Zaph|Audio - ZRT - Revelator Tower

These drivers are very well behaved and were relatively easy to work with. With their low distortion, a flat response gives us a neutral tonal balance that sounds great with no listening fatigue. The reverse null is deep, showing us that this is a traditional Linkwitz-Riley design. The low crossover point gives us a broad vertical listening window with good symmetry. The transfer functions are smooth with good tolerance to component variation. The design axis is ears level with the tweeter, and all plots are at 1 meter.

The impedance curve is relatively benign and should work well enough with tube amplifiers. Impedance for this 2-way version stays above 6.5 ohms, averages 8 ohms through most of it's range and does not exceed 14 ohms above the bass tuning peaks. No additional impedance control circuitry is needed.

2-way measured response

Measured frequency response on-axis, 10ms window

Actual measured response is very close to the modeled with the biggest difference being in the 400-500 Hz range. Seeing as the impedance curve with the crossover in place is exactly the same as the modeled curve, this is likely due to a mild difference between the individual driver measurements used in the model, and the system response curve. The model uses on-axis data with piston based off-axis simulation for a given mic location, while the actual measured system response is real. The mic location for the woofer response used in the model is 6" lower that the mic location used in the measured system response. Some floor bounce still comes through a bit in the nearfield woofer plots with diffraction added. This floor bounce is at a mildly different frequency and level for the measured system response compared to the driver curves taken for modeling, and the difference presents itself in the 400-500 Hz range. This is more than anyone wants to know, but I like to have a perfect grasp on differences between modeled and measured - even if it's only 1dB.

The system harmonic distortion sweep was done in-box at a further distance than my standard infinite baffle tests to allow proper summing between the tweeter and woofer. As such, the plot is not directly comparable to others on this site, but it does show trending between the drivers. The system is clean throughout it's entire useful range, but it is exceptionally clean in the bottom of the tweeter's range.

There are often some concerns about mounting a port on the front baffle. I used to believe the concerns myself. I often mount my ports on the back, but now this is more for baffle space considerations than midrange leakage or organ pipe resonances. I found that the midrange leakage of front mounted ports, particularly these flared ports, low enough to make front mounting an acceptable option. Above I show a near field port measurement. While not exactly smooth a couple octaves above tuning, there isn't much peaking to worry about and the raggedness is off the audible radar. An item of concern for measuring front mount ports is output from the woofer leaking into the nearfield port measurement, combining out of phase at some frequencies and causing additional raggedness. Some of this can be seen in the plot above. It's impossible to differentiate between midrange leakage though the port and interfering output from the woofer on the outside of the baffle. I'm not sure it matters anyway.

18W8531G Power Handling

Excursion vs SPL, full range, vented, 1,2,4,8,16,32,64,128 watts

Run full range, linear excursion in the vented enclosure is exceeded with about 30 watts. Be aware that the 18W8531G and most other Scan Speak drivers overload gradually and are capable of quite a bit more excursion than their Xmax spec may suggest, before distortion becomes objectionable. With an 80 Hz LR4 sub filter, excursion is exceeded at 85Hz with a few hundred watts. Realistically, I'd expect thermal and compression issues to show up well below that. A 50-100 watt amplifier should be a good match for this. These speakers have decent sensitivity, so a person with a low power amp could be made happy.

2-way enclosure

The "ZRT" in a Madisound MD38T enclosure

The tweeter location in this design is centered for manufacturability and the distance from the top was selected for a decent diffraction signature with a centered tweeter. Tuning calculations were done with T/S parameters that were averaged from 8 woofers. Qts averaged a little higher than spec, and it works out well using the full volume of the MD38T enclosure. Bass extension is extremely deep with a single vented 18W8531 in this box. My preference is the 30 Hz tuning, which gives a mild rolloff with good extension that seems to blend well with a typical room response.

