Peerless HDS164 (850438) + Vifa XT25TG-30-04
Cabinets made from 19 mm MDF and added a 4 mm black acrylic faceplate. Cool!

Well-known drivers. The HDS164 (850438) has been discontinued, but there may be a
few of these around and the shown crossover may be interesting for others.

The HDS164 (850438 and -39 = different chassis colour) has been replaced by the 830874 (pp membrane) or the 830875 (Nomex membrane), apparently having almost identical frequency response and impedance profiles, but whether these new drivers will fit in with this crossover is a very good question. If you would like to make a construction with the new drivers, lend me two or four drivers and I'll fine-tune the crossover to work properly.

Original crossover

Frequency response of original loudspeaker.

I shan't reveal the original crossover, but from the schematics it was supposed to have a 15 mH coil to the lower bass. On the coils were written 14 mH and the CLIO said 11.3 mH. Hmm...
Next the 0.8 mH to the upper bass would really mean that this driver should have quite some help from the lower bass in the upper bass/lower midrange, but basically we have a speaker with a weak bass and an upper midrange + treble some 4 dB above the basic sensitivity of the speakers. No wonder it sounded much too forward and aggressive on most recordings. I guess that if the cabinets were placed in two corners, the overall response might have been good, but with 1 meter to the rear wall and free of side walls the speakers just had way too much midrange. The midrange and treble was well balanced and the FR from 600 Hz to 20 kHz very flat. The crossover was a 6/12/12 dB set-up and the phase-tracking between upper bass and treble was quite poor despite making a smooth overall FR.

Above is seen a simulation of the original crossover and the FR is pretty close to the actual measured response.
The phase-tracking between upper bass and treble isn't too good.

New crossover:

Impedance measurements. Green: system imp. Red: One bass driver. Blue = tweeter. One of the things that made a week bass was the port tuning at 30 Hz (green graph). These drivers need a much higher vent tuning and the 50 x 150 mm vent was cut off to 70 mm length.

Above the frequency response of the HDS164 and XT25TG drivers without crossover.
Disregard what's below 350 Hz.

New crossover simulation

The new crossover is basically a 6/12/18 dB set-up. Above the LspCAD predicted response.
The lower bass has been added an impedance correcting RC circuit of 6.8 ohm + 10 uF.
All drivers have positive polarity. See actual response below.

Impedance of old and new system. Red(imp)/blue(phase)= old system. Green and yellow = new system.
Generally this is an easy load on any amplifier. 4 ohms minimum impedance and the phase profile is good too.

Two new crossovers ready for wiring.

Crossover layout.

The frequency response has been tilted to produce the BBC dip. Blue = minimum phase.

I'm not sure of the reduced level of the XT25TG above 5.5 kHz. I haven't seen this before, but I guess the symmetrical placement of the tweeter is responsible for this. Also there are some wrinkles around 2.7, 3.7 and 4 kHz not usually seen from this tweeter. If you want more treble, bypass the 1.5 ohms resistor in the tweeter section.

Cabinet:
The cabinet is 20 x 28 x 96 cm (WxDxH), 35 litres, and made from 19 mm MDF with three bracings between drivers. The vent is 50 x 70 mm (DxL).
Important!: Driver placement is 75, 230 and 430 mm from top and all drivers are centered on the front panel.
I would ad a base plate to the construction, lifting the speaker some 30-40 mm. The optimum listening height is a little low - for me. Allow the base plate to make the bottom panel of the speaker (see SP38) allowing the crossover to be attached to the front panel.

Take a look at the SP38, 2.5 clone, PL18 and other constructions to render details of cabinet construction. No more will be given here.

Sound:
This construction comes very close to what I had in mind for a 2½-way with the intended features. These speakers will play anything with a wide soundstage, some really good bass and with excellent midrange transparency. Getting the XT25TG tweeter level right is - as always - very important. I have found the 1R5 resistor to the tweeter to give the best results. Usually I start with a too high tweeter level and slowly increase attenuation until a level where it just seems to "disappear", i.e. it gels into the overall sound without making an act of itself. It's a pity Peerless has discontinued these sandwich diaphragms. I hope the new Nomex and PPB cones will work as good as these do.