Anatomy of a Polk RT3

[Guiria]

They likely weight 25-30lbs a piece.



Cabinet resonance is the measurable aspect of added bracing/internal dampening. The coloration from cabinet resonance does not show up in typical measured frequency response graphs. Rather taking multiple impulse response measurements at various angles in an anechoic chamber and averaging them to derive a waterfall plot will show resonance. Also, use of an accelerometer can give one a less accurate, but general idea of cabinet resonance.



The entire audible frequency range does not require equal construction methods to fully attenuate resonances. For example a subwoofer only requires simple bracing techniques that are common, but as the pass band rises the required dampening increases exponentially.

There are certain materials that are far more capable than others at converting energy into heat and thus minimizing resonances. Chris recommended peel n seal which is an asphalt based roof flashing material. Asphalt happens to be an excellent material for use in a viscoelastic constrain layer because it so efficiently makes the conversion previously mentioned. When such a material is placed between two materials such as plywood and cement board as pictures the viscoelastic layer's effectiveness increases greatly due to the added mass and created sheer forces created.

The use of a high grade acoustic damping material within the cabinet such as OC705 or 8lb mineral wool removes all internal reflections which also can cause such resonances.



If you take your time measuring and cutting as well as gluing it is a doable project. The reason for recommending use of steel is because it is far more dense than wood allowing for a less dense bracing matrix and a audibly inert cabinet. Also, this allows for a smaller cabinet to be build since less volume is dedicated to bracing.

If you decide to go with a wood bracing method use of oak stock would be ideal. Also, to achieve the same results you would need to construct a bracing matrix that leaves no gap larger than 3" (ideally 2") in any direction on the XYZ axes.

So with your cabinets there is no rockwool being used, only bracing correct?

Does the fibrous concrete count against your net volume?

So many decisions for a simple 2-way. I do want my midrange to sound ever so sweet though so perhaps I need to just take the plunge.

 

[TLS Guy]

I just can't stomach the thought of putting these new speaker components into the RT3 box. I can also offset most of my box costs if I sell the RT3s, even with one broken. If I hack up the RT3 and things don't end up good then the boxes will be as good as trash.

I am taking the box design you created a couple of posts back.

I plan to keep the front baffle as narrow as the polks (7.25") and get the extra volume in depth and height.

I will be mounting the tweeter, woofer, and port aligned down the center of the front baffle. Unless their is some audible reason to offset one against the other.

You are on the right lines. Yes, you can put the the drivers in line and the port also.

As far as bracing, I would use a vertical and a couple of horizontal braces.

I don't know about steel, for bracing. Steel has a high Young's modulus and rings. It rings at about double the frequency of bronze. I went up to my shop today to finish changing the manifold gasket on my Chrysler LHS. Seven hours work over three days. I have plenty of pieces of that box girder steel and it rings when you rap it. Unless it was filled with a dense foam, I think it could cause mid range coloration. Also if I were experimenting with steel I would use my Miller welding system. That type of steel is known to ring with similarity to the physics of the bell. All sorts of metal objects are included in percussion sections these days. An Anvil rings when struck with a hammer, and that is not hollow. Wagner even wrote a chorus for them, the anvil chorus in Das Rheigold, for the decent into Nebelheim. And a piecing terrifying HF ringing sound they make. I would be cautious about using a high Young's Modulus material for bracing.

For your first project I would do the bracing with carpentry.

Resonances cause peaks in the frequency plot. If they are narrow, then they have to be about 10db above average level to be heard. If they are broad however, 1 or 2 db is audible.

My advice is to keep your first project simple, and use MDF for cabinet and bracing. I would make the front speaker panel 1.5 inches thick. Don't forget to add the speaker, brace and crossover displacement to Vb.

You have a small speaker for your first project and one of the advantages of small speakers, is that they are less prone to objectionable cabinet resonances than large ones.

 

[avaserfi]

So with your cabinets there is no rockwool being used, only bracing correct?

Some sort of damping agent such as rockwool or OC705 is not an option, it is a requirement if a design goal is an inert cabinet. So I will be using 8lb rockwool.