I call for heavy stuffing at the bottom of the enclosure to reduce the lengthwise pressure node at 140Hz. Not doing so would cause some audible raggedness in the midbass. The difference between enough and not enough stuffing is easily seen in an impedance curve.

2-way options

The default design is the one that sounded best to me in my listening room, but not everyone has a similar room or similar tonal balance preferences. This design is intentionally easy to tweak, so don't be at all afraid to try new values, even if you're just curious how it will change the sound. It's relatively easy to swap components, provided you didn't glue the baffle on or glue the crossover to the bottom.

2.5-way design option

This 2.5-way design would work well sealed in the MD38T enclosure or a custom built vented enclosure of larger volume. The sealed bass extension of the 18W8531 woofer is surprisingly satisfying in 19 liters, which is half of the MD38T enclosure. In the sealed MD38T, the lower woofer can be located below the halfway point so that the grill can be put on without the center crossmember hitting the woofer frame or surround. This configuration is what I built, modeled and listened to. If you aren't going to use the grills, the lower woofer could be mounted close to the upper woofer with your own cuts in the blank baffle.

The 2.5-way crossover design is a cascaded configuration, with the lower woofer's baffle step compensation inductor placed between the woofers and the rest of the components effectively filtering both woofers. The crossover is close to a 3rd order configuration, and as such there is no reverse null. The individual driver rolloffs are close to 3dB down at the crossover point rather than the 6dB down that is a characteristic of Linkwitz-Riley crossovers. More discussion on this below.

A 2.5-way system is by nature a full baffle step compensation design. This design in particular requires a large room and at least 2 or 2.5 meters listening distance for the best summing between drivers. The modeled mic location is on the tweeter's axis and 2.5 meters distance, though I've verified good frequency response as close as 2 meters. At one meter, the response gets ragged due to the relative distance from the mic/ear to the lower woofer and the tweeter. The vertical polar response shows this design's 3rd order tendencies which favors on or above axis. Power response will be very smooth. The weak point in this and most 2.5 way designs is the lower woofer causing a droop in response above the axis. See the blue 750 Hz lobing for example. This is minimized either by mounting the woofers close together, or by maintaining a further minimum listening distance which is what this design calls for. There is a sole benefit of having the lower woofer spaced down a bit - a reduction in the floor bounce reflection. The two woofers have different floor reflection frequencies resulting in a smoother in-room response. Typical floor bounce is 300-400 Hz, thought it's not completely smoothed out with the lower woofer already starting to roll off.

If a builder of the 2.5 way chooses to mount the lower woofer right below the upper woofer, there will be a mild improvement in the vertical lobing centered on about 750 Hz. This is shown above, as 3dB less droop about 20 degrees above axis. The on-axis effect at 2+ meters is minimal however with only a change in the power response having a barely audible effect in-room. Which is better? It comes down to choosing better upper midrange response above axis with one or lower high midbass floor reflection with the other. Your choice.

A near field bass response of the woofer shows just how low the sealed enclosure goes. Not bad for a 7" woofer. Getting a useful anechoic response curve with a mic at 2.5 meters is just about impossible in a home environment, but a room response curve is certainly easy to do. (and often a dose of reality) A typical room response curve will always have a primary dip and peak corresponding to the room's largest dimension. This plot is no exception. Other than that however, it stays within a 5dB window for most of the response and tonal balance is excellent to my ears.

Summary
This is about as good as it gets for a 7" 2-way or 2.5-way. Not exactly cheap, but about as high performance as anyone could hope for. When the drivers like this have incredibly low distortion, and the design is well implemented, you get a non-fatiguing "listen all day" type of sound.

Enjoy!

Page done by John "Zaph" Krutke © 2008
Also visit -Zaph|Audio-

Related Speaker Designs

• SR71 - 2-way Bookshelf
• ZBM4 - MTM Design
• ZMV5 - Compact 2-way
• C500 - Tower Speaker
• DLR - Full Project