Background
On the web, I've seen negative comments about building loudspeakers with wood. Generally, people say that using real timber is difficult. The usual advice is to go with some type of composite (ply, MDF) and then to fake the appearance of real wood by applying veneer. The most extreme comment I've seen was this, which I quote from the High Efficiency Speaker Asylum:
"Solid hardwoods move around a lot with varying environmental conditions. Your solid hardwood speaker box will literally rip itself apart in a matter of a few months."
This got my interest. For me, real wood has never caused any problems. I thought I understood the theory about how timber can fail due to movement.
However, nothing that I've built with good, aged hardwood has cracked or split. No joint has ever failed. That seems odd.
Theory
So, here is some theory and some stats, followed by practical examples. If you don't live in Melbourne, Australia, these stats won't be spot on, but you can always look up your own climate and timber data.
Wood changes shape as its moisture content changes. Kiln or air drying timber from its original 'green' state reduces its water content by about 80%, and this transition is where the most dramatic shrinking will occur. Once the timber is fully dried, its moisture content can still change, in response to changing humidity.
Tangential movement is always greater than radial movement.
Timber types vary the amount of movement:
All of the above means that the least stable timber will have about 5 times as much movement as the best.
Some rules of thumb I found:
This lag will be greater if the timber is painted, waxed, oiled etc.
The second rule is very conservative (pessimistic); it corresponds to exposing a bad board to a 50% increase in humidity for a long time. That's a worst case scenario, like manufacturing a speaker in a arid region, then shipping it to the tropics. A better selection of board, or a smaller exposure to humidity variance, will surely improve on this 2% figure.
specifically useful to me, I found this info:
Most eucalyptus has moderate to poor stability - see http://www.wood-database.com
Most shrinkage is specified as one amount - % shrinkage as the wood is fully dried, without stating what water content that refers to - so it is hard to tell how a certain shift in water content correlates to shift in dimension. One source I found seems to give more detail: How Wood Shrinks and Swells, a U. S. DEPARTMENT OF AGRICULTURE Approved Technical Article. It gives some stats on shrinkage between green, 20, 6 and 0% moisture.
However, these numbers are just padding - they are just fractions of the final oven-dried value, assuming green was 30%
1) For Ironbark (Eucalyptus paniculata) the change is a 5.6% and 8.4% shrinkage.
2) In my climate, the average humidity of the wettest month is 15% above the driest month.
3) A handy chart tells me that as air humidity shifts by 15%, the total moisture content of a piece of dried, oil treated hardboard will shift by 2%.
In How Wood Shrinks and Swells the estimates for US cities vary between 0% (Dallas, San Francisco) and 5.5% (New York).
Combining these figures: a 2% decrease in the moisture content of ironbark should cause one 15th of the movement that complete drying causes. That's 0.37% radial and 0.56% tangential shrinkage.
Therefore, a 50cm width of fully exposed ironbark, when experiencing an extended (months long) 15% air humidity variation, should move by about 2mm.
In practice, actual movement appears to be closer to zero, possibly because:
1) All the stats I find are for the amount that timber shrinks when it is originally dried. Many woods have less in service variance than this, but I have never seen a firm figure for the difference.
2) The timber in a solid loudspeaker cabinet will respond to air humidity much more slowly than timber does when its stacked to dry:
- only some faces of the timber are exposed to room humidity
- the wood will should be treated (oiled / waxed) to reduce its permeability to moisture
, t
Wood changes shape as its moisture content changes. Kiln or air drying timber from its original 'green' state reduces its water content by about 80%, and this transition is where the most dramatic shrinking will occur. Once the timber is fully dried, its moisture content can still change, in response to changing humidity.
Tangential movement is always greater than radial movement.
Timber types vary the amount of movement:
- Species: teak and mahogany shrink tangentially by 3.4% when dried to 6% moisture content, whereas some types shrink more then 10%
- Old growth timber is slightly more stable than 2nd growth.
- Wood from near the heart of a tree is more stable than sapwood.
- Good cuts are more stable than flatsawn boards - expansion is more even / minimised across the board's width.
All of the above means that the least stable timber will have about 5 times as much movement as the best.
Some rules of thumb I found:
- Traditionally, for air-drying timber, one year is allowed per inch of wood thickness.
