Originally Posted by cattskinner
Great, thanks for the info. Appears to be very well supported by some folks who design, build, and measure.
Just to be clear, I'm not sure Mark Seaton was endorsing the Hornresp lossy Le tool, maybe he was, maybe not. I'm not even sure if he uses Hornresp or is aware of Hornresp's lossy Le tool. But he clearly favors considering lossy Le issues at least.
Yeah, it would appear I might be giving up a few dB on max SPL compared to modelling without allowing for inductance, not enough to sell the farm over IMHO.
It's not just the different frequency response to consider with lossy Le issues. The response becomes more damped due to the lossy Le losses. This indicates that perhaps a bigger box would be more suitable. When you consider that power compression will also make the response even more damped (if you run it hot enough to get into power compression territory) the response can end up REALLY bad.
If you want to simulate some power compression on top of the lossy Le losses, just add some extra Re. Very heavy power compression can result in a doubling of Re. So double your Re and see what happens to the response. It's not likely that you will run into heavy power compression with a horn like that in a residential setting, but it's something to consider for future designs, if you are interested in taking this hobby further.
(Simply adding Re is not a perfectly perfect method of predicting power compression response either, but at least it's a pretty good educated guess about what's going to happen).
If you look at some of Danley's stuff, his designs are usually very underdamped since they are meant to be used at very high power levels which will in turn lead to heavy power compression. In his underdamped designs the response will actually get flatter as it gets into power compression. When you get into heavy power compression the frequency response kind of "melts" away, leaving peaks where the impedance peaks are. The more Re you add to simulate power compression, the more the frequency response will start to look like the impedance curve. 2x Re is a practical limit though, at and beyond that limit you are into extremely heavy power compression and the voice coil is likely to be hot enough to melt.
Personally I like to design at least a bit underdamped even if I don't plan on getting into heavy power compression. It just seems the right thing to do. This was my thinking even before I became aware of lossy inductance and the issues it presents, mainly influenced by reverse engineering Danley designs.