1. Losses in Loudspeakers Claus Futtrup, MSc M.E. Chief Technical Officer

2. Traditionally loudspeaker designs are simulated with a standard electrical equivalent circuit following the Thiele/Small approach. To see how good or bad the model is, let’s build a box and measure, then compare to simulations. Test box – essentially lossless:

3. 3 measurements, compared to 3 simulations (Thiele/Small)

4. Simulations do not match real-world at all! Why not? What shall we do about it?

5. Richard Small on Ql, Qp and Qa - conclusions Qp – between 50 and 100 Qa – typically 100 or more Ql – between 5 and 20 Link to paper (PDF) page 5PDF Small is looking for an explanation in the system enclosure – but the actual cause shall be found in the transducer. Quote: The last result is surprising, because the enclosure tested well built and appeared to be leak free… Quote: … leads to the conclusion that the measured leakage in apparently leak-free systems is not an error of measurement but an indication that the actual losses in the system enclosure is not constant with frequency…

6. Losses, Ql = 7 : A mix of Qa and transducer admittance Λes

7. Good vs. bad models The conclusion is valid for all models, also State Variable Models, Finite Element Models, etc.

8. Transducer model including frequency dependent damping Model proposed by Thorborg and Futtrup

9. Measurements and simulations – using the proposed model

10. What does the box model look like? When the transducer includes frequency dependent damping then the box model must include effects of damping material. Source: http://www.aes.org/e-lib/browse.cfm?elib=15791 Link: (PDF)PDF

11. Model of damping material (mechanical domain)

12. Box model, Electrical Equivalent, with Rb

13. Volume expansion graph

14. Douglas Hurlburt (JAES, March 2000) Makes all the classical errors: 1)Combines the T/S transducermodel with absorption in the box. 2)Includes the Marshall Leach absorption model without a fill percentage. We can assume that Hurlburt didn’t verify his model, because we would have observed that the quality of the model vs. reality would have been worse – i.e. worse than simply assuming Ql = 7. Link to paper (PDF). See page 15.PDF

15. Mechanical model, Tarnow For standard polyester damping material. This is different from Bradburys tentative results (36) by a factor 2.5, published in the Journal of the AES, April 1976

16. Impedance, free air vs. in-box

17. Impedance, simulation of a ported (reflex) box Blue = 0% filling Green = 70% filling Red = 100% filling

18. Thank you. Questions? Claus Futtrup futtrup@seas.no For more info see: http://www.cfuttrup.comhttp://www.cfuttrup.com