Thoughts on speaker enclosure design

This is the "soon-to-be-mech-engineer" side of me coming out. Lying in a half-asleep state this morning, these thoughts came to me.

According to all I've read about speaker design, the ideal is to have a non-resonating enclosure, which explains the popularity of MDF as an enclosure material. The fact that it's cheap probably doesn't hurt.

Now, when it comes to cross-bracing inside the enclosure, I have yet to see any rhyme or reason other than "try to make it non-resonant." This answer has never really satisfied me, so I began to wonder what rhyme or reason might apply to this problem.

You know how similar vibrations that are out of phase with each other tend to cancel? Perhaps the box could be designed in such a way where the cross-bracing mechanically linked areas of the enclosure that are out of phase with each other, thus providing a nice, non-resonant enclosure.

Calculating these point would be a trick, and is something I don't know how to do (yet). I also think that how you do the crossbracing would depend on the crossover you choose, since the phase of the signals coming in can change depending on the crossover design, or at least that is what I've divined from all I've read. I could be wrong.

It also seems to me that the placement of the drivers on the front would also have a bearing on where you would want to place such crossbracing. This kind of design would definitely need some thorough planning out.

I also seriously doubt I'm the first to think of this.

So, is my thinking any good?

Thanks, all!

Yes alot of what you posted has been addressed.

Look for the Green covered AES book on loudspeakers. Writeups on the B&W speaker boxes. The loudspeaker design books by Vance Dickason

check the

http://www.partsexpress.com/ site in the book section.

Thanks, Gus. Time will tell how far I decide to go with speaker design. These were just random thoughts I had in a half-asleep stupor, so to speak. At least I know how to think in the right directions, though.

I'll definitely try to look up some of those books...

> the ideal is to have a non-resonating enclosure

Actually, to not-radiate the driver's backwave. In general, a wood wall trumps anything off a paper cone, but at resonance the wood wall "vanishes" and radiates freely. Lessened by dampiong, but increased by the large area of the walls compared to the cone.

> You know how similar vibrations that are out of phase with each other tend to cancel? Perhaps the box could be designed in such a way where the cross-bracing mechanically linked areas of the enclosure that are out of phase with each other, thus providing a nice, non-resonant enclosure.

You are being vague. For best result with minimum material, you connect the center of two opposing faces together. If done perfectly it raises the resonant frequency, I think 1.4 times higher.

But it still resonantes. And at a higher more audible frequency, though perhaps up in the range where 1"-2" internal fuzz reduces sound pressure on the back of the board. I don't think it gives any great reduction in resonant Q. You have not fixed the problem. only moved it.

And in fact, it never works. The forces are large and the displacement is small. For effective coupling of two sides, the brace must be VERY stiff. 1x2 pine is stretchy enough to be pointless. Think 2" steel bar or 6" log. These over-kills tend to be less attractive than other options.

> crossbracing would depend on the crossover you choose,

Rarely. Panel resonances in wood/wood-dust boards of typical dimensions are 150Hz-300Hz. Typical speaker crossovers are 800Hz or higher.

> I have yet to see any rhyme or reason other than "try to make it non-resonant." This answer has never really satisfied me, so I began to wonder what rhyme or reason might apply to this problem.

Take a 2-year specialty in vibration analysis.

Or do what has worked for 500 years: study violin making. Except do it backward. A fiddle maker starts with resonant (low-Q) wood (selected spruce) then shapes and thins it to give high Q at specific frequencies. The goal is to set the bottle-resonance, front-plate resonance, and back-plate resonance so they boost-up the bottom octave of the fiddle (where the body is traditionally too small for good radiation) evenly across all notes of the first few octaves.

All of this tuning is done by knuckles. Carla Hutchens et al have done a LOT of study and some theory, but working fiddlemakers go by tap-tones, pitch and ring-time.

You want to burn the spruce and find dead-sounding wood. Away from resonance, mass-law rules, so you want heavy and dead. Dust-board is slightly better than plywood, won't split like boards, is consistent, and very cheap. Cellotex is deader, and has a little surface absorption, but is so low-mass that it passes a lot of sound even at nonresonant frequencies. And it won't survive abuse.

With an assembled cabinet, tap each panel, near center. Note the pitch and ring-time. A typical speaker cabinet has 4 main panel resonances: left/right (should be same); top/bottom, rear, and front (front is different from rear due to holes and driver mass). These should be different pitches, and should decay rapidly.

