Hi all and preemptively thanks for all your help!
So I'm dipping my toe into the world of speaker design. I've ordered some books and got a few papers ready to read but in the background I was just hoping to ask some questions here to supplement.
I personally believe most things are easy when broken down into small chunks, I'm an engineer in the aerospace industry and while nobody knows everything about a whole engine, its too complicated, people learn to design small parts of a large assembly and you dont really need to know that much at all in the grand scheme of things.
So the first 'chunk' I wanted to delve into was the enclosure design, as being a mechanical engineer this is the bit I feel most comfortable with.
So my question is really, what is the list of requirements for a cabinet?
So my understanding is as follows:
I haven't done any dynamics since university so I'm a bit rusty so sorry there's probably loads of mistakes in there.
Would people be able to list what they believe are the main requirements for an enclosure so I start evaluating the best compromises to make.
So I'm dipping my toe into the world of speaker design. I've ordered some books and got a few papers ready to read but in the background I was just hoping to ask some questions here to supplement.
I personally believe most things are easy when broken down into small chunks, I'm an engineer in the aerospace industry and while nobody knows everything about a whole engine, its too complicated, people learn to design small parts of a large assembly and you dont really need to know that much at all in the grand scheme of things.
So the first 'chunk' I wanted to delve into was the enclosure design, as being a mechanical engineer this is the bit I feel most comfortable with.
So my question is really, what is the list of requirements for a cabinet?
So my understanding is as follows:
- High stiffness - which would be a material with a high Young's modulus and a structure with high 2nd moment of area (or something like that)
- High damping - the speed in which vibrations in the structure decay in this instance.
- Low internal reflectivity - which I think is to do with the ratio of acoustic impedances
- The shape of enclosure itself - this seems to me where it gets difficult, the others are all relatively straightforward. I understand you dont want straight edges in order to remove internal reflections. But say if the nautilus shape is idea, why doesn't everyone use it?
- Internal pressure - it would seem to me the ideal internal pressure (x area of the driver) would match the force required to overcome the resistive forces of the driver itself minus the force the driver can create itself. Or more realistically isobaric.
I haven't done any dynamics since university so I'm a bit rusty so sorry there's probably loads of mistakes in there.
Would people be able to list what they believe are the main requirements for an enclosure so I start evaluating the best compromises to make.
I have some experience gained over the last 50 years.
1/ i believe stiffness to be very important. If one can raise the (potential) resonances high enuff, with music they will not be excited. Playing into this, something others often disagree with me, is to get the Q of any resonances as high as possible. This means that a lot of energy in a very small bandwidth would have to be injected into the panel to get it to move, something that you will very rarely see at higher frequencies in music. If you can do this any resonance will never be excited and it will be as if the resonance does not exist.
MDF is often used but its only asset is that it is cheap (to buy, to use, to finish). The most commonly used material i recommend is quality plywood, but a step up is (stranded/fossilized) bamboo plywood, but more exotic materials (that i have not actually used) like carbon fibre, other composites, metals, stone/concrete (i suspect one needs to be careful there) should be even better at a much higher need for specialist tools and skills and higher cost.
2/ while panel damping should decrease the amount of time that an excited resonance is active, it is, i feel counterproductive given the goal above of moving any resonances high enuff that they will neve rbe excited. ie adding mass without adding stiffness.
The one place where this can be utilized is in a constrined layer enclosure. Doing it properly is difficult and gretaer cost. My experience is that this is that if you can achieve the goals in point 1, it is over the top and guilding the lily.
3/ mostly an issue at higher frequencies and (mostly) the purpose of internal air space damping material
4/ an appropriate shape — like a cyclinder — can greatly increase stiffness which aids point 1. And it can often be quite inexpensive, PVC or cardboard tubes, bucket subs as examples. Usually more useful as lower frequencies where the wavelengths are longer than the maximum dimension of the enclosure. One has to be careful, the shape needs to be fairly radical to be able to significantly reduce internal standing waves. This is where we see things like the Nautilus, the N-series mid enclosure which is a tear-drop shape outside but different inside, and the B&W DM302 back panel (i can dig up pictures if needed — notr all B&W speakers). In the big Natilus the enclosure is also a half-wave TL with the goal of absorbing the entire back radiation from the loudspeaker ao tht it does not come back thru the cone causing time smear. Shorter TLs for midrange or tweeters also have the same (or greater) benefits. For midranges (or midTweeters, as in the half-wave example below, from an as yet unreleased WAW Scott & i are working on) they can be quite simple. This can also be quarter-wave (ie open at the end).
