Victorian Boombox Mk III. A brand new start.

[Prof Marvel]

Great analysis J.

I vote for Magic Server Pixie Dust

https://www.youtube.com/watch?v=3nbEeU2dRBg

yhs
prof mumbles

[J. Wilhelm]

Great analysis J.

I vote for Magic Server Pixie Dust

https://www.youtube.com/watch?v=3nbEeU2dRBg

yhs
prof mumbles

Ha ha!

I think I'm willing to redesign the horn itself. I feel like I have gone as far as I can go with the driver enclosures. They can't be any cleaner or stronger than they are. I think that if the outlet slots were larger there might be more radiation of the standing waves. The recommended outlet to inlet area ratio is about 1:6  which is much larger than I have now.

I have a horn design in my head that I could execute, a variation of this design:

http://p10hifi.net/TLS/downloads/B139-TTL-map-100707.pdf

The advantage is not immediately obvious, but the horn is more "cylindrical in cross section if you use triangular cross sections and you will use less wood as well (though the wood will need to be thicker). It's not a thin slot slot like I made (though I have seen small slotted ones too). If I can make the horn behave more like a horn and less like an infinite baffle, then there might be hope to extract more performance out of this system...

[J. Wilhelm]

I have conducted what I think are the last design studies for the "elbow" or 90 degree ramped driver enclosures. I tested 5 different configurations - except the one which achieved "parity" with the Sony Bass Reflex (the one with the mesh stucco ramp, extension ring and a little batting). The configurations were 1) the solid ramp enclosure 2) The Mesh wire/stucco ramp with extension ring, no batting 3)The mesh/stucco ramp with no ring and no batting and 4) The mesh/stucco ramp with no ring but a little batting in the back of the driver. All in all 25 plots, which I'm sure you don't want to go through, so I consolidated the information in a single scatter plot.

The results are a bit disappointing. The performance for all configurations is significantly less than the "Sacred Geometry^TM," defined as the mesh/stucco ramp with the extension ring and a bit of batting in the box, which itself is a fluke and unexplainable analytically. Any other configuration is worse (see plot above) and sometimes much worse.

The Solid wall ramp in particular performed at the bottom of all the configurations, especially below the design frequency of 59 Hz, and around 57 Hz the solid ramp has a performance of 3 dB lower than the Sony Bass Reflex. At 67 Hz that deficit is only one single dB (which is good) however, and above those frequencies (77, 87 Hz) it's an equivalent match to the Sony Bass Reflex, but then again so is any other configuration. Given the good performance above 77 Hz, the battleground is now set for 47, 57 and 67 Hz for any future design changes.

The mesh wire ramp with no batting did marginally better when no extension ring was used, and as far as the spectrum analysis is concerned there's not advantage to a bit of batting and it is 2dB in deficit at 47 and 57 Hz, One dB in deficit at 67 Hz. So the mesh and stucco ramp radiates about 1 dB extra over the solid ramp , with or without batting, but still about 2dB below performance without that ring and no batting. As to why the extension ring only worked well with batting for the Sacred Geometry^TM is anyone's guess. Apparently the extension ring demands the batting and viceversa, otherwise the performance is low  

For design purposes. the only configuration that achieved "parity" with the Sony Bass Reflex was the "Sacred Geometry^TM Waveguide Cap with Stucco Passive Radiator"  If I decide to stop and build what I have it would be the driver box with the stucco/mesh ramp with batting, exactly as I made it, because it's impossible for me to optimize something that I don't understand off hand, and I can't redesign or improve unless I have a structural dynamics finite element program at hand (which I suspect Sony did have). Not that I can't do it - this fall squarely with the type of design practice I'd do as an undergraduate at school, but the exercise turns into something else because I need to find software, configure it and do a study that may take weeks if not months. Maybe I will do it later on, but I'm not convinced the horn itself is alright performance wise now, and I want a lot more. I don't think I can extract more out of these boxes.

As it stands now for any other ("non-sacred") configuration as shown in the plot above, a full 3 dB max deficit for the solid ramp is very disappointing and a 2 dB deficit of the mesh ramp is not an insignificant deficit, so my instinct is to turn now to the horn itself seeking to recover more than 2-3 dB, and try to go beyond the Sony performance. That can only happen with a horn redesign. As long as I can keep the 1/4 wavelength, I should be able to use the same pair of driver enclosures I made

+++++++++++++++++++++++++++++++++

[MWBailey]

Lots of waqter gone under the relevant bridge by now (horrible multilevel pun intended), but...

on the resonance of materials issue, it might be that less is more. I know about this because at one point several years ago, I was on the point of making my own bass viol (costs and parental prejudices killed the project back then, but I still might give it a go in the future). The pre-baroque English type was the one that caught my attention, because  they used an extremely thin glued-up, then carved-and-bent beechwood soundboard. By bent, I mean that it was steamed and bent directly on to the ribs and over a support frame that then became integral to the interior of the instrument, as the soundboard was glued to it. This was of extreme interest to me, as I had  seen inside several double basses of the laminate-soundboard variety, and noticed that some of them had a support member inside to which the laminate soundboard was glued, much the same way as many guitar soundboards, and had decided that my own viol might benefit from having a laminate soundboard to offset costs.

Instead of the wire-and-stucco arrangement, perhaps a thin, bent laminate arrangement might provide better sound? Just an idea.

[J. Wilhelm]

Lots of waqter gone under the relevant bridge by now (horrible multilevel pun intended), but...

on the resonance of materials issue, it might be that less is more. I know about this because at one point several years ago, I was on the point of making my own bass viol (costs and parental prejudices killed the project back then, but I still might give it a go in the future). The pre-baroque English type was the one that caught my attention, because  they used an extremely thin glued-up, then carved-and-bent beechwood soundboard. By bent, I mean that it was steamed and bent directly on to the ribs and over a support frame that then became integral to the interior of the instrument, as the soundboard was glued to it. This was of extreme interest to me, as I had  seen inside several double basses of the laminate-soundboard variety, and noticed that some of them had a support member inside to which the laminate soundboard was glued, much the same way as many guitar soundboards, and had decided that my own viol might benefit from having a laminate soundboard to offset costs.

Instead of the wire-and-stucco arrangement, perhaps a thin, bent laminate arrangement might provide better sound? Just an idea.

