Beyond the Ariel - Page 672 - 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 19th July 2010, 07:55 PM   #6711
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.
After the first few simus and measurements it may have downed to us that equalizing a CMP system isn’t exactly a task for "caps and coils".
Neither it is for all that fancy equipment that emulates those EQing by DSP power.
But hey – compared to the poor colleague souls in the RF department, who have to deal with the multipath issue we are lucky ! - having control over the source – or at least having *some* control over the source.

Taking as a first example to explore EQing of CMP systems, it might make sense to look at a frequency first, where there was the most consense to be found.
Its been unisono stated by Earl John and me:
“that it is NOT possible to EQ a null” – eg at a point of 100% destructive interference (though any of us had other things in mind of course)

For the 1ms delay doublet 100% destructive interference is the case at 1000Hz.

We remember what's been shown above (BROWN trace in particular):

Click the image to open in full size.


We see that for the first ms we have full signal, after that – silence..

As this CMP thing happens in the time domain rather than in the frequency domain we possibly best apply the EQ trick in the time domain – meaning – as obviously there is no more signal after the second part of the doublet arrives (due to 100% destructive interference) we simply push the signal after that time delay.
So having started out with a 1Vpp we have to apply a 2Vpp after 1ms in our case.

BUT
after 2ms – when the second part of the doublet arrives – we will be facing again 100% destructive interference
So – best we again push the signal by another 1Vpp to 3Vpp now.

Look here:

Click the image to open in full size.


I guess you get the picture – no ?

So, back to the question if a “total null” can be EQed:

Well – if we push and push and push - Yes, we can ! - but at one time we will run out of steam.
The interesting part here is that we may run out of steam because the signal to EQ *lasts too long* – not because we have to crank it up too much ! Quite a differnece !!!



Are you havin' fun guys?
We've just begun...

Michael
  Reply With Quote
Old 19th July 2010, 08:08 PM   #6712
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.
Looking at EQing of 100% constructive interference ( in our case at 1500Hz for example).

We remember what's been shown above (BROWN trace in particular):

Click the image to open in full size.


We see that for the first ms we have full signal, after that : +6dB


Now look here:

Click the image to open in full size.


As obviously here is tooo much of signal - we best “take a break” - then give some signal - and take another break - and so on and so forth...
At least - EQing of 100% constructive interference seems to be less *exhausting* than before - a part time job - so to say.



Michael

Last edited by mige0; 19th July 2010 at 08:13 PM.
  Reply With Quote
Old 19th July 2010, 08:44 PM   #6713
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.
And what about “below dipole peak” ?

Look here:

At 750Hz

Click the image to open in full size.
Click the image to open in full size.

At 500Hz

Click the image to open in full size.
Click the image to open in full size.

At 250Hz

Click the image to open in full size.
Click the image to open in full size.

At 125Hz

Click the image to open in full size.
Click the image to open in full size.






Michael
  Reply With Quote
Old 19th July 2010, 09:44 PM   #6714
diyAudio Member
 
Lynn Olson's Avatar
 
Join Date: Jul 2004
Location: Northern Colorado
As I suspected, it's a time-domain problem, and not just restricted to open-baffle speakers. Any loudspeaker with diffraction and reflections suffers from baffle or enclosure-induced time-domain artifacts, and frequency-based equalization only increases the time-domain error. As mentioned in the posts above, deep nulls are not correctable by any method, since full EQ would require kilowatts of power and would destroy the driver.

Frequency-based equalization is at it's best compensating for driver resonances, which are frequency-based problems, unlike the problems with baffles and enclosures, which are time-based. Different domains require different solutions.

Thank you, Michael, for doing this work. These are much deeper waters than I normally swim in, and I gladly leave it to others to do the swimming. I'm still curious about Gary Pimm-style resistive boxes, since a modest 6 dB of acoustic attenuation to the rear wave changes the whole picture, getting rid of the deep nulls and cancellations. I'm also happy that Gary Dahl is going ahead with his build, smoothing the way for my version.

Last edited by Lynn Olson; 19th July 2010 at 09:54 PM.
  Reply With Quote
Old 19th July 2010, 11:32 PM   #6715
diyAudio Member
 
john k...'s Avatar
 
Join Date: Aug 2004
Location: US
Quote:
Originally Posted by Lynn Olson View Post
As I suspected, it's a time-domain problem, and not just restricted to open-baffle speakers. Any loudspeaker with diffraction and reflections suffers from baffle or enclosure-induced time-domain artifacts, and frequency-based equalization only increases the time-domain error.
No offence intended Lynn, but that is typical of an old school view, and incorrect. If the response is minimum phase then minimum phase equalization corrects both frequency and time errors. The question is always whether or not the response is MP. The problem is that most multiway speakers are not MP because of the crossover thus MP amplitude eq can not correct the time domain errors created by the crossover. Certainly deep nulls can not be corrected. That is why I continue to say that a dipole response can not be used above the dipole peak. Above the dipole peak we must rely on driver directionality to LP filter the rear response so that deep nulls are not present.

However, you are correct in that the problems for a dipole or, more generally, an OB system, are no different than for a "box" speaker. In both cases the diffracted sound around or from the baffle edge is most closely related to the driver's 90 degree off axis response as it is the 90 degree off axis response that propagates over the baffle surface and is then diffracted at the edge. The difference between an OB system and a box system is the magnitude of the diffracted signal. For a box system it is a maximum of 1/2 the magnitude of the direct sound (which is why the baffle step is -6db) where as for an OB system the max is 1.0 which is why the OB response continues to roll off at 6dB/ocatve with decreasing frequency.