Side-side bracing seems to do very little, probably because we can not get adequate stiffness in the brace.

Heavy screw-blocks at panel joints can deaden a panel a lot. Free-edge resonance is very different from hinge-edge or clamp-edge resonance. (Fiddle resonances shift a lot when the plates are assembled to the rim; fiddle plate edges are thin and act roughly as hinge-edge.) I like 1.5"x1.5" pine blocks all the way around all panels, using Liquid Nails or other tough non-hard glue. This approaches "clamped" condition for 3/4" panels a couple foot across, but is less effective on 4-foot panels and consumes a lot of internal volume in small boxes.

If you are still concerned, increase panel thickness. This increases mass and damping, giving a lower-F lower-Q resonance.

If you are realistic, quite large boxes of 1/2" or 3/8" chipboard work very well. Panel resonance is not a big deal. I've had 3"x5"x10" boxes of 3/16" cardboard and 1/16" plastic that did not shout out their sides.

Okay, so basically, what I'm reading is "use thick MDF and some internal baffling, and you'll be fine." I figured as much, but I like to keep the idea muscles flexed regularly.

I'd still like to seek out some of those books that Gus mentioned.

Take a 2-year specialty in vibration analysis.

Unlikely, since my planned focus is alternative energy... :wink:

[quote author="Consul"]Okay, so basically, what I'm reading is "use thick MDF and some internal baffling, and you'll be fine." I figured as much, but I like to keep the idea muscles flexed regularly[/quote]

That's one way to do it. Another is to use 1/2" plywood made from multiple very-thin birch layers (called "Baltic Birch plywood" or "Finnply" in the US) for maximum damping from the glue layers, then damp the panels with "bituminous felt", basically felt impregnated with asphalt. It's sold as car-noise-reduction panels, at a horrendous cost, but it's available cheaply as "sidewalk separator" or "expansion joint" from large hardware stores. For a rectangular panel, run a brace lengthwise down the center, except not down the center but a bit off-center, so the two panel segments are unequal. Glue the sidewalk separator to the walls using Liquid Nails. Let dry several days before installing drivers (apparently the fumes from the curing adhesive are damaging). Boy, will you have dead cabinet walls.

This is the approach used by British manufacturers of the LS3/5a (that was small enough not to need the mid-panel braces), LS4/6 and similar BBC-designed monitor speakers, renowned for their relative low-coloration. Spendor and Harbeth used similar designs for "civilian" speaker cabinets. I've done it, with good-sounding results.

Peace,
Paul

Wow, thanks a lot Paul!

You know, I've read up on the design of those BBC monitors, and I've considered trying to make a set for myself. As I recall, there was special consideration for the woofer, and I was unsure if I could find a suitable current-production driver at the time. I'll need to revisit the paper.

That BBC R&D website is one of the best resources out there for the DIY audio person, in my opinion. That's where I got the idea for the portable room treatments, too.

And a big thank you to PRR, too! (Sorry!)

All the stuff in this thread and more are in books and things like the B&W ads. Look for reviews of KEF stuff. Krell used Aluminum for a speaker box. People have used concrete. Old warfdales used sand fill walls IIRC. I like High Q materal that can be damped well.

One of the best ways to start to understand this stuff is the wine glass and plastic cup. If you tap a wine glass it rings and if you touch the rim you get a very dead sound, If you tap a plastic cup and then try to damp it it sounds about the same a lower Q sound.
This applys to speaker cones there are ripples down the cone and depending on the surround and the cone materal energy is stored and released if different ways. The place the cone, suround and basket meet must be a mess to model.

I tend to like the plywood type Paul posted about. I also like Liquid nails. Plastic floor tile can be useful sometimes

[quote author="Consul"]Wow, thanks a lot Paul!

You know, I've read up on the design of those BBC monitors, and I've considered trying to make a set for myself. As I recall, there was special consideration for the woofer, and I was unsure if I could find a suitable current-production driver at the time. I'll need to revisit the paper.