That would be a path to take, provided one is willing to do the trial and error procedure of building ramp after ramp of different materials. It's basically a curved plate passive radiator. When you get the material with the right density, the sound will be 180 degrees out of phase with the sound waves inside the box, and the output from the box is maximized.

The complication if instead you want to plan ahead (instead of trial and error) before you build is that there is no mathematical description of the structure readily available, because of the odd shape of the ramp. Maybe if I could model the ramp as a bent rectangular plate with fixed boundaries. But there's no equation that I know of for that. The modern way is to use a Finite Element structural model. Draw a triangular element mesh/grid with the shape of the ramp and put it into a program like NASTRAN, where each triangle is straight plate of known physical properties. Then have it be loaded cyclically with pressure waves (assumed to be parallel to the plate for simplicity). Turn the crank, see what happens, change the density of the wall, turn the crank again.

If instead of relying on the ramp I used a dummy speaker, or a loaded circular membrane, it'd be easier to tune by adding or subtracting mass to the membrane to keep the vibrations in the right phase. Sounds crazy, but it just might work. The membrane could be on the side of the ramp, but it's not a good idea to have it next to the speaker, because higher frequencies could leak out that way to cancel out sound coming from the front of the speaker. I'm thinking the membrane should be further down the line where the horn has had a chance to eliminate higher frequencies.

Or instead of a membrane it's a rigid piston. Thick enough to keep higher frequencies inside, but able to resonate at the desired frequency.

[J. Wilhelm]

There is a new horn in the works. I will finish that soon. It has significantly less wood inside, and so the outside of the box will need to be thicker, made from ½ inch pine craft board. The interior baffles are made from ⅜ inch craft board.

The central baffle is the trickiest to make. The box doesn't have a square cross section but a rectangular cross section, and so the "diamond" shape baffle has a special symmetry and its edges have to be chamfered at different angles to fit the diagonal of the box. The two thinner pieces are meant to form a triangular tube which is one half of the folded horn (for each right and left speakers).

Because there is less wood inside I can afford to use thicker wood and even then this horn may be lighter than the last one. If I need extra length of the horn there will be external extensions to complete the length. The exact geometry of the horn outlet is to be determined, but the minimum area of the horn outlet will be 1/6th of the horn's input.

As a result of the design, it is very possible that the horn will separate acoustically from the driver enclosure to allow the enclosure to become a Helmholtz Resonator. I'll just have to run that risk, because there's no way of getting the horn output to be big enough without making the overall go horn be much bigger. If that happens, I may try to re-tune the driver enclosure to have it produce a resonance frequency together with the horn. This means (gasp) using acoustics equations. But one way of dealing with this is to assume the horn is the neck of the Helmholtz Resonator and make calculations that way...

[J. Wilhelm]

Horrible Non-Euclidean geometry that can drive a man to the edge madness...

[Prof Marvel]

Horrible Non-Euclidean geometry that can drive a man to the edge madness...

Ah.
you are building an Escher Speaker.

Will you play Non-Euclidean Victorian Marches through it?

yhs
prof non-einsteinian

[J. Wilhelm]

Horrible Non-Euclidean geometry that can drive a man to the edge madness...

Ah.
you are building an Escher Speaker.

Will you play Non-Euclidean Victorian Marches through it?

yhs
prof non-einsteinian

I read a description about it in the Necronomicon. It is said to be loud enough to open the gates of Heaven or perhaps Hell. Abdul Al-Hazradi (wrongly referred to as Abdul Alhazred) assures the reader that this was the shape of a horn described in the Dead Sea Scrolls, and which which is said to have been abandoned on a field during a terrible battle fought between angels and demons during the Second Temple Period. Some texts attribute ownership of the horn to the Demon / Fallen Archangel Azazel.

[von Corax]

The Music Player of Erich Zann.

[J. Wilhelm]

The Music Player of Erich Zann.

Guaranteed to continue playing after I die.

In other news I'm making good progress on the assembly of the horn, but it's a difficult (Non Euclidean) geometry to work with. I had to glue all the interior baffles because it's too difficult to drill pilot holes at those horrible angles.. The horn is turning out to be lighter than the last one, as I expected. I had to guesstimate how far to extend the triangular tube (the exit port of the horn) into the waveguide, so as not to obstruct the sound too much. The sound waves still get squeezed into a slot just before entering the bend of the horn before going into the triangular pipe.

One advantage of the triangular and hexagon-oid cross sections is that you have much less of a chance of establishing traverse standing waves, a problem that plagues rectangular boxes, and thus the need for so much batting. Like the older horn, I'm expecting not much if any batting will be needed.

Another thing, but very important point. I could not sleep well over the fact that the new horn by necessity will restrict the inlet of the horn abruptly, thus likely forcing the driver enclosure to become a Helmholtz Resonator. I did some back of the napkin calculations, and I was pleasantly surprised that with the current geometry of the horn input, the resulting Helmholtz resonance very much approximates the resonant frequency of the waveguide. This should not be a surprise that the longer the neck of the Helmholtz Resonator the closer it behaves to a closed pipe. But what is interesting is that as I restrict the inlet of the horn, the Area to Volume ratio in this actual configuration, with these measurements, causes the Helmholtz resonance to approach the 60 Hz mark. Very close indeed to the horn design resonance of 59 Hz. This was a complete fluke. Reaching design resonance even when modeling the new horn as a very long Helmholtz Resonator is not something I anticipated without a major extension or trimming of the length of the horn (which I was prepared to do by way of external PVC pipes).

So it's looking like the overall shape of the device will not change much. There will be small extensions of the triangular port on the outside of the horn, about 20 cm or likely less, how much depends on what the effective length of the horn is. It varies a lot this time due to the acute angles of the bend and where the baffles need to be to try to guide the sound as far as possible into the bend /elbow of the horn. Compared to the last design, it looks like I'll be able to eek out about 4 extra centimeters from the enclosure toward the effective length of the horn. So to reach a 59 Hz quarter - wavelength, I may need about 14 extra centimeters outside of the box. That's small enough to make a secondary horn "flare" mounted on the  box somewhere. So I'm planning to get creative. As an alternative, I might have a chance to redirect the port to outlet port through the driver enclosure itself, so the outlet port would be mounted right next to the driver.