But the question of time domain error and amplitude eq comes down to whether or not the system is MP. And of course, whether or not the system is constant directivity. If the response is truly constant directivity and MP over a given listening window then MP Eq will correct both time and amplitude errors over the same window. This is another reason why constant directivity is becoming recognized as more and more important. Still, it is inherent with diffraction, particularly at higher frequencies, that it is going to be position dependent. Thus, there will always be some degradation in the time and frequency response that is uncorrectable. Or should I say that the window of "correctability" will narrow with increasing frequency.

Until we have true omni-directional, full range point source we are just going to have to live with that.
__________________
John k.... Music and Design NaO Dipole Loudspeakers.
  Reply With Quote
Old 20th July 2010, 02:07 AM   #6716
gedlee is offline gedlee  United States
diyAudio Member
 
Join Date: Dec 2004
Location: Novi, Michigan
Quote:
Originally Posted by john k... View Post
Until we have true omni-directional, full range point source we are just going to have to live with that.
Or elliminate/minimize the "uncorrectable" diffraction (all form of diffractrion are uncorrectable). Omni is not the way to go, you should know that -certainly dipoles are a step in the right direction. Just not quite a big enough step for me.
  Reply With Quote
Old 20th July 2010, 05:15 AM   #6717
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.
John, *if* you would have accepted meanwhile that CMP happens in "true dipole department" as well, I would have asked you to propose a "caps & coil" filter (one I possibly could build for verification by simu and measurement) to completely correct CMP in this department as well.

Michael

Last edited by mige0; 20th July 2010 at 05:30 AM.
  Reply With Quote
Old 20th July 2010, 05:15 AM   #6718
diyAudio Member
 
slowmotion's Avatar
 
Join Date: Nov 2002
Location: Norge
Quote:
Originally Posted by Lynn Olson View Post
I'm still curious about Gary Pimm-style resistive boxes, since a modest 6 dB of acoustic attenuation to the rear wave changes the whole picture, getting rid of the deep nulls and cancellations.
Would an increase of the front radiation have the same effect?
  Reply With Quote
Old 20th July 2010, 06:49 AM   #6719
diyAudio Member
 
Lynn Olson's Avatar
 
Join Date: Jul 2004
Location: Northern Colorado
Quote:
Originally Posted by slowmotion View Post
Would an increase of the front radiation have the same effect?
Yes. The problem is the magnitude of the front and rear phasors are equal, leading to deep nulls at zero frequency and at a number of comb frequencies above the dipole peak. I've made some crude sketches that show the vector relationships of an open baffle compared to a 6 dB resistive-mesh absorber.

The first (left) diagram shows the vector magnitudes as voltage relationships between the front and rear waves at a large distance (several meters) directly in front of the baffle. Notice the magnitudes are equal; the only thing that changes is the phase relationship. As the frequency moves downwards, the fixed transit delay from the center of the driver to the edge of the baffle becomes a smaller and smaller portion of the emitted wavelength. At very low frequencies, say, 1 decade below baffle peak, the front and rear waves are almost completely out of phase with each other, requiring substantial equalization if flat response is desired. In practice, diffraction attenuates the rear wave, but this is only significant at higher frequencies, primarily above baffle peak.

The second (center) diagram shows the vector magnitudes of a resistive mesh behind the driver, and for simplicity it is assumed that it has a frequency-independent loss of 6 dB. With practical acoustic absorbers, absorption is greatest at mid and high frequencies, and decreases at lower frequencies. Note that the deep nulls no longer occur, because the front and rear magnitudes are unequal, thus preventing total cancellation. The baffle peak is smaller as well, due to the decreased magnitude of the rear wave. In practice the baffle peak would be quite a bit smaller, possibly as small as 1 dB, because real-world absorbers have much better than 6 dB absorption from 500 Hz on up.

There is a continuum of possible ways to attenuate the rear wave, ranging from:

1) a totally absorbent infinite baffle (the classic mount-the-driver-in-a-door-to-another-room approach)

2) various types of resistive-vent boxes, such as the Dynaco A25 and the Planet10 Fonken

3) the Gary Pimm open-ended cardioid box

4) the proposed resistive mesh, a series of concentric mesh cylinders that are open at both ends and covered with thick felt

5) flat open baffles with no attenuation for the rear wave, ranging in size from large, door-sized panels to hanging the bare driver from strings

Sorry about the crudeness of the drawings and the math - I'm sure the figures are off, but the trends are there, and at the limit, this is how the systems behave with ideal drivers. Above the baffle peak the phasor for the rear wave spins around in the other direction, creating multiple 180-degree nulls and the comb filtering shown in Michael's simulations and measurements.
Attached Images
File Type: jpg Vectors_OB.jpg (23.7 KB, 412 views)
File Type: jpg Vectors_LossyBox.jpg (21.8 KB, 413 views)
File Type: jpg Resistive_Loading.jpg (23.8 KB, 427 views)

Last edited by Lynn Olson; 20th July 2010 at 07:17 AM.
  Reply With Quote
Old 20th July 2010, 09:22 AM   #6720
diyAudio Member
 
slowmotion's Avatar
 
Join Date: Nov 2002
Location: Norge
Thanks Lynn!

And I think the drawings are very informative.

The resistive mesh box concept is interesting,
I'm just not sure how to get absorption low enough in frequency,
if one wish to do without a sub.

Last edited by slowmotion; 20th July 2010 at 09:26 AM.
  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 05:36 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