Embedded Files
This page details the construction of a concrete subwoofer enclosure for a speaker designed to augment a pair of full-range electrostatic loudspeaker panels - which are yet to be built. The ESLs are the Acorn Mk II from ER Audio in Western Australia which looks like a nicely resolved three-way system with a reasonable claim to pleasant but not very powerful or deep bass. The ideal subwoofer would be smallish and fast to match the lively electrostat sound.
The driver selected is the small but high-quality ScanSpeak Revelator 23W/4557T00 with a matching passive radiator . This driver has a very rigid aluminium cone and a rugged magnetic motor. It's rated at 225 W continuous or 350 W max power (whatever that is, although it has an IEC spec and could be looked up if anyone was interested). "Selected" is probably too strong a word to use here because these drivers just happened to be in my cupboard from an abandoned project. (They actually belong to my son, but he's in Vietnam and the speakers are here in Australia!)
It was intended to drive the sub with a plate amp of about 250 W and the amp was to have DSP allowing creation of crossover and some equalisation if that proved necessary. My first preference for the amp was the Hypex FA251, but supply in Australia was difficult. Instead a Dayton Audio SPA250DSP was purchased.
The PR is supposed to allow smaller volumes, and a web search had turned up a recommendation for a tiny 1 cu ft enclosure. (To those outside the rather unit-eccentric state of the US, that's 28.3 litres) I decided to make the cab from concrete because the high density and great stiffness of concrete seemed ideal for a sub enclosure panel. You could argue that concrete 'rings' like a ceramic and lacks the natural acoustic damping of MDF or other materials, and I agree, it may not be ideal for a mid-woofer or midrange, but for a sub I didn't think this would be a problem. In any case the cabinet was to be lined with dense felt as a dampening material which should suppress any tendency to ring. The model for the concrete construction was taken from a very old article in Wireless World from 1970 detailing construction of a concrete horn. I'd built something like that back in the 70s myself and the use of concrete didn't add as much difficulty as you might expect, but maybe surprisingly I've never seen any other project using the same methods published anywhere since. There are some concrete speaker boxes described on the web, but these are mainly cast in one or two pieces.
The construction method by 'Toneburst' in the 1970 article uses a simple timber frame which 'boxes up' each face of the cabinet. Faces are filled with cement-sand mortar, allowed to dry, then the cabinet frame rolled over to another face, and that face filled with mortar mix, until all faces are concreted. The timber frame becomes a part of the finished cabinet and is either covered up, or celebrated! Bonding between concrete and timber is poor, so PVA glue, or equivalent, is used to improve bonding. In this instance Bondcrete was used. Partially-inserted screws were used to assist with keying the timber to the concrete although the original 1970 construction used nails. Speakers were mounted to MDF donuts embedded in the concrete.
Design of the enclosure opted for a near-cubic shape with the driver on the front face of the cabinet, the PR on one of the sides and the amp on the back. The amp was to be in a small enclosure of its own, even though it looks like Dayton expect the amp to sit in the main volume of the enclosure. My logic here was that the amp may not benefit from being pounded with bass pressure waves. A separate little space for the amp required a former to mould the concrete around, so that was made of Styrofoam with a timber surround and a faceplate that copies the Dayton Audio amp fascia. The intention was for the timber to remain embedded in the concrete while the Styro and face plate were ripped out and disposed. To make this a bit easier the face-plate was glued with a sheet of paper sandwiched between it and the timber. The paper was to be a point of easy separation.
The plug for the amp cavity is shown. The heavy timber frame was to remain in the concrete while all else was to be broken up and removed.
The timber frame was to be machined to 25 mm square from 90 x 35 hardwood stock. It ended up roughly a millimeter oversized so was left at 26 mm.
The frame elements were just glued together using a white woodworking glue claiming to be suitable for both indoor and outdoor use. Some degree of moisture resistance would be necessary because the concrete will be wet when the panels are filled. No fasteners were used in the construction.
Rather than try to fasten the speakers to concrete the design called for MDF donuts to be embedded in the concrete. These were cut from 25 mm stock using a router and a circle-cutting jig. A CNC router would make this sort of work a lot simpler.
The router bit was a 6 mm diameter spiral upcut, and cutting was taken to a depth that left a half millimeter or so of MDF to hold the donut in place. A thin layer of MDF can easily be broken out and cleaned up. Some tear-out did take place, but this was only on one side, and that side could easily be turned inwards, out of sight.
Screws were partly driven into the side of the MDF to help key the timber to the concrete. Pilot holes were drilled first to prevent delamination of the rather fragile MDF.
Preparation for casting the concrete began with partially inserting wood screws into the frame where it keys to the concrete. The screws are in place of the 1 1/2" nails recommended by Toneburst in the original article. Hammering nails into a flimsy frame was thought to be a bit violent, so screws were placed with an electric impact driver instead. As for the MDF donuts, pilot holes were drilled. Considering the holes were drilled more or less without measurement it was surprising how many of the screws collided inside the timber. About one screw in 5 struck another screw and was stopped in its tracks. The speaker donut was held in place with some thin strips of hardwood, the worktable was covered with a sheet of plastic - I used some high-impact polystyrene sheet (HIPS) left over from an earlier project, and at that point we were ready for the first face to be cast.