[J. Wilhelm]

Horn is nearly operational. I've decided to install the ports on the front and back with the idea to add PVC elbows and tubes, should I need an extension. No screws or nails have been added so far. I'm leaving that to the end, because the horn is sealed and I can't get into it anyhow. I could add batting if needed with a bit of imagination, but I'm no longer going to vary the acoustic length from the inside.

Non Euclidean beauty! Can you figure out those horrible angles?

[J. Wilhelm]

Do the evolution...

I have encouraging results from the first preliminary tests. The horn, however, is turning to be a difficult problem to crack. So I don'tknow if  I'll ever be able to present a "Eureka" moment.

The good news. The horn is producing copious amounts of sound from the outlet port this time. Resonance is present, not at one frequency but various frequencies, which means that overtones are present. That is proof positive of resonance. You can feel the vibrations on your hand over the outlet port this time. And below 80 Hz, the port produces equal or greater amount of sound than the driver.

The bad news. Figuring out the exact resonance frequency is proving difficult so far, there is no "aha" moment in my observations whether it's playing music or single sine wave tones. You can only hear the overtones as you sweep the frequency and guesstimate its position. So far as ¼ wave resonance goes, the fundamental frequency should be around 59 Hz. Based on the overtones, I'm guessing it's more like 70 Hz, but it's difficult to know with more precise tests.

I was wrong about my estimate of the Helmholtz Resonance. It seems the Helmholtz Resonance is even lower at what could be one overtone lower around 30 Hz, so I don't even know if the Helmholtz Resonance equation is even applicable in this geometry. There's no indication of resonance around 30 Hz. Also if you look at the Helmholtz Resonator equation, it's really difficult to change the resonance frequency to move it up to 60 Hz, on account that the frequency is proportional to the square root of the Area of the outlet port, divided by the length of the neck and the volume of the driver box. You need very drastic changes (over 50%) in these dimensions to even move the frequency up a decade. So my impression is that the horn in this shape is not a Helmholtz Resonator, and instead it's dominated more by the quarter wave resonance equation (like a flute) whose resonance frequency is linear with respect to the length.

At the behest of my roomate, who told me that the bass in the new horn was still "a little flat" and knowing what I explained above, I decided to add two reflectors inside the driver box, to basically "extend the horn inside the driver box." I'm not sure if it's helping at all, but theory states that should be good for the overall shape of the horn. In my imagination, I liked the sound of the horn better without them, but the spectrum analyzer reads that the frequency response between 60 and 80 Hz is basically very strong, definitely better than before. But I can't give you numbers yet because I have to calibrate and compare to the Sony speaker.

Any shortcomings are not from a lack of volume inside the horn. (read post below). The horn by all accounts increased a whopping 30% in volume compared to the previous design, mostly because the diagonal slat design saves a lot of wood. I am however using the solid ramp enclosure, which means there's no passive radiation. If I get desperate enough I may investigate passive radiation, but not yet.

The port is pumping a good amount of air now the port hole fits 1½ inch PVC elbows, which should allow me to experiment a bit with pipes.

[J. Wilhelm]

Eureka?
Or if you can't beat them, join them!

Do you notice something different?

I have a big development. I could not sleep at the suggestion that I had not cured my horn's woes after all that work. The reflectors I added (above) left me - not satisfied. I did not hear any measurable improvement in the sound, at least to my ears.

Desperate, I started researching passive resonators. I had the idea of using a wooden plate as a piston. But long hours of burning the midnight oil left me with two impressions 1) I've forgotten everything I knew about Bessel Equations and most scholarly papers were based on more advanced plate deformation theory than what I had studied as an undergraduate engineer (Timoshenko)  2) The results for thick wooden plates as opposed to membranes returned non physical values ( 10^-4 Hz), so the membranes needed to be thin, not thick plates.

I stumbled upon a little article that read that a design method is to copy the parameters of a Helmholtz Resonator (where have I read that before?  ) Basically, the ideal mass of a passive resonator is equivalent to the mass of the air in the neck of a Helmholtz Resonator (air column).  And then it dawned on me. Why do I need a passive resonator if it's just a substitute for a Helmholtz port? I realized that I may be able to use a pipe to radiate the sound I was trying to recover.

Quarter Wave Horn, or Helmholtz Resonator?

Then I realized that according to my calculations, the resonance of the horn if you assume it's a Helmholtz cavity is about 30 Hz for the given half-length of the horn. But that is the "undertone" or fundamental corresponding to a first overtone of 60 Hz, nearly the design frequency of the horn!  Why not leave the horn as-is, and add a shorter pipe turning the horn into a two-pipe Helmholtz resonator, with one pipe resonating at 30 Hz (Horn) and the other at 60 Hz (a new port punched into the driver enclosure)? The two overtones should reinforce each other, pretty much the same way that Sony used a near-120 Hz resonance of the drivers (130 Hz) to stimulate the bass reflex enclosure at the fundamental frequency of 60 Hz. Further calculations revealed a 3.5 cm PVC pipe and 7.5 cm of it was all I needed to add to the driver enclosure, which easily fit's into the box. If Sony can have a "Secret Sauce," so can I.

The ambiguity is of course that the horn now may either be resonating as a quarter wave resonator at 60 Hz, or as the second neck of a Helmholtz cavity at 30 Hz, depending on which model you use - and guesstimating which part of the horn is the Helmholtz cavity and which is the neck. But since 60 Hz is the overtone of 30, both frequencies reinforce one another.  It doesn't matter which the horn does - probably both.

So today I replaced the ramp cap on the driver enclosure, with a simple plate with a 7.5 cm pipe sticking out. And guess what happened? Not only did the second Helmholtz pipe come alive with a flow capable of putting out a candle, but the horn's port also started producing more sound and copious wind - not as strong as the Helmholtz port, but much better than yesterday. I was so encouraged by the positive results, that I made up my mind to add the Helmholtz port permanently to the enclosure facing forward next to the tweeter and main cone.

It seems that the problem with the horn is that it's cross section is too small. The air is too restricted. I got a first improvement with the new horn design because it opened the narrow outlet of the horn to 1/6th the area of the inlet as recommended by transmission line speaker builders. BUT that is not enough. By the time the inverted horn is done whittling out all undesirable frequencies, there is simply not enough energy left to pump through the output port of the horn. Transmission lines only work if the cross sectional area is sufficiently big, apparently. And I should have been clued to that when I read that people are filling up to 86% of the horn with batting. Batting, you see, increases the apparent volume of the horn. The batting slows down the speed of sound in an enclosure, so sound waves take longer to reach the walls, making the enclosure seem bigger than it really is. So the inverted horn transmission line may be a great device to "clean up the sound" but it's an energy hog, and very stingy with its output. Thus, you need a lot of cross sectional area to favor the formation of standing waves, because the horn has swallowed a lot of energy acting as a near infinite baffle.

