Obviously a loudspeaker design canít be successfully completed without making actual measurements of the design while youíre in the process of perfecting it. Usually more work is associated with simulating and fine-tuning than with actually building them. I usually measure the individual drivers first and create an impedance and frequency response plot prior to starting simulation. I compare these plots to the ones supplied by the manufacturer, Seas in this case, and if theyíre close I know Iím on the right track. I then use these, or the ones plotted from the actual datasheet, to do the first simulations in SpeakerWorkshop. I try to get to a point where the simplest configuration, i.e. the fewest parts, will give me the best response possible, i.e. the flattest frequency curve with the least amount of phase changes without taxing the drivers too much and keeping a close eye on resonance peaks.
SpeakerWorkshop plots showing the frequency response and impedance curve for the loudspeaker.
Once Iíve gotten to that point it is time to construct a prototype as simulating will not get you much further. Youíll need actual data from the prototype to be able to see if it is close to the simulation, or miles off. From my experience this yields a loudspeaker which is about 75% done. There will be some inconsistencies with the simulation but at this point the tool to use is the MLSSA and 3D waterfall plots to finalize the design. Now it usually is a question of changing values of the parts, adding more damping, or removing some, or sometimes trying a different filter scheme. I initially started off with a 2nd order series-filter for all drivers, but as it turned out an asymmetrical approach yielded far better results, which resulted in the crossover schematic on the previous page.
Actual measurements of the loudspeaker using FFT, showing the frequency response and 3D waterfall.
By looking at the frequency response and 3D waterfall plots the loudspeaker has a flat response within +/-2dB, that's a lot better than what most commercial designs are specified at, they usually are within +/-6dB or +/-3dB. Since dB is a logarithmic scale a difference of 0.5dB can be picked up by the human ear and will influence the character of the loudspeaker. The 3D waterfall plot shows that this is a loudspeaker with excellent rise and fall times, especially the tweeter is exceptionally fast and void of any resonances. Baffle step compensation also works out well, and the fact that the tweeter is sunken, and hence itís voice coil is offset by about 2-cm, does wonders for the phase response with no visible drop in the frequency curve around the crossover frequency. Note however that due to the limitations of my measurement microphone the curve drops rapidly after 15kHz, this is not due to the loudspeaker.
Discuss This Article (28 Comments) - If you have any questions, comments or suggestions about the article and/or its contents please leave your comments here and we'll do our best to address any concerns.
Rate This Product - If you have first hand experience with this product and would like to share your experience with others please leave your comments here.