Horses for courses - Optimal boxes for different passbands

I'm going dipole for my midrange. With the right size baffle I can apply zero baffle step compensation, and get an 87db/watt efficient speaker.

I'm also keeping the crossover extremly simple. 2nd order passive for the lowpass, and 2nd order line level passive for the highpass.

Dan

Hi Rick,

Your suggestions will move the resonances to different frequencies and this will help,... somewhat. However, ultimately you would only get partial results from just moving the panels resonance.

Every panel will resonate at some frequency. There's no way around that fact. Getting back to basics, there are only 3 ways to deal with a resonance; move it to a different frequency as you suggest by changing the values of the mass or spring or both, removing the excitation which you suggest by adding damping in the box, or applying damping.

For speakers boxes, moving the resonance is only a partial solution as the panel will still resonate at the new frequency. So unless you move the resonance completely out of the audible range, it will still be an issue to one extent or another.

It's impossible to remove all the excitaions because the excitations are coming from the speaker itself through the inside of the box, transmitted through the box panels, and through the air from the other speakers.

The most effective method is to damp the panels and the excitation source where possible (like your damping inside the box). This is where the material from which the panel is made can make a huge difference. Materials with high internal damping like carbon fiber can work very well in this regard. I've seen suggestion on the forum here to use various plastics. This too could work very well as many plastics have high internal dampings. I know there are some that don't like it but MDF has relatively high internal damping and is easy to get, relatively cheap to use, and easy to make into boxes.

One approach the takes advantage of each of the 3 methods to a greater or lessor extent is to use different shapes like curved panels. The analysis of a curved panel would be much more complex but the resonance would be moved upward, the panel would receive the excitation at different angles/times so the effects of the excitation would be less, and I'm not certain but I believe the damping would be effectively increased due to the complexity of the resonance modes in the curved panel.

Hi

When I built my last speaker system I had a box for the mid and treble and a separate bass unit.

I found that the critical box was the mid box. Here as you state in post 1 you need neutrality. I found this best achieved by using a concrete/corian box for the mid/treble unit. I shaped mine as a foward sloping pyramid. I left the interior walls rough and did not use any filler material.

For the base unit I used a rigid box 30mm walls with bracing made of hdf. I have used condrete in the past for a base unit but I did not find that the extra rigidity helped much. ( And the lighter weight of wood/hdf certainly does help when it comes to moving them around!!!)

Don

Rick,

Even though the designs that copied the BBC monitors had good sound for thier day they are bettered by other more modern techniques. Looking for the best approach you must look at the really expensive speakers. Basically you have to use exotic materials or exotic construction techniques or heavy overbuild using MDF and Plywood.

As far as I can say it works better for the mid/tweeter cabinet to be high mass/high internal dampening and rigid. By doing this you can move the resonances high into the overtones range and out of the fundementals where they do the most damage. This is also easy to accomplish with thick MDF and dampening and bracing. ALso, with acoustic foam and dampening pads attached to the walls some of the resonances can be attenuated by turning the air motion into physical motion in the foam which can attenuate the resonance by turning it into heat. Also the dampening can attenuate panel born resonances.

Also, cabinet designs like aperiodic ones work fantastic for the midrange driver because it lowers the internal air pressures in the cabinet which helps to keep internal air pressures from exciting the cabinet. For fantastic midrange sound avoid the tradition of putting the driver in a small sealed housing unless it's large enough for the driver to be critically damped or near critically damped.

For the bass speaker rigidity and strength is more important than mass. So I would use plywood in the bass cabinet with lots of bracing. The better quality the plywood the better since the better grades are more rigid. But I have found that doubled up plywood in the bass cabinet still sounds better than single sheets due to being more rigid. Now if you can move the panel resonances in the bass module high enough that the driver SPL is way down at those frequncies then there is a whole lot less problem and the sound quality improves a lot. And if the bass cabinet is decoupled from the mid/tweet cabinet it also helps to keep the bass resonances out of the mid/tweet cabinet.

