Originally Posted by Ricci
So...How do you feel about the results so far? The horizontal polar looks a little odd with the recurve in output as you get further off axis. Is that an interaction between the ht and mid units radiation? How about a 1w sensitivity measurement on the mid horn? Looks like a 3 ohm minimum.
Well the proof of concept is definitely there, and the thing sounds awesome....this is the first time I've had a DIY build exceed expectations. I was actually expecting to throw the prototype away, but I think it's going to work out nice for this application. We've been doing the sweeps at ~115dB to 120dB at 1m and the distortion is a good 45dB down, which is like 0.5% distortion...when cranking it up, the XTi amps actually crap out way before the speakers do.
We pretty much hit the horizontal polar target, which was 6dB down at ~45 degrees (for a 90 deg horizontal). I'm pretty sure the -4dB dip at ~25 degrees is because the nose doesn't come together to a point - at least that's the current theory. Early on we did some measurements with and without a makeshift nose and could see that energy come up in the off-axis by a good 3-8dB depending on frequency. I think the color scheme is a bit misleading to how dramatic that dip is because it's certainly not very audible when walking around (though some of that could be all the reverb in the gym - trying to find a chance to listen outside to isolate that variable). I know, it's -10dB overall, but that's only -4dB relative to -6dB, and getting that 4dB back would result in additional energy from the summing with the energy already there. I am definitely going to run some experiments with a nose before we fly the speakers, so stay tuned. Also, I think the polar response of the tweeter falling flat on its face also makes it look worse too. The wideness at 1-2kHz should hopefully be addressed with new tweeters.
Right at 1kHz, there is a little interaction between the tweeter and midbass unit that is the result of non-ideal xover interaction. The XTi amps don't offer Bessel filters so I haven't figured out the math yet on how to make the Butterworth work in the scenario. I can adjust the tweeter delay to move that energy at 45degrees into the 25 degree hole, but then it really sucks out as you go wider in angle. Right now I'm at a 3rd order butterworth, which was a significant improvement over the 4th order L-R. Here's a plot of the L-R polar response:
I'll be sure to keep you guys updated with the xover experiments too. It takes a while to run quality polar response sweeps and the angles we're measuring now aren't perfectly exact. We're working on a turntable for the speakers so that we can get more data points with more accurate angles. I'm also really curious how the horizontal polars look at different vertical angles.
It sounds like we'll be dragging these guys down to the Klipsch pilgrimage so if anyone wants to hear them, that'd be a great opportunity (plus the pilgrimage itself is a blast):http://www.klipsch.com/pilgrimage/
Btw, another comment on the polars...If you stick this thing into a room corner (which was the original design intent), the polar response makes a nice transition to the 90degrees of the room corner...so your horizontal is effectively 90 degrees wide over the entire bandwidth. The LF corner also extends down to 50Hz. There's a good enough corner in the gym, so I'm planning to make some measurements just to verify the model correlates to real world performance.
RE Penn, the wavelength at 200Hz is "only" 5ft and the mouth on this thing is around 4ft when splayed, so you're getting into the 1 wavelength piston polar response, which offers some directivity. You've also got the mouth diffraction helping you out a bit too.