Dayton Reference RS225/RS28A
Completed 12/2005
Part 1
Introduction
This project developed to satisfy my curiosity on a number of issues. Having done a number of tests on the Dayton RS series, as well as others, I was curious to see how well my measurements could be used to help guide crossover design. I was also curious to see if I could get almost as good bass extension with an 8" driver, in the same size sealed box that would typically be used for a ported 7" driver. I also wanted to try out the Cauer/elliptic filters to see how easy they are to implement, and to look at the off axis behavior of this type of crossover compared with a more typical LR4 crossover.
Specs
2 way sealed
Cauer/Elliptic filter 4rth order electrical topology following an eighth order LR acoustic topology, sort of**
Woofer/mid Dayton Reference series RS225
Tweeter Dayton Reference unshielded series RS28A
Box-PE 0.75 cubic feet non MTM
**I followed the notched filter topology currently being made popular by Jon H who you'll usually find at www.HTGuide.com. However, his topology is only one of a number of notched filter topologies. This is reviewed by Theile in his paper, Loudspeaker Crossovers with Notched Responses, in the Sept 2000 JAES. John K will soon have some detailed excel spreadsheets on his site, www.musicanddesign.com with multiple variations on this theme. Some of the other variants look very promising, at the expensive of another component or two. I hope that Bohdan and Ingemar will soon incorporate these targets into SE and LspCAD. You can do this currently, but with a bit more effort.
Schematic
Impedance curve and TS parameters
The impedance curve is fairly benign, except for the 2.8 ohm minimum. Oh well, I wasn't going to use a one watt SET amp anyway.
=praxis= Thiele/Small Parameters
______________________________________________________
RS225 in sealed PE 0.75 cu box
______________________________________________________
Qts = 0.618 Total Q
Qes = 0.936 Electrical Q
Qms = 1.819 Mechanical Q
Fs = 47.282 Hertz, Free Air Resonance
Res = 5.709 Ohms, DC resistance
Ls = 3.533m H, series inductance
Lp = 1.177m H, lossy series inductance
Rp = 1.474 Ohms, loss across Lp
Dia = 162m meters, effective
______________________________________________________
Frequency Response
The overall on axis 1m frequency response is shown below. It is about +/-1.5dB, although the majority of the curve lies within a +/-1 dB range. Praxis normalizes the spl to 1v. You have to add 9dB to get to 2.83v. So the sensitivity is ~83-84dB/2.3v. This is almost complete baffle step compensation, 5-6dB.
This is smooth. Really smooth, especially considering it's in a standard PE baffle with only a minimal roundover. The following is a comparison to two other sealed projects I have. The first is a 2 way based on the venerable SS8545 and 9500. The other is the Usher 2 way documented elsewhere on this sit.
Now, for all sorts of off axis plots
Below is a graph of the off axis performance off axis to the speaker's left, from 0-15 degrees.
Now, the corresponding 0-15 degree off axis plots to the speaker's right.
Overall, it's pretty smooth, but you can see that the off axis plots to the right are a little less smooth than the ones to the left. What's interesting is that the smoother off axis plots are to the tweeter's "long" side. That is, the off axis performance is smoother with the tweeters pointed outward as opposed to inward. This goes against the conventional wisdom of placing the left and right speakers such that the tweeters are inward. I guess the moral is that you should measure more and spout conventional wisdom less.
Now, praxis has the ability to average multiple frequency responses, so what you get when you average all the curves of the plots above is a kind of listening window average. It's a listening window average; the direct sound field between +/-15 off axis.
We can do the same type of averaging to look at a quasi kind of power response. Below you'll see an average of horizontal FR curves from 0 to 60 degrees. It's not a real power response, but it gives an idea.
This is pretty typical for a 2 way box type speaker. How about vertical response? After all, these higher order crossovers have weird vertical polar response patterns.
It's a bit difficult to make definitive conclusions, because neither is probably exactly follows the textbook crossover, but at first glance, it does appear that the off axis issues are at least a little worse for off axis points above the axis. Below axis, as shown below, it's less prominent. And, as you'll see with some further averaging, maybe the difference is very small, if it even exists.