- For timber in use: allow for 2% movement in the wood.
This lag will be greater if the timber is painted, waxed, oiled etc.
The second rule is very conservative (pessimistic); it corresponds to exposing a bad board to a 50% increase in humidity for a long time. That's a worst case scenario, like manufacturing a speaker in a arid region, then shipping it to the tropics. A better selection of board, or a smaller exposure to humidity variance, will surely improve on this 2% figure.
specifically useful to me, I found this info:
Most eucalyptus has moderate to poor stability - see http://www.wood-database.com
Most shrinkage is specified as one amount - % shrinkage as the wood is fully dried, without stating what water content that refers to - so it is hard to tell how a certain shift in water content correlates to shift in dimension. One source I found seems to give more detail: How Wood Shrinks and Swells, a U. S. DEPARTMENT OF AGRICULTURE Approved Technical Article. It gives some stats on shrinkage between green, 20, 6 and 0% moisture.
However, these numbers are just padding - they are just fractions of the final oven-dried value, assuming green was 30%
1) For Ironbark (Eucalyptus paniculata) the change is a 5.6% and 8.4% shrinkage.
2) In my climate, the average humidity of the wettest month is 15% above the driest month.
3) A handy chart tells me that as air humidity shifts by 15%, the total moisture content of a piece of dried, oil treated hardboard will shift by 2%.
In How Wood Shrinks and Swells the estimates for US cities vary between 0% (Dallas, San Francisco) and 5.5% (New York).
Combining these figures: a 2% decrease in the moisture content of ironbark should cause one 15th of the movement that complete drying causes. That's 0.37% radial and 0.56% tangential shrinkage.
Therefore, a 50cm width of fully exposed ironbark, when experiencing an extended (months long) 15% air humidity variation, should move by about 2mm.
In practice, actual movement appears to be closer to zero, possibly because:
1) All the stats I find are for the amount that timber shrinks when it is originally dried. Many woods have less in service variance than this, but I have never seen a firm figure for the difference.
2) The timber in a solid loudspeaker cabinet will respond to air humidity much more slowly than timber does when its stacked to dry:
- only some faces of the timber are exposed to room humidity
- the wood will should be treated (oiled / waxed) to reduce its permeability to moisture
, t
Example one: crappy old speakers
I still own some loudspeakers that I built about 20 years ago ...I forget exactly when I made them: sometime in the late 90s. They were one of my first woodworking projects. They were built from old (salvage) timber, without:
The fronts were solid slabs of red gum, from a tree that had been winched out of a farmer's dam somewhere. The rest could be redgum or possibly jarrah, but it is hard to be sure.
Since then, these speakers have been exposed to freezing and searing temperatures (in various parts of Australia), varied humidity and several house moves, including being dragged around and hauled up and down staircases. They currently languish in a badly-constructed shed, which leaks. I did pull them out a couple of months ago to test. They still played fine, and seemed normal on testing (with DATS and REW).
In summary: badly built and badly treated. But they are still totally intact, as solid now as they were in the 90s.
If these are fine, then anything that uses good hardwood should be fine. It doesn't require great equipment, highly skilled joinery or any special treatment.
- power tools
- any special joinery techniques
- any knowledge of how to 'baby' timber to prevent it splitting apart in response to expansion
The fronts were solid slabs of red gum, from a tree that had been winched out of a farmer's dam somewhere. The rest could be redgum or possibly jarrah, but it is hard to be sure.
Since then, these speakers have been exposed to freezing and searing temperatures (in various parts of Australia), varied humidity and several house moves, including being dragged around and hauled up and down staircases. They currently languish in a badly-constructed shed, which leaks. I did pull them out a couple of months ago to test. They still played fine, and seemed normal on testing (with DATS and REW).
In summary: badly built and badly treated. But they are still totally intact, as solid now as they were in the 90s.
If these are fine, then anything that uses good hardwood should be fine. It doesn't require great equipment, highly skilled joinery or any special treatment.
Some less crappy new speakers
In 2014, I scored a stack of old floorboards for free. I used some to build a pair of speakers which I gave to a friend.
The techniques were still basic: butt joins and PVA glue. One improvement was having power tools, another was using dowel joints. I don't have a special tool for the dowels - I just drill a hole, squeeze in some glue, then bang a piece of dowel in, exactly like a nail.
