Directivity without Waveguides or DSP - diyAudio
Go Back   Home > Forums > Loudspeakers > Multi-Way
Home Forums Rules Articles diyAudio Store Gallery Wiki Blogs Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

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 26th May 2017, 05:24 PM   #1
diyAudio Member
 
Patrick Bateman's Avatar
 
Join Date: Oct 2006
Location: San Diego
Default Directivity without Waveguides or DSP

Here's an idea I had, to make a loudspeaker with directivity, that doesn't use DSP processing or waveguides.

It offers the following advantages:

1) Higher efficiency that a U-Frame cardioid
2) Much much smaller size than a loudspeaker with a waveguide
3) Lower cost and higher efficiency than a DSP cardioid

It was inspired by these two threads, so the idea will make more sense in that context:

Cerwin Vega Cardioid Horn

Beer Budget Beam Steering

Click the image to open in full size.
Click the image to open in full size.
As detailed in the two threads above, it is possible to make a loudspeaker directional with the use of an "end fire array." Basically you do this:

1) Pick a frequency, for instance 85Hz. (4 meters)
2) Divide that frequency by four (1 meter)
3) Put the two loudspeakers 1 meter apart (or whatever distance you have from step two)
4) Delay ONE of the loudspeakers by one meter (or whatever distance you have from step two)

Hope that makes sense.

The reason that this works is that the DSP delay focuses the energy in a single direction. Easiest way to visualize this is with the wavefront simulator in Hornresp. (pictures in the thread linked above, the thread titled "beer budget beam steering.)



If everything above makes sense, then we take this to another level using all-pass filters. Here's how this works:

Click the image to open in full size.
A conventional DSP delay adds a fixed amount of delay across the bandwidth. For instance, if we had two loudspeakers in an end-fire array and they were a meter apart, we'd want to delay the FRONT loudspeaker by one meter, or 2.94ms. With an all-pass filter, we can create a delay that's VARIABLE. By tweaking the Q and the frequency of an all-pass filter, we can have a delay of 2.94ms at 85Hz, and a smaller delay at higher frequency.

Easiest way to visualize how this works is to load up the spreadsheet here:
Car Audio | DiyMobileAudio.com | Car Stereo Forum - View Single Post - Advanced filters (Allpass/LT etc) with MiniDSP

If these two different cardioid arrays make sense, then you're ready for the next one...
  Reply With Quote
Old 26th May 2017, 06:28 PM   #2
diyAudio Member
 
Join Date: Mar 2007
Location: California
It's maybe just an example case, but your 4-driver array is 3m (over 9 feet) long. I hope you have a very large listening space to back away from it!

Sure, this works, but you will (at some frequency) get an on-axis null. It's no different than any other multiple source array - a "flower petal" directivity pattern develops as frequency increases. I modeled your system as 4 drivers spaced 1m apart and with delays of 8.7ms, 5.8ms, 2.9ms, and 0 msec for drivers when seen looking end-on at the "front". The first on-axis null is at 140Hz.

If you are going to come up with some arrays like this you should probably model both the on axis response as a function of frequency and the polar response at a variety of frequencies. You want to know both where nulls will develop and what the off-axis response looks like. You can get both from lots of polar plots done at a variety of frequency but a single on-axis FR plot is helpful to get the overall picture as well.

I've looked at this kind of stuff before in detail (although using only 2 drivers) and you can do some interesting things...
__________________
Visit my Audio Web Page <<--CLICK TO LEARN MORE-->> Get my LADSPA plugins
  Reply With Quote
Old 26th May 2017, 06:49 PM   #3
diyAudio Member
 
Patrick Bateman's Avatar
 
Join Date: Oct 2006
Location: San Diego
Using the ideas detailed in post #1, you can get similar results without DSP and without waveguides. You do it by combining a horn and a sealed box.

