Delta Mk II
Opera is when a guy gets stabbed in the back and, instead of bleeding, he sings - Ed Gardner

March 2006 This construction of my Leeloo surrounds has given me the opportunity to address the tweeter issue in Delta. As mentioned in my listening notes, I feel that the top octave or 2 in Delta may not be as good as it could be. My plan then, is to take the Morel Supreme 110 from Delta and replace it with a Dynaudio d21/2. The Supreme 110 will then be used in Leeloo.

Why the old classic Dynaudio d21/2 I hear you ask? Well, my requirement is for a 18mm soft dome driver in a 110mm chassis, and that is probably the only high end driver available that meets those specs. It's also a tweeter I've always had a huge admiration for since hearing it's previous model, the Dynaudio d21af some 20 years ago. The top end extension was truly amazing, and I never forgot the rendering of cymbals with the life and air that you experience in a live performance. 20 years is a long time though, and I was hoping my memory and ears weren't playing tricks on me.

Further changes to Delta MkII are in the internal stuffing of the midrange enclosure and further refinement to the mid & tweeter crossovers.

The Dynaudio d21/2 datasheet can be seen here. Below is it's quoted frequency response, showing the 40kHz bandwidth & just enough sensitivity.
Here's one of the drivers, thankfully surviving a long trip from my Ebay seller in Germany.

When it came to fitting the driver into the tweeter recess left by the Morel Supreme, I naievely thought it would be just a straight swap. But then, in DIY world, things rarely go as smoothly as planned. I soon discovered that the Dynaudio required a wider internal cutout, different terminal recesses & different mounting holes! To make matters worse, while trying to file & sand the required areas, I ended up damaging the veneer! This was turning into a major headache. I then bit the bullet and decided to strip all the veneer on the satellite enclosure (which I had always thought could have been a little tidier) and totally re-veneer, stain & finish them. So here we go ...

First I stripped the boxes using a regular iron and scraper. Not as hard as I imagined.
Next I reapplied the iron-on teak veneer, painstakingly hand-sanding the overlapped driver holes. I then wiped on 1 coat of waterbased gel stain.
After 3 coats of hand-painted matt polyurethane, I left it to dry 1 week and then rubbed down with Scandinavian Teak oil & 800 grit sandpaper followed by 000 steel wool.

After 2 weeks work, I'm really happy with the final outcome. It's a far neater job than my original work so it seems the initial troubles were meant to be. Now I'm rewiring & stuffing the enclosure with wool-felt, long-haired sheepskin, and polyester / wool blend.

Finally after installing the drivers, I can begin measurments. I'm now using JustMLS, with a Soundblaster Audigy PC card, set to 48000 sample rate, & 32768 sample length. I measure each driver on it's axis, at 1 metre, right-angled from the sloped baffle plane. The response at any listening location is then determined by entering the relevant x,y values into LspCAD.

d21/2 measurements

Lovely flat response. Ignore the noise below 200Hz.

15h52 measurements

LspCAD simulations

The crossover is little changed from Delta Mk1. Basically it's first order electrical which, when combined with the driver's natural response, produces a 6dB slope around the crossover point, asymptoping to 18dB/octave further away the crossover point.

Here's a clearer look at the 15h52 midrange & d21/2 tweeter with their filters. I deliberately add a slight notch on the tweeter at 1.2kHz to reduce the output at it's resonant frequency. That little hump just below 15kHz I won't bother notching as I doubt that it would be audible, and it would disappear off-axis anyway.
Below is the LspCAD predicted low frequency in-room response (floor & wall behind speaker), showing the floor bounce dip at 200Hz.
The measured frequency response of both channels, measured at 2 metres. Below 1kHz, the response is affected by room reflections.
Measured total impedance.
Measurements at 2 metres with & without grill & 30 degrees off axis without grill.
Oct 2006 - Above. Rear wiring after crossover moved to the internals.