Bending 3 mm plywood

[woodbrains]

Hello,

The whole point in plywood, is that it is designed to remain flat! Wanting to do something with a material that is designed to do precisely the opposite to your intention is just daft, IMHO. Use bending plywood with constructional veneers, or aero ply between the laminations and what ever face veneers you choose on both outside faces. This will make a very stiff structure that will stay in shape and not require much, if any an understructure. 3 layers of 5mm bending ply is good.

You could use 2mm MDF.

Mike.

 

[profchris]

How big is the structure? If it's domestic loudspeaker size then you have two options.

1. 3mm ply can be bent over a hot pipe - I've done that to make musical instrument cases. A pipe around 2 inch diameter in a vice with a heat gun blowing through it works well. It takes practice to bend precisely, but it's not that hard.

2. Cheaper 1.5mm ply is available via eBay - you can usually pick up 10 12 inch square sheets for about £15 delivered. These could be bent and glued cold.

If stiffness is the important criterion then I don't think you'd need to go to 18mm thick. I'm reasonably confident that a piece of 3mm ply bent to a 150mm radius might well hold my weight if I stood on it (if the ends were anchored), because curves are vastly stiffer than flat, and I'm sure that a two layer (6mm) instrument case would do so.

 

[h.g.]

The whole point in plywood, is that it is designed to remain flat! Wanting to do something with a material that is designed to do precisely the opposite to your intention is just daft, IMHO. Use bending plywood with constructional veneers, or aero ply between the laminations and what ever face veneers you choose on both outside faces. This will make a very stiff structure that will stay in shape and not require much, if any an understructure. 3 layers of 5mm bending ply is good.

Thanks for the input but I thought I was proposing to use the same construction method as that for wooden aeroplanes: a wooden frame and a plywood skin although in my case the skin uses more layers. I would like to use 1.5mm 3 ply aero plywood but cannot afford it (unless someone knows where I can get it cheap?) and so I am looking at 3 mm sheets of 3 uniform plys plus more messing about. High stiffness is a requirement which is why there is a frame and the more straightforward options of MDF, flexiply, kerfing and the like have been set aside. An attractive surface finish is not important because it will not be visible.

 

[h.g.]

How big is the structure?

It is about 1' x 1.5' x 2.5' and is the size of a domestic loudspeaker.

1. 3mm ply can be bent over a hot pipe - I've done that to make musical instrument cases. A pipe around 2 inch diameter in a vice with a heat gun blowing through it works well. It takes practice to bend precisely, but it's not that hard.

Thanks. That sounds exactly like what I am after. Did you use the cheap 3 mm ply with 2 thin outer layers and a thick core or the more expensive 3 mm ply with uniform thickness?

2. Cheaper 1.5mm ply is available via eBay - you can usually pick up 10 12 inch square sheets for about £15 delivered. These could be bent and glued cold.

Thanks again. I had seen a few of these but could not find anything large enough. My panels are about 650 x 850 mm. I have sent a couple of emails asking about larger sizes but with no response yet.

If stiffness is the important criterion then I don't think you'd need to go to 18mm thick. I'm reasonably confident that a piece of 3mm ply bent to a 150mm radius might well hold my weight if I stood on it (if the ends were anchored), because curves are vastly stiffer than flat, and I'm sure that a two layer (6mm) instrument case would do so.

The requirement for stiffness is not really to do with strength but a need to raise the lowest resonant frequency higher than it would normally be in speaker using flat panels.

 

[profchris]

I used the cheap ply - this was only an instrument case after all. But the size of your panels seems too big for a hot pipe. 1.5mm ply is expensive in large panels, it's only the offcuts which you can get cheaply.

If it helps, I've also bent 3mm MDF over a hot pipe (dry bending only) and that works OK too, and I suspect you could bend it over a form using a heat gun. But I doubt that layering up MDF would be as stiff as you want.

