Dayton Reference RS225/RS28A
Part 2
Measurements-Unit 2
First, on axis 1 meter frequency response. Unit 2 in red.
Next, impedance.
Now, here's one I don't do much. It's an on axis RTA. It averages the on axis spl, power, and room response. Pretty good, although it does tend to suggest that I have a dB or two of excess baffle step, depending on how you want to look at the whole baffle step/power response conundrum.
Now, some nonlinear distortion graphs.
Just some general comments about the above. Basically, distortion is as good or better for unit two. FR linearity is also excellent between the two. The impedance curve doesn't have any glitches, and a nearfield of the woofer shows only one possible small FR irregularity at around 350 Hz, suggesting well damped midrange internal reflections.
One last graph. It's an excess group delay comparison of my SS8545/9500 two way and RS22528A. Note that the SS has a stepped baffle, as well as an LR4 xover. The lack of a stepped baffle, deeper cone of the RS228A, and steeper slope filter do make the GD worse for the RS22528A.
Listening impressions
Music included Bela Fleck, Turtle Creek Chorale, DeeDee Bridgewater, Ramsey Lewis Trio, a bit of Peter White and Chris Botti (too much low end in the mix, but I happen to like them.) I also listened to a number of classical pieces from the Cincinnati Pops and the Minnesota Orchestra. Not to mention Dire Straights, Keb Mo, and others.
In general, I listened to them in a large family room, with the speakers out from walls and obstructions >3 feet.
Of note, they have the full amount of baffle step, ~5-6dB. It sounds about right out in the middle of a room, but would definitely sound heavy in the lower midrange if placed against a wall. I had wanted 4dB, but had intentionally targeted, 6dB. Huh? Well, my last couple of designs seemed to under correct for baffle step. I guess this time my measurements and modeling were more on target. Oh well.
First off, let's get it out of the way. There really is not any significant distortion from the low tweeter crossover point that I can tell. In general, the RS225 midwoofer loses definition before the tweeter distorts. That's not to say that the RS225 distorts. The midrange performance of the RS225 outperforms both the SS8545 and the Usher sealed 2 ways. It's not surprising, really, since the RS225 is an 8" driver. Bass performance is very good, although the lowest octave is definitely a bit subdued. Since there is no port, you don't hear the occasional 30 Hz bass note as you do with a speaker that has a resonant pipe in it.
What about the treble? Well, honestly I have a lot of trouble distinguishing the RS28A from the SS9500 and the Usher. The tweeter seems transparent and doesn't draw attention to itself. Of course there are differences between these, but it's very hard to pinpoint the origin sometimes.
Image stability is excellent. The soundstage is wide and three dimensional, though not particularly tall.
I won't blubber on with meaningless subjective impressions too much more. But a few final thoughts. Overall, the RS22528A two way has better low end and midrange performance than either of the other 2 designs. There appears to be no detrimental effects of a Cauer type filter on subjective polar response/imaging. There also does not appear to be any ill effects from a 1.2kHz crossover in this particular circumstance, using an appropriately chosen and tested tweeter. The speakers sound "bigger" than the typical 2 way, perhaps because they can play louder overall in the 50-1k range than the typical 5.5-7" midrange. The high end is similar to a typical high end unit using, say, the SS9500.
I'm not sure how much importance to place on the somewhat worse GD numbers. Overall, the RS22528A is subjectively better than the other two, but so many other factors are involved. It becomes very difficult to ferret out which factors are most important. However, I'm quite convinced that nonlinear distortion is very important. Because of the RS225's Sd advantage, as well as less octaves to spread because of a lower crossover point, IM distortion is notably better. My next project will be a 3 way, and it will have a midrange driver with significant Sd, more that the typical diy'er might use, based on my measurements and listening impressions.
I'd consider this experiment a success.
Thanks to everyone on the MAD and PE boards!