Also, if one can keep all cabinet dimensions smaller than the bandwidth of the driver used in that space, it effectively removes any standing wavesfrom occurring.
5/ A potential issue at the lower frequencies, much less an issue with boxes that have a hole in them (i call them low-pressure boxes), and called ballooning when talking about subs. Stiff material and adequate bracing are the easiest ways to deal with this. As far as cabinet panel resonances a very small amount of input into exciting any resonances.
Bracing: i have 2 broad rules i apply here
1/ after Tappen: any brace should create subpanels with higher aspect ratios than the panel they are dividing.
2/ braces should not be placed at integer multples of the panel being braced (ie don’t put one in the exact middle of the box)
3/ there should be sufficient bracing that any subpanel does not have too large a (minimum) dimension
4/ i like to rigidly brace the back of the driver to more than one panel (typically called a holey brace and seen in the illustration above). Given that the baffle, with all those big holes in it, is typically the weakest and this shares the energy more directly thru more panels reducing the possibility of exciting panel resonances. One could take this further by coupling the front of the driver to another panel as well (for instance by using ready-rob instead of bolts or screws to afix the driver to the baffle).
The above also illustrates a very useful principal (primarily with woofers) so mount them push-push and rigidly coupled to achieve a large degree of active reaction force cancellation greatly reduceing the energy getting into the enclosure walls. 1st described to me a long time ago, the 1st commercial example i saw was in the KEF 104 with bandpass woofers, and later in an elegant NAIM loudspeaker and others since. I use it everytime it is practical to do so.
My father was a Mechanical Engineer so i learned a lot from him theu osmosis, and i stand on the backs of those brfore plus a whack of real-world experience.
I look forward to your views from a mechanical engineering POV (it is also of sideways interst that the deigner of most of my favorite drivers is a mechanical engineer who helped with much of the Saudi oil infrastructure)
dave
Attachments
A brief contribution from me! Looks like Dave has covered all the bases while I was thinking about the topic, but I haven't read his contribution yet! 🙂
Certainly you had the basics covered. An enclosure should be stiff, well damped and have a very high resonant frequency.
Of course, factors such as cost, machinability, ease of surface finishing and weight also come into the equation.
Certainly you had the basics covered. An enclosure should be stiff, well damped and have a very high resonant frequency.
Of course, factors such as cost, machinability, ease of surface finishing and weight also come into the equation.
I'm interested in the idea of 3d printing an enclosure, if I can get someone to allow me to use the 3d printer at work that would be ideal but otherwise I might try see if I can use one at the local university. The interesting thing about 3d printing is you can produce almost any shape with huge complexities with very little added cost. So what you could do is create a lattice structure to increase stiffness while keep mass down (resonant frequency is proportional to stiffness/mass - you want high stiffness, low mass).
The other interesting thing about being able to create any shape is I could recreate this tapered cone of the B&W nautilus at low cost. While the 3d printer at work could actually create something as large as the bass section I dont think I could get away with using it for such a long time or afford the materials produced in that manner so unfortunately I'd have to go with something more similar to the diamond series with a normal cabinet for the bass section and a tapered tube for the mid and high frequency units. but its certainly possible!
I've seem a similar setup to what you have above but in push pull to create an isobaric chamber. Not seen them rigidly coupled like that, do you know what the name of the design is so I can read up on it? Also do you know the name of the design which creates the isobaric chamber actually?
The other interesting thing about being able to create any shape is I could recreate this tapered cone of the B&W nautilus at low cost. While the 3d printer at work could actually create something as large as the bass section I dont think I could get away with using it for such a long time or afford the materials produced in that manner so unfortunately I'd have to go with something more similar to the diamond series with a normal cabinet for the bass section and a tapered tube for the mid and high frequency units. but its certainly possible!
I've seem a similar setup to what you have above but in push pull to create an isobaric chamber. Not seen them rigidly coupled like that, do you know what the name of the design is so I can read up on it? Also do you know the name of the design which creates the isobaric chamber actually?
3D printed boxes are something i would really like to play with.Saving up for a Form3L.
Someone did print one of my designs, but pretty much the same as one you would build out of wood..
dave
Someone did print one of my designs, but pretty much the same as one you would build out of wood..
dave
An isobaric chamber is formed when two speakers share a single sealed volume of air in between them, and the cones move in unison. Basket orientation makes no difference as long as the cones are in phase. Since the cones move together, the air in between isn't compressed/rarefied. Therefore it's iso-baric. Create that chamber however you see fit, but typically smaller is seen as better.