Find the above four graphs difficult to make sense of? So did I. So, I averaged all the vertical measurements into two sets of measurements. One for vertical off axis measurements above the design axis and one for vertical off axis measurements below the design axis.
If you also keep in mind that the Usher only has ~3dB of baffle step and incorporate that into the above two plots, the vertical off axis performance is very similar.
Horizontal off axis performance is also quite similar. The take home message is that the off axis performance of these quasi eighth order filters is similar (if not marginally worse) in performance to a standard 4rth order LR, as best as I can tell. Of course, neither has an ideal power response.
Well, that's enough on and off axis plots, eh?
Although, if you want to be adventurous, praxis can save files as a polar plot. They're fun to play with. At least I think so. The following links will download a small praxis file. Then, download Praxis (the software is free) and play around with it.
RS22528A praxis horizontal polar response file
RS22528A praxis vertical polar response file
Nonlinear distortion measurements
Below is a five tone test, with the usual tones at 20, 28, 44, 64, and 92 Hz. The RS225 is clearly better. The SS8545 does very well. Still, the SS can't beat out the extra 30% Sd of the 8". Of course, I'm sure that a closed box with an SS8555 would probably have better low end brute force excursion compared with the RS225 as my other tests suggest. At four times the price of an RS225, one would hope it would be better.
Is it fair to compare and 8" to a 7"? Sure. The RS225 is ~$40 and would go in the same size box as a ported 7". If you can put a larger, cheaper driver in the same size box and get better performance, well, nothing unfair about that...
At 80 hz, I've shown comparisons to the SS8545 and Usher. The RS225 is slightly better than the Scan, and significantly better than the Usher.
At 150 Hz, the RS225 is better than both the SS8545 and the Usher, as you can see below. For the SS8545, the difference is small, however.
Also at 400 and 800 Hz, the RS225 is superior to the SS and the Usher.
From the above, you can see that the RS225 has overall significantly better performance than either the SS8545 and the Usher units. The SS8545 does surprisingly well, but the bass and midrange performance winner up to 800 Hz, is the RS225. What happens at 1k and above? Well, see below. The response at 1k is excellent, perhaps slightly better than the SS8545, which has a more traditional crossover just over 2k.
Here's the high stress point, the approximate crossover frequency, 1.3kHz. You can see the tweeter is just starting to stress a bit, with the typical higher order pattern starting to appear. It's still controlled, with the overall higher order levels at -80 dB or below.
Below is the same data, with the RS22528A and the SS reversed. Any red is excess RS22528A distortion.
Here's a plot compared with the Usher 2 way. Again, the RS22528A is slightly worse.
The higher order RS225 excess distortion is still present at 1.5k, but almost gone. Still excellent.
This particular RS28A is ever so slightly worse than the 9500, but the difference is so close I doubt you could hear it. And, the data could easily be reversed with a different RS28A or 9500.
Here is the comparison to the Usher two way. A dead heat.
And, a comparison at 3k. Again, small, probably statistically insignificant differences.
Part 1 Preliminary summary/thoughts
Memorable one line (one long line) thoughts--
The extra Sd of an 8" driver clearly makes a difference. The RS225 beats the SS8545 in low frequency reproduction
The 8" driver has outstanding nonlinear performance throughout the midrange all the way to 1k. In this frequency range, an 8" mid is clearly better.
This RS28A does not seem to suffer any ill effects from a crossover at 1.2kHz. The nonlinear distortion pattern at 1.3k, though just beginning to show some stress, probably won't be audible. Nonetheless, in retrospect I would feel more comfortable with a slightly higher crossover, probably around 1.4k.
Using a lower crossover point allows a larger driver, and subsequently, more improved bass performance compared with a typical 6.5-7". And, larger drivers usually have lower Fs, making a sealed box a more viable option. In this case, you get an Fs below 50 Hz with a more gradual low frequency roll off and improved internal box resonance control with the stuffing.
The quasi-eighth order Cauer filter topology is easy to work with and does not seem to have marked off axis abnormalities compared with a more traditional fourth order LR, although there may be some small differences.
Part two coming in December. Then I'll really have all the measurements...just kidding. Probably some outdoor low end measurements, pair matching, and listening impressions.