Another technique that I have considered for the mid/tweet cabinet is to make the basic cabinet from a mix of epoxy resin and lead shot. Poured and cured in a cheap foam mold. And then skin both the inside and out with fiberglass or carbon fiber and epoxy resin. Thicker walls the better for the inside of the wall. The outer skin adds a lot of rigidity and can be made as thick as is practicle.

The molded part can also be made of epoxy and sand or crushed and washed rocks. THe more dense the better.

rick57 said:

Could you explain the difference between “adding damping in the box, or applying damping”

Click to expand...

No difference. I used the term "applying" in a general sense.

There are a lot of good points made in the link you provided in your last post. Many of which are basic to mechanical design but might be a bit counter-intuitive. Ideas like how stiffness without damping can make a vibration/noise problem worse.

One point that Art Ludwig makes is in regards to the energy applied to a speaker. Basically, the Watts in has to equal the Watts out of the speaker. Most drivers are less than 1% or 2% efficient. So for every 100W of power in, a typical speaker is only converting 1W of that energy to sound/music. Luckily 1 acoustic Watt of sound is quite loud (~115db@1m). However, that means the other 99W that is not converted to sound has to be converted to heat or go to excite the various components of the speaker. Much of that energy will go to heat in the driver(s) the rest will go into vibrating the components of the speaker (walls, braces, other drivers, crossover components, etc.). This vibrational energy will in turn, be coverted to sound/noise or to heat (through damping). So, the higher the damping, the more sound/noise energy is turned to heat, the better the enclosure "sounds". As you can see, damping is very important.

Rick,

Regarding CLD. I have not used this technique but I did read a paper from Wilson Audio on how they tested different CLD techniques. The best form of CLD is to use twin identical panels of the same mass and stiffness. Like two sheets of equal thickness MDF or Plywood. Then believe it or not the glue between the sheets has to remain soft and plyable but not too thick. I believe the ideal thickness range was between 5 - 10 thousands of an inch thick if memory serves.

I was surpriced to find out that the kind of adhesive had a larger effect than you would think. And the ldeal thickness was only a few thousands of an inch thickness. Appearantly when the two panels which have equal mass and stiffness vibrate they counteract each other by turning motion into heat at the adheasive layer. But if the glue is too thick it acts too much like a slinky kind of spring which in a worse case scenerio can amplify the resonance.

The interesting thing also is that Wilson does not seem to use much CLD anymore in their speakers. Maybe they do in the front baffle. But I have seen pictures of thier cabinets being built and they seem to use only 1/2 - 3/4 inch thick material. However the material is a very expensive composite aerospace panel made of phonolic resin and compressed linen or paper. THis stuff is lighter and stiffer and stronger than steel. And costs more than the drivers. It also must be machined with carbide tooling on CNC machines. In fact this material is so tough that it is used to make gears in applications were metals cannot be used. However it is not so easy to glue together. Which is why companies like Rockport and other have chosen a different cabinet construction method.

So in a way the Wilson's cabinet design may fall under the category of "low mass". And yet the panels are very dense. The material is heavier and dense than MDF for the same thickness.

One company that makes very high end sound speakers but uses rather common materials and methods is Thiel. They use thick MDF panels with lots of bracing and thick concrete or MDF baffles. THe concrete baffles sounded the best but they started using MDF for other reasons. Less costly to ship mostly.

Well sorry for the ramblings. But I think in the end the best way to get great sound is to attempt to optimize every issue that you have control over and not focus on only one thing. This is the approach I have used and made fantastic sounding speakers.

Here are a couple of ideas that I have had in the past but have not done myself due to time constraints. For the bass cabinet you can make your own stiff and dead composite panel by using two equal thickness plywood or MDF panels with lots of round holes sunk all over the panel at regular close intervals. The holes are not through holes. Just there to hold short dowels to space the panels apart that act as structural members. Then cut up a whole lot of 1-2 inch dowels (what ever size you are using), to fit into the hole recesses and space the panels 1/2-3/4 inch apart. Then you will have to have a edge piece sandwiched between the panels around the parimeter to seal the inside so the empty space can be filled with dry sand.