Does the fibrous concrete count against your net volume?

Yes, you subtract the concrete volume from the cabinet volume.

So many decisions for a simple 2-way. I do want my midrange to sound ever so sweet though so perhaps I need to just take the plunge.

If you want optimal sound quality it isn't a hard choice .

 

[WmAx]

I don't know about steel, for bracing. Steel has a high Young's modulus and rings. It rings at about double the frequency of bronze. I went up to my shop today to finish changing the manifold gasket on my Chrysler LHS. Seven hours work over three days. I have plenty of pieces of that box girder steel and it rings when you rap it. Unless it was filled with a dense foam, I think it could cause mid range coloration. Also if I were experimenting with steel I would use my Miller welding system. That type of steel is known to ring with similarity to the physics of the bell. All sorts of metal objects are included in percussion sections these days. An Anvil rings when struck with a hammer, and that is not hollow. Wagner even wrote a chorus for them, the anvil chorus in Das Rheigold, for the decent into Nebelheim. And a piecing terrifying HF ringing sound they make. I would be cautious about using a high Young's Modulus material for bracing.

Like anything, the resonance that occurs is a complex sum of parts/behaviors. Used in the capacity I specified, the steel can not resonate; it's directly grounded to the heavily dampened constrained layer walls. As for the steel tubing in the cross brace section, it's as simple as applying very small strips of Dynamat or similar(Peel N' Seal) if you are paranoid; but this is not likely to be any cause for concern either. I can assure you, I have subjected all of such suggestions to rigorous testing, and the structure I recommended will have panel resonances on the order of -20 to -30 dB lower than standard moderately braced MDF construction that is usually used by DIYers and hi-fi speakers.

F
Resonances cause peaks in the frequency plot. If they are narrow, then they have to be about 10db above average level to be heard. If they are broad however, 1 or 2 db is audible.

This is not true in the capacity of many resonances, such as those usually caused by a speaker cabinet. The resonance will reside below the frequency response of the drivers usually, and be masked by the driver SPL in measurements, but still be completely audible, or even slightly above, but you can not practically measure it accurately without a broadband anechoic chamber, or at least, by using specialized measurement methods. Standard MLS gating techniques will not work. You must have a far field response, and you must be able to aquire/analyze the common range that cabinet resonances occur with high resolution. With this being he 150Hz-600Hz rane primarily, it is not possible to analyze this with any degree of resolution according to the requirements as set forth above with standard DIY measurement methods. The actual audibility of said resonance is easily determined if you can derive a separate relative SPL curve of the direct signal vs. resonance signal. The works by Toole and Olive in perceptual sciences have allowed for an accurate reference body to exist that you can compare relative dB and Q of the resonance response to the main signal response and determine audibility under different conditions(acoustically and with different signals and music).

My advice is to keep your first project simple, and use MDF for cabinet and bracing. I would make the front speaker panel 1.5 inches thick. Don't forget to add the speaker, brace and crossover displacement to Vb.

It really makes little difference if that front panel is 0.75" thick of 3" thick. If it's connected to a resonant cabinet, that cabinet is going to be excited by the direct vibration/energy transfer and from the acoustic backwave firing in the cabinet.

-Chris

 

[Rickster71]

Speakers have always interested me, and I have much respect for all the speaker guys on these forums, The Doc, Chris, and Andrew.

I've got a question that has been gnawing at me since I first saw the cabinet with the steel bracing.

With All the parameters that have to be met to make a truly good speaker, and cab resonance being one of them.
How can a non-resonant speaker be made, and still meet all the other qualities of a feasible, and marketable product?
It looks to be a bit labor intensive; for the box to be produced in numbers that would be profitable.
If that is indeed the ultimate goal.

(I want to stop here and clarify. I don't mean this as a criticism in any way. I'm truly interested in speakers, and this process)

The first thing that comes to mind after the labor considerations, is the box weight, and shipping costs.

Is there any way, that some type of spray foam; like the insulation in an Igloo cooler would meet 90% of the resonance requirements, and be a bit more feasible, to produce and ship?