Here's how this works:

In an end fire array, the loudspeaker at the front of the array is delayed. For instance, an 85Hz end fire array will have the loudspeaker at the front delayed by one meter. (2.94ms)

But there's a way to delay sound without using DSP, and that's to bury the woofer in a horn. The time-of-flight from the mouth to the throat creates a delay.

Voila! End-fire-array with no processing.

Here's an example:

Click the image to open in full size.
For my front loaded horn, I use a Alpine SWS-10D2. Sells for $130 online.
Click the image to open in full size.
For my sealed box, I use a Tymphany FSL-1530R01-08. Sells for $84 online.

Click the image to open in full size.
Here is the frequency response of the Tymphany woofer in a sealed box and the Alpine in a FLH. Note that the power had to be 'ramped up' to four watts on the Tymphany to keep the output levels similar. Keeping the output levels similar is VERY important; if one enclosure is louder than the other, it will reduce the cardioid effect.

Click the image to open in full size.
Here is the group delay of the two boxes. To create a perfect end-fire-array at 80.69Hz, we want the boxes 1.05m apart with a delay of 3.09ms. A combination of a sealed box and a FLH gets us a delay of 5.5ms. That delay isn't perfect, but you get the idea. To improve upon that delay, you would use a different woofer than the Alpine. For instance, a prosound eight can be put in a shorter horn, and a shorter horn will yield less delay. Less delay gets us closer to a "perfect" cardioid, at least for this specific frequency.

Click the image to open in full size.
Here is the group delay of the two boxes. To create a perfect end-fire-array at 60Hz, we want the boxes 1.42m apart with a delay of 4.17ms. A combination of a sealed box and a FLH gets us a delay of 3.74ms. That's just about perfect; it looks like the Alpine FLH and the Tymphany would make an excellent cardioid at 60Hz. I think this sim illustrates the relationship between the depth of the horn and the spacing that's required to make a passive cardioid using a sealed box and a front loaded horn.
  Reply With Quote
Old 26th May 2017, 07:41 PM   #4
diyAudio Member
 
Patrick Bateman's Avatar
 
Join Date: Oct 2006
Location: San Diego
Quote:
Originally Posted by CharlieLaub View Post
It's maybe just an example case, but your 4-driver array is 3m (over 9 feet) long. I hope you have a very large listening space to back away from it!

Sure, this works, but you will (at some frequency) get an on-axis null. It's no different than any other multiple source array - a "flower petal" directivity pattern develops as frequency increases. I modeled your system as 4 drivers spaced 1m apart and with delays of 8.7ms, 5.8ms, 2.9ms, and 0 msec for drivers when seen looking end-on at the "front". The first on-axis null is at 140Hz.

If you are going to come up with some arrays like this you should probably model both the on axis response as a function of frequency and the polar response at a variety of frequencies. You want to know both where nulls will develop and what the off-axis response looks like. You can get both from lots of polar plots done at a variety of frequency but a single on-axis FR plot is helpful to get the overall picture as well.

I've looked at this kind of stuff before in detail (although using only 2 drivers) and you can do some interesting things...
It looks like it scales fairly easily. Becasue the concept is the same whether we're making a subwoofer array or a midrange array; the idea is that we're PHYSICALLY delaying one of the two drivers by running it through a horn.

For instance, I just knocked out a sim using a fifteen inch woofer and a four inch woofer. That may sound like an odd combination, but the horn loading raises the efficiency so high, you really need two drivers with vastly different efficiencies. (Obviously you could just dump more power into one of them also, since raising the power doesn't change the phase or the group delay.)

Anyways, the combination of a 4" woofer in a horn combined with a fifteen inch woofer in a sealed box yielded a combination which should be cardioid from about 175Hz to 350Hz. With that combo, you could have a box that measures about 15" x 15" x 15". Not big at all.

Click the image to open in full size.
Picture something like the EAW Otto, but smaller.