Inside my horn, however, I had observed before a great buildup of pressure. Lifting the driver enclosure from the horn I could hear the back end of the Sony driver screaming loudly, and yet the output of the horn was always very meek, even though I was plugging all the little gaps to make sure that over-pressure didn't create jets of air at the joints. So I started suspecting that my horn was too thin sometime ago (I wrote about it above, somewhere). The horn, basically is "choked." How could it be that with all that pressure inside you could not get anything out? Even if your impedance calculations were way off. That's when I decided to start researching passive radiators to try to recover some of that energy. But it turns out that passive radiators are only a substitute for Helmholtz Resonators. The membranes are too thin to keep the higher frequency content in the speaker enclosure, so sound cancellation of mid and high frequencies is a real problem.

There are several observations I have made. about this setup when making preliminary measurements.

1. The horn is fairly oblivious to the effect of punching a hole in the driver enclosure - in spite of the fact that it likes the ramp to direct energy into the horn's inlet. You can cover the new port with your hand, and at best you'll get a 1 dB increase of pressure at the horn's outlet. In contrast, you lose about 4 dB of pressure when you cover the new Helmholtz port with your hand. Hence there is an important level of energy recovery made by the new Helmholtz port. In other words the energy going to the horn is fairly independent from the energy powering the shorter Helmholtz pipe. This is most likely the evidence for the horn being choked.

2. The horn is resonating now audibly down to 30 Hz. Both the Helmholtz port and the horn's port produce a good amount of sound. The sound is not entirely clean however, and at those frequencies you get a lot of other spurious frequencies creeping in, especially through the Helmholtz resonator ports. The Sony Bass Reflex produces a strong whizzing sound to go along with the signal (read below).

3 The new port can quite possibly blow out a candle during a drum solo. There is a substantial amount of turbulence associated with Helmholtz Resonator  ports. You can hear it as a "whizzing" sound if you place your ear next to the port. The Sony Bass Reflex does the exact same thing. You can mitigate the whizzing sound by putting some batting at the entrance of the port or by adding a PVC elbow at the entrance of the port.

4. At the design frequency of 60 Hz or less, the Helmholtz port takes all the low frequency energy away from the paper cone of the Sony driver - about 4 dB of sound pressure, and the cone quiets down to near nothing. When you cover the port with your hand, the cone re-emits those 4 dB and starts vibrating very strongly. This is consistent with my observation that the horn was producing low frequency sounds through the Sony driver, though the horn's port was "quiet." The Helmholtz port is much better at letting that energy out; otherwise it finds its way out by way of the cone.

5. I'm using the sold wood ramp driver enclosure. The cap is removable and I compared output with/without the ramp. The resonating hybrid Horn/Helmholtz device is prone to garbled sound when no ramp is used. The ramp completely cleans up the sound, probably because the ramp eliminates sound bouncing between parallel surfaces in the box and directs it toward the horn.

Later today I'll start more serious tests on the new horn arrangement. At this point I consider the horn to be a hybrid device, with the horn resonating, but mostly working as an infinite baffle, and the Helmholtz port changing the phase of some of that sound from the back of the Sony driver and pumping it outside. The horn's port is pretty much doing the same thing it was before, in addition to the output of sound coming from the new Helmholtz port, but now I can hear 30 HZ signals coming from both ports, so both modes of oscillation are operating.

[J. Wilhelm]

I can't do tests tonight, because my roommate next door is asleep. But now that I feel like I'm getting closer to the end of the speaker build, it's time to start thinking about Steampunk design. The shape of the SoundBeam™ is entirely dictated by function and it has all the grace of a... well, wood beam! The style could be very modern or anything else you want.

So I'm thinking that one way to start decorating is to consider surface wood carving. I do have some rudimentary experience with wood carving I did back in high-school in the mid 1980's. And I feel comfortable carving simple shapes like leaves on pine wood. Maybe nothing as fancy as the carving below, but it's giving me ideas.

I would like to pursue some ethnic style. Since Lt. Gen. Bahlmann and Adm. Wilhelm proceed from Germanic countries, perhaps I could do some themed wood carving that would go along with that identity. Since AdmiralWilhelm has also a Texan Bavarian background and since Bahlmann has ties with Austria, Belgium and Mexico, perhaps I could even blend in some Mexicana into the German - which perhaps could sound quite bizarre to an American or even a Brit, but historically is very accurate within the context of the Second Mexican Empire, during the 2nd French Intervention of Mexico during the 1860s.

A discussion on Mr. Harrison's new abode has reminded me that cobalt blue and white ceramic blends very well with wood tones. And keeping in the theme of Austrian Prince / Mexican Emperor Maximilian, and hiswwife, Empress Carlotta of Belgium, some Dutch ceramic elements (easily reproduced with white Sculpey - or perhaps find some real white and blue ceramic) would not be out of place...

I still have the white horns I made for the small speakers of Mk II; perhaps I can repurpose them for the ports of the horn.

[Prof Marvel]

Ah My Dear J

Persevere!

l have, long ago, forgotten nearly all the math they forced me to learn in college. Lack of use.
No need for all that engineering math in software, I fear.

But keep up the good work! It is most gratifying to be strecthing the little grey cells as I follow your endevour!

and I still like the little white floral horn.

yhs
prof marvel

[J. Wilhelm]

Ah My Dear J

Persevere!

l have, long ago, forgotten nearly all the math they forced me to learn in college. Lack of use.
No need for all that engineering math in software, I fear.

But keep up the good work! It is most gratifying to be strecthing the little grey cells as I follow your endevour!

and I still like the little white floral horn.

yhs
prof marvel

I'm quite ashamed of my poor memory, actually.I'm havingaa hard time recalling simple things like phasor diagrams. Granted, acoustics was an elective which I found boring, but electrical engineering was such an easy side dish for me!