Thanks guys! I find this stuff fascinating.
Rick

 

[avaserfi]

With All the parameters that have to be met to make a truly good speaker, and cab resonance being one of them.
How can a non-resonant speaker be made, and still meet all the other qualities of a feasible, and marketable product?
It looks to be a bit labor intensive; for the box to be produced in numbers that would be profitable.
If that is indeed the ultimate goal.

This is one reason why there are so few inert cabinets on the commercial market. It typically is not feasible to produce these units at large scale and low cost. Some manufactures have found less labor intensive methods through development of specialized materials that are not reasonable for DIYers to use (think the 802D midrange enclosure).

The easiest way to minimize resonance while maintaining low weight and intensiveness is just implementing a viscoelastic constrain layer properly while using a bracing matrix slightly more intensive than normal. There are some possibilities that could result in a relatively low cost inert cabinet if access to a CNC is had, but these would likely still have to retail at at least $3000-4000 per pair. If I had access to a CNC I would find out for sure though .

As far as an easy one stop solution there isn't one, but there are some materials that would work well in given situations such as glass. Glass is a high Q resonant material. Due to the nature of human hearing high Q resonances are often not audible. Thus using a glass enclosure (such as a 1/4" thick vase) would be a possible idea for a midrange enclosure. I have plans to make a bookshelf speaker that looks almost identical to the 802D (minus the fact that it is a bookshelf) using such an enclosure I just don't have the time or money to actually build it.

 

[Guiria]

If you decide to go with a wood bracing method use of oak stock would be ideal. Also, to achieve the same results you would need to construct a bracing matrix that leaves no gap larger than 3" (ideally 2") in any direction on the XYZ axes.

Is that solid oak you are talking about? Doesn't solid wood resonate more so than ply and mdf?

Given my experience I'm going for the simpler route of wood bracing. I guess the beauty of DIY is if I really get the itch I can always build new cabinets.

Using wood bracing, I plan on one down the center and two or three horizontal braces. I think rockwool or OC705 (whichever I can find locally) would be worth using as well. The question is, how much and where to put it? From what I can tell behind the driver on the backwall seems to be the recommendation in a configuration like this. Does the size (not density, I know I want the 8lb stuff) of rockwool matter?

I thought of building a chamber beneath the actual speaker cavity to house the crossover incase I ever need or want to tinker with it I can get to it without opening the box. I don't want to end up soldering something wrong and then have to open the box up or try and pull the crossover out of a bracing matrix.

 

[WmAx]

Since now wood is considered, here is the wood version of the constrain layering combined with minimum bracing requirements(note: it's very effective - but not quite as effective as the concrete/steel variant - the wood version takes far more volume, of course):

http://www.linaeum.com/productinfo/other/infinity_primus160_ultra/braces_back_s.jpg

http://www.linaeum.com/productinfo/other/infinity_primus160_ultra/braces_front_close_s.jpg

-Chris

 

[WmAx]

Is that solid oak you are talking about? Doesn't solid wood resonate more so than ply and mdf?

Let's first address this point. Never depend on the single resonant behavior of any single material. When you use these materials in and directly adhered to the constrained layer system discussed, this energy absorption layer over-rides the natural resonant properties of the single materials, for all the intents/purposes discussed here, at least. The oak is far stiffer than the other material(MDF), thus actually being more effective in creating an effective constraining layer. The two opposing plates of the constrain layer should be as stiff as possible, and molecular excitation of the constraining channel should be maximized.

Given my experience I'm going for the simpler route of wood bracing. I guess the beauty of DIY is if I really get the itch I can always build new cabinets.

Using wood bracing, I plan on one down the center and two or three horizontal braces. I think rockwool or OC705 (whichever I can find locally) would be worth using as well. The question is, how much and where to put it? From what I can tell behind the driver on the backwall seems to be the recommendation in a configuration like this. Does the size (not density, I know I want the 8lb stuff) of rockwool matter?

I thought of building a chamber beneath the actual speaker cavity to house the crossover incase I ever need or want to tinker with it I can get to it without opening the box. I don't want to end up soldering something wrong and then have to open the box up or try and pull the crossover out of a bracing matrix.

