Delta
Gattiweb
Crossover
Driver Measurements

The following responses were measured in the enclosure using JustMLS, using a calibrated Panasonic microphone & preamplifer from Bamberg Labs.
I mostly used a 48kHz sample rate, 32768 MLS length, and 7mS sample length for measurements.

Farfield was taken at 1 metre, and nearfield with the microphone 20mm from the diaphragm. Both were merged at around 500Hz, with 6dB baffle step simulated using the JustMLS facility, causing the dropoff below 200Hz. This is, in essence, the free-field response of the speaker.

I simulated minimum phase for the frequency response using LspCAD's active filter simulation.

Morel Supreme 110

This was measured with the mike 83cm off the floor (on the midrange axis).
Curiously, the resonance is at a much higher frequency than the spec 700Hz.

CQuenze 15H520613 SDK

Note the slight dip at 1.2kHz, which correlates with the little blip in impedance. The peak at 5kHz will requre a moderate notch filter.
The rolloff above 5kHz is apprpximately 3rd order (according to Per Skanning).

Scanspeak 18W8531G00
Scanspeak 26W8861

Above shows how I use LspCADs active filter design to simulate the driver responses and hence determine the minimum phase response.

Crossover version 1.7 (updated April 2006)

* Coil equivalent series resistance.

Lspcad predicted free-space frequency response using measured driver responses.

Combined woofer responses appoximates a 1st order rollof above 300hz. Low frequency performance is into 4pi space.

Complete system measurements

The total frequency response was obtained by taking the 90cm response for frequencies above 500Hz, and merging it with the low frequency measurement taken with the microphone placed about 25mm from the baffle, between the 2 woofers.

Below is a magnified view showing the measured responses at 30 degrees off axis and the effect of the grill. The falling response above 15kHz is due to the measurment limitations. What is important is the difference betweem the plots.

The open frame balsa grill has less than a 1dB effect across the spectrum.
The 30 degree off-axis curve tracks the on-axis plot nicely with no sudden jumps and is 7dB down relative to the on-axis plot at 15kHz.

Below is the in-room low frequency measurement taken using a Tandy SPL meter at the listening position 3 metres away.
This suggests that a subwoofer may be useful for home theatre to fill in that bottom 20-40Hz octave.
The hump at 55Hz is a room mode.

Input impedance reaches a minimum of 3 ohms at 130Hz. The image below that compares the LspCAD predicted impedance to the JustMLS measured impedances.

Being a linear phase speaker, Delta should be able to reproduce a reasonable facsimile of a square-wave, so I thought it would be
fun to put this to the test. This was done using Speaker Workshop to generate the square waves and freeware (and hence coarse)
version of Osciiloscope for Windows to capture the waveform.
This test was peformed in my living room with the microphone at a distance of 2 metres and 75cm from the floor
- i.e less than ideal conditions. Still, the results (below) seem to meet expectations.

500Hz
1kHz

Listening.

Since it's about 6 months since I completed construction of Delta, I thought it's a good time to finally put down my listening impressions.
I'm using a Sony XA5ES CD player, and initially a 60 Watt Metaxas Ikarus amplifier *.
The room is fairly small living room 4m x 2m x 3m (whd) , moderately damped and carpeted, and I'm seated 9 feet from the main speakers. The external crossover allows me to disconnect drivers individually to listen for specific characteristics.

Initial impression of "Delta" is in a word, natural. It feels right - nothing jumps out at you, no obvious frequency response or other aberrations. No fatiguing "hi-fi" sound. I wouldn't call them overly-analytical, nor unduly warm, just natural and relaxed. Poor recordings are obvious, but usually still bearable (i.e they are slightly forgiving). There is no congestion, even at very high levels and the constrained layers, heavy bracing and felt gaskets have been very successful in virtually eliminating all vibrations on the baffle and sides.

Bass is excellent - powerful, tuneful, clean and extended. Extension is very good, reaching down to 25Hz (tested with tones), and no hint of boom. If anything it's slightly dry but that's my preference. The 10" carbon/paper Scanspeak in a sealed enclosure is very solid and mates very well with room gain. I think the partnership of 10" and 7" woofers has come together exactly as expected. A subwoofer would be virtually unnecessary for music, though I may add one later purely for ultimate movie effects.

The top end from the Morel Supreme is very dynamic and detailed. Extension and air at the extreme top end could maybe be a bit better, but that's probably an artifact of off-axis response rather than outright extension. Maybe an 18mm dome would have made a better choice, or I should reduce it's attenuation to compensate? **

Midrange is lovely, natural & dynamic. Ambience and low-level detail from the 5" Audiotechnology mid is excellent.It deserves it's reputation as one of the finest. The lower mid from the paper Scanspeak seems to be a touch warm, but very natural and full sounding. Male & female voices, the hardest & most important area, come across beautfully and truly sound "in the room". Coherency of vocals, a weakness in my old speakers, is excellent, a function of 1 driver handling the 300Hz-5kHz range. I've watched several movies and I've satisfied myself completely that a centre channel is superfluous. Ambient levels are almost up to my old Dynaudio D52af dome which excelled in this area. Imaging is excellent though height strangely appears to be that of the bass/mid driver rather than the design axis at centre of the 5" midrange. I suspect this is due to the downward tilted lobe of the 1st order crossovers.

In regards to amplifier loading, my little 60W Metaxas was struggling a bit at high levels. The 3 ohm load must be sagging its power supply.
The sound would easily harden and become edgy (clipped) when pushed. I've now purchased a much more powerful 100W Primare A30.1 and the improvement at higher levels is very substantial. Bass is more powerful, dynamics are greater still and the clipping at high levels is completely gone. Delta defintely needs 100W (high current) to show it's best, and so far I haven't been able to produce any audible distortion when driven with the Primare.

The grill has a definitely audible supression of high frequencies, and so is always removed for serious listening.

So, to sum up I'm very pleased with the performance, and at this stage really don't know how I could significantly improve on them.

* Later changed to a Primare A30.1 100W/8ohm, 180W/4ohm.
** I've now made a 1dB increase in level to the tweeter which has provided a little more liveliness.

May 2006. Delta Mk II is born. Click here.

Home