AR-4x Cabinet Damping

[johnieo]

Some time ago we discussed differences in how fiberglass (FG) and polyester (PE) stuffing behaved in older speakers. Ken Kantor commented that he had not found an exact replacement for FG in older speakers for which FG was designed. It is itchy to handle and best done with gloves and mask (or outside on a nice 88-degree day like today!) As Ken commented, speakers can achieve the same performance with polyester, if other parameters are changed. The issue faced by those who restore old AR speakers is that there appears to be no direct replacement for fiberglass that will produce the desired low Qtc. Attached are data for the AR-4x that illustrates the nature of these differences. Some time ago I started to collect these data manually, but crunching the numbers became a pain so the idea was abandoned. Ken kindly donated Woofer Tester 1; voila, we crunched numbers. Due to scatter in the data, each Qtc and Fc tabulated in the attachment was obtained by averaging ten runs. Next a frequency scan from 30-80 Hz was obtained for each density using WT’s arbitrary frequency scan function. Some comments:

1. All data were taken on the same AR-4x cabinet and woofer at about the same temperature. Changes in Fs, Qts, and temperature affect Fc and Qtc. I observed about –1/4 Hz per degree F. Also, the original coating on the woofer’s cloth surround was absent in many places; we removed much of the remaining material and re-coated the cloth obtaining Fc = 33 Hz after an 8-hr break-in

2. Qtc = 1 was obtained with either stuffing, but with Fc = 59 Hz (FG) and 64 Hz (PE).

3. It was not possible to obtain Fc = 55 Hz or Qtc = 0.86 with PE.

4. The AR-4x had two x-o / stuffing combinations. Older units: #4 inductor (1400 Hz x-o) and 18 oz. FG. Later units: #5 inductor (1200 Hz x-o) and 12 oz FG. These data were taken with the #5 coil. Results depend on coil size. A #4 inductor with 18 oz. FG yielded Qtc ~ 0.7.

5. It is OK to extrapolate the concept illustrated here to other speaker models.

6. It is NOT OK to use these data to determine how much FG or PE stuffing to use in a different speaker model!

7. Resonance and damping values depend on the electrical and mechanical properties of the woofer and the mechanical properties of the cabinet; please do not scale “pounds/cu-ft” from the AR-4x to another AR model! Qtc and Fc depend on many more factors than stuffing density. Stuffing density is not a “conservable quantity!”

In spite of its itchiness, I am still using FG in these old beasties!

[dynaco_dan]

>Some time ago we discussed differences in how fiberglass (FG)

>and polyester (PE) stuffing behaved in older speakers. Ken

>Kantor commented that he had not found an exact replacement

>for FG in older speakers for which FG was designed. It is

>itchy to handle and best done with gloves and mask (or outside

>on a nice 88-degree day like today!) As Ken commented,

>speakers can achieve the same performance with polyester, if

>other parameters are changed. The issue faced by those who

>restore old AR speakers is that there appears to be no direct

>replacement for fiberglass that will produce the desired low

>Qtc. Attached are data for the AR-4x that illustrates the

>nature of these differences. Some time ago I started to

>collect these data manually, but crunching the numbers became

>a pain so the idea was abandoned. Ken kindly donated Woofer

>Tester 1; voila, we crunched numbers. Due to scatter in the

>data, each Qtc and Fc tabulated in the attachment was obtained

>by averaging ten runs. Next a frequency scan from 30-80 Hz was

>obtained for each density using WT’s arbitrary frequency scan

>function. Some comments:

>

>1. All data were taken on the same AR-4x cabinet and woofer at

>about the same temperature. Changes in Fs, Qts, and

>temperature affect Fc and Qtc. I observed about –1/4 Hz per

>degree F. Also, the original coating on the woofer’s cloth

>surround was absent in many places; we removed much of the

>remaining material and re-coated the cloth obtaining Fc = 33

>Hz after an 8-hr break-in

>

>

***What did you re-coat the surrounds with and how many coats?

>

>

>2. Qtc = 1 was obtained with either stuffing, but with Fc = 59

>Hz (FG) and 64 Hz (PE).

>3. It was not possible to obtain Fc = 55 Hz or Qtc = 0.86 with

>PE.

>

>4. The AR-4x had two x-o / stuffing combinations. Older

>units: #4 inductor (1400 Hz x-o) and 18 oz. FG.

>

>

***What was the capacitor value of the earliest AR-4X and up to what serial number?

>

>

Later units:

>#5 inductor (1200 Hz x-o) and 12 oz FG. These data were taken

>with the #5 coil. Results depend on coil size. A #4 inductor

>with 18 oz. FG yielded Qtc ~ 0.7.

>

>

***Do you know what the starting serial number of this coil mod was?

>

>

***Was the capacitor value changed or what value were they and starting with what serial number?

>

>

>

>5. It is OK to extrapolate the concept illustrated here to

>other speaker models.

>6. It is NOT OK to use these data to determine how much FG or

>PE stuffing to use in a different speaker model!

>7. Resonance and damping values depend on the electrical and

>mechanical properties of the woofer and the mechanical

>properties of the cabinet; please do not scale “pounds/cu-ft”

>from the AR-4x to another AR model! Qtc and Fc depend on many

>more factors than stuffing density. Stuffing density is not a

>“conservable quantity!”

>

>In spite of its itchiness, I am still using FG in these old

>beasties!

>

>

>

***This has been an itchy subject for quite a while now. lol

John, this is a most valuable piece of research you have undertaken, thank you very much.