This is a good idea. I used to build such separate chambers when I used to use passive crossovers. It can be difficult to deal with placement of crossover in a densely braced matrix. In fact, I would recommend building an entirely separate cabinet for the crossover and use separate binding posts on the speaker cabinet for each driver. This makes it easy to later go to a full active crossover, which is going to be far superior alternative to a passive crossover.

-Chris

 

[TLS Guy]

Is that solid oak you are talking about? Doesn't solid wood resonate more so than ply and mdf?

Given my experience I'm going for the simpler route of wood bracing. I guess the beauty of DIY is if I really get the itch I can always build new cabinets.

Using wood bracing, I plan on one down the center and two or three horizontal braces. I think rockwool or OC705 (whichever I can find locally) would be worth using as well. The question is, how much and where to put it? From what I can tell behind the driver on the backwall seems to be the recommendation in a configuration like this. Does the size (not density, I know I want the 8lb stuff) of rockwool matter?

I thought of building a chamber beneath the actual speaker cavity to house the crossover incase I ever need or want to tinker with it I can get to it without opening the box. I don't want to end up soldering something wrong and then have to open the box up or try and pull the crossover out of a bracing matrix.

I think you are on the right lines. I find MDF easiest to work with. If you brace it as planned. Yes, put the Rockwool behind the driver. I think it is a good idea of yours to put the crossover under the speaker.

There was a recent paper in audio x-press showing that the larger share of cabinet resonance is induced by direct contact between driver and cabinet and not just the back wave.

Decoupling the speaker from the cabinet with a rubber gasket and using rubber retentive nuts, these are nuts embedded in rubber and available in most hardware stores, drastically reduced cabinet resonance. I have not tried this yet but I intend to. The measurements with accelerometer were persuasive.

We do seem to have had some mission creep on this project. We were aiming to resuscitate and improve some existing speakers, now we have a totally new speaker. Hopefully it will be a good speaker considering it will be using pretty small drivers. I would not let the project get too far out of hand. Spread your wings with the next one.

I have no idea of your application, or whether you will be using it with a sub or not. I think that is important, as if you will always be using a sub with 80 Hz crossover then the optimal box is 0.217 cu.ft, with a 1.5 inch port 7 inches long flared at both ends. That will give the best spl, with an F3 of 74 Hz. If you are not using a sub, or will not always be using one, I would build the 0.35 cu.ft. box, and get the best compromise of spl and bass extension.

 

[WmAx]

There was a recent paper in audio x-press showing that the larger share of cabinet resonance is induced by direct contact between driver and cabinet and not just the back wave.

Decoupling the speaker from the cabinet with a rubber gasket and using rubber retentive nuts, these are nuts embedded in rubber and available in most hardware stores, drastically reduced cabinet resonance. I have not tried this yet but I intend to. The measurements with accelerometer were persuasive.

It's induced by both. You can not ignore either one. I have worked extensively with de-coupling drivers, and it does help substantially. But it takes more than a simple thin rubber gasket to optimize this technique. To get a high degree of isolation, and through a broad bandwidth, you have to create an effective suspension with a low Fs. This means using very thick, very soft closed cell materials, and having the driver mount with very low pressure, as to avoid stiffening the suspension.

The easiest method that is effective without going to more extreme mounting techniques, is to route an additional 0.5" into the driver recess hole. Use thick vinyl closed cell foam applied as one layer. Now cut a solid ring out to sit on this layer. You can use 1/8" metal, or use 1/4" hardwood ply and increase the recess depth to compensate for the ring. This will act as a stabilizer for the 2nd layer of foam. Now mount another layer of the closed cell vinyl foam. Mount the driver over this double layer of stabilized foam. Use a long screw and use metal washers, with some foam rings cut out to use as spacers between the screw/washers/driver. Tighten the screws enough just to hold the driver in place and get a good seal.

-Chris

 

[TLS Guy]

Speakers have always interested me, and I have much respect for all the speaker guys on these forums, The Doc, Chris, and Andrew.