I have to go back and try to recover my notes from yesteryear. I never finished my project of scanning all my notes. I only have fluid mechanics and Supersonic and Hypersonic Aerodynamics scanned with a math formulary thrown in for good measure. But since I don't know where I'll end up living, I need to start thinking what I'm going to do with a storage unit that has all my books plus 15 boxes of photos dating to the 1890s of my French great-great grandparents and the 1930s 8mm film of my Italian /Basque great grandparents in CDMX.

I feel like I've wasted so much of my life.

[J. Wilhelm]

Ah My Dear J

Persevere!

SNIP

But keep up the good work! It is most gratifying to be strecthing the little grey cells as I follow your endevour!

and I still like the little white floral horn.

yhs
prof marvel

Well I think I have definitely bested the Sony speakers, but I'll let you be the judge by how much...

Right click to zoom in all images (Mozilla)

OK I'll tell you right now   I have gained 2 decibels at 40 Hz and 80 Hz *over* the Sony Bass Reflex, and I have matched every single other performance point below 90 Hz Additionally, as promised, the hybrid horn is now resonating at 30 Hz - and neither the Sony Bass Reflex, nor the "pure horn" can produce ANY sound at 30 Hz whatsoever   So this is cause for celebration, however paltry my advantage may be over Sony's performance. Please note the ridiculously sharp cutoff frequency of 90 Hz, and keep in mind that the "pure horn" without any Helmholtz port basically has the *same* frequency response curve below 100 Hz, as Sony's Bass Reflex Helmholtz resonator! That is utterly amazing when you consider that the horn is a very different animal altogether, in shape and size, with the exception perhaps of the interior volume, which I tried to match. This explains why the horns sounded so close to the Bass Reflex, even though I could always tell the horn was "a little flat downstairs" with my ears.

So... how do you compare the "pure horn" performance now that you have punched a hole in the horn, you say? Easy, I just plug the Helmholtz port! The horn is fairly insensitive to the change of volume impinged by the Helmholtz port. So all you have to do is plug it! A pill bottle was all I need to perform the conversion from "pure horn" to "hybrid horn."    I have taken 144 data points with the spectrum analyzer to obtain the frequency response charts shown here for the Sony Bass Reflex, the "pure horn" (the current configuration with the Helmholtz port plugged shut) and the new Hybrid Horn-Helmholtz resonator, shown as yellow, orange and red lines respectively.

One thing that you will notice (chart below) is that there is a LOT more going on above the cutoff frequency of 90 Hz between the horns and the Sony Bass Reflex. The Sony speakers have a huge dip in performance in the frequency range around 300 Hz (about 4 to 5 decibels of deficit below the horns), THIS range is where the Helmholtz resonator borrowed the energy to pump down below 90 Hz. You have to steal the energy from somewhere higher in the frequency spectrum to pump into the lower bass frequencies. In general both horns perform much better than the Sony Bass Reflex by borrowing a lot less energy from the 300 Hz range, and so between 80 and 800 Hz both horns give you a much flatter frequency response that the Sony speakers. Because the hybrid horn also has a Helmholtz resonator there is also a small dip in that 80 - 800 Hz range, though not nearly as bad as the Sony Bass Reflex. That is a major improvement of the horn design over a Helmholtz resonator type. So the horn was not just "sounding dead down below" without actually giving me much better performance somewhere else in the spectrum! It's just that I couldn't notice because I was all preoccupied with getting resonance below 100 HZ!

If you look at the chart at the top, you will also see another deep dip in performance for the Sony speakers around 3000 Hz. That is odd, and I only have a vague idea of what to make of it, but also notice by how both horns also bridged that gap very nicely - about that is comforting, because it shows the horns have naturally a flatter frequency response than bass reflex systems. Horns are good, folks! It may not be obvious at first, but they are!

And then further up the spectrum between 8 kHz to 14 kHz, we have veritable alpine territory with peaks and valleys that rival the Rocky Mountains. I have to be very careful to ascribe any meaning to these measurements, though, because in the thousand Hz range, it is the tweeter and not the full range driver that is generating the sound. The tweeter is an encapsulated unit with no opening toward the speaker cavity, so it's doubtful that the dip or lack thereof has anything to do with what we've been talking about all along about horns and whatnot. In fact, I found that higher frequency measurements were *very* sensitive to sound reflections in my room, including the position of my body in the range 10 kHz to 20 kHz. You need to take those higher frequency measurements with a big grain of salt. Maybe one day I'll look into it, but since I can't hear anything above 10 kHz , I don't see the point  . Maybe youngsters will tell me "hey dude! Your speakers suck in the higher range!   But regardless, the fact that the horns and the bass reflex speaker show a huge drop in performance above 8 kHZ and below 20 KHz, means that there is a greater performance issue, probably related to phase inversion of the tweeter (common to many speakers) of just a plain performance shortcoming due to the type of tweeter used by Sony, with just a capacitor as a crossover.

Either way, I'm very happy, and I honestly don't think I can do much better over what I have done, and I was going to talk to you about noise issues at low frequencies, but I'm tired and chocolate pie awaits me right now... Maybe tomorrow I will, but I'm very anxious to push this project forward!

Cheers,

J. Wilhelm

[J. Wilhelm]

I'll just drop these here and let it be a teaser... I'm doing some experiments to see if I can improve performance after matching the Sony speakers. My goal is to reach much further... which may mean looking at adding sub-woofer drivers, if I can find them compact enough. I may even have to buy or design a crossover, but I'm dead set on keeping the current horn geometry, because it is now resonating at 30 and 60 Hz.

Trying my hand at adding another speaker - a wooden cap with a mini subwoofer and a port - results will be shown later


The horn with the removable cap allows me to experiment with different setups for the horn without having to fabricate other enclosures

[J. Wilhelm]

So today was "second hand store find" day and I netted a couple of very good items. The first one is a wireless Samsung subwoofer and the second one I'll keep as a surprise should I ever need it

The first find, the Samsung Subwoofer cost me $6 USD

Samsung wireless subwoofer Model PS-WK360

The amplifier in this subwoofer is not worth mentioning and I'm not using any wireless features either. Instead I was looking for a quality speaker. The inside of this subwoofer enclosure is quite plain - juts a rectangular Helmholtz resonator. The neat part about the box is that really nice plastic horn, which I think I might want to repurpose toward the end of the build. Notice also that white rubber "lip" around the Helmholtz neck/pipe. I think that was done to lower turbulence noise. I would be well advised to do the same for my port pipes.