I did make the total noob / chump mistake of coating the inside of each box with liberal amounts of old house paint, pouring it in and sloshing it about, using at least a couple of litres per box.
The idea was to seal and deaden the interior, but the downside was that the boxes absorbed a lot of water from the paint. This was about the worst thing you could do to a speaker box if you were worried about moisture destabilising it. It was the equivalent of leaving them out in heavy rain for a few days.
So what happened, in response to this travesty?
Some of the timber did swell: no joints opened, but a couple beams ended up a tad fatter than their neighbors, so they stood proud of them by 0.2 of a millimeter or so. I wasn't at all keen to start over, so I sanded them for ten minutes, let them dry, and sanded them again. 'Problem' solved. About 8 months later, these speakers are still perfectly intact, with no movement visible, all the the joints are as tight as they were last year.
So I guess that they could literally be left out in heavy rain without splitting, particularly now that they have been oiled.
The techniques were still basic: butt joins and PVA glue. One improvement was having power tools, another was using dowel joints. I don't have a special tool for the dowels - I just drill a hole, squeeze in some glue, then bang a piece of dowel in, exactly like a nail.
I did make the total noob / chump mistake of coating the inside of each box with liberal amounts of old house paint, pouring it in and sloshing it about, using at least a couple of litres per box.
The idea was to seal and deaden the interior, but the downside was that the boxes absorbed a lot of water from the paint. This was about the worst thing you could do to a speaker box if you were worried about moisture destabilising it. It was the equivalent of leaving them out in heavy rain for a few days.
So what happened, in response to this travesty?
Some of the timber did swell: no joints opened, but a couple beams ended up a tad fatter than their neighbors, so they stood proud of them by 0.2 of a millimeter or so. I wasn't at all keen to start over, so I sanded them for ten minutes, let them dry, and sanded them again. 'Problem' solved. About 8 months later, these speakers are still perfectly intact, with no movement visible, all the the joints are as tight as they were last year.
So I guess that they could literally be left out in heavy rain without splitting, particularly now that they have been oiled.
Wee subwoofer
I picked up a HT sub up from a garage sale for a few bucks. Eventually, I got around to peeling off the vinyl covering and cladding them in hardwood. I'm still not 100% done with the finish. I might get onto that, and possibly sell them off, some fine day.
The first bit of good news is that this is a pretty sweet and simple construction method. Just glue and clamps again, made even easier because I was clamping to an existing object. The MDF core also guarantees against leaks. I'd happily use this technique from scratch, possibly outsourcing the MDF inner box to some CNC cutting place - so they can cop the noise, dust and sinonasal cancer on my behalf.
The second, predictable bit of good news is that the box is stable. Again, no rocket surgery, just dumb butt joins and glue. I haven't even oiled it / done anything to seal the timber against moisture.
The first bit of good news is that this is a pretty sweet and simple construction method. Just glue and clamps again, made even easier because I was clamping to an existing object. The MDF core also guarantees against leaks. I'd happily use this technique from scratch, possibly outsourcing the MDF inner box to some CNC cutting place - so they can cop the noise, dust and sinonasal cancer on my behalf.
The second, predictable bit of good news is that the box is stable. Again, no rocket surgery, just dumb butt joins and glue. I haven't even oiled it / done anything to seal the timber against moisture.
Tables
Loudspeakers are my focus, obviously, but I am including these tables for completeness / cos they are in my home.
TABLE A) I bought a table in 2003 which I believe uses new timber. It's built in a kinda butcher block style. The maker preferred to use slab timber, he seemed to regret having made this table. I think it is a one off. Anyway, 12 years on, the joins are still perfect, there is no warp or cupping. As far as I can tell, anyone with a good saw could replicate this:
1) slice up the source timber (railway sleepers or whatever)
2) invert every second strip, so the direction of any tendency to warp is also inverted, and averages out to nothing
3) clamp and glue
4) oil and/or seal the finished object
I took a photo of the end grain, to show that the maker probably did something along those lines.
TABLE B) This is something I slung together to use as a standing computer table, partly because I don't want to embrace the Australian cultural norm of becoming a fat bastard as one slides into middle age. This is the only project to which I made any concession to the hygroscopic nature of timber: I've only fastened the wood to the steel base at one point, so it is free to stretch out without buckling against the merciless steel.