I can see that you must have started your research from scratch.

[JKent]

>I can see that you must have started your research from

>scratch.

oooooo!! ;-)

seriously tho, I recently insulated my walls with Johns Manville FG insulation from Lowes (I think). It was white and not very itchy. Don't know if it has the same acoustic properties as the pink or yellow though. I was thinking of using some of the leftover in some speakers I'm working on now.

[dynaco_dan]

>

>>I can see that you must have started your research from

>>scratch.

>

>oooooo!! ;-)

>

>seriously tho, I recently insulated my walls with Johns

>Manville FG insulation from Lowes (I think). It was white and

>not very itchy. Don't know if it has the same acoustic

>properties as the pink or yellow though. I was thinking of

>using some of the leftover in some speakers I'm working on

>now.

>

Hi Kent;

Good point.

Now that you mention it, I have seen several versions of insulation being sold here.

Ten years ago, or so, I would have just walked into a lumber yard and bought a bag of Pink 15" x 48" batts or a roll of tar paper backed fiberglass.

Now I see several different plastic bag batts and wonder what the difference might be as far as our purposes are, rather than just as a household insulation.

[Carlspeak]

John:

Your results at 18 oz. FG stuffing level were very close to those reported my earlier, very emperical study. My AR4x cabinet measured about .625 cu. ft volume, thus yielding a stuffing density of 1.8 lbs/cu ft.

I believe the diameter of fiberglass is much smaller than the PE you used in your study. Consequently, there are many more fibers per lb of FG stuffing. So, is this a fair test? I don't know. What is the exact effect of the number of fibers/lb? I can only wonder.

Is the 23 um PE tri-lobal PE the PE standard used for commercial loudspeakers? Is that Acousta-stuff?

What if you used microdenier PE - something in the range of 5-10 microns like fiberglass is?

There is another variable operating here as well - the density of the basic polymer. Glass is higher the PE for sure. Fibers/lb estimates must take this into account for best accuracy.

So, where does all this leave us? A few hz difference in Fc here and there I guess.

Carl

[Carlspeak]

About how much fiberglass do you reckon should be used in an AR3a?

It's all about the music

Carl

Carl's Custom Loudspeakers

[johnieo]

Vern, Kent and Carl: too many questions and only one tired person here!

*** The only capacitor I have ever seen in any AR-4x is 20 uF.

*** The inductor value was increased to reduce the crossover frequency from 1400 to 1200 Hz.

*** IMO, the stuffing was reduced from 18 oz to 12 oz for the purpose of increasing the cabinet Q from about 0.7 to about 1.0 I belive these two changes to be independent and unrelated, but I could be wrong. Have no date for change from 18 oz to 12 oz FG.

*** After swapping data with three of you, a list of 25 serial numbers ranging from 84,xxx to 364,xxx was generated. All cabinets with serial numbers below 174,068 (14 July 1968 parts date) contained a #4 inductor; all those with serial numbers greater than 227,334 (Jan. 1969 parts date) contained a #5 coil. Perhaps other speaker nuts, I mean enthusiasts, can narrow this range.

*** I did not want to mess with the cloth surround at all, but it was leaky in spots -- like fly screen. First, I tried a thin, dilluted coat of Parts Express "latex surround glue." It increased Fs; not good! Individual solvents seemed unable to remove it, so made a mixture of about a "splash" each of Stodard Solvent (paint thinner), lacquer thinner, acetone, and MEK. (A "splash" is a measure of how much wine one adds to stir-fry.) Working outside, of course, on a nice hot day in March with the breeze behind and rubber glove, we tried removing the stuff by rubbing it off with cotton balls. When each turned black, a new one was used. In time, the thicker parts were becoming more compliant and rather than all coming off, it was being spread around the surround and filling in where there was none. Let it dry for 24 hr, and then noticed Fs was still drifting, so used WT's 25-Hz break-in feature over the night. After that it stayed at 33 Hz provided the temp stayed constant. Messy, but it worked. Wouldn't want to do that for a living.

*** We used "yellow fiberglass" just removed the tar?/paper backing. Its fiber diameter was measured to be 2.8-5.5 um. Never measured pink or white. Send me a tad in an envelope and it will be measured.

*** All AR-4x interior dimensions are the same. Early cabinets were 5/8-in thick; later were 1/2-in. However the outside dimension changed (for example, 19-1/4" --> 19")

*** The 23-um trilobe was found in many AR speakers (AR-11, AR-18, AR-94 for examples). I believe it to be a representative AR polyester fiber.

*** Of course comparing small diameter FG to large diameter PE is a fair test! Those were the materials that were used! The two materials have different properties and fiber diameter is ONE of the properties that makes them different. We are comparing reality and showing why one cannot be a substitute for the other.

*** Do not have enough Acousta-Stuf to fill a cabinet, so don't know its stuffing property. However, it is a 52-um diameter trilobal Nylon fiber.

*** I would think it best not to quote density of stuffing in units of weight/volume. How about we use dimensions of, for example, "12 oz FG per AR-4x." Otherwise someone is going to try to scale density to another AR model and get into deep doo-doo.

*** One parameter used to calculate the acoustic impedance of stuffing was the percent free air. Free air is total volume minus stuffing volume. Here we note 6 oz of PE and 12 oz of FG both gave a Qtc = 1. The density of PE and FG are about 1.1 and 2.25 g/cc, respectively, so both yield about the same VOLUME of fiber in the stuffed cabinet.