You might try the experiment of attempting to bend 3mm ply over a form using a heat gun - try a narrow 150mm panel using a big saucepan as your form. Clamp one end to the saucepan, clamp the saucepan down somehow (handle in vice?), heat the ply where the bend is to start (you want it hot enough so you can't rest your hand on it, but short of scorching!) and ease it into shape, moving the heating area as you go. You will need to go slowly, applying only light pressure, or it might bend unevenly (you can actually feel the wood turn plastic with a bit of practice).

If this works you could build a form the right size and go for it. This would also tell you how many 3mm layers you'd need - my guess is two or three.

 

[woodbrains]

Hello,

Are you sure you are not putting too much empasis on the strength of the material you want to use. I've built speaker boxes before, and have an interest in such. I don't believe the system resonance can be altered by the rigidity of the box material. I think you need the material to a) not vibrate at the sound volume the speaker is used at and b) not have any voids of loose plies within the structure and c) not have any unintentional holes in it. Resonance is modified by the compliance of the air in the box and size shape and length of any ports in the box. In fact some enclosures like labyrinths are non-resonant. As long as the speakers are mounted on a strong enough baffle the air envelope does not know whether the box is made from MDF, ply or solid timber. In fact it is a good idea to line the insides of the box with a yielding material anyway such as bitumastic sheet.

For this reason I would avoid saw kerfing, as you already have disqualified. But I would also avoid trying to bend ordinary plywood. If it can be done at all it will only be a slight bend and you will find the pieces will 'saddle' which not only is not the curve you want, but it will layer up with voids, as each piece will be slightly different to another and not nest.

Bending ply will be fine for what you need, and since you need to know the radius of the curve to be able to calculate the internal volume accurately, bending ply will give a more accurate shape, without spring back. Trying to bend plies that are too thick will result in a lot of spring back.

Mike.

 

[h.g.]

But the size of your panels seems too big for a hot pipe.

A flue pipe perhaps but I had thought to bend a sheet of metal and heat that possibly over the internal frame for the speaker or a bit tighter depending on the degree of spring back.

You might try the experiment of attempting to bend 3mm ply over a form using a heat gun - try a narrow 150mm panel using a big saucepan as your form. Clamp one end to the saucepan, clamp the saucepan down somehow (handle in vice?), heat the ply where the bend is to start (you want it hot enough so you can't rest your hand on it, but short of scorching!) and ease it into shape, moving the heating area as you go. You will need to go slowly, applying only light pressure, or it might bend unevenly (you can actually feel the wood turn plastic with a bit of practice).

Thanks I hadn't considered hot and dry but if it works it should be quicker and less messy. Well worth an experiment and can always fall back to hot water and towels and/or steam if not.

If this works you could build a form the right size and go for it. This would also tell you how many 3mm layers you'd need - my guess is two or three.

Tapping the speaker with a hammer and looking at the frequency spectrum of the response will show me if I need more layers. The BBC used to use 9mm plywood to get LOW resonant frequencies in some of their famous old monitor designs (they did one or two other things as well to get low resonant frequencies).

 

[DennisCA]

I got some 4mm birch plywood and it bends like heck in one direction. with unsteamed ply I made this from 4 laminations for my kids:

I didn't specify bendy plywood or anything, but that's what it does.

 

[h.g.]

Are you sure you are not putting too much empasis on the strength of the material you want to use. I've built speaker boxes before, and have an interest in such. I don't believe the system resonance can be altered by the rigidity of the box material. I think you need the material to a) not vibrate at the sound volume the speaker is used at and b) not have any voids of loose plies within the structure and c) not have any unintentional holes in it. Resonance is modified by the compliance of the air in the box and size shape and length of any ports in the box. In fact some enclosures like labyrinths are non-resonant. As long as the speakers are mounted on a strong enough baffle the air envelope does not know whether the box is made from MDF, ply or solid timber. In fact it is a good idea to line the insides of the box with a yielding material anyway such as bitumastic sheet.