TABLE A) I bought a table in 2003 which I believe uses new timber. It's built in a kinda butcher block style. The maker preferred to use slab timber, he seemed to regret having made this table. I think it is a one off. Anyway, 12 years on, the joins are still perfect, there is no warp or cupping. As far as I can tell, anyone with a good saw could replicate this:
1) slice up the source timber (railway sleepers or whatever)
2) invert every second strip, so the direction of any tendency to warp is also inverted, and averages out to nothing
3) clamp and glue
4) oil and/or seal the finished object
I took a photo of the end grain, to show that the maker probably did something along those lines.
TABLE B) This is something I slung together to use as a standing computer table, partly because I don't want to embrace the Australian cultural norm of becoming a fat bastard as one slides into middle age. This is the only project to which I made any concession to the hygroscopic nature of timber: I've only fastened the wood to the steel base at one point, so it is free to stretch out without buckling against the merciless steel.
Appeal to authority
From Brittanica.com:
"Coatings do not reduce the quantity of moisture the wood can hold, but they slow the exchange of moisture between wood and atmosphere and, therefore, reduce the magnitude of dimensional changes of the wood in use. Most dimensional problems are caused by the use of wood with excessive moisture content. Instead, at the time of use the wood should have a moisture content at the approximate midpoint of the expected range in a particular location. This practice minimizes moisture content changes and, therefore, the adverse effects of shrinkage and swelling."
These seem to be the two important points. Use well dried wood. Slap some oil on it. Done.
"Coatings do not reduce the quantity of moisture the wood can hold, but they slow the exchange of moisture between wood and atmosphere and, therefore, reduce the magnitude of dimensional changes of the wood in use. Most dimensional problems are caused by the use of wood with excessive moisture content. Instead, at the time of use the wood should have a moisture content at the approximate midpoint of the expected range in a particular location. This practice minimizes moisture content changes and, therefore, the adverse effects of shrinkage and swelling."
These seem to be the two important points. Use well dried wood. Slap some oil on it. Done.
Odds and ends
Salvage timber is widely available. I recommend it.
Where I live:
- a quick web search shows >10 specialist timber salvage businesses. There are more if I include general demolition yards.
- salvaged timber gets put up for sale online by randoms.
- some council operated rubbish dumps sell timber as part of their recycling efforts.
A web check of small cities in England and the USA shows that similar timber resellers exist elsewhere; they are not just a local oddity.
Most of the above options are cheaper than new timber. Free is also possible. I've picked up free salvage timber:
- simply by asking. Just while cycling about, I sometimes stop and ask people who have a skip out / demolition underway (I currently have ~100m of floorboards from this)
- from friends / trades (e.g. helping a friend with a demolition / renovation).
- and others have had success online
Even slightly warped, cupped, off-square or grimy old wood can look great once a small amount is stripped off the surface. You can mount a router to a home-made guide / platform to strip & level beams, if you don't have access to more specialised tools. I'll try to find a photo of this.
You can source benchtops from hardware stores, or tabletops from Ikea that are of similar construction to the table pictured above: made from many small blocks of solid wood. I assume they are similarly stable. Their price is OK, similar to good ply.
Where I live:
- a quick web search shows >10 specialist timber salvage businesses. There are more if I include general demolition yards.
- salvaged timber gets put up for sale online by randoms.
- some council operated rubbish dumps sell timber as part of their recycling efforts.
A web check of small cities in England and the USA shows that similar timber resellers exist elsewhere; they are not just a local oddity.
Most of the above options are cheaper than new timber. Free is also possible. I've picked up free salvage timber:
- simply by asking. Just while cycling about, I sometimes stop and ask people who have a skip out / demolition underway (I currently have ~100m of floorboards from this)
- from friends / trades (e.g. helping a friend with a demolition / renovation).
- and others have had success online
Even slightly warped, cupped, off-square or grimy old wood can look great once a small amount is stripped off the surface. You can mount a router to a home-made guide / platform to strip & level beams, if you don't have access to more specialised tools. I'll try to find a photo of this.
You can source benchtops from hardware stores, or tabletops from Ikea that are of similar construction to the table pictured above: made from many small blocks of solid wood. I assume they are similarly stable. Their price is OK, similar to good ply.