Hey, Wyn, thanks for bringing your experience and joining in on this. I read the KEF page and, after some further internet digging, found some of what must be some of the earliest discussion of damping factor...by Gilbert Briggs (the founder of Wharfedale). He seems to have been either the originator, or at least early popularizer, of some of the notions about loudspeaker damping and amplifier damping factor that are what people like Augspurger and Toole (over on this side of the pond) later went to some lengths to debunk.
KEF's discussion starts off with
The thing is, these are three conflicting statements; not more than one of them can be true. It's late and I'm not going to try getting any more deeply into that tonight though. (It turns out physics is a complicated subject that is difficult to explain to...anyone.)
In this thread you have people quoting sloppy statements like the first and third of the quoted sentences, and spreading confusion, and people sticking to the 2nd sentence, which at least is correct (since it is a definition).
I do understand that KEF's blog is for doing marketing and that its also not possible for them to get down to full technical detail. But I don't think you can hold that up as a good reference. Do you have something on the subject with a little more technical meat?
Ok. The KEF description is a good non-technical reference and indeed all three statements are true, if limited.
Yes, they do seem to conflate "damping factor" with "output impedance" and they certainly can be criticized for that, but the general ideas are reasonable.
Frankly, I am reluctant to approach this from a highly technical sense less those who shall remain nameless get their "knickers in a twist" (yes, I am a Brit), and increasing the technical level enough to be consistent and informative without raising someone's ire has proven to be impossible in the past.
This is a reference for the methodology that I sometimes use to model the driver units.
Electrical Model of Loudspeaker Parameters | PROJECT RYU
When modelling the behaviors of my amplifier/speaker cable/loudspeaker interface I use different levels of complexity, depending on circumstances.
1. a "lumped equivalent model" where I represent the amplifier, speaker cable and loudspeaker as impedance elements.
This works well for my current loudspeakers- ML Montis, as the Bass unit has a built in power amp designed for that task, and the remaining element is an electrostatic panel.
The interconnect impedance function is lumped, but it includes elements for RLC and skin effect.
Sometimes speaker impedance curves are provided by the manufacturer, sometimes HiFi mags provide them as part of their evaluation, sometimes programs like REW or Virtins can be used to measure the unit. I've used all three methods at various times.
This is very useful for simply evaluating the frequency response of a system, including resonances. It makes inserting transmission line elements for the lumped elements in the interconnect cables easy and avoids some of the convergence issues that appear when trying to simulate combinations that are near instability.
2. A "full" model including elements derived as in the white paper above. and an opamp like model for the driving amp with feedback, and in some cases non-linear, load dependent elements that allow me to perform transient simulations. These are very good for evaluating dynamic behavior, but often have convergence problems that limit their use in cases where the system is marginally stable.
One final comment. The AHB2s have very low output impedance/high damping factor, and Benchmark has explained the value of this by focusing on the frequency response implications just due to the voltage divider aspects of the amp output impedance/interconnect cable/speaker load impedance combination.
Deviations in frequency response as low as 0.25dB over broad regions are audible and can change the character of the system. This implies that the "damping factor" should be in the order of 50 or higher over the entire audio spectrum- whatever the speaker load presented is.
If the load, at some frequency, is 1ohm, then the output impedance should be 20mohms which would give a damping factor of 400 into 8ohms.
If the 1ohm is at 20kHz, then achieving such values is quite challenging.
This has gone way off topic