As I was writing in previous posts, I'm trying to add an extra driver to the horn. I have decided that this is as far as I can take the blue Sony cones, being that at best the cutoff frequency could not be possibly be any lower than 80 Hz. Sony took the driver as far low as they could, and the evidence for that is the super-sharp response curve at cutoff frequency of 90 Hz.

When looking at crossover maths, I remembered that the slope of a downward profile in a first order high pass filter is -6dB per octave, for a second order filter is -12 dB, and so on... thus, a very sharp curve on a frequency profile means that the curve "has been pulled up" as far as it can go at the cutoff frequency. Like tieing a string at the cutoff frequency and pulling up. the higher the order of the high pass filter, the steeper the curve (rolloff below the cutoff) and the sharper the corner at the cutoff. This electrical phenomenon is also analogous to acoustical devices, and so it stands to reason that the 60 Hz Helmholtz resonator Sony used in their bass reflex enclosure already had pulled the blue driver down to its absolute limit from it's (woofer+tweeter) fundamental frequency of 120 Hz.

So looking at that situation, I started experimenting with adding a 3rd driver to the horn. I have not decided whether I want that to be powered by the same line (ie an extra driver for each channel right and left), or if I want to use a separate dedicated subwoofer amplifier (mono or stereo I have not decided). But I would prefer to stay away from using a mono amplifier setup if I could, and even better with no extra amplifier. I can always provide more power by way of the main amplifier. I also know that I don't want to use an extra driver enclosure. I already have a horn resonating at 30 and 60 Hz, and it should be enough!!

My Second Order Crossover setup for the Logitech subwoofer scavenged from the MkII.
I wrapped my own coils around 1.2 cm PVC pipe and found a couple of suitable capacitors

At first I started looking at simple first order passive filters, and when I realized that hardly did anything I started experimenting by building a second order filter. My problem is that the driver you see in the picture above is a very low quality driver scavenged from the now infamous Mk II boombox. Power rating is unknown. The magnet is small. Safe bet that it will not withstand the 100 W power rating of my Sony "test mule amplifier." It's clear to me now that every single element of that Logitech computer system was absolute garbage. One of the reasons I want to stay away from cheap subwoofer systems for the Mk III. I could get some low frequency sound out from the Logitech cone, but very little, and only by using that awful amplifier around which they built the system (which from now on I will refer to as the "kareoke machine" because that is probably what it was). 

First Order Crossover, +-6 dB per Octave
(high pass filter for the tweeter and low pass filter for the woofer)


By the way, Sony used a first order high pass filter for the tweeter, and no crossover for the woofer.

Second Order Crossover, +-12 dB per Octave

And so I now introduce you to my find at the second hand shop today. The Samsung subwoofer has a really nice 5" mini subwoofer, not as large as the one I used for the Mk. I - and besides, I am planning to stay away from mono channels if I can. Ideally each channel should have it's own subwoofer - at least that was my approach with the horns - to purposefully stay away from giant subwoofer boxes. But the real beauty of this driver is in the back. I was pleasantly surprised to see a rather hefty magnet in the back. Similar samsung cones hava ratings of up to 200W and the voice coil in this driver looks a lot like that of a much bigger subwoofer driver. There's a reasonable chance I will not blow up this driver.

The Samsung AH59-02735A  5-inch driver for the wireless subwoofer system Model PS-WK360

Preliminary tests are showing encouraging results - with the second order filter shown above. I will build a new "cap" for the horn's driver enclosure (with no resonating pipe) to see if feeding from the same line is effective - and measure by how much. If all goes to plan, I might mount the new woofer with a 3-way second order crossover permanently on the side of the enclosure, and pointing down to the ground under each driver enclosure. The downside, is that if I use this driver, I will be forced to find another identical driver somewhere on the Internet. That's not going to be fun at all. And if I'm forced to make due with a single driver, I might have to use a third channel, a second driver enclosure, or a second amplifier for the single driver. Or else, I might have to resort to the "Nuclear Option^TM" which is related to my second find at the second hand shop... and which shall remain a mystery for now!

Cheers,

JW

[Prof Marvel]

Great score on the subwoofer!
the thrift stores are awesome sources for "old stuff".

good luck!
prof amrvel

[SeVeNeVeS]

I am still watching with much interest here, may I say, most of the technical bits are over my head, but superb woodwork skills! Neat and precise, like it so far.

[J. Wilhelm]

I am still watching with much interest here, may I say, most of the technical bits are over my head, but superb woodwork skills! Neat and precise, like it so far.

Thank you Mr. Seveneves! The woodwork I confess I did so as to make minimal effort (letting the plank width decide the overall girth of the horn) because I knew I was going to do a lot of sawing and cutting, over and over again. To be honest, I'm a bit sick of making boxes, and I want to get to the next phase of the project, but I can't yet rubberstamp it yet, because the Mk I boombox still grossly outperforms the horn (not above 100 Hz though). The horn has surpassed the Sony system, but it's falling short compared to the Altec Lansing Subwoofer. Since I have no use for the Sony Bass Reflex speaker any longer, I'm tempted to disassemble the second Sony enclosure and just build the mirror image of the horn's driver enclosure shown above, but since I don't know how many more modifications I will need in the present box to fit in the subwoofer(s), I don't want to find out later that I need to make two new boxes instead of only one. Plus I'm listening to music in stereo while I am planning my next steps.

Throughly recommend these which I've been using to hear the mid-range performance of the horn.
Chopin Nocturnes
Chopin: Polonaises
Debussy | Clair De Lune and other piano pieces of Claude Debussy with impressionist paintings

[J. Wilhelm]

So this is my train of thought in the last couple of days. I've been contemplating using an external amplifier for the subwoofer, but I'm increasingly dissatisfied with my choices, namely adapt a poorly made subwoofer amplifier, or build one myself from scratch, with all the attached complications. like power supply, etc. So I'm gravitating more and more toward building a crossover and simply connecting the subwoofers to the same line that feed the horn drivers. Listening tests, tell me that I'm getting very little volume from the Logitech 30 W amplifier. The Sony amplifier can give me a lot more, but makes things difficult by only pumping surround sound signals to the rear left and rear right channels and the center channel. And there is no powered subwoofer output, only a line level signal, meant to be amplified externally by whichever subwoofer you are using. Consequently, the only way to get the volume I want and the frequency profile I want is to turn the speakers into a 3-way system with integrated subwoofers.

