Beyond the Ariel - Page 288 - 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 16th December 2007, 11:53 PM   #2871
diyAudio Member
 
john k...'s Avatar
 
Join Date: Aug 2004
Location: US
Quote:
Originally posted by mige0
Hi





Its not possible with simple 6 dB filters ? what else is needed ?


Hi Michael,

Sorry it has taken so long to get back to you. 2nd order LP are required. I have been working on this today and will post more later.
__________________
John k.... Music and Design NaO Dipole Loudspeakers.
  Reply With Quote
Old 17th December 2007, 12:14 AM   #2872
diyAudio Member
 
Lynn Olson's Avatar
 
Join Date: Jul 2004
Location: Northern Colorado
Data from the Shred Muzik site about the Madison drivers. The efficiency was calculated from Fs, Vas, and Qes with this handy site. Note: 1% conversion efficiency is roughly equal to 92 dB/metre/watt into a half-space.

10 Knight
-------------------
Fs=59 Hz
Vas=0.626
Qes=0.79
Qts=0.75
Xmax=0.25
SD=63.61
Eff=0.44%
$69

10 Warrior
-------------------
Fs=65 Hz
Vas=0.555
Qes=0.915
Qts=0.867
Xmax=0.25
SD=63.61
Eff=0.45%
$49

12 Executioner
-------------------
Fs=52 Hz
Vas=1.505
Qes=0.569
Qts=0.547
Xmax=0.375
SD=95.033
Eff=1.0%
$179

12 Knight
-------------------
Fs=50 Hz
Vas=2.25
Qes=0.665
Qts=0.64
Xmax=0.25
SD=95.033
Eff=1.14%
$89

12 Warrior
-------------------
Fs=41 Hz
Vas=4.12
Qes=0.681
Qts=0.649
Xmax=0.25
SD=95.033
Eff=1.13%
$59

15 Executioner
-------------------
Fs=46 Hz
Vas=3.69
Qes=0.753
Qts=0.692
Xmax=0.375
SD=153.93
Eff=1.29%
$229

15 Knight
-------------------
Fs=35 Hz
Vas=4.48
Qes=1.15
Qts=1.006
Xmax=0.25
SD=153.93
Eff=0.45%
(something funny going on here, probably an error in the other measurements)
$129

15 Warrior
-------------------
Fs=36 Hz
Vas=8.29
Qes=0.888
Qts=0.826
Xmax=0.25
SD=143.13
Eff=1.18%
$69

Comment: I think the Xmax figures are fictitious, and might be the max-destruction-limit instead. Maybe not even that. The only real way to know is to measure the difference between the gap height and the length of the voice-coil. If these two numbers are the same - which is common in quitar speakers - then Xmax is zero, or in other words, there is no linear region!

With guitar speakers, this is actually a plus, not a negative, since the tonal character of the speaker then changes significantly with drive level (T/S measurements are taken at a very low level and may not be affected). Since the Madison site appears to be aimed at heavy-metal bands, it's probably safe to conclude low distortion is not a major design priority.
  Reply With Quote
Old 17th December 2007, 12:44 AM   #2873
diyAudio Member
 
john k...'s Avatar
 
Join Date: Aug 2004
Location: US
Here is a little more of what I have considered.

It comes down to how you view the dipole roll off. Is it a 1/2 empty or 1/2 full thing? If you look at the roll off as decreasing efficiency as frequency goes down it may appear 1/2 empty. If you look at the dipole effect as efficiency increasing with increasing frequency it's 1/2 full.

I think the basic consideration is just the physics: How much volume displacement is required to produce the required SPL at the low frequency cut off and how are you going to get that. This will not change regardless of how you flatten the response (line level active or passive eq, or rolling off augmentation drivers). This is really a driver/baffle consideration. Properly specified ether approach should result in the same driver/baffle compliment and the same power requirement at low frequency. Then, assuming identical, multiple drivers, rolling off some of the drivers as the frequency rises to equalize the response is really just throwing away the increasing efficiency with rising frequency. It will place greater excursion demands on the main driver as the frequency increases and place greater demand on the power amplifier as well. Running all the drivers full range and using line level eq on the other hand is superior here since it means the increase in efficiency of the multiple driver configuration and the increase in dipole efficiency as frequency rises is retained and less amplifier power is required at the upper bass frequencies to produce the same SPL. With a given driver/baffle complement the power required at the system low cut off will be the same. Since IM distortion is basically related to the low frequency excursion, it will also remain about the same for both configurations. However, running all the drivers full range (i.e. up to 200 or 300 Hz) will reduce excursion in the upper bass range and therefore should result in lower HD at those frequencies. Running all the drivers full range also has the advantage that all drivers are in phase at all frequencies.