I've got a question that has been gnawing at me since I first saw the cabinet with the steel bracing.

With All the parameters that have to be met to make a truly good speaker, and cab resonance being one of them.
How can a non-resonant speaker be made, and still meet all the other qualities of a feasible, and marketable product?
It looks to be a bit labor intensive; for the box to be produced in numbers that would be profitable.
If that is indeed the ultimate goal.

(I want to stop here and clarify. I don't mean this as a criticism in any way. I'm truly interested in speakers, and this process)

The first thing that comes to mind after the labor considerations, is the box weight, and shipping costs.

Is there any way, that some type of spray foam; like the insulation in an Igloo cooler would meet 90% of the resonance requirements, and be a bit more feasible, to produce and ship?

Thanks guys! I find this stuff fascinating.
Rick

Rick, why on earth did you have to ask why we do it?

I really don't care what the commercial guys do. I'm not aiming to make money out of it all. This gives me a huge advantage. I can do things that are totally impractical commercially.

The only company that I can think of that sold good TL speakers over an extended time was TDL.

I like the TL, and always have. I find that they have very low coloration, even though the bracing has to come largely from the pipe configuration. If you put braces across the pipe it won't work. I have found TLs to have by far the most natural speech reproduction, and very accurate reproduction of tymps, bass strings and piano.

I have decided to continue the heritage of Voight, West, Bailey, Fried, Wright and Rogers. Also not forgetting Peter Walker's contribution to the BBC TL monitors. Those speakers, now sadly parted out, where the very stuff of legend. This lineage has pretty much died out commercially, and few now have a chance to hear the total package. The commercial fruits of all this were the Radford and IMF monitors, and the line of speakers produced by John Wright of TDL until his untimely death. You just about never see a TDL offered for sale. A lot were sold, but their owners just won't part with them. I know there have to be a huge number in use, especially in the UK.

I had an audio enthusiast from the Twin Cities here yesterday afternoon.
He has not built a set of speakers for 20 years. He visited the Hi-Fi dealers and decided it was time to get building again.

He had just come from hearing the $50,000 per pair Wilson series 2 MAXX. He is an organ enthusiast. He said my system was far more realistic. We played a lot of Bach and he could not believe the definition and articulation of the bass pedals. On orchestral music he really noted the realism of the typms and bass strings. After listening to the Klais organs from Cologne on SACD, he said he could he could feel he was in the cathedral. He said he really felt the notes hang in a big space with the organ high in the distance. He could really feel the big bass pipes.

So Rick, my motivation is to build something that would never see the light of day commercially, and not have to live with the constraints that imposes. I want to have system that will satisfy me for years and years to come.

 

[TLS Guy]

It's induced by both. You can not ignore either one. I have worked extensively with de-coupling drivers, and it does help substantially. But it takes more than a simple thin rubber gasket to optimize this technique. To get a high degree of isolation, and through a broad bandwidth, you have to create an effective suspension with a low Fs. This means using very thick, very soft closed cell materials, and having the driver mount with very low pressure, as to avoid stiffening the suspension.

The easiest method that is effective without going to more extreme mounting techniques, is to route an additional 0.5" into the driver recess hole. Use thick vinyl closed cell foam applied as one layer. Now cut a solid ring out to sit on this layer. You can use 1/8" metal, or use 1/4" hardwood ply and increase the recess depth to compensate for the ring. This will act as a stabilizer for the 2nd layer of foam. Now mount another layer of the closed cell vinyl foam. Mount the driver over this double layer of stabilized foam. Use a long screw and use metal washers, with some foam rings cut out to use as spacers between the screw/washers/driver. Tighten the screws enough just to hold the driver in place and get a good seal.

-Chris

That sounds like a good idea. With my plasma cutter, 1/8 inch steel rings would be easy to cut.

What do you use for the stabilized foam?

 

[Guiria]

We do seem to have had some mission creep on this project. We were aiming to resuscitate and improve some existing speakers, now we have a totally new speaker. Hopefully it will be a good speaker considering it will be using pretty small drivers. I would not let the project get too far out of hand. Spread your wings with the next one.