Sound is radiated by the drivers (wanted) and from the cabinet (not wanted). My concern is reducing the latter sufficiently to be inaudible in the presence of the former. If you have a separate enclosure for the low frequency woofer/s which is isolated from the enclosure for the midrange and tweeter then a high stiffness can push the lowest resonant frequency of the woofer cabinet above the highest frequency handled by the woofers and so the resonances will not got driven. If there are no resonances then the cabinet requires very little damping which is a significant win because effective damping has a significant cost. Subwoofer cabinets are usually designed like this but I am wanting to do it for a woofer cabinet which is pushing things a bit but hopefully not too much.

The midrange cabinet needs to be designed differently because it will contain cabinet resonances in the passband. A high stiffness will not do much that is beneficial and it is high damping that will lower the resonant peaks. Which is just as well because effective damping tends to come at the cost of a reduced stiffness.

 

[woodbrains]

Hello,

Just make the box from thicker material. Stiffness goes up markedly with quite moderate increases in thickness. It seems to me that finding out the resonant frequency of the plywood or whatever would be virtually impossible, anyway, since it would vary with size for any given thickness. An 8 by 4 sheet is less stiff than a 4 by 4 sheet of the same thickness. Just make the box thicker for the woofer cabinet by as much as you think is more than you need. Maybe use 5 layers of 5 mm bendy ply. Otherwise there is no way around using the expensive materials!

Alternatively, if you can design out the curve and make the enclosure from flat panels. Makes construction easier and less expensive. Why are you wanting a curve? Is it to prevent standing waves? If so a many sided polygon will do as well, especially if it isn't parallel to the speaker baffle.

Mike.

 

[h.g.]

Just make the box from thicker material. Stiffness goes up markedly with quite moderate increases in thickness.

Stiffness goes up but so does mass and, unfortunately, the resonant frequency scales only with the square root of the stiffness/mass ratio so we get an increase but not by anything like the amount the stiffness has gone up.

It seems to me that finding out the resonant frequency of the plywood or whatever would be virtually impossible, anyway, since it would vary with size for any given thickness. An 8 by 4 sheet is less stiff than a 4 by 4 sheet of the same thickness.

The resonant frequencies can be predicted using Finite Element Analysis with the CAD model of the structure. There are one or two things which lead these predictions to being close rather than precise but they are normally good enough to answer most things. Once the structure has been built then tapping it with hammer, recording the 'ping' it makes and then creating the frequency spectrum for that 'ping' will show the frequencies of the resonances.

Alternatively, if you can design out the curve and make the enclosure from flat panels. Makes construction easier and less expensive. Why are you wanting a curve? Is it to prevent standing waves? If so a many sided polygon will do as well, especially if it isn't parallel to the speaker baffle.

The woofers are in the side of the cabinet and the sound needs to pass around to the front without encountering any sharp edges which would cause it to diffract and act like a second source which interferes with the direct sound. In addition it is part of the waveguide on the tweeter.

Approximating a curve with a series of straight sides would work better than a single sharp edge but not as well as a curve at the frequencies handled by the tweeter. At lower frequencies there probably wouldn't be much difference. If I had a CNC machine to play with then it might be easier than bending some plywood but by hand my guess is bending plywood will be easier.

 

[woodbrains]

Hello,

Bendy ply is lighter than birch ply.

Unless you find a compromise, you can't make what you want. All design is compromise. Always. You can't bend 3mm ply, so what are you going to do, no one here can make it happen by talking about it, that is for sure. Just make a box solid enough that does not vibrate. All your calculations are just so you can make a box that does not vibrate, however much you make a song and dance about it. Calculations do not make a speaker sound good, making it sound good does. Just do it, and good luck.

Mike.

 

[h.g.]

You can't bend 3mm ply, so what are you going to do, no one here can make it happen by talking about it, that is for sure. Just make a box solid enough that does not vibrate. All your calculations are just so you can make a box that does not vibrate, however much you make a song and dance about it. Calculations do not make a speaker sound good, making it sound good does.