*** I don't think fiber/pound is meaningful. The heat exchange that produces box size increase is dependent on total fiber surface area, and the fiber's heat capacity. Here FG wins hands down on both numbers. Also, IMO, the increased reactance that reduces Fc with FG-stuffed cabinets is due to FG's small fiber diameter and high elastic constant. Think of the stiffness of a glass fiber compared to a piece of thread!

*** I do not know of any PE fiber made in 2-5 um diameter, but have some melt blown polypropylene that size. Unfortunately, it needs to be fluffed to be useful as stuffing. I would be quite interested in seeing how it performed compared to FG. FG has small diameter, high stiffness, polymers have lousy stiffness at any diameter.

*** AR-3a: 28 oz of FG yields Q ~ 0.7 in my Alnico-magnet woofers. I believe Roy C has found that 18-20 oz of FG gives Q ~ 0.85--0.9 with a ceramic-magnet woofer; its not just cabinet size!

If anyone complains about using FG, tell them to go scratch.

Cheers,

[dynaco_dan]

Hi John;

I've been itching to thank you.

Thank you very much for all of your efforts.

I have more than several pairs of AR-4X's that have never needed opening for repairs, yet.

I will document everything that I can for you when I should enter into the itchy zone.

This will be all models that I have, not just the 4X's.

Thank you again, John.

[Carlspeak]

>>

Thanks John for your patience with all of our questions. If you look back at the AR4x stuffing study I posted earlier, you'll see that MB PP was included in it. Bunched up sheets of nonwoven filter material were used and thus, it wasn't fluffed as I also would have liked. I tried it because of it's small diameter similar to FG. At the same stuffing weight as original rock wool (520g) measured Fc was 17 dB lower (51 hz) and also 5 dB lower than OC FG. However, measured efficiency decreased - not sure why. With MB PP density even lower than PET, I suspect the volume of fiber was significantly greater than FG. IMO, the relatively lower results of MB pp suggest the 'tortuosity' path presented to the woofer's back wave is significant. I can only wonder what the results would be if the nonwoven was fluffed up!

Also suggest, if you haven't already, check out Vance Dickason's emperical box stuffing study included in the 4th ed. of his Loudspeaker Design Cookbook (pp 23-31). He freely used stuffing rates to characterize FG in particular. Also, it appears more sophisticated instruments were used to measure impedance and SPL response. I saw a marked difference in the shape of the impedance curve at Fc (with different materials and stuffing densities) between your WT2 and Dickason's Audio Precision System 1 analyzer. His curves shift to the left (lower Fc), as yours did, but tend to remain more point shaped.

It's all about the music

Carl

Carl's Custom Loudspeakers

[johnieo]

>you'll see that MB PP was included in it. Bunched up sheets of

>nonwoven filter material were used and thus, it wasn't fluffed

>as I also would have liked. I tried it because of it's small

>diameter similar to FG. At the same stuffing weight as

>original rock wool (520g) measured Fc was 17 dB lower (51 hz)

>and also 5 dB lower than OC FG. However, measured efficiency

>decreased - not sure why. With MB PP density even lower than

>PET, I suspect the volume of fiber was significantly greater

>than FG. IMO, the relatively lower results of MB pp suggest

>the 'tortuosity' path presented to the woofer's back wave is

>significant. I can only wonder what the results would be if

>the nonwoven was fluffed up!

Carl:

Since the bulk density of polymers is about half that of FG, I would suspect that about 6 oz would be near the desired point. E.g., I would begin with 3-, 6-, 9-oz and see what Q and Fc resulted. Then if one had not reached the point where Fc turned around and increased, I would add more.

The MBPP I obtained is in large, dense, fluffy pieces -- if that is not a contradiction in terms. I am trying to figure out how to improvise a mechanism that will turn it into fluff like cotton. No time to conjure right now. The kind I obtained is used to stuff long "socks" that are floated on the sea to absorb oil from tanker spills. Fluffy, but not fluffy enough. It measured about 3.5 lb/cuft as received. I have been able to fluff manually a small quantity to about a pound/cuft but that is not enough.

I too also intrigued by its fine fiber density. Its impedance-frequency characteristic would reveal which property (fiber diameter or stiffness) is important in obtaining both low Q and low Fc. Then, if it worked for cabinet resonance, one would have to measure its sonic properties in the upper woofer range -- as Ken Kantor explained -- a totally different set of conditions.

>Also suggest, if you haven't already, check out Vance

>Dickason's emperical box stuffing study included in the 4th

>ed. of his Loudspeaker Design Cookbook (pp 23-31). He freely

>used stuffing rates to characterize FG in particular. Also, it

>appears more sophisticated instruments were used to measure

>impedance and SPL response. I saw a marked difference in the

>shape of the impedance curve at Fc (with different materials

>and stuffing densities) between your WT2 and Dickason's Audio

>Precision System 1 analyzer. His curves shift to the left

>(lower Fc), as yours did, but tend to remain more point

>shaped.

Ken explained that one needs to use the same box, crossover and woofer before attempting meaningful stuffing comparisons. Please note that the AR-4x cabinet used here was a later version (1/2" MDF construction) and used an AR-4x crossover. Dickason's box was larger (1 cuft) and made from 3/4" particle board. His raw driver's characteristics were not like those of an AR-4x. His system had no crossover network. His empty box measurements showed higher Zo and Q indicating stiffer box/driver properties. Ditto when adding stuffing, which in many cases was not uniformly distributed, but lined the box. My knowledge of these measurements is not great, but I would assume that we are comparing apples and oranges. Too many differences.