I've been looking at practical crossover circuits with design values that are representative of what I have on my hand. I just wanted to see what kind of component values I need. The crossover will be relatively independent from the type of subwoofer I use, because the cutoff frequency is so low anyway. It turns out that since the Sony blue drivers have such a wide frequency range, there is hardly any need to place a crossover circuit on the blue cone (there is none), and only the tweeter got a crossover, in the form of a single capacitor. To see why look at the circuit below.  This is what I'm looking at design wise with a full 3-way 2nd order Butterworth crossover:

For the tweeter at the top of the graphic, the inductance needed is 0.09 mH. That's actually smaller by a factor of 10 than the inductance of the tweeter voice coil itself, so that explains why Sony didn't use anything other than the capacitor.

For the midrange (the blue woofer in the Sony Bass Reflex) you also need a very small value ~ 0.11 mH to cut the high frequency range so by the same token that companion capacitor is not needed as well, and only the low pass part of the circuit (C3 and L3) have any worthwhile values needed to make room for a subwoofer. But since the rolloff in response (I assumed a 100 Hz cutoff needed to make room for a subwoofer) is so steep in the lower range for the Sony driver, then it stands to reason I don't need C3 and L3 either! Doing something else on either end of the Sony blue driver (high/low frequency) would change the frequency profile, in the mid  and high range, which I don't want to do.

The only thing that's left to do is use a low pass crossover to limit the frequency response of the subwoofer.  But how much do I need to attenuate?

Listening tests tell me that attenuations of 6 dB per octave (first order filter) and even 12 dB (second order filter) are not enough. You hardly notice the difference, and the subwoofer is small enough to work well as a mid-range speaker - it's just a bit muffled in response compared to the Sony blue cones. But it actually makes a good center channel speaker by itself around human voice range say 500 Hz to 1 kHz. But I don't need a center channel.

So the thought was to treat the Sony frequency rolloff as if it was the rollof of a high pass filter. I mean, for all intent and purposes the Sony drivers are an electromechanical band-pass filter, so why not measure the frequency rolloff and try to match that with a filter?

Assuming a perfect speaker, the frequency response of the subwoofer would be the black line. The frequency response rolloff of the horn was guesstimated to be about 20 dB per octave. The closest I could do is assume a cutoff frequency of about 80 Hz and use a 4th order low pass filter. It's basically two second order filters back to back, which gives me a 24 dB/Octave rolloff and zero phase shift between the input and output of the circuit.

There is free software online that will actually  use the frequency response of individual drivers from data sheets provided by manufacturers, and combine with the low pass transfer function (black line above) to give you the exact frequency response profile (estimated) for your speaker setup. I don't have data sheets for the Sony drivers though, and I guess I could prepare a spreadsheet table with my measurements for the combined tweeter/woofer. But that may not be necessary unless I hear something very wrong with my setup (the frequency range is so low anyway, that its all icing on the cake. It's not like I'm going to get a flat response from 20Hz to 100 Hz. But I may do that later on.

https://en.wikipedia.org/wiki/Low-pass_filter

Unfortunately, at the moment, I'm severely limited in my electronic component supply. I looked around and I don't have the capacitor values  necessary to make a filter with an 80 Hz cutoff frequency. This is when you curse the day that Radio Shack declared bankruptcy. Fry's electronics is not only difficult to get to in my city, it's also pretty much devoid of... well anything, nowadays (is Fry's still open? Or have they "bought the farm" yet?)! That means it's all mail order for me. Weeks if not a even a month will pass before I can even get anything to start building and testing. I hear there's another electronics shop in town, but I forget the name - I'll have to look it up.

The next best thing is to find some cutoff frequency that I can realize with the components I do have. As it turns out I can make a 120 Hz filter - which is not ideal, but should be low enough to give some usable results. So the idea is to build the lower portion of this graphic:

Building the inductor coils is very easy, I can make any value I want with magnet wire. I have 47 μF capacitors to make a 94 μF capacitor and I have 470 μF capacitors. I don't have enough components to make two sets of speakers, though. I would need to find another identical subwoofer or alternatively a pair of identical subwoofers from another brand (at $40-$60 a pop, I would not buy them new), and I need to order online an wait for the capacitors, or try to find an electronics shop near me.  

But this is what I need to build now to make a proof of concept. If all goes well, I should be able to go back to making sound tests soon...

[J. Wilhelm]

This last week has been consumed by my efforts to gather the electronic components needed to make the 4th order filter. It has not been easy, Starting with the fact that when I wrote the post above, I was under the misguided notion that I needed inductors in the order of micro Henrys (μH) and in reality I needed those values in MILLI Henrys (mH) *facepalm*.

I began the week by looking for capacitors at the one electronic supply shop that I know in Austin. After a good 20 minute walk from the nearest bus stop, I went in only to find out that the supplier for the shop had gone out of business. I could find very few things in the shop, giving me the impression that COVID 19 is ending just about every other business on sight. Images of Cuba, with people driving 1940's cars, and mechanics winding their own alternators flashed in my mind. When I was a kid, I saw an alternator be re-wound by a mechanic at a shop when we had our RV break down in the town of Taxco in Mexico in 1980. It was a fascinating process.

I managed to scrounge up a few capacitors to wire in parallel to make the 500 μF and 100 μF values I need, but if I need capacitors for a second woofer. I'll be out of luck. With no magnet wire to be found anywhere in town, I realized I could only make my own solenoids for inductors.

But the jump from μH to mH is a difficult one. It's the difference from making your own coil out of less than 50 turns of wire on a plastic tube, to needing a ferrite or iron core, and needing to wind at least a few hundred turns.  And 10 mH is a value large enough to need something akin to a small electric transformer with an iron core. Given that this is a 4th order low pass filter for a very low cutoff frequency (100 Hz or lower) means that the inductor will be relatively big and heavy, not something you usually see for speaker crossovers. Some crossovers used by the industry will have values in the order of 1mH and are made by very large numbers of turns of copper wire (literally a golf ball size roll of copper wire). And I need ten times larger than that. This means an iron or ferrite core is mandatory to concentrate the magnetic flux in the solenoid (multiply its efficiency).