Be aware that the two drivers literally work as one. You deliver power to two drivers, and only get output from one.
The smaller the coupling chamber is the greater the potential HF extension. The prime reason to us eisobarik is to halve the box volume for a given response (+ the coupling chamber).
dave
- acoustical loading for drivers
- sufficient mass to minimise movement of cabinet by drivers reaction force
- appropriate stiffness and/or damping to make cabinet radiation inaudible
- shape to provide smooth and controlled directivity/diffraction for radiated sound
- visuals
- other things I have missed
Not really. The desired mass, damping and stiffness properties of a driver's enclosure tends to follow from the overall configuration (i.e. what is isolated from what) and whether resonances overlap the driver's passband.
Again, in some cases this is the dominant required material property whereas in others it is unimportant.
Ratio of which impedances? What is required at the low frequency end of the driver's passband which may involve wanted controlled resonance/s to extend the low frequency response is usually different to that at the high frequency end.
What internal reflections? At low frequencies the air volume/s often behave as a lumped mass modifying the cone motion without causing problems. At high frequencies stuffing can absorb sufficient sound for internal reflections not to be a problem. There is no significant engineering reasons for long lines but if done well they don't cause problems other than increased size and cost.
What requirement is this addressing?
If you wish to combine a line with additive manufacturing then a well engineered cabinet shape might look something like this. The overall speaker has made one or two questionable design decisions but it suggests how to improve cabinet design when less constrained by flat sheets.
My (very limited) understanding is that this is only half the ideal solution - it is great to have one layer as you described to provide overall stiffness but ideally it should be coupled to another layer with a resonance at a very different frequency, or no resonance at all or sandwich a damping layer in between two stiff layers, i.e. constrained layer damping.
I think you're definitely right with it being only half the solution. The stiffest materials you can use are ceramics, yet I dont see many people making speakers out of them. I had wondered whether you could layer ceramics on to a damping layer like you'd suggested above to create a damped stiff structure, I suppose it would also create a very fragile speaker.
Could someone tell me if there is a quick way to quote again?
Andy19191:
You say stiffness and damping always appropriate, can you give me an example of where it isn't?
The ratio of the acoustic impedances of the material the sound sound is travelling within (air usually) to the material which creates the boundary in which it interacts with determines how much of the sound is reflected.
You seem to argue internal reflections aren't an issue but everything else I've read argues to the contrary and my understanding of the science would also argue to the contrary. So if you can model vibrations through air at low frequencies as a lumped mass, why does that then make internal reflections not an issue?
The requirement internal pressure is addressing is the the spring damper like system which is created when the speaker moves back and forth against a sealed box of air. This would make it more difficult for the speaker to accelerate or deccelerate to it's desired position. Or at least thats my understanding of it.
Tsmith1315/planet10:
Yes sorry I read over my comment last night and it didnt make sense. I had read that the isobaric chamber doubles the low frequency response for a given size but I was just hoping someone would know the name of the theory so I could research it. It definitely has a name as I've read it before.
Planet10:
Yes you should definitely experiment more with additive manufacturing/3d printing as the possibilities are much more creative than I think many people realise. For instance aeroplanes now use a lattice structure on their seats to create a stronger structure without the added weight. You can create all sorts of complicated designs with little added cost. If you live in the UK Nottingham university and Loughborough university have some advanced additive manufacturing departments. Im actually going to contact my old lecturer and nottingham who's in charge of the research there, he's a musician so may actually really appreciate this project and help me.
I don't know what the situation in Nottingham is, but in my neighbourhood it is relatively easy to collect discarded Ikea furniture from the street when it's moving time. So when I want to try out a speaker cabinet idea I collect garbage off the sidewalks and use screws to put panels together once cut. This doesn't give me perfect cabinets but there is also no hesitation to cut a hole somewhere or board it up again or change a panel when it's too messed up until I get something useful. Often it has been interesting/educational to just try something that in theory is totally wrong to connect the numbers from the simulation and measurments to a objective experience of "sound".
Once that is done moving on the a proper build with better materials makes sense. Going through the time and expense of 3d printing a prototype enclosure that might have issues might turn out frustrating.
Once that is done moving on the a proper build with better materials makes sense. Going through the time and expense of 3d printing a prototype enclosure that might have issues might turn out frustrating.