My remaining interest is to see if the type of MBPP I have can be -- figuratively speaking -- "run through a cotton gin" and fluffed uniformly to a low density.

Cheers,

[Pete B]

I recently took another look at fiberglass as compared to a few readily available poly stuffing materials. I spoke to a few of the manufacturers to determine what type of materials were available.

I incresed the stuffing density until Fc no longer decreased or started increasing. One particular poly material (Morning Glory Premium from Walmart) was within a few percent on Fc as compared to fiberglass. This seems to be fine replacement for fiberglass, although I see no reason to change the fiberglass in a sealed speaker.

It is important to note that the Qtc reduction is strongly dependent on how close the material is to the back of the woofer frame. Qtc is typically higher when the material is packed in away from the frame, and lower when right up against the back of the driver.

Pete B.

>Some time ago we discussed differences in how fiberglass (FG)

>and polyester (PE) stuffing behaved in older speakers. Ken

>Kantor commented that he had not found an exact replacement

>for FG in older speakers for which FG was designed. It is

>itchy to handle and best done with gloves and mask (or outside

>on a nice 88-degree day like today!) As Ken commented,

>speakers can achieve the same performance with polyester, if

>other parameters are changed. The issue faced by those who

>restore old AR speakers is that there appears to be no direct

>replacement for fiberglass that will produce the desired low

>Qtc. Attached are data for the AR-4x that illustrates the

>nature of these differences. Some time ago I started to

>collect these data manually, but crunching the numbers became

>a pain so the idea was abandoned. Ken kindly donated Woofer

>Tester 1; voila, we crunched numbers. Due to scatter in the

>data, each Qtc and Fc tabulated in the attachment was obtained

>by averaging ten runs. Next a frequency scan from 30-80 Hz was

>obtained for each density using WT’s arbitrary frequency scan

>function. Some comments:

>

>1. All data were taken on the same AR-4x cabinet and woofer at

>about the same temperature. Changes in Fs, Qts, and

>temperature affect Fc and Qtc. I observed about –1/4 Hz per

>degree F. Also, the original coating on the woofer’s cloth

>surround was absent in many places; we removed much of the

>remaining material and re-coated the cloth obtaining Fc = 33

>Hz after an 8-hr break-in

>2. Qtc = 1 was obtained with either stuffing, but with Fc = 59

>Hz (FG) and 64 Hz (PE).

>3. It was not possible to obtain Fc = 55 Hz or Qtc = 0.86 with

>PE.

>4. The AR-4x had two x-o / stuffing combinations. Older

>units: #4 inductor (1400 Hz x-o) and 18 oz. FG. Later units:

>#5 inductor (1200 Hz x-o) and 12 oz FG. These data were taken

>with the #5 coil. Results depend on coil size. A #4 inductor

>with 18 oz. FG yielded Qtc ~ 0.7.

>5. It is OK to extrapolate the concept illustrated here to

>other speaker models.

>6. It is NOT OK to use these data to determine how much FG or

>PE stuffing to use in a different speaker model!

>7. Resonance and damping values depend on the electrical and

>mechanical properties of the woofer and the mechanical

>properties of the cabinet; please do not scale “pounds/cu-ft”

>from the AR-4x to another AR model! Qtc and Fc depend on many

>more factors than stuffing density. Stuffing density is not a

>“conservable quantity!”

>

>In spite of its itchiness, I am still using FG in these old

>beasties!

>

[johnieo]

>I incresed the stuffing density until Fc no longer decreased

>or started increasing. One particular poly material (Morning

>Glory Premium from Walmart) was within a few percent on Fc as

>compared to fiberglass. This seems to be fine replacement for

>fiberglass, although I see no reason to change the fiberglass

>in a sealed speaker.

JoAnne Fabrics and WalMart buyers would specify price and softness; the only consistent properties of big-box pillow stuffing are likely to be low price and sleeping comfort. There is no assurance of this product's long-term or marketing region consistency. The JoAnne stuffing I purchased was on sale, because it was being discontinued.

>It is important to note that the Qtc reduction is strongly

>dependent on how close the material is to the back of the

>woofer frame. Qtc is typically higher when the material is

>packed in away from the frame, and lower when right up against

>the back of the driver.

and that is why meaningful comparisons require careful attention to this and other details.

[Pete B]

>>I incresed the stuffing density until Fc no longer

>decreased

>>or started increasing. One particular poly material

>(Morning

>>Glory Premium from Walmart) was within a few percent on Fc

>as

>>compared to fiberglass. This seems to be fine replacement

>for

>>fiberglass, although I see no reason to change the

>fiberglass

>>in a sealed speaker.

>

>JoAnne Fabrics and WalMart buyers would specify price and

>softness; the only consistent properties of big-box pillow

>stuffing are likely to be low price and sleeping comfort.

>There is no assurance of this product's long-term or marketing

>region consistency. The JoAnne stuffing I purchased was on

>sale, because it was being discontinued.

>

>>It is important to note that the Qtc reduction is

>strongly

>>dependent on how close the material is to the back of the

>>woofer frame. Qtc is typically higher when the material

>is

>>packed in away from the frame, and lower when right up

>against

>>the back of the driver.

>

>and that is why meaningful comparisons require careful

>attention to this and other details.

And so where do you get your stuffing material Johnio?

