10th December, 2007.
The CA15RLY and W15LY001 versions from this project are finished and I'm left with the task of writing something about these drivers? Hmm...they're all good, but for various reasons. From the CA15RLY you get a dynamic little speaker that I'm sure will satisfy most if you know what you can expect from 75 cm^2 cone area. The W15LY001 costs more than twice the CA15RLY price and as always we apply to the law of diminishing return. The W15LY001 is a better driver. It provides a tad deeper bass and it can be driven a harder without noticeable distortion. But some may prefer the more vivid presentation from the cheaper brother.

I've used the 22TAF/G tweeter in other - not published - constructions, and once more I like what I hear from this small 3/4" dome with the wide surround. I haven't made the W15CY001 version yet and it's to be heard if this tiny fellow can match the HIQUPHON OWI tweeter used in the W15-OWI construction. They're both 3/4" domes but one is a classic, coated fabric dome and the other a hard dome made from aluminium. Due to the small diameter of the alu dome the inherent cone break-up seems to be as high as 30 kHz or more. And the wide surround makes us want to treat it as a 1" dome although we have to realise that the smaller voice coil won't take as much heat as a 1" voice coil, but from a 3rd order filter, this is not a problem as we won't play any of these three speakers excessively loud. Small minis are not for Saturday parties.

The CA15CY001 system is the "Joe-Cool" of the gang presented here. With 83 dB sensitivity it's slightly less dynamic compared to the other two, but the lack of colouration from the magnesium cone is hard to ignore. This speaker tells you very much what's on the recording, but in a non-aggressive way as can be the case from speakers with an elevated upper midrange or treble voicing.

Overall there's a steadily progression in clarity and naturalness going from the CA15RLY over the W15LY001 to the W15CY001 driver, with a wider gap between the CA15 and the W15LY compared to W15LY and W15CY. The latter two are close with the increased sensitivity of the W15LY being the key difference.

For TS data "added volume" method has been used. As always some discrepancy between claimed data and actual measurements. Dependent on actual signal applied to the units during measurements, we can have different results. No surprise.

For the W15LY001 (nextel coated driver) the correlation is good. The magnesium driver is well used and still maintain Fs = ~50 Hz. The CA15RLY is a lightly coated paper cone and I wonder if this driver really has higher Mms compared to W15LY001, as claimed by SEAS.
None of these drivers have particular low Qt and the magnesium and nextel drivers are close to being "closed-box-only" drivers. But these are all small drivers and we need the extra bass from the vented enclosures.

The port: I've used Jantzen Audio #900023, 50 x 120 mm (Ø x L) for all drivers.

Cabinet front panel dimensions and driver placement. If you make your own cabs, calculate cabinet depth to make 11.5 litre internal volume, e.g. from 19 mm MDF approx. 200 x 350 x 270 mm outer dimensions (W x H x D).
Use bracing in the middle of the cabinet and add e.g. 4 mm bitumen pads on all internal panels except front panel.
I've used 10 mm polyester foam for damping + additional MDM3 on top and bottom of cabinet. Some MDM3 is used to cover the crossover attached to the rear panel.
Don't forget to experiment with the MDM3 (acoustilux). This 11.5 litre cabinet does take some damping material. Initially I had used too little damping and the W15LY001 got too boomy. Experiment and listen carefully. A significant part of the voicing of a speaker is determined by amount of damping material used.
Preliminary webpage on damping: http://www.troelsgravesen.dk/cabinet-damping.htm

Left: SPL from drivers at 2.8V, 1 metre distance. Blue = W15CY001 (new), red = CA15RLY, green = W15LY001.
The 300-400 Hz region gives us an idea about what system sensitivity we can expect from these drivers. The W15CY001 (magnesium) doesn't allow more than 82-83 dB/2.8V, where the other two suggest 85-86 dB system sensitivity. Let's see where it all ends.

Right: SPL from "new" and "old" W15CY001 magnesium driver. I had a mail from Finland suggesting differences between former and present 5" magnesium drivers and so it appears. I've got a pair of both. The newer drivers have a minor bend to the cone where it meets the voice coil, allowing a larger area for gluing - and this is pushing break-up nodes towards higher frequences - but also with higher amplitude. From initial modelling this doesn't make much difference to the crossover, but minor adjustments to the notch-filter have to be made. Actually the "old" driver appears easier on modelling due to lower impedance where we need correction to be done.

Left: The W15LY001 has an extended frequency range, all the way up to 10 kHz. A minor peak at 5.7 kHz didn't produce any problems during crossover construction. Tweeter response equally manageable with a flat response from 3-22 kHz.
Right: Red = summed response from drivers driven from crossover and merged at 300 Hz with nearfield bass response. Blue = summed response with inverted tweeter polarity. Point of crossover is 2.8 kHz.

Left: Red = summed response of drivers driven from crossover. Blue = bass response. Green = tweeter response.
Right: System impedance reaching a 4 ohms minimum at 300-400 Hz. Overall an easy load on the amplifier.

Left: System response from left and right speaker at 1 metre distance, 2.8 volts input. System sensitivity is around 85 dB/2.8 volts.
Right: System response (red) and minimum phase (blue).

Left: System response at 1 metre distance, 2.8 volts input. System sensitivity is around 86 dB/2.8 volts, slightly higher than the W15LY001 system.
Right: System response in red and blue = response with inverted tweeter polarity.

Left: Purple and green display response of individual drivers driven from crossover.
Right: System impedance reaching a 3.8 ohms minimum at 400-500 Hz. Still an easy load on the amp.

Left: Red = summed response from system. Blue = tweeter with inverted polarity.
Right: Left and right speaker from final crossover.

Left: System impedance. Lowest impedance is 4.7 ohm at 400-500 Hz.
Right: "New W15 driver": Significance of RC-circuit across series coil. An easy way to reduce cone break-up.

Left: Red = summed response of drivers driven from crossover. Blue and yellow = bass and treble driven from crossover. Point of crossover = 2400 Hz.
Right: Same thing, here with the "old W15 driver". Actually this is a tad better than the new one as predicted from LspCAD simulation.

Left: System response at 1 metre distance, tweeter height and 2.8 V input. System sensitivity = 83 dB/2.8V.
Right: Tweeter attenuation. Red = 8R2, blue = 10R, green = 12.2 ohm. I finally settled on 11.2 ohm, 2 x 5R6 in series.