DIY Speaker Project PAD

Years ago I decided to rebuild my DIY Ariel transmission lines. The rebuild journey took much longer than anticipated.

I'll spare you the details on how this project came to be and the path taken to get to this point and get to the details.

The ArielZ: A modern interpretation of a DIY classic.

I kept the name Ariel out of Homage for the original design and designer. A system that I loved and who's mids still had, at last listen, something special about them, an open laid back effortlessness that was instantly attractive to my ears. I swapped the s for a Z in the name because this new design uses the Zaph ZA14W08 and because in using metal drivers the sonic signature obviously deviates somewhat from the original sound produced by the Ariel.

I searched long to find drivers that would work in the Ariel cabinets. In the end I chose the:
ZA14W08 in large part because it physically fit in the Ariel cabinets (with some mods).
Dayton RS28A Also for it's physical similarity to the original SSD2905/9000 tweeters.
http://www.zaphaudio.com/ZA14W08/
http://meniscusaudio.com/dayton-rs28a4-p-1321.html

Both drivers were also chosen because of their reputations for being very solid performers.
The Zaph, because of it's reputation for great midrange performance, and when rebuilding the Ariel, IMO the mids is where you must start. In having a nice midrange the ZA14 is similar to the original P13WH. Neither driver is famous for their bass production, but that's what subz is for. And acoustically the RS28 was selected for it's all around performance and real lack of any complaints. Once again, similar to the D2905, the RS28 does most things very well and very few things wrong.

There are plans for the Ariel cabinets out there, though some changes would be necessary to accommodate the new drivers. The flange on the ZA14's is slightly smaller than that of the original P13WH. And the RS28A flange is a bit larger than that of the D2905. Other than that, no cabinet changes should be necessary, my particular cabinets are Ariel Mark 2. Though I've not tried the other Ariel Versions, I'm fairly sure any version of the cabinet would work.

The ArielZ, at the paint shop, with primer.
Acoustically they turned out nice enough that I decided to have a professional paint them.

The original Ariel had a bit of a sag in the measurements, possibly to keep the sound from getting aggressive. I've opted for a slight downward tilt to the overall response to achieve similar goals. Stop the sound from being bright or aggressive. To my ears I've achieved this goal. A slight change in series resistance will raise level of the level above 2k somewhat, for those that prefer prefer a flatter response.

This image shows the effects of swapping resistors in the network, from 1 ohm to 1.4 ohms. The effects are visible from 2k and up. It also shows the predicted reverse phase notch for the flatter (1 ohm) version. Using the softer sounding 1.4 ohm resistor results in a slightly shallower predicted notch.

Mr. Olson created a simple, effective network for the original, using very smooth drivers. The Network in the ArielZ is, of necessity, a bit more complicated. In their passband both drivers in the Z are very smooth, but being Metal both drivers suffer metal cone breakup. Fortunately this is easily tamable in the Za14's and not considered an issue in the tweeter. A simple .33uf cap bypasses the main BDC inductor to tame the metal cone peak in the ZA14's. I stuck a resistor in series with that cap to tame the ultrasonic impedance dip that results with just the capacitor. The tweeter breakup is high enough that it's normally not tamed but another issue arises due to the tweeter location on the cabinet: the cabinet introduces a small lump centered on 3khz in the tweeter response. The series LCR in the tweeter circuit tames that lump, and is actually optional. It does make a small improvement in the sound so for my pair I left it in.

The Z's network itself is a slightly asymmetrical 4th order LR. The asymmetry is there to provide a good phase match between the drivers. A 1.8khz crossover frequency was chosen to keep a smooth power response. The RS28A can handle even lower crossover frequencies and the ZA14's are barely starting to roll off in the off-axis so 1.8k seemed like a good choice.

Some years ago (yes, this has been in development for that long) Paul K designed a proper MLTL for this system. Thank you Paul. I'll add those plans here soon, for those that wish to have a much simpler cabinet design. Baffle width and height are the same in that design and I would also suggest keeping the same double thick baffle. The 3/4" round over isn't simple and I've not measured either version without those.