Also, the horn loaded woofer doesn't have to terribly small. I used a 4" woofer because I was trying to raise the F3. But you could also use something like a prosound 8" woofer. Basically you want the horn loaded woofer to have a low QES and a high FS, to keep the F3 from being TOO low. If the F3 is TOO low, you wind up need a box that's three meters long! (As seen in post #2)

I'll post the sims later.
  Reply With Quote
Old 26th May 2017, 10:02 PM   #5
diyAudio Member
 
Join Date: Mar 2008
When talking about directive (woofer) designs, there is always a directivity vs bandwidth (across that desired directivity or DI can be met) tradeoff.

And that interesting "efficiency" has to be measured at the lowest frequency of operation IMO to be useful.


See e.g. "Endfire Array":

https://www.google.de/url?sa=t&rct=j...Md42Sw&cad=rja
__________________
Oliver, RFZ believer (?)
www.dipol-audio.de

Last edited by LineArray; 26th May 2017 at 10:26 PM.
  Reply With Quote
Old 26th May 2017, 10:28 PM   #6
diyAudio Member
 
Patrick Bateman's Avatar
 
Join Date: Oct 2006
Location: San Diego
Quote:
Originally Posted by LineArray View Post
When talking about directive (woofer) designs, there is always a directivity vs bandwidth (across that desired directivity or DI can be met) tradeoff.

And that interesting "efficiency" has to be measured at the lowest frequency of operation IMO to be useful.


See e.g. "Endfire Array":

https://www.google.de/url?sa=t&rct=j...Md42Sw&cad=rja
Click the image to open in full size.

Yes.

That's one of the interesting things about using an all-pass to do the delay, instead of a constant delay.

With an all-pass, the delay gets higher and higher as you go lower in frequency. So it allows you to discard some efficiency on the low end of the design, to get directivity, while retaining efficiency at the upper end of the passband.

The method illustrated in post 2 of this thread doesn't behave exactly like an all-pass filter, but it is similar. You can see that the group delay of the horn-loaded woofer has multiple peaks in it. If you replaced the horn loaded woofer with a bandpass woofer you would have a single peak.

Although I think this scheme could work with a bandpass woofer instead of a front loaded horn, I need to sim it to be sure.

To know for sure I really need to build the thing. If I'm not mistaken, the 'spikes' in the group delay that you see in Hornresp don't happen in the real world. If I understand it correctly, the 'spikes' in group delay are related to the 'spikes' in the frequency response of the horn, and we've all seen that horns DO have peaks in their response, but they're nowhere near as severe as hornresp predicts. In the real world, those 'spikes' are much smaller, so I'd assume that the 'spikes' in the real world group delay are also less severe.

TLDR: I need to build one of these.

Last edited by Patrick Bateman; 26th May 2017 at 10:32 PM.
  Reply With Quote
Old 26th May 2017, 10:34 PM   #7
diyAudio Member
 
Join Date: Mar 2008
Using (multiple) horns for "delay" (with a multitude of transducers) is problematic, because at the "lower end" sound pressure and phase has much ripple, as the mouth opening usually is too small to obtain constant (resisitive) load over frequency.

Thus a lot of DSP compensation would be needed to compensate for the (always too small) horn anyhow, to maintain the desired radiation pattern (made up by that "mixed sized" driver/horn multitude).

To do away with that horns is the most "easy" way, to get a well behaved (and also efficient) woofer design IMO.

However to lift a common CB's acoustic/mechanic impedance at the lower end, it can always be replaced by a well suited BR design ...
__________________
Oliver, RFZ believer (?)
www.dipol-audio.de

Last edited by LineArray; 26th May 2017 at 10:50 PM.
  Reply With Quote
Old 26th May 2017, 10:46 PM   #8
diyAudio Member
 
Patrick Bateman's Avatar
 
Join Date: Oct 2006
Location: San Diego
Quote:
Originally Posted by LineArray View Post
Using (multiple) horns for "delay" in a multitude of transducers is problematic, because at the "lower end" sound pressure and phase has much ripple, as the mouth opening usually is too small to obtain constant (resisitive) load over frequency.