Thanks for the input but I think we may be at cross purposes somewhere. I may not be experienced in how to do it, which is why I am posting to the forum, but I am pretty confident one can bend 3 mm plywood if it is wet and heated. I agree it will not bend cold and dry to the radius I want and flexiply possibly will but flexiply is unlikely to be stiff enough to do the job required. A near miss with the stiffness is a big fail because the resonances will then get strongly driven, in a box that is effectively undamped and at a frequency where the ear is relatively sensitive.

All speaker boxes vibrate. Good speaker boxes use the appropriate combinations of mass, damping and stiffness to minimise the cabinet vibration caused by the drivers. How to do this effectively changes with frequency which makes the construction of a good cabinet an interesting challenge. Solid in the sense of thick walls is a widely held belief among speaker DIY folk but it is not particularly effective at midrange frequencies. A church bell for example is pretty solid but it would not make a good speaker cabinet.

 

[Beau]

Speaker cabinets are complicated old things. Was privileged to be shown around the Heybrook speaker factory some years ago by it's founder. They had done a lot of research into the best materials for the conventional box and chipboard was the preferred material as it did not "ring" .They did not rate solid wood, ply, or any of the MDF offerings. Sorry no help for your project but understand that there is more to it than strength.

 

[Racers]

Well h.g. we have given you some answers but you don't seem to like them!

My speakers are MDF boxes with plastic feet that rock about but sound very good indeed (they retail for about 4K) so there is more than one way to skin a cat.

Why do you need the sound directing to the front? my speakers fire upwards at about 45 deg.

Pete

 

[h.g.]

Well h.g. we have given you some answers but you don't seem to like them!

I don't know why you would think that. I came to the forum somewhat uncertain about bending plywood and have been given advice that has both set my mind at rest and also given me one or two things to try which, if they work, will be better than what I was initially intending to do. Good stuff.

My speakers are MDF boxes with plastic feet that rock about but sound very good indeed (they retail for about 4K) so there is more than one way to skin a cat.

The design of commercial speaker cabinets is a balance of a wide range of concerns. A bit like a family car being a lot more difficult to design than a racing car where the objectives are much better defined. In my case the speaker is a bit unusual in being an on wall design to be placed against a light partition wall which will not take it's weight and would act like a sounding board if the speaker vibrated like a conventional speaker. The fun lies as much, if not more, in the challenge of getting things to work well rather than the construction of the speaker itself.

Why do you need the sound directing to the front? my speakers fire upwards at about 45 deg.

If I understand your question correctly, the speakers are designed to be conventional in the trade off between imaging and spaciousness. The directivity is smooth with a beam width that is a touch wider than a good studio monitor.

 

[Racers]

I have problems with speakers that shout directly at you I much prefer Semi-omnidirectional speakers.

I can understand why you wouldn't want the wall exciting.

The best mod I ever did to an old pair of speakers was to paint the inside with car underseal it damped the cabinets very well.

Pete

 

[sploo]

I used to be heavily into loudspeaker (and sub) building, but it's been quite a few years, and memory of specific values/frequency/material/diffraction issues has faded - so, it's entirely probable some of my info may be wayward, but...

It is about 1' x 1.5' x 2.5' and is the size of a domestic loudspeaker.
...
The requirement for stiffness is not really to do with strength but a need to raise the lowest resonant frequency higher than it would normally be in speaker using flat panels.

Linkwitz recommends that you don't have any panel area larger than 4"x4" unbraced, in order to push the resonant frequency high enough such that (at those frequency levels) there won't be enough energy to excite them. If you're going for a bent lamination then I assume you might require some bracing/skeleton inside the cabinet. Could you get anywhere near that level of bracing, and then likely make the resonance issue irrelevant?

Sound is radiated by the drivers (wanted) and from the cabinet (not wanted). My concern is reducing the latter sufficiently to be inaudible in the presence of the former. If you have a separate enclosure for the low frequency woofer/s which is isolated from the enclosure for the midrange and tweeter then a high stiffness can push the lowest resonant frequency of the woofer cabinet above the highest frequency handled by the woofers and so the resonances will not got driven. If there are no resonances then the cabinet requires very little damping which is a significant win because effective damping has a significant cost. Subwoofer cabinets are usually designed like this but I am wanting to do it for a woofer cabinet which is pushing things a bit but hopefully not too much.