Diluted Bondcrete was used as the water in the mix in the hope that this would make a stronger mix. The cement-sand bagmix was beefed up with a small amount of added cement. Guessing, but probably at about 10% of the volume of the mix. Just enough water was used to make a stiff mix with the texture of plasticene, or Play-Doh or whatever modelling clay stuff you used in your childhood. Undiluted Bondcrete was brushed into the timber where it would contact the concrete, using a part of the work table away from where the concrete would be placed into the frame - didn't want to get too much Bondcrete between the box and the plastic surface in case the Bondcrete adhered a bit too well to the surface. .
Scooping some mix into the frame around the speaker donut I poked the stuff into all the corners and edges with a small piece of scrap wood. The mix was a bit thick but still workable. Naturally I had to make up more mix because I'd underestimated the quantity. The strips that secure the donut made the work a bit awkward, but hopefully there won't be too many voids. After the mix was prodded and pushed into place the work table was rapped moderately firmly with a rubber mallet to settle the mix. Nothing you can do after this, so just walk away and come back in 24 hours.
Disappointing amount of tear-out when removing the concrete box from the HIPS plastic surface. Looks like the Bondcrete had better than expected adhesion to the HIPS. The next strategy to reduce adhesion was to spray the HIPS surface with kitchen spray oil as you'd use with baking. Despite this there was still some deep removal of the top of the concrete. After this one side was poured using baking paper to separate the timber frame from the HIPS. The baking paper produced a sort of elephant hide wrinkling that was at the same time both annoying and interesting. Still looking for a good material to use for to allow easy separation of the casting from the HIPS, aluminium foil was used for the next face. Results were mixed, with no wrinkles, but a tightly granular effect. It was decided to use the baking paper for the rest of the cabinet and either fill the wrinkles or to celebrate the texture.
The amp cavity plug was prepared for casting with screws driven partly into the sides and Bondcrete applied. A loose mix was used for the first part with the aim of forcing the concrete into the spaces around the lip of the amp cavity. This was followed by a stiffer mix to go up and over the body of the plug. No effort was made to locate the plug, and as a result it ended up ever so slightly rotated. I'm not sure now irritating that will be in the longer term. Maybe the plug could have been wired in place with some sort of sacrificial wire that would be left in place in the concrete.
After the final face was cast it was time to clean things up a bit - the bits where the concrete was proud of the timber frame should be rubbed back and the tear-out filled. If a defined texture like the 'elephant hide' finish produced by the baking paper had been applied to all faces from the start for example, that could become a finished feature. Instead we had a variety of textures as each face had been an experiment. Because of the lack of a consistent featured texture there was only one possible way through - the surface of the box should be made as smooth as possible.
Turns out it's possible to 'sand' concrete using a concrete rubbing block. This is a coarse abrasive block that can grind away the concrete slowly. Takes a certain amount of rubbing, and is done dry. This was done to all faces that were not still 'green'. The rub back also opened up a few more cavities as air bubbles were exposed. At this stage a little PVA woodworking glue was smoodged into any concrete-timber junction that still looked a bit open, in particular around the speaker donuts.
Choice of a filler material was decided by the need for a hard surface to keep options open for finish materials. It wasn't settled at this stage what the finish was to be: paint, a plywood cladding, even timber veneer. The filler chosen was a 'water putty' with a good spreadability and capable of being sanded. All sanded faces were filled and allowed to dry for 24 hours.
Quite a bit of sanding and rubbing back and a smooth, hard finish started to emerge. The water putty was very hard and ate quite a few sheets of sandpaper.
The plug for the amp was removed and the cavity checked for size. A good fit it turned out. The paper between the timber frame and the faceplate shape worked very well, separating quite cleanly. If I was to do it again I'd make that faceplate a few mm deeper than the real plate to allow it to be more deeply recessed.
Felt was glued to the inside of each face using ordinary woodworking glue, small triangles being cut to fit around the driver donuts.
A test fitting of the driver and passive radiator was done because - well, it's just hard not to find out what it sounds like after all that work!
Box and driver came in at 38 kg, and was rolled around the house on a dolly for testing. Just a bit heavy to be easily tossed around.
Connected a single channel of the amp to one of the speakers in the study - Troels' Bookshelf speakers. Good output down to maybe 60 Hz. The DSP in the sub amp was set up for a 65 Hz crossover point with 24 db/octave and about -8 dB gain.
No measurements yet, but bass seems deep and strong. Surprised to note the box shows some vibration on the deep notes, but I guess that's what happens with a 250 W amp and a hefty driver. I'd be surprised if the sides are flexing at all. Can't tell readily by feel.
If it's any indication, four tracks of Pink Floyd left me feeling a bit overwhelmed and had to turn it down and find some quieter jazz. It's not quite the same as the twin 18" subs with NX1000 amp in the other room, but still solid.
Next will be finishing the boxes. Thinking at the moment is a timber top with iron-on timber veneer on the sides?
So was it worth it? Geez, hard to say. It's a really heavy, solid enclosure. Acoustically I think it's outstanding, but in its unfinished state it's not great to look at. Putting a smooth finish on concrete turned out to be quite laborious with the materials I chose - lots of sanding and recoating then sanding again. If a rough finish was acceptable then it would have been a very easy build. Toneburst in the 70s was happy with wrinkly concrete just painted glossy black; and in my bachelor days that might have been fine ...
There's another aspect to this enclosure: to touch it is to be amazed by its sense of strength and mass. It's impressive in the same way as any other really sturdy object - a tractor, or a stone statue for example.
Page updated
Google Sites
Report abuse