Last edited:
Yes when university is over here people go to the university areas with trolleys and fill them with stuff because so much stuff is thrown out. Shocked me how wasteful my fellow students were. Too many rich kids with too much stuff. Good idea actually.
I get your point but unfortunately you can't trial a tapered tube idea with old furniture. I was planning on modelling it first though. We had a post grad from imperial college london come and write some code for us as part of his PhD to model sound. He was using FE to model sounds interaction with things smaller than the wave length more accurately than the software we currently use which is based on the Huygen's principle. He said he'd give me the source code and adapt it to speakers for me but I wonder if it would be any better than the current free software around.
I get your point but unfortunately you can't trial a tapered tube idea with old furniture. I was planning on modelling it first though. We had a post grad from imperial college london come and write some code for us as part of his PhD to model sound. He was using FE to model sounds interaction with things smaller than the wave length more accurately than the software we currently use which is based on the Huygen's principle. He said he'd give me the source code and adapt it to speakers for me but I wonder if it would be any better than the current free software around.
This was a trend some decades back. Floor tiles were recommended.
It makes sense, of course. The thing is that damping material is fairly good at managing response variations from internal standing waves. Of the handful of different cabinet noise methods this one is fairly easy to fix.
You speak of impedance, it is worth seeing the modal nature of these standing waves, so that basic discussion of the fundamental box function and rolloff has its own considerations.
Might have read completely over it, but I see nobody mentioning the following. The outside (shape) of an enclosure is at least as important as is the inside shape and construction. That is, if the desired frequency range corresponds with wavelengths that are relatively small in comparison with the size of the enclosure. No news though.
A stiff small woofer cabinet can be constructed to have the lowest frequency resonance well above the passband of the woofer. If it is acoustically isolated from the midrange cabinet it requires no damping because no resonances are driven. Competent subwoofer cabinets will be stiff and fairly heavy but not damped.
A midrange cabinet however will inevitably have resonances in the passband. This makes stiffness relatively unimportant and damping important in pushing down the magnitude of the radiated cabinet resonances below the sound radiated by the driver and in making the cabinet swiftly stop resonating when the driver stops radiating sound. There are pros and cons in having a strongly damped floppy cabinet vs a strongly damped medium stiffness cabinet but a stiff cabinet will offer poor performance.
Tweeters and waveguides have their own requirements when it comes to keeping radiated sound at inaudible levels.
I am familiar with the concept. My question was more about which interfaces were causing issues that required addressing.
Please discuss or reference the science because it is counter to my understanding of what is or is not relatively important. This is not intended as a rebuttal since I am happy to be persuaded and learn more.
I am afraid that many that are enthusiastic about speaker DIY are not engineers and do not view things as an engineer would. They are happy to suggest all sorts of strange and wonderful things based on fairly casual supporting evidence.
I suspect in the near future a few DIY folk will start showing 3D simulations of the sound radiated from vibrating cabinets and provide a solid quantitative basis for reasoning. I started to do this a while back but still need to finish implementing a method to handle the nonlinearity with frequency introduced into modal analysis by viscoelastic materials. Unfortunately given a typical PC a full 3d model of the damped vibration of drivers and cabinet requires modal analysis in order to efficiently perform a design study. Simply performing thousands of single frequency calculations would require too much computer time.
Yes it changes the motion of the cone but far from making things more difficult it can extend the low frequency response, raise the output, lower distortion by limiting deflection, etc... depending on the frequency and type of rear loading. So if you opt to drop all these quantitatively significant benefits how large would the benefit be from a theoretically perfect rear acoustical impedance presumably to match that of the front?
A practical problem with simulating the acoustics of porous media is being able to specify the required empirical constants (which vary between models) for your particular foam, wool, cloth, fluff or whatever. One approach is to build a line, stuff it with various materials and densities and measure. Perhaps another reason to consider breaking up some discarded furniture?
I don't think your question has been answered.
Click on the 'Multi-Quote' button which is immediately to the right of the 'Quote' button.
Do this for each of the posts from which you want to quote, except for the final one when you should click on the quote button.
The best thing you can do for accuracy is to model the correct thing to begin with, isolate the issues effectively and to interpret the results properly. Easy to talk about, harder to do.
This will give you fresh quote tags.
Attachments
Clay plant pots and ceramic chimney linings are often used as diy enclosures. Can’t recall the name, but an east asian company builds and sells a speaker with ceramic enclosures.
3D printing is definitly on my mind.
dave