IRC The issue is usually that the frequencies produced by a mid driver are the ones that tend to excite the panels on a larger cabinet. Therefore, building a small mid/treble cabinet and an isolated larger woofer cabinet prevents that transmission. AFAIU the frequencies produced by a woofer are usually too long/low to excite all but a really huge panel (even though there's plenty of energy). If I read your comments above correctly then you're trying to stiffen the woofer cab to protect it from woofer frequencies?

The woofers are in the side of the cabinet and the sound needs to pass around to the front without encountering any sharp edges which would cause it to diffract and act like a second source which interferes with the direct sound. In addition it is part of the waveguide on the tweeter.

I can't recall the radius vs frequency formula, but you do tend to need some pretty large radii to alleviate diffraction at anything but the highest frequencies. I'm just bringing that up in a "how practical will it be to get a large enough radius?" question.

I'd very much like to see any designs or finished boxes though - I'm still very much "in" to speaker builds, even though I don't tend to do them myself these days.

 

[h.g.]

Linkwitz recommends that you don't have any panel area larger than 4"x4" unbraced, in order to push the resonant frequency high enough such that (at those frequency levels) there won't be enough energy to excite them. If you're going for a bent lamination then I assume you might require some bracing/skeleton inside the cabinet. Could you get anywhere near that level of bracing, and then likely make the resonance issue irrelevant?

Pushing resonances high works if the driver creating those frequencies is elsewhere and not exciting the cabinet. The energy to excite the resonances comes from the drivers and will be just as much present at midrange and high frequencies as at bass frequencies. What is different is that it takes less movement of a panel to create a given sound pressure level at high frequencies than at low (e.g. large woofer and small tweeter).

The inside of the cabinet is braced with unsupported areas slightly larger than 4"x4". Extensive bracing does not make resonances quieter, it raises the frequency of the lowest ones which may or may not be desirable depending on how the noise is being tackled. The BBC for example aimed to lower the frequency of the resonances in their old monitors. What reduces the size of resonances is damping. Exactly on a resonance it is the only force that reduces the motion because the forces due to stiffness and inertia, although large, cancel each other.

If I read your comments above correctly then you're trying to stiffen the woofer cab to protect it from woofer frequencies?

Sort of. If the resonances are above the passband of the woofer then the largest force (of stiffness, damping and mass) is stiffness. The cabinet is "stiffness controlled" and so making it stiffer reduces the motion making it quieter. If you were to make the cabinet heavier keeping the same stiffness the motion would be largely unaffected.

I'd very much like to see any designs or finished boxes though - I'm still very much "in" to speaker builds, even though I don't tend to do them myself these days.

I have a CAD model but want to check things first. The design seems a reasonably straightforward thing to do and yet it doesn't seem to be done. The obvious reason for that is that I have missed something and it doesn't work well but I am hoping the reason is a less probable one.

 

[sploo]

Pushing resonances high works if the driver creating those frequencies is elsewhere and not exciting the cabinet.

I don't quite follow that? Surely if the bracing pushes the resonances high then it works just as well if the driver is in another cabinet (that's transferring energy to that cabinet) or is directly mounted in that cabinet. Obviously the energy will be higher with the driver in the cabinet, but if the cab is braced sufficiently then there should be relatively little resonance in the lower frequencies (due to the bracing).

The inside of the cabinet is braced with unsupported areas slightly larger than 4"x4". Extensive bracing does not make resonances quieter, it raises the frequency of the lowest ones which may or may not be desirable depending on how the noise is being tackled.

Indeed. But the question surely is: if you've removed a lot of resonance issues, are the remaining resonances a genuine problem?

I ask that question because - ignoring the downright crazy audiophool stuff - there's a lot of effort in the audio world that seems to be spent on reducing problems that aren't actually detectable by the human ear (measurable doesn't mean audible).