Most of the industrial suppliers that I spoke to would not

sell in small quantities, do you have a source?

You state the obvious regarding attention to detail, would

elaborate on that so that I can be sure to understand you

correctly.

[soundminded]

Ah, we're back to talking about stuffing and it isn't even time to roast a Thanksgiving turkey. Well, the recipe is for speaker stuffing anyway so it hardly matters.

The role the stuffing plays in an acoustic suspension loudspeaker system insofar as woofer performance is concerned is critical, more critical probably than for any other design. As I've said in other postings, the motion of the woofer in this or any system is very nicely described and understood by using Newton's second law of motion as applied to forced oscillation which has been very long established and has excellent correlation with experimental results. This equation relates the moving mass (cone, voice coil, and former), springiness (both mechanical of the driver and the air trapped inside) the cone works against, and the viscosity or damping factor the driver works against as well. I'm not going to give the solution again but it can easily be found in any college first year physics text or text on mechanics/dynamics. An approximate solution is also given. This describes the system's propensity for mechanical resonance, its resonant frequency and the size and width of its resonant peak. Therefore, getting flat response is a matter of carefully tuning the three elements optimally. Electrical resonance is another matter and IMO opinion should be considered as a separate problem and solved after the mechanical problem is solved. The electrical problem is the easy part. In theory, any driver could be tuned to any resonant frequency with any sized resonant peak by adjusting these three variables, m, k, and b. In practice, it's a matter of efficiency. Villchur's analysis relied on the thermodynamics of the phenomenon and while correct, it does not tell us anything about the motion of the cone the way Newton does. The thermodynamic loss comes from the frictional loss the air squeezed between the fibers causes which is converted to heat. Thermodynamics generally deals with quantities of heat but does not necessarily tell us much about the specifics of the mechanics of the process. (BTW, the quantity of heat generated this way is very small compared to the I squared R heating losses of the voice coil.)

So what does all of this have to do with the stuffing in the box? The cone is forced to work against the aerodynamic drag of pushing and pulling air between the fibers. The greater the loss, the lower the resonant frequency and the lower the peak. Optimal is a Q of 0.707 which is what the best AR speakers are designed for. This gives the lowest FR extension without a rise in FR and is the best design choice for high fidelity. And not only does this hold true but the falloff below resonance is a linear 12 db per octave which can be compensated for to a degree by electrical equalization extending bass response even further. Original and Series II Bose 901 extended bass this way for at least 2 1/2 to 3 octaves. So why not just keep adding stuffing to a small box? If too much stuffing is added by packing it tightly, there is less space between the fibers for air to circulate through and the stuffing itself displaces air and less air means a higher springiness and a higher resonant frequency so there is a tradeoff. Ported systems are clearly inferior. They usually have a large resonant peak, that peak recurs every octave with a trough in between peaks, and falloff below resonance is at 24 db per octave making equalization to extend response impractical. The irregular repeating peaks and dips make flat bass response over more than about a octave or so also impractical. Generally the resonance frequency of a ported speaker is its effective low frequency cutoff.

In all probability, the optimal tuning for the size of the box and the amount of fiberglass for the type of fiberglass used in AR's designs was determined emperically in a lab at the factory. When the speaker was laid in, in all likelihood, the baffleboard was set to one side already wired up and wired to the controls on the back, the fiberglass which was pre cut was neatly laid in, and the baffleboard glued into the cabinet neatly over the fiberglass. Anyone who has had to work on the crossover network in one of these speakers by removing the woofer from the baffleboard and pushing the stuffing to one side or more likely removing it through the woofer hole and then replacing it later knows it is next to impossible to get it to go back in exactly the way it came out. This will effect a change in woofer response. An even greater change will likely be caused by substituting another type of stuffing material, even a different fiberglass with different length and diameter fibers and different packing compression. Sad but true, it's often the only way to repair these speakers if you have to replace a capacitor or clean a potentiometer.

[johnieo]

As Ken Kantor has described, the cabinet resonance is determined by many factors, the cabinet, the driver and the stuffing.

>When the speaker was laid in, in all likelihood, the

>baffleboard was set to one side already wired up and wired to

>the controls on the back, the fiberglass which was pre cut was

>neatly laid in, and the baffleboard glued into the cabinet

>neatly over the fiberglass. Anyone who has had to work on the

>crossover network in one of these speakers by removing the

>woofer from the baffleboard and pushing the stuffing to one

>side or more likely removing it through the woofer hole and

>then replacing it later knows it is next to impossible to get

>it to go back in exactly the way it came out.

I assume you were trying to be humorous. Cabinets were fabricated before delivery to AR. The back board and baffle board were rabbeted into the four sides, so must be glued along with the four sides in the same operation. Even the lowly AR-4x was fabricated in this manner. Next, a completed crossover assembly was hot glued and stapled over the back opening, then stuffing and drivers.

Anyone working carefully can stuff a cabinet properly; it's not rocket science; it was done routinely on hundreds of thousands of AR speakers.

[johnieo]

>Most of the industrial suppliers that I spoke to would not

>sell in small quantities, do you have a source?

No, as you found, you can purchase railroad cars full of anything you specify. For the purpose of stuffing one particular pair of speakers, I measured fiber diameters and shapes of a few over-the-counter materials in order to find something close to the polyfil found in AR speakers. The 2 bags of material I purchased at JoAnne a couple of years ago were close to AR material. I have no idea what they are selling today. No more need for it.

