JohnConover_medium.jpg 
john@email.johncon.com
http://www.johncon.com/john/

Loudspeaker Design


Home | John | Connie | Publications | Software | Correspondence | NtropiX | NdustriX | NformatiX | NdeX | Thanks



home.jpg
john.jpg
connie.jpg
publications.jpg
software.jpg
correspondence.jpg
ntropix.jpg
ndustrix.jpg
nformatix.jpg
ndex.jpg
thanks.jpg

The loudspeaker Thiele/Small parameters can be used with the Spice analog electronic circuit simulator to provide behavioral analysis of loudspeaker enclosures. Of particular interest is the frequency response, phase response, and group-delay of loudspeaker enclosures, each of which can be optimized for maximum performance.

As an example, the Dayton Audio ND90-4 will be utilized in a design suitable for high quality computer monitors. The ND90-4 is rather unique in that the optimal, (optimized for low frequency response,) ported enclosure is nearly the same size as the optimal, (maximally flat midrange,) sealed enclosure, permitting an empirical comparison of the sound produced by both optimizations using the same enclosure for both. The ND90-4 driver is readily available, for under US $25, from Parts Express, Amazon, Ebay, etc.


The Design

gEDA Electronic Design Automation tools, particularly, Gschem(1), was used for schematic capture and Spice electronic circuit simulation netlist extraction. Ngspice was used for the Spice circuit simulations

The design work sheets and data bases for the speakers are available in a single tape archive format, (TAR file, using RCS for version control,) from nd90-4.tar.gz.


Optimal Designs

The optimization programs are avalable from the Subwoofer Speaker Design project as Calc(1) scripts:

The ported-box.calc script indicates that a ported enclosure volume of 10 liters, (which may be too large for most computer stations-the box optimal internal dimensions are 13.7" X 8.5" X 5.2",) which would have a low frequency -3 dB of 32 Hz., and the sealed-box.calc script indicates that the 10 liter sealed enclosure would have a low frequency -3 dB of 75 Hz., and a Q of 0.78, (a Q of 1 / sqrt(2) = 0.707 is optimal maximally flat, which is fairly close.)

The Characteristics


gain.jpg

Figure I

Figure I shows the frequency response of the various optimal ported and sealed enclosure sizes utilizing the ND90-4 driver.

Now that the optimal enclosures have been determined, the analysis can begin-remember, two designs, ported and sealed, are being developed.


The Ported Design


ported/nd90-4.jpeg

Figure II

Figure II, (1600X1200),(3200X2400), is the ND90-4 symbol used for the ported design.

And descending down the hierarchy of the ND90-4 symbol:


ported/ported.jpeg

Figure III

Figure III, (1600X1200),(3200X2400), is the schematic of the ND90-4 ported design using the Thiele/Small parameters.


ported/gain.jpg

Figure IV

Figure IV is the frequency response, simulated using Spice, of the ported design.


ported/phase.jpg

Figure V

Figure V is the phase response, simulated using Spice, of the ported design.


ported/transient.jpg

Figure VI

Figure VI is the transient response, simulated using Spice, of the ported design. Note the ringing at the enclosure's resonant frequency, Fb ~ 42 Hz.. This is a result of optimizing for maximum low frequency response. The enclosure volume is about 10 liters-about twice what the manufacturer recommends, (most computer sound systems available at the big-box stores use the same optimization.)


ported/group-delay.jpg

Figure VII

Figure VII is the group delay, simulated using Spice, of the ported design.

Note that the group-delay of the ported design is acceptable, (at least by ported enclosure standards-but it is a controversial issue,) and is at the threshold of audibility.


One of the advantages of using Spice simulations including the Thiele/Small parameters is the ability to electronically modify an enclosure's responses, (i.e., analog, DSP, etc.) For example, consider, (as a thought experiment,) using another ported ND90-4 driver's characteristics to modify the frequency response curves of the ND90-4.


ported/nd90-4-Hs/nd90-4.jpeg

Figure VIII

Figure VIII, (1600X1200),(3200X2400), is the implementation of using a second ND90-4 driver to modify the frequency response curves of the ported ND90-4.

Notice the subtle point: the second ND90-4, X2, is connected backwards, and since the negative input of the operational amplifier, X3, is a virtual ground, and summing the currents to the virtual ground for the X3's output voltage, Vox2:



          Vox2 * H(s) / R2 = -Vin / R1

          Vox2 / Vin = - R2 / (R1 * H(s))

          R1 = R2

          Vox2 / Vin = - 1 / H(s)

        

Where H(s) is the transfer function of the second ND90-4, X2, and 1 / H(s) is the inverse transfer function. The Thiele/Small parameters were used to construct a prototype software filter that provides an overall gain of unity, flat frequency response, and with a phase shift of zero as the output of the X1 ND90-4. Remember that this is a thought experiment, (negative group delay?) and this circuit-and variants thereof-are unstable in the real world; however, small signal simulation can be used for verification with the theoretical operational amplifiers used, etc.


ported/nd90-4-Hs/gain.jpg

Figure IX

Figure IX is the frequency response using a second ND90-4 driver to modify the response curves of the ported ND90-4. Note that the ouput of the X1 ND90-4 is unity, and not frequency sensitive.


ported/nd90-4-Hs/phase.jpg

Figure X

Figure X is the phase response using a second ND90-4 driver to modify the response curves of the ported ND90-4. Note that the output of the X1 ND90-4 phase is zero, and not frequency sensitive. It is not clear that a Bilinear Transform is applicable to the circuit due to stability issues, and frequency warping may limit the applicability of the circuit to woofers at digital music sampling frequencies.