Without going much into the science of it, how effective their ability to store energy as a magnetic field depends on the material inside the coil; you're looking at increasing the magnetic field line density inside of the coils so you can keep the amount of wire and the overall size to a minimum, and substances like iron can multiply the flux density value of a solenoid anywhere from a few hundred times (most ferrous metals, ferrite) to 100000 times (pure iron with no oxygen around it). In general terms, in equations (see below) μ is the permeability, that is the resistance to forming a magnetic field which is proportional to the ability to store energy inside the coil, depending on the nature of the material in the core of the coil. Often you will see this term,

μ = μ_oμ_r,

where μ_o is the permeability of vacuum, and  μ_r is the "relative permeability," a multiplying factor over vacuum depending on the material in the core. Air has a relative permeability of μ_r~1, so it's about the same as vacuum. Metals have a relative permeability ('multiplying factor") between 100 and 100000 which (unfortunately) depends on the strength of the magnetic field, making it difficult to design inductors without good test data on metals.

Relative permeability μ_r for mild steel as a function of flux density, B

The kind of inductor sizes I'm looking for, from 1/10th H to 1 H and beyond, really are values you see in power transformer winding for tabletop electronic appliances, so it gives you an idea of the volume and weight you're looking at when making one of these inductors. This is about 10 times bigger than the values you see in speaker crossover components - not too different in value, but a 10 mH coil made of copper with only an air core is something I haven't seen inside a speaker, as it would require thousands of loops of wire. The only way to reduce it to a more manageable size is to use a ferrite or iron core.

The general formulas to calculate the induction L of a solenoid are:

Linear inductor


Toroid inductor

This is where the Steampunk gene helps you the most. I found a steel ring bearing on the road once , which I kept because it looked interesting. "The ring would be good for a torus inductor," I thought, so I used all my heavy gauge magnet wire good for about 150 turns, but I ran out of magnet wire. I needed about 175 turns - so I couldn't even make something better than 9 mH - I needed 11 mH. You can see my initial mistake in the picture below  

About 150 turns of Gauge 22 magnet wire over a torus of hard bearing steel I found on the ground once (right side)
It's about 9 mH. Compare to a 11.2 μH air-core solenoid I made previously (left side)

Not satisfied with the situation, yesterday, I decided to scrounge parts at the hardware store, hoping to find something I could make into inductors - a big chink of steel like a rod should be easy to find, either raw or in the form of a big fence bolt. Maybe I can find magnet wire? If not standard wire? A transformer to adapt?

*I can hear cuban music, smell cigars and feel the dirt and cobble stones under my feet. A 1940 Packard passes me by*

You can make solenoids out of steel door springs

I found several types of steel rods I could use, but nothing as large as 2 cm diameter, which I was looking for to limit the number of coil turns to 75 (I was still thinking of using some left over magnet wire in a thinner gauge), but the though had occurred to me that maybe springs could be used as coils. I didn't take the spring as coil -thing seriously, because I could not remember if would find springs that were thin and long enough. But I did!  A few punches in the calculator revealed that less than 200 turns with a 1 cm core of steel would get me close to the values I needed.  The core I used is a rod of 3/8 inch zinc plated steel - cold rolled mild steel. The springs are just wide enough to fit over the rod plus insulation (masking tape - electrical tape is too thick).

Making the solenoids was a bit harder than I expected. I had envisioned that I could stretch the springs and then use epoxy or enamel to insulate the wire before releasing the spring. I bought a large 20 cm long spring and a smaller 10 cm spring which I thought would be able to be used with minimal stretch - they were almost the perfect size according to my calculations. That was a big error. I ended up ruining the large spring using that approach. The spring kept getting stuck on the insulation I used for the core - it was a mess, and to pull it off the iron rod, I had to destroy the spring - I managed to save a little piece which I used (see below). Good word of advice, stay away from very stiff springs, only use the softest you can find. Also buy the longest springs you can get, and plan to cut them to size later.

Destroying the large spring reminded me of a trick I learned for my Steampunk business years ago, which was to stretch the springs far enough that they deform plastically, that is, far enough they hold their stretched shape. It's a very tricky thing to do, because you don't know how far each spring will stretch. Too far and the spring will be too "open." Quality control in those type of retail supplies is poor. Each spring will have a different stiffness, so  every time you do the stretching the spring will look a bit different at the end - which means that I can't make two identical solenoids. Each is a "one-of." To make matters worse, the more open the spaces between the turns, the weaker the inductance becomes, so you need an extra length of coil if you stretch too far. In other words, make a length in excess and then plan to cut, with calculator at hand to cut at the right place.

Three very fancy inductors. Chromed high carbon steel coils over a core of zinc-plated mild cold rolled steel,
served over a bed of masking tape with vinaigrette
From top top bottom, 1.6 mH / 0.8 Ω,  5.4 mH / 1.2 Ω, 8.9 mH / 1.4 Ω, assuming μ_r = 300
It is possible μ_r may be half as large, but then I'll need test, and double the number of inductors

Unfortunately, as I explained above, it is difficult to know the permeability of the core without testing. And that means I don't really know the inductance values of the solenoids. In the absence of magnetic data sheets, it's very possible the value of the relative permeability μ_r, may only be half as large (cold rolled steel goes down below μ_r=150). If I just build my 4th order circuit as is, and μ_r=150, the cutoff frequency will be a bit over 300 Hz, instead of the 100 Hz I want.  I am currently trying to device a method to test the inductance, but because the resistance values are low, I can't use my phone as the signal generator, and I'm having a hard time using the old trick of wiring a potentiometer (variable resistor) to a mains to 12 V transformer (60 Hz), in series and dividing the voltage between the solenoid and the potentiometer evenly  to find the equivalent impedance. I may try again tonight without starting a fire or something like that. But in any case, if the relative permeability turns out to be one half than I need,  I can always make more solenoids until I complete the required specs.

Normally you don't use high carbon chromed coils to make solenoids, because speaker designers are looking for pure values of inductance with as little resistance as possible. But in my case, since I am connecting a 4 Ω speaker to an amplifier that expects 8 Ω, i need to add resistance anyway. According to my measurements if I add the resistance of the three coils I made, I would be adding 3.4 Ω to the 4.4 Ω of the speaker, so the total (resistive) impedance would be 7.8 Ω, pretty darn close to 8 Ω, seems to me. Just what I need.

Oh well. Time to stop writing!

Cheers!

J. Wilhelm