Bending 3 mm plywood

[h.g.]

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).

Bracing stiffens the cabinet and pushes the frequency higher at which resonances start to occur. It does not reduce the sound pressure level of those resonances at the listening position which I suspect is what you are expecting to happen. At high frequencies a panel needs to deflect less to be equally loud as at low frequencies. To see that consider how far a large woofer has to move to be equally loud as a small tweeter. In practise they tend to be roughly the same level but we are able to perceive resonances below the signal easier at higher frequencies and so the net result is likely to be to make things slightly worse not better although not by a great amount.

So what you need to do to make things work well is put the woofer and midrange in separate cabinets that are isolated from each other and build those cabinets differently. The woofer cabinet needs to be stiff so that lowest resonances are in the midrange passband and will not get driven because the midrange is elsewhere. The midrange cabinet needs to be heavily damped and less stiff so that the lowest resonances are in the woofer passband and not driven and the higher order resonances that are in the midrange passband are strongly damped.

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).

Yes I am removing resonance issues in the way outlined above and I am doing it a pragmatic way without using 2" thick mdf, excessive bracing and similar favoured by many DIY speaker enthusiasts. There are a small number of audibility studies of cabinet resonances around which show, not surprisingly, that audible levels are around 30 dB down in the same ball park as other audible grunge.

Audiophile hardware is designed to get bought by audiophiles. If audiophiles want, for example, multiple binding posts for biwiring/biamping or whatever then speakers that do not have them are not going to get bought. It doesn't matter that biwiring/biamping or whatever isn't detectable. It doesn't matter that the manufacturers know this. A manufacturer needs to provide multiple binding posts because that is what the customer wants.

In the 1970s hi-fi enthusiasts used to give a lot of weight to small differences in completely inaudible levels of total harmonic distortion of a 1 kHz sine wave into an 8 ohm resistor when deciding which amplifier to purchase. Today audiophiles make fun of such foolish consumers but then base their purchasing decisions of DACs on small differences in jitter that are even further below audibility thresholds. In both eras the manufacturers were perfectly well aware of the irrelevance to real world performance but they are businesses giving the consumer what they want to buy.

 

[profchris]

Reading all this (mostly mysterious to me!) a thought occurs which might be stupid, but ...

What about an inner and outer skin of 1.5 mm ply (or aluminium, or plastic) with expanding foam to fill the gap? That should be rigid with high damping, but also give you a smooth reflective surface. Easy to build, compared with all that bending.

 

[h.g.]

What about an inner and outer skin of 1.5 mm ply (or aluminium, or plastic) with expanding foam to fill the gap? That should be rigid with high damping, but also give you a smooth reflective surface. Easy to build, compared with all that bending.

If things work out the next one may well use composite panels, double curvature and be more CNC orientated. Single curvature was a compromise to ease construction and was not in the initial paper design. I did briefly consider honeycomb but I need to rout recesses for the drivers, a conical shape for the waveguide, blending chamfers, a lot of tapped holes for movable panels (it is something of a prototype and I do not know the optimum configuration),... By hand this seemed more straightforward with plywood compared to making a lot of insets for a composite panel.

 

[SteveF]

what about just glue up a load of boards flat and shape by hand?
you would not need many and would not have to be full size boards as you will be removing a lot of them in the shaping?

Steve

 

[sploo]

Bracing stiffens the cabinet and pushes the frequency higher at which resonances start to occur. It does not reduce the sound pressure level of those resonances at the listening position which I suspect is what you are expecting to happen.

Nope - I understand the concept of pushing the resonance higher vs reducing sound pressure level - my question was whether there would be enough energy at those higher frequencies to be audible.

Now that said...

So what you need to do to make things work well is put the woofer and midrange in separate cabinets that are isolated from each other and build those cabinets differently. The woofer cabinet needs to be stiff so that lowest resonances are in the midrange passband and will not get driven because the midrange is elsewhere. The midrange cabinet needs to be heavily damped and less stiff so that the lowest resonances are in the woofer passband and not driven and the higher order resonances that are in the midrange passband are strongly damped.

I was meaning make "the" cabinet (or all cabinets) braced sufficiently to push resonances higher (thus it wouldn't matter where the drivers were). However, building separate cabinets targeted for the specific resonance issues, makes good sense and is a pragmatic bit of engineering.

Yes I am removing resonance issues in the way outlined above and I am doing it a pragmatic way without using 2" thick mdf, excessive bracing and similar favoured by many DIY speaker enthusiasts. There are a small number of audibility studies of cabinet resonances around which show, not surprisingly, that audible levels are around 30 dB down in the same ball park as other audible grunge.

Audiophile hardware is designed to get bought by audiophiles. If audiophiles want, for example, multiple binding posts for biwiring/biamping or whatever then speakers that do not have them are not going to get bought. It doesn't matter that biwiring/biamping or whatever isn't detectable. It doesn't matter that the manufacturers know this. A manufacturer needs to provide multiple binding posts because that is what the customer wants.

In the 1970s hi-fi enthusiasts used to give a lot of weight to small differences in completely inaudible levels of total harmonic distortion of a 1 kHz sine wave into an 8 ohm resistor when deciding which amplifier to purchase. Today audiophiles make fun of such foolish consumers but then base their purchasing decisions of DACs on small differences in jitter that are even further below audibility thresholds. In both eras the manufacturers were perfectly well aware of the irrelevance to real world performance but they are businesses giving the consumer what they want to buy.

Indeed. It's a world that contains more nonsense and non-science than just about anything else I know :wink:

Random other thought; I'm looking into 3mm sheets of PETG plastic for a forming/bending application. Have you considered using something like that instead of ply? Lots of different engineering issues granted.