I have no idea of your application, or whether you will be using it with a sub or not. I think that is important, as if you will always be using a sub with 80 Hz crossover then the optimal box is 0.217 cu.ft, with a 1.5 inch port 7 inches long flared at both ends. That will give the best spl, with an F3 of 74 Hz. If you are not using a sub, or will not always be using one, I would build the 0.35 cu.ft. box, and get the best compromise of spl and bass extension.

I will be using these with a sub which is currently crossed over at 80 Hz. I would prefer to build them with solid performance in mind given the range of the driver and not maximize the total extension in sacrifice of spl. If I get an f3 of 74 with best spl @ .217 cu. ft then I will go in that direction.

If I like the sound of these speakers then future plans would be using these speakers as surrounds and building larger speakers for my mains. That won't happen for quiet some time though since other audio priorities would be first such as dual subs and acoustic treatments in my media room.

 

[TLS Guy]

I will be using these with a sub which is currently crossed over at 80 Hz. I would prefer to build them with solid performance in mind given the range of the driver and not maximize the total extension in sacrifice of spl. If I get an f3 of 74 with best spl @ .217 cu. ft then I will go in that direction.

If I like the sound of these speakers then future plans would be using these speakers as surrounds and building larger speakers for my mains. That won't happen for quiet some time though since other audio priorities would be first such as dual subs and acoustic treatments in my media room.

Your getting the hang of it, and catching the contagion!

 

[Guiria]

I'm having trouble finding a flared both ends 1.5" port. Will a 2" port be too large? What would be the length for a two inch flared port?

 

[TLS Guy]

I'm having trouble finding a flared both ends 1.5" port. Will a 2" port be too large? What would be the length for a two inch flared port?

A 2" flared vent would be 12.5" long. If you don't have room for that then use non flared cardboard tube. 1.75 diameter needs to be 8.4" long and a 1.5" tube needs to be 6" long. Use the largest diameter you can fit, as that lowers port velocity. Vent velocities are 15, 21 and 30 m/sec respectively. The lower the better.

 

[Rickster71]

So Rick, my motivation is to build something that would never see the light of day commercially, and not have to live with the constraints that imposes. I want to have system that will satisfy me for years and years to come.

Didn't mean to imply that you were all interested in mass producing speakers.
I was always under the impression, that Chris's goal was to produce a quality speaker, and someday sell it; along the lines of a Jim Salk or Linkwitz Lab.
Don't mean to put words in anyone's mouth.

Just wish I lived closer to you guys; I'd come over and learn something.

It's induced by both. You can not ignore either one. I have worked extensively with de-coupling drivers, and it does help substantially.

Paradigm has a mass production version of that, with their Signature series, they call it an IMS / Shock Mount; along with a rear damping chamber around the tweeter and mid drivers.

Chris,
going back to the steel reinforced box. Would I be safe to assume, it's an experiment / prototype to see how far you can take a
'no-resonance' cabinet?

Thanks All,
Rick

 

[WmAx]

Chris,
going back to the steel reinforced box. Would I be safe to assume, it's an experiment / prototype to see how far you can take a
'no-resonance' cabinet?

Thanks All,
Rick

No, it's not an experiment to see how far one can take a cabinet. In fact, it is not a totally inert box. It is far more so than normal cabinets, yes, but I consider the specific example I provided earlier to be a minimum build quality for no audible resonance, or at least almost no audible resonance. That design with the concrete and steel was meant to preserve as much volume as possible using commonly available materials with no special tools, yet minimizing resonances within reason. One could take it to the next level, and use thicker concrete, more visco-elastic dampening, more steel, etc.; but I do think going farther will start to effect the law of diminishing returns.

-Chris

 

[WmAx]

That sounds like a good idea. With my plasma cutter, 1/8 inch steel rings would be easy to cut.

What do you use for the stabilized foam?

The grey vinyl foam, the closed cell type, used for weatherstripping will work, in appropriate thickness. As for life of this material/method: I have one speaker with this type of suspension, using this material, that is approximately 6-7 years old, and the suspension still works fine and has no air breach.

-Chris