On the other hand, the idea of using a low Q driver for the upper ranges and a high Q driver with LP filter to augment the low frequency is really just turning the woofer system into a two-way low frequency system. The lowest frequencies will be reproduced primarily by the high Q driver(s) and the upper bass by the low Q driver(s). Since the low Q drivers won't contribute too much at low frequency this will place greater demands of the high Q driver(s) requiring either greater excursion or even more area. At the same time, the upper bass being handled by only the low Q drivers will requires higher excursions than if all drivers were full range with the possibility of increased HD (just as was the case for identical drivers). But since the low Q drivers will have limited excursion at lower frequencies there is a potential for reduced IM in the upper bass. How the LP filtering of the high Q driver is performed will determine whether advantage is taken of the increasing dipole efficiency with rising frequency.

Of course, the final result will depend on the exact driver/baffle configuration but no matter how it is set up the most judicious application of amplifier power is obtained when we take advantage of the rising efficiency of the dipole response and apply line level active or passive eq to minimize the power requirements on the amplifier. If the driver/baffle compliment is correctly chosen for the required low frequency max SPL and excursion limitations the line level eq should always be though of not as boosting the power applied at low frequency but rather reducing the power required as the frequency rises. This is because once the driver/baffle compliment is set the low frequency power requirements are as well.
__________________
John k.... Music and Design NaO Dipole Loudspeakers.
  Reply With Quote
Old 17th December 2007, 01:29 AM   #2874
diyAudio Member
 
Lynn Olson's Avatar
 
Join Date: Jul 2004
Location: Northern Colorado
Of course, direct-radiator drivers are not constant excursion devices. If the response is flat, they are constant-acceleration, and excursion increases at a rate of 12 dB/octave as the frequency is lowered. Thus, excursion in the upper bass and midrange is visibly quite low. Distortion mechanisms then shift from excursion (which dominates at low frequencies) to magnetic and inductive nonlinearities. At the highest frequencies, cone breakup mechanisms dominate.

Thus, three distortion regions, depending on frequency: lowest frequencies, excursion and resulting nonlinearities as the VC traverses the gap structure and its nonuniform magnetic fields, combined with spider nonlinearities. At higher upper-bass and lower-mid frequencies, magnetic nonlinearities (nonlinear Le) dominate, so refined polepiece-shielding methods are beneficial here. At 1 kHz and higher, diaphragm breakup becomes more important, and narrowband peaks in distortion appear.

The other drawback of a large array going into the lower mid-frequencies, of course, is dispersion. Whether the array is vertical or square, dispersion starts to matter in the 500 Hz ~ 1 kHz region, especially in the crossover region, where lobing can affect tonality and voice quality.

The point of matching the radiating area to the frequency is twofold: offset the very rapid increase in excursion below the baffle peak (particularly if compensating boost EQ is applied), and retain favorable dispersion characteristics above the baffle peak.
  Reply With Quote
Old 17th December 2007, 10:01 AM   #2875
soongsc is offline soongsc  Taiwan
diyAudio Member
 
soongsc's Avatar
 
Join Date: Mar 2005
Location: Taiwan
Quote:
Originally posted by Lynn Olson
...
Thus, three distortion regions, depending on frequency: lowest frequencies, excursion and resulting nonlinearities as the VC traverses the gap structure and its nonuniform magnetic fields, combined with spider nonlinearities. At higher upper-bass and lower-mid frequencies, magnetic nonlinearities (nonlinear Le) dominate, so refined polepiece-shielding methods are beneficial here. At 1 kHz and higher, diaphragm breakup becomes more important, and narrowband peaks in distortion appear.