This panel should have very high stiffness and very high dampening. And be cheap to manufacture. Only drawback is complexity and time required to make the panels.

rick57 said:

Thanks Rodd,

an off topic diversion from one of your points, but apart from damping, if other things were equal (which they are alas not) -

“So for every 100W of power in, a typical speaker is only converting 1W of that energy to sound/music. . .
However, that means the other 99 W that is not converted to sound has to be converted to heat or go to excite the various components of the speaker.”

a reason why HE drivers could sound better - if they are 2% instead of 1% efficient: 50% less inducement of box coloration?

Cheers

Click to expand...

Yes, but only in small part. The main message in my ramblings (which I guess I didn't come out and say) is that there are many more Watts that must be turned to heat by damping in the enclosure or it will turn to noise, ...not an easy task. It's not to say that stiffening and bracing and adding mass or changing materials, et al won't help. The energy in/energy out thing is just a way to say that damping must always be considered in any enclosure design effort.

For instance, if you stiffen the panel and by this you raise the resonant frequency of the panel, you must do so because the damping you're applying is more effective at the new resonant frequency of the panel not because you heard its a "good" thing to do.

If you always consider the 3 basic ways to deal with a resonance and understand the need for damping and how damping works, it becomes much easier to evalueate various enclosure construction suggestions. Knowing, of course, the finished product will always be a tuned set of compromises.

Hi Rick,

I wish I had a link to that article. A few years ago I built a pair of quazi Watt/Puppy clones. I scoured the internet for everything I could find from Wilson's site or anywhere else on the internet. They are now very closed off in what they talk about in regards to thier technology. But I knew someone that knew one of Wilson's old speaker designers and gleaned a few things from here and there. One thing I do remember though was that when they tested a single sheet of the phonolic composite aerospace panels they were orders of magnitude better than the best (production reasonable) CLD panel that they could devise. The aerospace panels just don't hardly store any energy which I suppose is one reason the Wilsons are so dynamic.

I will read up on your CLD thread since it's been a while since I have been digging into the cabinet design literature. One thing is that I'm not sure if Wilson even considered the massively overbuilt multi-layer CLD method since it drives the productions costs up very high. The methods that Wilson uses with the panels they employ allows them to assemble cabinets rather efficiently. So even though the material is expensive as were the initial tooling costs. Now they can built a lot of speakers with only a few personel.

But as I recall, the massively multi-layer CLD method can damp things pretty well but it is weak in the speed department. It damps resonances well but it is not as fast at settling down. A critical factor in panel dampening.

I'm not sure what Thiel is now using for their baffles but a few years ago they were using a very dense concrete baffle four to six inches thick in the upper line speakers. They switched to thick MDF in the lower price speakers because they saved 50- 80 pounds of shipping weight and the sound is nearly as good with MDF.

But you are right in that for the DIY'er CLD is a very realistic approach to take. These exotic materials that Wilson uses are a bit out of reach for the DIY'er. However, the method used by Rockport can be done by the DIY'er on a smaller scale with say the upper cabinet mid/tweeter.

I am in the process of designing a new 3.5 way medium size floor stander from some drivers that I have and I have half a notion to try and assault the state of the art with inexpensive DIY cabinet construction techniques since this is the only area that I can really push the envelope in.

I already have the drivers and this cannot be a high budget project but I have the most to gain from trying out some new techniques with the cabinet construction. At this point if I go beyond the overbuilt MDF/plywood route I am going to built the upper small cabinet out of epoxy resin/lead shot and fiberglass. The bass cabinet will use my plywood/dowel/sand construction process. I will post a thread on the process if I decide to go this route.

Hi Rick,

The Rockport cabinets are supposed to be even more inert than the Wilsons but I have never auditioned them. Stereophile accelerometer testing showed the most inert cabinets ever tested as I recall.

Basically they make an inner and outer skin out of fiberglass/epoxy or carbon fiber/epoxy which represent the inner and outer surfaces of the cabinet. This is maybe anywhere from 3mm to 10 mm thick. Then a pourable composite is poured in between the mold cavity to make an extremely stiff, dense and internally lossy panel. Something like pourable Corian. The panel is anywere from 1 to 2 inches thick I believe and can be varied in thickness where needed. They are massively heavy.

