gating of 4-6ms is good enough.

For woofer, at under 100hz I don't think you can get any farfield useful, even raised 2m. I just do a nearfield (1-2cm at mouth), indoor.
__________________
http://gainphile.blogspot.com
  Reply With Quote
Old 8th July 2013, 12:31 PM   #128
DrDyna is offline DrDyna  United States
diyAudio Member
 
DrDyna's Avatar
 
Join Date: May 2007
Blog Entries: 3
Quote:
Originally Posted by Rudolf View Post
Some observations of my Neo3. First a comparison of my equalized system (free hanging Neo3 on top from 2.5 kHz upward). On-axis measurement on tweeter height from 40 cm distance (top) and from listening position (bottom):
Attachment 358684
You see how the 7 kHz dip gets more shallow with distance/reflections and off-axis. At 30° the response is quite smooth without any EQ:
Attachment 358685
Measurements are 0-90° at 10° intervals. 30° is pink.
If REW is trying to equalize that 6-8 kHz hole - don't allow it. This EQ will be a severe source of sibilance.
Listen to the response hole without thinking about the curve.
Yeah, at first I allowed it to EQ the tweeter flat, but the sibilance drove me crazy..they were hissing at me! So I went into the DSP and started bypassing filters until the dip came back, and I even added a little more dip by creating a 1db drop from 6k to 7.5k. Now they just sparkle, and Eva Cassidy doesn't make me want to jump from the window

It would seem from your measurements though, I could actually just adjust the listening axis a few degrees and get the same effect. That is on the list of stuff to try today

I have a lazy susan downstairs, I'm going to try and do some off-axis measurements once the sun is fully up today.

Quote:
Originally Posted by gainphile
Yes, the ironing board

Correct, only for midrange. Especially to find that pesky frequency, depth, and Q of that midrange dipole peak!

For tweeter, indoor with gating of 4-6ms is good enough.

For woofer, at under 100hz I don't think you can get any farfield useful, even raised 2m. I just do a nearfield (1-2cm at mouth), indoor.
Awesome, thanks. I'll see what types of measurements I can get today, and I also plan to try some other measuring software (tolvan) to see if it can tell me anything additional.

Edit: Rudolf, one other thing, do you have a version of your page that's in english? Google translate fails to translate it

Last edited by DrDyna; 8th July 2013 at 12:37 PM.
  Reply With Quote
Old 8th July 2013, 12:54 PM   #129
Juhazi is offline Juhazi  Finland
diyAudio Member
 
Juhazi's Avatar
 
Join Date: Mar 2007
Location: Jyväskylä
Drdyna, based on my experience, you should use milder (12/24dB/oct) xo slopes. They come sort of naturally with dipoles and they help to mix drivers' polar responses more evenly.

When measuring outdoors - set the speaker on a stand and on grass to minimize "floor" reflection! 3-5' is good distance, IRgate 3ms for tweeter -9ms for mid-20ms for the low end of the mid or woofer section. Tell us what IRgate you used in the measuremnts that you show us. Generally for measuring the speaker alone - use short IR (3-12ms) and when checking/ eqing room response - use long gate (60-200ms)

Otherwise - keep going! About the tweeter - it's dispersion is important about how it sounds. I like mine when direct is flat and 15¤ response rolls off a little. But the amount/degree of this depens on your tweeter's type/characteristics (and your ears and room reflections).
__________________
AES Associate Member / My DIY speaker history: -74 Philips 3-way, -82 Hifi 85B, -07 Zaph L18, XLS10+PR/Hypex, -08 CSS125FR, -08 Hifitalo AW-7, -08 TangBand FR, -09 MarkK ER18DXT, -13 PPSL470, -13 AINOgradient
  Reply With Quote
Old 8th July 2013, 12:59 PM   #130
DrDyna is offline DrDyna  United States
diyAudio Member
 
DrDyna's Avatar
 
Join Date: May 2007
Blog Entries: 3
Quote:
Originally Posted by Juhazi View Post
Drdyna, based on my experience, you should use milder (12/24dB/oct) xo slopes. They come sort of naturally with dipoles and they help to mix drivers' polar responses more evenly.
I actually changed the slopes to 24db/oct yesterday, but I haven't measured it yet today. We shall see!

Quote:
Originally Posted by Juhazi View Post
When measuring outdoors - set the speaker on a stand and on grass to minimize "floor" reflection! 3-5' is good distance, IRgate 3ms for tweeter -9ms for mid-20ms for the low end of the mid or woofer section. Tell us what IRgate you used in the measuremnts that you show us. Generally for measuring the speaker alone - use short IR (3-12ms) and when checking/ eqing room response - use long gate (60-200ms)

Otherwise - keep going! About the tweeter - it's dispersion is important about how it sounds. I like mine when direct is flat and 15¤ response rolls off a little. But the amount/degree of this depens on your tweeter's type/characteristics.
I'm hoping to be able to use Tolvan's Sirp today to do some gated measurements, I'm looking forward to what I can learn from it.
  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
Active Crossover for Open Baffle 2-way mortron Analog Line Level 17 9th December 2012 09:40 PM
Open Baffle active Xo/EQ/Amp design Zicdave Multi-Way 0 15th May 2012 10:55 AM
Please recommend a 4 or 5" FR for Open baffle w/ active bass augmentaion JandG Full Range 19 9th July 2010 06:54 AM
Single Ended Active Crossovers for Open Baffle Speakers ljfont Solid State 7 9th October 2009 12:24 AM
Open baffle design help fin1bxn Full Range 4 5th February 2007 04:07 AM


New To Site? Need Help?

All times are GMT. The time now is 04:33 PM.


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