I disagree with your statement that the polyfill you found is within a few percent of FG in cabinet resonance, therefore a suiable replacement. The fact that it was similar for the cabinet-driver-crossover you measured should not be generalized. It certainly was not acceptable for the AR-4x. In this case, Q = 0.7 at Fc = 59.1 Hz were measured in a uniformly FG stuffed cabinet. I could not lower Q below 0.9 even when the cabinet was tightly stuffed with polyester.

The upper woofer range response is, as Ken Kantor explained to me, dependent on the absorption properties of stuffing. PE and FG have very different absorption and dispersion.

[soundminded]

There seems to be a bit of confusion. The topic says "cabinet resonance" but the discussion is really about system resonance. The resonance of the cabinet itself depends on the type of wood, its thickness, the dimensions, the way it was assembled. The goal for all except a few iconoclasts like Peter Qvortrop is to make the cabinet as non resonant as possible because it is not desirable for the vibration of the cabinet to contribute to the sound of the system. The system resonance consists of two parts, mechanical and electrical resonance. For the mechanical resonance in an acoustic suspension design, the three factors I described m,b, and k are defined by the moving mass of the driver, the springiness of the driver suspsension and the air trapped inside (in a true acoustic suspension design, the driver springiness is virtually zero, the system depends entirely on the air inside) and the viscosity of the air being pushed and pulled through the fibers which creates velocity related dampening by aerodynamic frictional drag. Have I left anything out you can think of? The reason it is possible to create speaker systems with exceptionally extended low end response using this principle is that it is easy to understand and easy to tune for the desired result. It is both surprising to me and unfortunate that as evidenced by most speakers on the market today, the overwhelming majority of designer either don't know how or don't care.

After about 40 years inside the box, I have found it impossible in the half dozen or so AS speakers I've disassembled to get the fiberglass out and put it back into the box in any way resembling the way it was before I disturbed it. If it is true that the box was originally stuffed by pushing the fiberglass through the woofer hole, that is really a shame. It seems to me that had the front or back been held in with screws and a neoprene gasket used to make the cabinet airtight, this would have been a far preferable method of manufacture especially since Kloss decided that he had to put the crossover network inside a sealed box. Laying the fiberglass as a single cut piece in over the crossover network and then glueing on the front baffleboard would have been a second best method, at least there would have been assurance of uniformity of manufacture. As I've said, for all of Henry Kloss' ingenuity and skills, he is not anywhere near the top of my list for world's greatest manufacturing engineers. Ease of repair is not just an issue for hobbyists 40 years later, it's an issue for in-warantee service which is a cost directly hitting the bottom line of any company.

[dynaco_dan]

Hi there;

Just to add to the stuffing information.

In the AR library there is this photo from AR England (number 24).

This photo is just to show the weighing of poly stuffing or at least a spot check of the weight, prior to stuffing an enclosure.

Note that it is not a precison industrial digital scale but appears to be a typical older kitchen or fish weighing scale.

The accuracy was probably quite satisfactory at that time for their purpose.

[Carlspeak]

>Hi there;

>

>Just to add to the stuffing information.

>

>In the AR library there is this photo from AR England (number

>24).

>

>This photo is just to show the weighing of poly stuffing or at

>least a spot check of the weight, prior to stuffing an

>enclosure.

>

>Note that it is not a precison industrial digital scale but

>appears to be a typical older kitchen or fish weighing scale.

>

>The accuracy was probably quite satisfactory at that time for

>their purpose.

Thanks Vern for the dose of reality. After looking at the photo I must concur with John O. that box stuffing is not rocket science. But, then again, this has been my opinion all along. Stuffing speaker boxes back then must have been the apitomy of drudgery.

Like the inductor manufacturing tolerances, they most likely had an allowable tolerance when weighing out the stuffing. However, the technique used to insert the stuff must have varied. I'm sure the pot adjustments at the end of the line made up for numerous variables in driver efficiency, power response, stuffing, etc. etc....

It's all about the music

Carl

Carl's Custom Loudspeakers

[Guest Bret]

>The upper woofer range response is, as Ken Kantor explained to me, dependent on the absorption properties of stuffing. PE and FG have very different absorption and dispersion.<

Something tells me there is more going-on here than meets the sliderule. As you showed, the capacitance of the stuffing material makes a difference in the system response.

There are several interesting questions to be asked. If the stuffing is returning energy to the system, what is the "net result" of all those individual glass fibers vibrating around in there? What are they adding back? Do they add at some frequencies and cancel at others? Is the net result a "suck-out" point of sorts? What frequency does the FG "return" if it is fed, oh, 100Hz for instance? Since you couldn't make a 12" strand of glass "sing" hitting it with a 22' wave, why does it have any effect on the backwave if there is no "wind?" ETC, ETC, just add nausea and a solid state amplifier.

Bret

[dynaco_dan]

Hi there;

If I remember correctly, Audio magazine had an article, in the '60's, about adding a sub or just a plain woofer to the Quad electrostatic speaker system.

My big bros built that electronic crossover from scratch, for a friend of his, in this article used to add this woofer to the Quads at around 70 hz.

A single AR-3 woofer was added to the Quads, very, very nice in a very large room, particulary with Also Sprach Zarathustra.

I wrote about this demo before.

I was feeling sick to the stomach with that drawn out drone, loved it though.

After the tone stopped I felt relieved.

This was a derived center channel fill if I am correct.

As a side note I went to our local library to look up old Audio magazine articles from the mid '70's recently.

