My new speakers almost finished, and ALWAYS check the simple things first

[madisonears]

After years (I mean that) of test construction and trials and a lot of mistakes, my new speakers are finally nearing completion. They're very complicated: no six rectangles of MDF cut at Home Depot and glued together for me. These babies are high tech, and they will be the best speaker I know how to make.

Last night, I finally discovered why they just didn't sound quite right. The passive xovers for mids/highs were designed by Madisound, based on actual measurements of these drivers in my enclosure. I wasn't happy with the recessed vocals and weak, cloudy highs, so I spent the better part of this weekend tinkering with the xovers. I eventually realized, after hours of listening and analysis, that one of my tweeters was wired out of phase. How can this dumb stuff happen? I don't have any test equipment, so the discovery was made by listening to test tones from a CD.

Check the simple stuff first!

A brief description: 3-way system to be used with a subwoofer, already built. Peerless tweeter (now Scanspeak, but they were Peerless when I bought them a couple years ago) with two Peerless 4" poly mids. Passive xover with Sonicap and Mundorf SIO bypass for tweets, Solen inductor for mids. The baffle is cast polyurethane, the box is 3/4" plywood. Enclosures are lined with 1/4" felt and 1" recycled cotton batting, and stuffed with wool. They play flat within +/- 1.5dB from around 160Hz to 15K or so.

Woofers are two Peerless 8" Nomex, purchased before they were discontinued. Bass bins are sealed, with walls that are 3/4" rough ply with another 3/8" finish layer over that. All joints are glued and screwed with backing screw blocks. Baffles, tops, and bottoms are two layers of 3/4" ply. Walls are lined with 1/2" felt. I'm still playing with the stuffing.

Xover between tops and bass bins is active, built from a kit from ESP. I recommend it. I chose 350Hz as the xover point.

Here are a few pics. I'll post more as I make progress. It won't be long now!

Peace,
Tom E

[studiotech]

Good man there using the Bonded Logic insulation! Enjoy.

Greg

[madisonears]

The cotton batting is more effective at absorbing mid frequencies than any other type.

Some more pics of the upper enclosures.

Peace,
Tom E

[studiotech]

Originally Posted by madisonears

The cotton batting is more effective at absorbing mid frequencies than any other type.

Some more pics of the upper enclosures.

Peace,
Tom E

Yes, we use if for absorber panels at the studio and I use it in all my builds now.
How did you cast the baffles? They look super cool. How do you plan on finishing them?

Greg

[madisonears]

Baffles were cast from high density polyurethane. Pics of the molds below. This was a real experiment for me, although I am familiar with casting metal and some other materials. The mold is made of plaster, which was cast first using a wax model of the baffle. That was carved from a solid block of paraffin wax that I also cast in a plain wooden box.

So, the full procedure:

Design baffle. After attempting a couple open baffle designs that didn't work the way I wanted them to, I set about designing an enclosure to feature low diffraction distortion, using edges curved with a large parabola, not just a simple radius. The box is triangular to reduce internal reflections.

Choose materials. A wooden baffle would entail too much carving and not allow easy duplication. I chose poly because of its density, stiffness, and relative ease of use. It's expensive: each baffle uses about $60 worth of poly.

Design mold. You must know a little about casting to make the mold properly. It must release the part without too much difficulty. It must be near net shape to reduce waste and machining.

Build mold. This was not as difficult as I expected, but it took a long time to make it right. I built a plain rectangular box slightly larger than the baffle. I put a flat piece of plywood in the bottom of the box to represent the flat front of the baffle, then poured a couple pounds of paraffin wax around it. I carved the wax into a model of the front of the baffle. That was a lot easier than expected. Carving wax is kind of fun, but very messy. Then I built another box quite a bit larger than the baffle. I set the wax model into it and filled the cavity with plaster. After some sanding and patching porosity, the hardened plaster formed a negative of the front of the baffle. The internal mold was considerably more difficult because it had to include the rough openings for the two midrange drivers. I elected to add some stiffening ribs to the back side of the baffle, and make beveled reliefs around the rear of the drivers. I also included a small pad where each driver fastener would be bolted. Not easy to achieve that level of complexity and accuracy with woodworking tools, but a lot of patience and forethought paid off.

Now I had a plaster drag (lower half of a mold) with the negative shape of front of the baffle, and a wooden cope (the upper half of a mold) that had the exact negative shape of the inside of the baffle. I bought a couple gallons of poly and some dye and poured my first baffle. That's a bit of a challenge, as poly is a wee but fussy to work with. Things went well, until it came time to remove the part from the mold. It hung up, and I put a pretty good size crack in the casting. Some epoxy fixed that (amazing stuff!).

I machined the face of the casting using my radial arm saw and a milling attachment. I sanded the baffle face, finished the tweeter mounting hole, then drilled and tapped the perimeter of the baffle edge so I could bolt the baffle to the flange of the box.

This all took about three months of pretty steady work. Finally, with one box and one baffle finished, I could mount drivers and hear if all the effort was worth it. Yes, it was.

Peace,
Tom E

[_henry_]

wide baffles are nice, i like them

cheers
henry

[_henry_]

hi tom,

u should come up with a name for them (as wilson did), 3months is alot of time and attention

bravo
henry