THe internal material is proprietary but is probably something like mineral filled epoxy.

I have thought about different materials and have concluded that several would work. Some perhaps better than others.

Sand filled epoxy, crushed rock filled epoxy, lead shot filled epoxy, lead shot mixed with high density polyurethane expanding foam.

Last night I was thinking about this and wondered how strong and stiff the panel would be but still lossy and highly damped in the center if you used lead shot and high density expanding foam.

This method is difficult to make large enclosures for the DIY but it is possible. However small enclosures while requiring some ingenuity would not be so difficult. And more time consuming and expensive than MDF but not out of the ballpark. In essense you can make an organically shaped enclosure which performs as well or better than the aerospace panels which Wilson uses. OF course these speaekrs are very expensive because of the time it takes to built them.

http://www.soundstage2.com/tours/rockport_200702/

Here is a link to a brief factory tour.

Here is another link to an older Stereophile article:

http://www.stereophile.com/floorloudspeakers/904rockport/

Here is a link to the page where Stereophile measured the cabinet resonances. On the Rockport Antares. The lack of energy in the cabinet is unpresidented.

http://www.stereophile.com/loudspeakerreviews/644/index6.html

And to think that we can build a small cabinet nearly as good as this one with a little effort. The only thing we probably can't get is the special epoxy formulation that was custom designed for Rockport. But maybe we can get it if we try hard. After all Rockport is not a chemical and polymer manufacturer.

the following will probably ramble a bit.... and i'm sure there will be some that disagree ...

To my mind -- short of going nuts -- the best midrange box is going to be a low pressure design (aperiodic is ideal here -- an aperiodic TL amounst the best (think Nautilus) along with open baffle (if we call an OB a dipole, the well designed aperiodic TL is the monopole analog), and you want to get the panel resonances up as high as you can. This means very stiff, low mass panels that are not very big at any point. the amount of energy to excite panel resonances decrease as you go up in frequency -- heavy MDF panels take you in the wrong direction and have significant energy storage causing time smear.

Having all non-parallel walls is very helpful as well. Sometimes it just needs to fit into a rectangular space -- a heavily tapered triangulated TL fits those requirements. Sufficient damping -- light at the fat end becoming increasingly denser towards the terminus will make it aperiodic and very non-resonant.

The energy to excite the walls comes from 2 places -- the acoustic energy from the back of the driver, and mechanical energy directly coupled to the box. The 1st is largely ameriolated by the low pressure box -- the energy is directed to the outside world or fairly effectively turned to heat by the damping material. The 2nd is the one we most need to worry about. Coupling as much of the driver to as much of the box as possible helps but having the energy required to excite panel resonances well above the frequency of the majority of the mechanical energy generated, coupled with panels that have high Q and release any energy quickly deal with most of the letter

As an example the Fonken satisfies many of the criterion listed (1st entry on this page http://www.planet10-hifi.com/boxes-fostex.html) with thin, stiff walls and no panels exceeded ~4" wide) and does an admiral job of being a midrange enclosure (one with 65-70 hz up to more than 15 kHz range to boot)

Bass is actually a bit easier to deal with... it is fairly easy to deal with acoustic energy ,,, using a low pressure box helps here too ... you just need to make sure that panel resonances are above the passband, and yu need to make sure that the box won't balloon.

The mechanical energy of a bass speaker is quite a bit higher thou and can be a problem -- i'm sure that almost everyone has heard stories of or seen subwoofers "walk". The same techniques used in the mid enclosure work to a large extent, but ... Fortunately a simple trick allows us to overcome a huge amount of this energy in one simple swoop. Since bass speakers are effectively omni-directional, with no drawbacks except having to pay for 2 woofers instead of 1, we can actively cancel the majority of the mechanical energy by mounting a pair of drivers on opposite panels, rigidly couple them and create a push-push pair. This is so effective that i have to have a really good reason to only use a single woofer.

dave

PS: none of this touches on the optimum external shape.... (more important for the mid-tweet box)