 

[h.g.]

what about just glue up a load of boards flat and shape by hand?
you would not need many and would not have to be full size boards as you will be removing a lot of them in the shaping?

It is an option and had I stuck with the original double curvature or decided bending ply was going to be a problem it is what I would most likely have done. It is a fairly popular method for small scale manufacture of expensive speakers using a CNC machine to cut the stacks and was how the local woodworking firm suggested they would make the original paper design with double curvature panels.

 

[h.g.]

my question was whether there would be enough energy at those higher frequencies to be audible.

The rate of energy pumped into the cabinet is the product of the reaction force on the driver frame to the accelerations of the cone and the velocity in direction of that force by the cabinet. Broadly whatever is required to be done on the air to generate a given sound pressure level is equally done on the cabinet in reaction. So yes there is broadly the same proportion of energy available to generate a matching sound pressure level by the cabinet as the driver. Note that the relative amount of energy at high and low frequencies is not relevant, it is the amount relative to the driver that is relevant.

Having said that, in practise it is easier to control vibration at higher frequencies than low and so they tend to be less of a problem. Damping is more effective and, for example, a tweeter with a reasonably heavy frame does not need the mass of the cabinet to react against in the way a woofer usually does and soft rubber grommets could be used to isolate the vibration of the body of the tweeter from the cabinet.

I was meaning make "the" cabinet (or all cabinets) braced sufficiently to push resonances higher (thus it wouldn't matter where the drivers were).

With normal construction and materials it is possible to push the resonant frequencies into the midrange passband but not into the tweeter passband and certainly not above the tweeter passband. A speaker cabinet will resonate at audible frequencies and the task is to make those resonances sufficiently low in level to be inaudible.

Random other thought; I'm looking into 3mm sheets of PETG plastic for a forming/bending application. Have you considered using something like that instead of ply? Lots of different engineering issues granted.

Not at this stage. I would expect plywood to be stiffer, lighter and often cheaper when bought as sheets. Plastic mouldings have a significant presence in studio monitors and so it is more a question of the cost and effort of the tooling rather than the material. They can be designed to have significantly more damping than most wood products which is useful.

 

[sploo]

The rate of energy pumped into the cabinet is the product of the reaction force on the driver frame to the accelerations of the cone and the velocity in direction of that force by the cabinet. Broadly whatever is required to be done on the air to generate a given sound pressure level is equally done on the cabinet in reaction. So yes there is broadly the same proportion of energy available to generate a matching sound pressure level by the cabinet as the driver.

Good point. I guess there's potentially also a larger radiating area of cabinet than driver (though off the top of my head I don't know how that would affect the unwanted resonances).

Having said that, in practise it is easier to control vibration at higher frequencies than low and so they tend to be less of a problem. Damping is more effective and, for example, a tweeter with a reasonably heavy frame does not need the mass of the cabinet to react against in the way a woofer usually does and soft rubber grommets could be used to isolate the vibration of the body of the tweeter from the cabinet.

Linkwitz tends to mount mid/bass drivers by the magnet in order to reduce transmission to the baffle. Would that be helpful/effective in your application?

Not at this stage. I would expect plywood to be stiffer, lighter and often cheaper when bought as sheets. Plastic mouldings have a significant presence in studio monitors and so it is more a question of the cost and effort of the tooling rather than the material. They can be designed to have significantly more damping than most wood products which is useful.

It (PETG) is something I've only just come across myself (no experience with working with it) but just wondered. Most of my builds were MDF, with the occasional bit of ply. I'll admit I've generally gone for the heavy and well damped approach, as I've not had need to create something lighter.

 

[h.g.]

Linkwitz tends to mount mid/bass drivers by the magnet in order to reduce transmission to the baffle. Would that be helpful/effective in your application?

Yes for the woofers because they are in force cancelling pairs. Unfortunately for reasons of cost in the first build I have opted for modestly priced midwoofers that do not have a way to bolt to the frame on the magnet and so I will simply use metal rods between the normal mounting holes on the frame. Not ideal and one of the things to be addressed in the next version if there is one.

Yes for the midrange because it sits behind a waveguide that isn't deep enough to take bolts for rear mounting. However, the chosen midrange also doesn't have a means of bolting to the frame around the magnet. This has forced mounting from the front in the conventional way with a bit too much wood blocking the cone at the rear and, worse, the need for a conical ring to be inserted after the driver to fill the gap from the lip of the cone to the wall of the hole required to insert the driver. An unwanted complication caused by the need to be able to assemble the speaker!

 

[sploo]

Yes for the woofers because they are in force cancelling pairs. Unfortunately for reasons of cost in the first build I have opted for modestly priced midwoofers that do not have a way to bolt to the frame on the magnet and so I will simply use metal rods between the normal mounting holes on the frame. Not ideal and one of the things to be addressed in the next version if there is one.

It's a good approach. I've built dual "boxer" style subs and I'm always amazed how little vibration you get with opposing pairs - even with large excursion.

Yes for the midrange because it sits behind a waveguide that isn't deep enough to take bolts for rear mounting. However, the chosen midrange also doesn't have a means of bolting to the frame around the magnet. This has forced mounting from the front in the conventional way with a bit too much wood blocking the cone at the rear and, worse, the need for a conical ring to be inserted after the driver to fill the gap from the lip of the cone to the wall of the hole required to insert the driver. An unwanted complication caused by the need to be able to assemble the speaker!

I've previously used a split 'O' bracket made from good plywood, which bolts together and the compression holds the magnet. Not practical with all drivers granted, and it indeed makes putting everything together more of a hassle.