Note that this was a thought experiment, but we can approximate the response of the second ND90-4 with two all-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.


ported/all-pass/nd90-4.jpeg

Figure XI

Figure XI, (1600X1200),(3200X2400), is the ported ND90-4 with two all-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.


ported/all-pass/gain.jpg

Figure XII

Figure XII is the frequency response of the ported ND90-4 design with two all-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.


ported/all-pass/phase.jpg

Figure XIII

Figure XIII is the phase response of the ported ND90-4 design with two all-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.


ported/all-pass/group-delay.jpg

Figure XIV

Figure XIV is the group delay of the ported ND90-4 design with two all-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.

Note that the all-pass filters are a simplified model of the ported ND90-4 and enclosure.

A more complex model would use high-pass filters, with Q's other than maximally flat, where Q = 1 / sqrt(2) = 0.707, or Bessel, where Q = 0.577215665, e.g., Euler's Constant.


ported/high-pass/nd90-4.jpeg

Figure XV

Figure XV, (1600X1200),(3200X2400), is the ported ND90-4 with two Sallen-Key high-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.


ported/high-pass/gain.jpg

Figure XVI

Figure XVI is the ported ND90-4 frequency response of the ported ND90-4 with two Sallen-Key high-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.


ported/high-pass/phase.jpg

Figure XVII

Figure XVII is the ported ND90-4 phase response of the ported ND90-4 with two Sallen-Key high-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.


ported/high-pass/group-delay.jpg

Figure XVIII

Figure XVIII is the ported ND90-4 group delay of the ported ND90-4 with two Sallen-Key high-pass filters to shift the phase of the ported ND90-4, addressing the issues of group delay.

Note that the high-pass filters are a simplified model of the ported ND90-4 and enclosure.

There are limits to the gain that can be used in compensating circuits-specifically, the maximum power that the driver can handle.


ported/maximum.response.jpg

Figure XIX

Figure XIX is the maximum output power, at a specific frequency, that the ND90-4 ported design can handle.


The Sealed Design


sealed/nd90-4.jpeg

Figure XX

Figure XX, (1600X1200),(3200X2400), is the ND90-4 symbol used for the sealed design.

And descending down the hierarchy of the ND90-4 symbol:


sealed/sealed.jpeg

Figure XXI

Figure XXI, (1600X1200),(3200X2400), is the schematic of the ND90-4 sealed design using the Thiele/Small parameters.


sealed/gain.jpg

Figure XXII

Figure XXII is the frequency response, simulated using Spice, of the sealed design. Note that the ported design has a substantially lower cutoff frequency.


sealed/phase.jpg

Figure XXIII

Figure XXIII is the phase response, simulated using Spice, of the sealed design.


sealed/transient.jpg

Figure XXIV

Figure XXIV is the transient response, simulated using Spice, of the sealed design. Note that the ported design has substantially more ringing.


sealed/group-delay.jpg

Figure XXV

Figure XXV is the group delay, simulated using Spice, of the sealed design. Note the the ported design has substantially more group delay.


One of the advantages of using Spice simulations including the Thiele/Small parameters is the ability to electronically modify an enclosure's responses.

sealed/simple.filter/nd90-4.jpeg

Figure XXVI

Figure XXVI, (1600X1200),(3200X2400), is the sealed ND90-4 design with a two pole shelving filter addressing the issue of low frequency response.


sealed/simple.filter/gain.jpg

Figure XXVII

Figure XXVII is the gain response, simulated using Spice, of the sealed design with a two pole shelving filter addressing the issue of low frequency response. Note that the filters, (which have similar characteristics to the tone controls on power amplifiers,) have two zeros at the same frequency as the ND90-4's two low frequency poles-but the frequency response shapes around the zeros and poles are different, creating a bump in the overall frequency response.

Note that extending the frequency response down to 40 Hz., (competitive with the ported design,) is not practicable due to power limitations of the ND90-4.


sealed/simple.filter/phase.jpg

Figure XXVIII

Figure XXVIII is the phase response, simulated using Spice, of the sealed design with a two pole shelving filter addressing the issue of low frequency response.

What Was This Project For?

It was to investigate the issue of group delay in ported speaker systems. Both the sealed and ported designs are optimal, (or nearly so,) using the same driver in a simple enclosure box, allowing one to place/remove a hand over the port and listen to the difference under different program scenarios. Various simple schemes to improve the ported design's poor group delay, (that's why the components were chosen,) were implemented, also.


License

A license is hereby granted to reproduce this software for personal, non-commercial use.

THIS PROGRAM IS PROVIDED "AS IS". THE AUTHOR PROVIDES NO WARRANTIES WHATSOEVER, EXPRESSED OR IMPLIED, INCLUDING WARRANTIES OF MERCHANTABILITY, TITLE, OR FITNESS FOR ANY PARTICULAR PURPOSE. THE AUTHOR DOES NOT WARRANT THAT USE OF THIS PROGRAM DOES NOT INFRINGE THE INTELLECTUAL PROPERTY RIGHTS OF ANY THIRD PARTY IN ANY COUNTRY.

So there.

Copyright © 1992-2015, John Conover, All Rights Reserved.

Comments and/or problem reports should be addressed to:

john@email.johncon.com

http://www.johncon.com/john/
http://www.johncon.com/ntropix/
http://www.johncon.com/ndustrix/
http://www.johncon.com/nformatix/
http://www.johncon.com/ndex/



Home | John | Connie | Publications | Software | Correspondence | NtropiX | NdustriX | NformatiX | NdeX | Thanks


Copyright © 1992-2015 John Conover, john@email.johncon.com. All Rights Reserved.
Last modified: Tue May 26 12:34:30 PDT 2015 $Id: index.html,v 0.0 2015/05/26 19:35:04 conover Exp $
Valid HTML 4.0!