...
This is quite interesting. I have a wide range driver that never sounded right with a shorting ring. But I think it may not have been implemented with the proper calculations.
__________________
Hear the real thing!
  Reply With Quote
Old 17th December 2007, 01:10 PM   #2876
diyAudio Member
 
john k...'s Avatar
 
Join Date: Aug 2004
Location: US
Quote:
Originally posted by Lynn Olson


The other drawback of a large array going into the lower mid-frequencies, of course, is dispersion. Whether the array is vertical or square, dispersion starts to matter in the 500 Hz ~ 1 kHz region, especially in the crossover region, where lobing can affect tonality and voice quality.


Yes, agreed, and this is one reason, along with IM distortion and others, why we build 3-way systems. In each band we must conside the max SPL required at the low frequency cut off of that band and the dispersion characteristics at the upper limit, along with the impact of distortion. This is true for any system speaker system.

The point of matching the radiating area to the frequency is twofold: offset the very rapid increase in excursion below the baffle peak (particularly if compensating boost EQ is applied), and retain favorable dispersion characteristics above the baffle peak.

I have no problem with this. Again, it's not different than what is done for a conventional speaker. But there are several ways to accomplish the task. One it to combine the drivers in such a way as to achieve flat response without equalization, that is, constant sensitivity vs. frequency (regardless of speaker type, conventional, dipole or other). The other is to design in such a manner as to take advantage of the frequency dependent sensitivity and apply line level equalization to compensate for the frequency dependent sensitivity, yielding flat response. For get dipole systems and consider a simple 2-way box speaker. Maybe the woofer has a 2Pi sensitivity of XXdB and the tweeter YYdB. If we apply full baffle step compensation in a constant sensitivity system the system sensitivity will be XX-6 dB. Since the peak above the baffle step in a conventional speaker is about 3 dB we would be throwing away about 9 dB sensitivity in the midrange around that peak. For the tweeter we would need to add an L pad or some other attenuation throwing away YY - (XX-6) dB of sensitivity. Why? We can retain this sensitivity variation and compensate with line level Eq as I proposed in my Hybrid Design article: http://www.musicanddesign.com/HybridDesign.html and save a lot of amplifier capability in reserve as shown in red (see articale for details). Click the image to open in full size.

Now some audiophile purists may object to having a line level active (or passive) equalization stage between the amp and preamp, but when considering the dynamic capabilities of the overall system it just makes sense.

Lynn, you keep on referring to the low frequency eq as boost. This was my earlier point about full vs. empty. Once you determine the driver compliment for the SPL requirements at the low frequency cut off consider that as the 0dB point. It is the defining point on the sensitivity curve. Then look at the eq as attenuation, not boost, to compensate for the increasing sensitivity as frequency increases (dipole or baffle step), thus holding amplifier potential in reserve. The same argument applies to using a high efficiency tweeter with lower efficiency woofer (like my design example). you can attenuate the input to the tweeter after the amplifier, dumping power in resistors or what ever, or correctly attenuate the input to the amplifier using line level eq so that on what is required by the tweeter is delivered.
__________________
John k.... Music and Design NaO Dipole Loudspeakers.
  Reply With Quote
Old 17th December 2007, 09:10 PM   #2877
Radian is offline Radian  Germany
diyAudio Member
 
Join Date: Jun 2004
Location: Wiesbaden
Hi Lynn,

I have not read the whole thread, so I don't know if you would consider building a horn midrange with "out of production" drivers.
Probably the best mass production driver ever made was the the Yamaha JA6681 and it's predecessor. Meyer Sound used them with a "modification" that made them more durable but stole them some fidelity.
I bought 2 of them after speaking with three people that have used them and compared them against other drivers. Two guys could compare them directly to their own TAD 4001 and said the Yam sounded smoother and more natural. It's a 1inch exit driver with
107db 1W1M. It can be fit with a phenolic (6603)or the original aluminum diaphragm. The aluminum diaphragm is not detectable as such, it just sounds uncolored, in the right horn of course. The aluminum version goes to 12000hz and then gradually goes down and the phenolic version goes to 8000hz. What is really amazing is that it can be crossed as low as 350hz in a big enough horn, which as far as I know is not heard of from any other 1inch driver.