Thus a lot of DSP compensation would be needed to compensate for the (always too small) horn anyhow, to maintain the desired radiation pattern.

To do away with that horn is the most "easy" way, to get a well behaved (and also efficient) woofer design.
The simulations show ripple. In the real world, I don't think there's a whole lot.

Click the image to open in full size.
Here is my measurement of a commercial Kef subwoofer, along with my Alpine front loaded horn. Note that the phase is well behaved within it's passband. There's less than ninety degrees of phase shift within the horn's passband. The Kef sub has a 180 degree rotation, of course, because it's a vented box.

Unless I'm missing something, it looks to me that the horn can be treated as an almost "pure" delay, because the phase varies so little in the horn's passband. This particular horn has a pathlength of a little over two meters.


In case anyone's doubting that horn phase measurement, here's a measurement from Danley of one of his products:

http://www.danleysoundlabs.com/danle...pec-sheet4.pdf
  Reply With Quote
Old 26th May 2017, 10:58 PM   #9
diyAudio Member
 
Join Date: Mar 2008
Does the mechanical impedance (at the "lower end") seen by the driver in any of those ("horn" called) designs really compete with a well suited BR ?

Will the phase ripple - whithout any smoothing - compete with a well suited BR or be more easy to compensate for ?

Low frequency horns at compact (mouth) size are just little more than a myth to my view ...

And if they "work", they are no horns but quarterwave or BR constructs in fact.



Edit: And btw. forming a radiation pattern from multiple "phased sources" is highly dependent on meeting that phase relations between the sources (also) in a broadband manner.
__________________
Oliver, RFZ believer (?)
www.dipol-audio.de

Last edited by LineArray; 26th May 2017 at 11:07 PM.
  Reply With Quote
Old 26th May 2017, 11:44 PM   #10
ICG is offline ICG  Germany
diyAudio Member
 
Join Date: Jul 2007
Location: Sound Cloud 9
Quote:
Originally Posted by Patrick Bateman View Post
For my front loaded horn, I use a Alpine SWS-10D2. Sells for $130 online.
The driver got a very low spl (82dB), a very high Mms (230g), a huge Xmax (15mm), low EBP (57,4). For a front loaded horn it got ideally a high spl, low mms, medium xmax, high EBP; that's a car-subwoofer for very small BR or sealed enclosure. There are a lot of problems, the horn does not nearly get you as much spl as a direct radiating (non-horn) driver would give you, you can't get back 15-23dB spl advantage with power. And the power is the next thing, this driver depends on the excursion for cooling - but in a horn it does not do that, the horn greatly reduces the excursion. That means, you can't neither use the maximum power nor the excursion at all because the ends of the voice coil are burning up with a lot lower power than the drivers are rated.

An Eminence DELTA PRO 10MR-8 or a B&C AMX 250 (or others at the same price range) would perform a lot better in a front loaded horn.
__________________
After all is said and done, usually more is said than done.
  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
B&O Beolab 90 - adjustable directivity by DSP Quip Multi-Way 150 19th January 2017 03:02 PM
S15: Econowave DSP - a Constant Directivity vs Dipole study gainphile Multi-Way 172 23rd October 2016 06:52 AM
HTPC DSP versus dedicated DSP platforms for loudspeakers diyaudnut Digital Line Level 16 21st June 2016 07:13 AM
DSP midrange directivity control aka kinda cardioid natehansen66 Multi-Way 104 26th January 2016 11:54 AM
RS-232 cable for DSP-Amp Digisynthetic DSP-1400 Sabbelbacke Parts 0 9th May 2009 09:33 AM


New To Site? Need Help?

All times are GMT. The time now is 10:23 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2
Wiki