I was informed that only after 1980 is available.

I used to have the drawing from that magazine many years ago.

I believe they used an eliptical KEF woofer but it is not that which I am concerned with.

They designed an enclosure from scratch and it used wool only as a stuffing.

It was not an acoustic suspension but labrynith (spel) or similar type.

They used long strands of wool, as a curtain, at the rear of the enclosure and smaller pieces as filler behind the curtain.

This is the only memory I have but it was specially moth-proofed, etc.

I don't think wool is cheap today.

Another comment about poly fills.

When I used to be involved with the outdoors, camping etc, I did a lot of reading before buying a Taiga Goose Down sleeping bag.

A typical sporting goods store would have cheap poly bags for well under $25.00 and the down were about $100.00 and up.

My point is that there was, then, various poly's, Hollowfill 2 was the hot poly, if I remember correctly.

This fibre is tubular in design, like hollow spaghetti, as were some other competive poly's.

For our purposes this might be a bad poly, for the reason that if the fibers were kinked, then they would be an air filled tube which can compress and spring back, not just a mass of material.

Just a thought late at night when I can't sleep.

[dynaco_dan]

>>Hi there;

>>

>>Just to add to the stuffing information.

>>

>>In the AR library there is this photo from AR England

>(number

>>24).

>>

>>This photo is just to show the weighing of poly stuffing

>or at

>>least a spot check of the weight, prior to stuffing an

>>enclosure.

>>

>>Note that it is not a precison industrial digital scale

>but

>>appears to be a typical older kitchen or fish weighing

>scale.

>>

>>The accuracy was probably quite satisfactory at that time

>for

>>their purpose.

>

>Thanks Vern for the dose of reality. After looking at the

>photo I must concur with John O. that box stuffing is not

>rocket science.

.

.

Hi Carl;

No cabinet stuffing is not rocket science, but as we all know now, very important and very critical to a small degree.

Dynaco classic A-25's, for just one example only, used pre-cut pieces of fiberglass in their enclosures.

I am certain that they engineered the correct amount and just ordered carloads of that size already cut to just drop into the enclosures.

I believe that they did that with the vent openings pieces as well, after engineering the design.

As with anything mass produced, they probably pulled a sample at random and tested to see that it met their specs to a close standard.

Looking at the photo #24 from our UK library shows that it would be a boring and time consuming job to arrive at the correct weight from a bulk pile.

.

.

But, then again, this has been my opinion all

>along. Stuffing speaker boxes back then must have been the

>apitomy of drudgery.

>

>Like the inductor manufacturing tolerances, they most likely

>had an allowable tolerance when weighing out the stuffing.

.

.

.

.

A local store here sells or at least sold, a Chinese manual coil winder, that had a set of metal dial counters to keep track of the number of turns.

A chap I know in the service business uses a, motorized, with speed control homemade coil winder.

It has a pulley to analog counter for the number of turns.

.

.

>However, the technique used to insert the stuff must have

>varied.

.

.

.

In the Heath/AR speaker kits they inserted fist-sized pieces of fiberglass into their enclosure.

I just received a copy of a Heathkit AS-103A and also AS-2 owners manuals to replace the original AS-103 one I mis-placed somewhere here a year or two ago.

After having it so carefully preserved over all these years.

.

.

.

In my heating experience I would, over a period of time, see modifications for furnaces or boilers where a piece of foilbacked fiberglass was used as a precaution.

The piece might have been only 1" thick pre-cut to a certain size.

.

.

I'm sure the pot adjustments at the end of the line

>made up for numerous variables in driver efficiency, power

>response, stuffing, etc. etc....

>

>It's all about the music

>

>Carl

>Carl's Custom Loudspeakers

[kkantor]

You got it right, John. The kind of unit-to-unit variances that occur between the Qts of woofers is much higher than any variance from a reasonable stuffing procedure.

For some reason, this reminds of a story:

In the late 1970's, I was attending a roundtable group interview of a young technician candidate for a position in AR's research group.

Questioner: "So why is it that you want a job here at AR."

Candidate: "Oh, I really love audio. It's my passion."

Questioner: "Son, saying that you want to work at a loudspeaker manufacturer because you love audio is like saying that you want to work at a slaughterhouse because you are a gourmet!"

-k

http://kkantor.spaces.live.com

http://www.amazon.com/Cult-Amateur-Interne...e/dp/0385520808

[Carlspeak]

I'm reviving this thread once again in light of an understanding I've reached. That is, the inherent density of stuffing materials, which has not been given any significant discussion here, has and does play a role in affecting the Fc of a speaker (in addition to other factors).

In a number of posts, Johieo has given us stuffing amounts (with calculated rates) in AR4x's and 3a's and noted changes in the amounts that have occurred along with inductor and/or woofer changes. http://www.classicspeakerpages.net/IP.Boar...c=1929&st=0. What I don't know is: did the change from rock wool to the more fluffy yellow FG occur at about the same time as the change in coils and woofers?

Recently I worked on an AR4x and 3a. The 4x was stuffed with 12 oz yellow FG and had a #5 coil. The 3a was a very early model with Alnico woofer and had 28 oz rock wool and a #7 coil. So here, I had a later 4x and an earlier 3a with stuffings the same as johnieo's stated amounts. It occured to me as I handled each of the stuffing materials that the Rock wool stuffing seemed heavier per handfull than the yellow FG. I took a small box (0.22 cu. ft) and zeroed it on a scale and lightly filled it with each of the materials and weighed each separately. I tried to make the degree of filling similar to what occurs in speaker boxes. The rock wool calculated density using my little informal measuring device was 1.73 lbs/cu. ft. The yellow FG ended up at 0.94 lbs/cu. ft. That's a significant difference. I took my little test protocol a step further and filled my test box with poly-fil pillow stuffing and got a density of 0.71 lbs/cu.ft.