That BBC R&D website is one of the best resources out there for the DIY audio person, in my opinion. That's where I got the idea for the portable room treatments, too.[/quote]

Unfortunately, it's not really possible to duplicate the BBC designs. First, the drivers are no longer made. And second, even if you can find some old drivers, the ones used in the BBC monitors (at least the LS3/5a's) are specially selected for certain characteristics, most notably the magnitude of a particular resonance in the midrange, and to do the selection you really need some pretty fancy test gear, not to mention a bunch of drivers. The ones they used are nowhere near the center of the bell-curve; the ones that didn't meet the BBC's specs were, I'm almost certain, the ones that got sold by KEF as the "Constructor" series. Probably the same applies to the drivers from the LS4/6.

On the other hand, you can take a set of really good drivers and design a system from scratch around them, using ideas from the BBC designs (like the cabinet walls), and get excellent results.

Peace,
Paul

My feeling is that baltic birch is not as stiff and dead as the same thickness MDF. Just like PRR said, if you take a piece of baltic birch and a piece of MDF of the same size just try to tap test them with your knuckles. As a luthier I do this everyday....except with tonewoods, not MDF:wink:
If anyone would like to try this tap test, here's how you do it:
Cut two equal size pieces -1 MDF & 1 Baltic birch say 3" wide by 24" long
measure 6" down the side....this is where you hold the piece lightly by the edges between thumb and forefinger
Hold the piece up high so that the flat area is facing your ear and tap with one knuckle near the bottom (on the flat part)
My experience is that you can produce a much stronger joint with baltic birch however, therefore if the enclosure will be moved around a lot (ie. touring) a baltic birch cabinet is more durable. But if I were to make home or studio loudspeaker enclosures that don't get moved around I'd use MDF.

Cheers,
Freddy G :sam:

[quote author="pstamler"]On the other hand, you can take a set of really good drivers and design a system from scratch around them, using ideas from the BBC designs (like the cabinet walls), and get excellent results.[/quote]

Thanks again for the reply.

This was actually my line of thinking. Like I said earlier, I don't know how far I'll take any ideas from this thread to make a set of monitors, since speaker design was never really my field of interest. I would much rather DIY a proven design.

However, I am intrigued by many of the ideas from the Beeb, and I think they might be worthwhile to pursue.

Freddy: Based upon what you just said, I think I can understand why the BBC chose the construction method they did. Most of their work was field work, involving lots of equipment moving around. There was definitely need for high-quality as well as very durable gear, including speakers.

In another field, the BBC has also done plenty in the areas of acoustic treatments. They have a membrane-style bass trap design that is portable, inexpensive, easy to make, and effective that I plan to build for my own portable setup.

Here's the BBC R&D web site: http://www.bbc.co.uk/rd/index.shtml

Most of these papers, including for the monitors, are in the R&D Reports section: http://www.bbc.co.uk/rd/pubs/reports/index.shtml

Thank you to all who have posted thus far.

One correction. I was wrong. Increasing panel thickness increases resonant frequency. (Assuming uniform thickness.... in fiddle-making they taper the plates and can shift resonance downward.)

> Okay, so basically, what I'm reading is "use thick MDF and some internal baffling, and you'll be fine."

You'll be fine, but that's not what I said.

Go over to Freddy's shop. Watch and Listen to what he does. Anything that is "good" for Freddy's work is probably "bad" for a speaker. If it has "a nice tone", you don't want it. When Freddy shaves the wood for an even nicer tone, you want to do the opposite.

Plain home-building plywood is bad for other reasons. I admit I have used a lot of it, but most generic plywood is made from defective veneer with a lot of splits. The resulting voids affect strength very little, affect stiffness somewhat, but once in a while there are loose strands inside the void. These will buzz horribly. You can cut panels and knock them around, listen for buzz, and discard the buzzy ones. But buzz may not be audible until after a few thousand power-chords. You can dribble Elmo's Glue into exposed voids and maybe firm-up stray strands. Overall it is better to use good solid plywood than to fool with sheathing-stuff.

Home Depot sold me some 7-ply oak faced plywood that seemed very solid and void-free. I built a 2x2x4' subwoofer. On tones it buzzes, but I think that is sloppy joints and cheap drivers (and every loose thing in the room). This was for knock-about use so I did not expect particleboard to last.

> then damp the panels with "bituminous felt", basically felt impregnated with asphalt. It's sold as car-noise-reduction panels, at a horrendous cost, but it's available cheaply as "sidewalk separator" or "expansion joint" from large hardware stores.