Greets,
Klaus
  Reply With Quote
Old 17th December 2007, 09:14 PM   #2878
mige0 is offline mige0  Austria
diyAudio Member
 
mige0's Avatar
 
Join Date: Jan 2005
Location: Austria, at a beautiful place right in the heart of the Alps.
Hi


Quote:
Originally posted by Lynn Olson
Of course, direct-radiator drivers are not constant excursion devices. If the response is flat, they are constant-acceleration, and excursion increases at a rate of 12 dB/octave as the frequency is lowered.

To put this " excursion increases at a rate of 12 dB/octave " into perspective:


Taking a BMS 8N315 for example and inserting Sd=222 and Xmax=8
we get ~ 120 dB at 200 Hz.
( calculating with the SL form / http://www.linkwitzlab.com/spl_max1.xls / D set to 483 )

To keep 120 db at 100 Hz which is 1 octave down - we would need Sd=1000 at Xmax=8
To keep 120 db at 50 Hz which is 2 octaves down - we would need Sd=4000 at Xmax=8

Meaning we need 4 times the area each octave down to not run into intermodulation distortion .
Also meaning we need at least 12 db / octave high pass filter to not run into intermodulation distortion when setting the XO point at the lowest possible SPLmax frequency .


Or, if seen from JohnK's perspective once the SPLmax requirements at the lowest frequency are met, intermodulation ( or linear-max excursion likewise ) isn't a problem for any higher frequency - given you don't go more extreme than 4 times Sd each octave ( which is hardly ever the case ).

Greetings
Michael
  Reply With Quote
Old 17th December 2007, 10:08 PM   #2879
diyAudio Member
 
john k...'s Avatar
 
Join Date: Aug 2004
Location: US
Quote:
Originally posted by Lynn Olson
Of course, direct-radiator drivers are not constant excursion devices. If the response is flat, they are constant-acceleration, and excursion increases at a rate of 12 dB/octave as the frequency is lowered.
Just a minor point. First, think volume displacement not excursion. Then also think for an open baffle operating in the dipole range it increase at 18dB/octave for flat response. So for 2 octaves at constant excursion the area would have to increase 64 times. Or if we set the area based on 8 mm and 120 dB at 50 Hz then at 200 Hz rhe excursion is down by a factor of 64. But 120 dB at 50 Hz, as I indicated in a previous post, is going to be rather intrusive in size.

Also, the IM is a function of the low frequency excuirsion, not the high frequency and the frequency ratio. That is, for a driver radiating two frequencies, Flow and Fhigh the IM will be dependent on the excursion the driver would experience id reproducing Flow alsone and the ratio of Fhigh/Flow. An article I wrote on IM distortion can be downloaded from my old web site at http://www.geocities.com/kreskovs/Doppler1.html
__________________
John k.... Music and Design NaO Dipole Loudspeakers.
  Reply With Quote
Old 17th December 2007, 10:36 PM   #2880
Banned
 
Join Date: Feb 2004
Location: Portal 2012
Quote:
Originally posted by Lynn Olson
Data from the Shred Muzik site about the Madison drivers. The efficiency was calculated from Fs, Vas, and Qes with this handy site. Note: 1% conversion efficiency is roughly equal to 92 dB/metre/watt into a half-space.

...

Comment: I think the Xmax figures are fictitious, and might be the max-destruction-limit instead. Maybe not even that. The only real way to know is to measure the difference between the gap height and the length of the voice-coil. If these two numbers are the same - which is common in quitar speakers - then Xmax is zero, or in other words, there is no linear region!

With guitar speakers, this is actually a plus, not a negative, since the tonal character of the speaker then changes significantly with drive level (T/S measurements are taken at a very low level and may not be affected). Since the Madison site appears to be aimed at heavy-metal bands, it's probably safe to conclude low distortion is not a major design priority.

I say all the numbers are wrong, not just sensitivity and x max = LOL I'm glad i got the 16 ohm drivers when they closed them out. Found some new Eminence NEO magnet 12's for really cheap/ I will use 12 per channel in my new panels.

Twenty four 12's with a little eq should move some air
  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



New To Site? Need Help?

All times are GMT. The time now is 02:39 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