So, now I have to wonder how much of the weight changes may be due, at least in part, to the differences in inherent density's of the fiber.

What's obvious to me is that if anyone tried to stuff a 3a cabinet with 28 oz. of poly-fil, they would have a very difficult time getting it all in there. So, one has to back off a bit on the poly-fil to get a reasonably stuffed box. With less fiber and fibers of relatively larger diameter with the poly-fil (12-13 uM vs 3-5 uM for FG) poly-fil inherently won't perform as well as FG. This has been shown in numerous tests.

[johnieo]

I'm reviving this thread once again in light of an understanding I've reached. That is, the inherent density of stuffing materials, which has not been given any significant discussion here, has and does play a role in affecting the Fc of a speaker (in addition to other factors

Hello Carl:

Thanks for reviving this thread.

Yes, I agree that density is important. However, I think the value of the opimimum density would depend on the kind of stuffing, and as well, the woofer's design.

To go back to your first question- I do think the coil change and stuffing change were made at about the same time, but perhaps not for the same reason? I think the stuffing reduction was made to increase the Q a bit to enhance AR in the showroom. Could the coil change have been made to deal with small woofer changes near crossover? e.g. the new 3a with foam surround used the #9 coil, whereas the cloth surround used the #7 coil, the AR-3a-Limited (Asian) used a stiffer woofer and a 2.6 mH coil.

Back to stuffing... One thing to keep in mind about rock wool in paticular is that it is refinery slag and contain large lumps of glass. These make RW heavier per unit volume than uniformly-spun FG. These lumps most likely don't enhance cabinet acoustics.

In last couple of years I have been making some back of envelope calculations about the roll of fiber properties in determining cabinet resonance. Historically, the role of fiber stuffing is said to be due to an adiabatic/isothermal conversion, wherein the pressure change caused by cone movemennt in and out is mitigated by the fiber's absorbtion of heat on the in stroke and release of heat on the out stroke, thus supposedly keeping the cabinet temperature constant. Next we observe that the replacement of fiberglass with polyfill resulted in a slight increase in the low frequency cabinet resonance--e.g. compare AR-3a to AR-11--using same woofer; PE not as good as FG for reducing low frequency cabinet resonance.

Then some quick calculations of how fast heat can flow into and out of polyfill and FG shows that the thermal diffusion constant is much faster in FG than PF, but both are fast enough to account for keeping temperature constant in the 20-60 Hz range. Also, if one calculates the surface area of the two times the indiffusion depth, there is sufficient heat capacity to absorb and the required calories in either material for the small temperature change. Therefore, the notion of thermodynamic effects causing the box to appear larger as its physical size doesn't appear to be on solid theoretical ground.

Now we look at the formulas I used to calcluate the resonance of a cabinet. These formulas were developed as part of an Air Force contract on noise reduction in fighter cockpits in the mid 1980s at UT Austin by a well known acoustics professor. The main variables in these formulas are (1) porosity of the cabinet (how much space is air and how much is fiber), (2) density of the material, and (3) fiber diameter.

The results for propagating a plane wave from a cone to a reflecting wall located at the back of the speaker cabinet show that the acoustic impedances are: Z(FG) = 2400 - j1000 Rayls/m, Z(PF) = 115 + J50 Rayls/m. This says that both fibers are absorbing and storing energy on each cycle. The real part of the impedance represents loss (heat dissipation), whereas the reactive part represents energy storage and release (fiber vibration). But, note that FG is 20 times more effective than PF in both absorbing and storing energy!

These differences in energy storage and release in the form of vibrations of individual fibers can be viewed in a different way by looking at the vibration frequency of a beam. Each "free" part of a fiber (between the points at which they touch) represents a miniature "beam" with a resonant frequency that is determined by the fiber's length, stiffness, and diameter. FG is much stiffer and of smaller diameter than PE. The "average" length of a free FG fiber is 60 um, whereas the same for PE is ~250 um. the boundary layer for air drag at 40 Hz is about 200 um, so our stuffing situation is crudely analogous to the difference in air resistance of wind blowing through fly screen versus chicken wire -- the widely spaced PE does not vibrate (absorb and release) nearly as much energy as does the FG. Said another way, the difference in natural resonance frequency between a glass fiber and a PE fiber can be observed by suspending pieces of thread and FG in one's fingers, tweaking them and observing their natural resonance!!

My current view is that the mechanical properties of ordinary fibers - properties such as Young's Modulus (stiffness), fiber diameter, fiber density, and cabinet porosity deterimne how well it vibrates in sympathy with the acoustic signal. Therefore it would be interesting to examine other non-glass fibers of similar diameter as PE, but with increased stiffness. Hmmm, got any other fibers to test, Carl?

The natural or uncompressed fiber density which you descrbe, Carl, is important and is contained in the above terms (diameter, density, cabinet porosity).

It is also likely that the optimum density depends on the woofer, so that the best density for an AR-4x is probably not the same as for a 3a--but I know little about woofer design! Thus, I have tended to use units of "pounds per model x cabinet". :-)