
Home  Forums  Rules  Articles  diyAudio Store  Blogs  Gallery  Wiki  Register  Donations  FAQ  Calendar  Search  Today's Posts  Mark Forums Read  Search 

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 

Thread Tools  Search this Thread 
18th April 2018, 12:38 PM  #1 
diyAudio Member
Join Date: Apr 2010

Spice Modeling Acoustic Properties of Speakers
Did you know that you can model electroacoustic properties of loudspeakers? I use LTSpice for my simulations. I create a circuit based on the electromechanical properties of a woofer, and an enclosure. It takes the form of a high pass filter of sorts. Using ThieleSmall parameters, you can convert them into a circuit. I am attaching a screenshot example acoustic suspension speaker system with a system Q of .707. You can plot frequency response, impedance, phase or anything you wish. This model has an amplifier with a damping factor of about 300. I also model ported systems. I can show an example of that if there is any interest.
Jim 
18th April 2018, 12:55 PM  #2 
diyAudio Member
Join Date: Feb 2008

I've never tried to include the enclosure in the model, but I did mess about with this in AIMSpice many years ago. Without using the T/S parameters, I just adjusted the various RCL values until the impedance plot matched a driver I had (Pionneer A11 IIRC) and played around with adding crossovers, zobels and Zpeak correction circuits. It was quite educative.

18th April 2018, 01:27 PM  #3 
diyAudio Moderator

Jim  you showed a pretty simplified model of the loudspeaker, how close would it match the actual measured results? Dr. Leach back in 1991 was already working with more complex models, which no doubt were subsequently incorporated into the modern speaker design programs.

18th April 2018, 02:47 PM  #4 
diyAudio Member
Join Date: Apr 2010

The model can be made as complex as you want, but it will have little effect on the box volume and/or tuning condition IMHO. Thiele showed more complex models in 1971 but opted for the simpler model for design.
The ported model is similar to the one shown but contains 2 more elements that are of consequence. Since there is a vent capacitance, there is also a resistance associated with it. This is used to calculate the box leakage Q. Tuning conditions for ported designs are selected based on Butterworth, Chebyshev, or whatever filter you prefer. Again, the simple model is best for these calculations. The other elements will have little effect on box and tuning conditions. More complex models can be created based on actual measurement of specific drivers. From this you could get more accurate impedance/phase plots. Jim 
23rd April 2018, 12:29 PM  #5 
diyAudio Member
Join Date: Apr 2010

Examples of a ported system.

23rd April 2018, 01:11 PM  #6 
Got Foam?
diyAudio Member

It’s good to do this to show that all this theory comes down to networks of RLC lumped elements. It works of course, but Spice really is best left for the design of the electronics and amps.
Akabak is a scripting language for modeling acoustic systems using lumped element and is basically doing the same thing but has all the functionality of modeling speakers built in. It also handles the diffraction, wave propagation in horns waveguides, velocity, etc. In the end, T/S parameters are just RLC equivalents of a driver. Sure we can do it in a Spice, but let’s use the optimal tool for the job. One can use a pipe wrench as a hammer, but we have hammers for the job. Akabak also has electrical elements like RLC, and opamps, etc and ok to use for occasional BSC or XO. But a power amp should be modeled on LTspice. Same thing. Akabak
__________________
XRKaudio https://www.etsy.com/shop/XRKAudio Last edited by xrk971; 23rd April 2018 at 01:13 PM. 
23rd April 2018, 02:19 PM  #7 
diyAudio Member
Join Date: Apr 2010

I think you missed the point of this. A.N. Thiele solved the mystery of ported loudspeaker design by converting the electromechanical properties of a speaker system into a circuit. The idea being that once these parameters are known, the values can be calculated for known filter types. As an example, a ported system can be designed as a Butterworth filter as quasi third order (QTS less than .383) or a forth order if Qts is .383. The poles are zero for orders 1 and 2 in first case and zero for 1,2 and 3 in second case. SPICE is great for modeling filters.

23rd April 2018, 03:06 PM  #8 
diyAudio Member
Join Date: Aug 2002
Location: victoria BC

If Thiele solved the mystery of ported loudspeaker design, why are there still so many horrid sounding examples inhabiting modern showrooms?
Perhaps there’s an element of art to balancing the numerous comprises involved in creation of a musically pleasing sounding system than just crunching numbers?
__________________
like my mind, this space intentionally blank 
23rd April 2018, 03:10 PM  #9 
diyAudio Member
Join Date: Apr 2010

Maybe you should you read the paper . There are an infinite number of "designs" that can be derived from the mathematical model.

23rd April 2018, 03:16 PM  #10 
diyAudio Member
Join Date: Aug 2002
Location: victoria BC

Jim, perhaps you missed my point, but there’s far more important things for all of us to worry about.
__________________
like my mind, this space intentionally blank 
Thread Tools  Search this Thread 


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Indoor acoustic modeling  Binely  Software Tools  0  15th May 2015 08:20 PM 
Spice modeling of speakers, other newbie questions  parallax  MultiWay  12  15th March 2013 12:17 PM 
Borbely Fet Follower SPICE modeling  pjacobi  Solid State  42  16th June 2008 02:32 PM 
New SPICE modeling freeware from Tex Inst  jackinnj  Everything Else  2  12th July 2005 10:11 PM 
SPICE modeling  EternaLightWith  Tubes / Valves  4  22nd April 2003 09:50 PM 
New To Site?  Need Help? 