Speaker Analysis for Armchair Critics


Hello everyone,
There’s a very important discipline called "Speaker Analysis" or "Speaker Testing" which though complicated, is brilliantly illustrated in this breakdown of the B&W 685.


http://www.audioexcite.com/?page_id=6070

Speaker analysis is to measure each of the components both separately and as they come together in a complete system. It is a part of creating a new loudspeaker, but it can also be used to analyze an existing speaker, to understand it and perhaps to make it better.  I prefer the term Analysis because it better reflects that the goal is not merely quality assurance, but to build a complete electro acoustical understanding of the system as a whole so changes can be considered, and their final results predicted.


This particular article does just that, and comes up with a couple of suggestions for re-working the crossover to end up with hopefully a better end result. At the very least, it is a significantly different speaker at the end, and achieves a far greater level of change than cables can.


I share this with all of you just as an example of the work that goes into making a loudspeaker from parts, and the tools, and how much of what we hear has to do with choices made in the crossover.


Best,

Erik
erik_squires

Showing 4 responses by audiokinesis

Erik wrote: " I share this with all of you just as an example of the work that goes into making a loudspeaker from parts, and the tools, and how much of what we hear has to do with choices made in the crossover. "

Your post and that article are a welcome respite from repeated accusations by one person on this forum of being a scammer because I’m a loudspeaker designer. Thank you Erik.

"... the goal is not merely quality assurance, but to build a complete electro acoustical understanding of the system as a whole so changes can be considered, and their final results predicted."

Imo a complete acoustic understanding would include a great deal more off-axis data. This is just my opinion.

" This particular article does just that, and comes up with a couple of suggestions for re-working the crossover to end up with hopefully a better end result. "

I agree with the direction of all of the changes the author made. I probably would have reduced the output of the tweeter in the crossover region even more that he did, to partially offset its excess off-axis energy at the bottom end of its passband.

Duke
"... for consumers, this is a decent introduction, not a comprehensive guide."

It is a FAR MORE THAN DECENT introduction!

The process I use is pretty much what the author did. Once the first iteration is built, even if the measurements look good, I use my ears to tell me whether it’s "right" or not. And if not, then I gather data to figure out what isn’t right, which in turn points me towards what to do about it.

Duke

Xyobgyn wrote: "Quite possibly we should have crossovers with some caps and or resistors we could swap easily to adjust the sound to where we want it."

I have been making crossovers with an external resistor that functions as a "tilt/level" control; that is, it effects the tweeter’s level more up at 10 kHz than it does down at 2 kHz, the theory being that this would be more useful in dealing with real-world variations in room acoustics than a level-only L-pad.  Also, L-pads are not readily available at the same quality as say a Mills resistor. 

I also like to build in some low-end adjustability in the form of multiple pluggable ports.

Ime often changing capacitors = re-designing the crossover, as there is often a rather precise balance between capacitor values and inductance values in a circuit, such that a change in one often calls for a change in the other.

Duke
Kenjit wrote:

"which is why you need to stop tinkering with these horrible passive crossovers."

The fallacy here is that your statement assumes facts which are not in evidence. You have not established that passive crossovers are horrible, nor that mine in particular are horrible.

Duke