Good panel-damping tar-felt is loaded with clay and VERY heavy. You need a "stationary" reference for the damping to work against. At these mechanical impedances there is no solid reference, so they use mass as a best approximation (and also to lower the panel res.freq.). Expansion joint is not loaded with clay and in the same thickness won't work as well. However it is SO much cheaper. Price-per-pound, much cheaper than boutique damping panels. If volume is not critical, buy a big roll of expansion joint and lay it in thick. You probably want the damping to be at least as heavy as the raw cabinet: if the cabinet is 40 pounds, pick up 40 pounds of expansion-joint. (In portable work, you probably would not want that much dead mass to move every night.)

PRR: "You can cut panels and knock them around, listen for buzz, and discard the buzzy ones. "

Finally a worthwhile application for NXT inertial drivers!

Seriously, I'm surprised (and maybe just ignorant of work out there) that some high end company hasn't resorted to active inertial vibration cancellation of loudspeaker cabinets.

I remember when Newport, a manufacturer of fancy optical test benches and such, at that time a lot snootier in their marketing than they have since become, introduced a product that established an inertial reference frame in all three axes, whether the vibration was external or induced by motion of equipment on the platform itself. This was accomplished with actuators, sensors, and electronics. Of course there was some low-frequency cutoff.

But they named it EVIS, for electronically vibration isolated system iirc. Clearly they chickened out as it could easily have been called ELVIS

Yikes, there's a lot to absorb here! I thought I was just posting an off-hand idea to see if anyone's done anything like it before.

Given all that I've read, I'm thinking making my own monitors for mixdown is going to be out of the question. For a setup that needs to be portable, I need small usable speakers, and I've heard lots of good things about the Yamaha MSP5s and the Yorkville something-or-others. Even such inexpensive small speakers will probably serve me better than anything I could put together.

(As a side note, I've heard the MSP5 monitors and think they definitely would do the job. They don't sound bad and they don't sound fantastic, which I would think is exactly what you need for mixing. I could add in a small subwoofer to check the very low end of the mix.)

But this thread isn't about choosing a small monitor, it's about building speaker cabinets that are non-resonant. All I can say is, someday I will build my own speakers for casual listening use, along with a spiffy amp (maybe I'll DIY a Leach amp). That said, anything I learn here will serve me well.

Thank you again to Paul, PRR, Gus, and others for the input!

Ach! Okay, I'm sorry for the tone of my last post. I really am here to learn, but really, my original post was only about how to make vibrations cancel each other out, and to find out if anyone had done any work in that area before (and I'm sure someone has). It was never really about how to build speakers, or how to choose small monitor for mixing. I was just being a student of mechanical engineering, like when I'm in class.

That being said, I do have to wonder if I could do better building my own small monitors than buying. I know Marik has a kit he recommends for the purpose, and I've heard Marik's recordings and know them to be good, so I'm willing to bet he knows a thing or two about what works. So I probably shouldn't dismiss building my own speakers out of hand just yet.

How about getting more practical? Here's what I think I would like in a set of near-field monitors:

• Sealed enclosure
• Flat enough response to be usable for mixing
• Barring the above, I'd at least like excellent midrange imaging
• Bi-amped, with amps built in for ease of use
• Not too heavy

The idea, of course, is to have something that can be dragged places and set up quickly.

Here's how I think it could be accomplished:

• An active fourth-order filter before each driver and amp.
• Use those good-quality power opamp chips for the amps (LM3886 or similar).
• MDF enclosure, 1/2" would probably be fine along with some bracing.
• 6.5" mid-woofer with 1" or 1 1/4" dome tweeter.

Am I even close to the right track?

[quote author="Consul"]
Bi-amped, with amps built in for ease of use
[/quote]

Bi-amped is a very good idea, but I actually would discourage from built in option. Vibrations are not very good for any piece of equipment, and you want to isolate the amp from them, as well.

Hey, Marik!

Okay, I see what you mean. Basically, I could put the crossovers and the amps in a separate rack case, then send wires to speakers? I'll bet that would work well. Thanks!

I wish I was still living out that way, so we could get together and play with mics some more. :grin:

BTW if somebody asks strangeandbouncy very nicely .....
He used to work for Roger Quested and has built and sold passive 8inch monitors

You know, this thread really belongs in the "DIY Monitor" thread now. Ethan, any chance you could merge these together? Thanks!

As for my first question, I believe I received satisfactory answers from the great people of this forum. :thumb:

And I would NEVER dream of asking strangeandbouncy to give away any professional secrets. :wink: Unless he really wanted to... :green: