Hi there and welcome!
Here’s my story and construction guide for
building epoxy “ball-shaped” loudspeakers:
The Borque Balls
Chapter 1 – Introduction
Welcome, my name is Bernie Borkent, I am a sound engineer @ the CatharijneStudio Utrecht as you might have noticed on the banner above. I am also a passionate speaker builder from childhood until today. I have built quite a lot of loudspeakers (also commissioned by others), everything from failure to victory.
I’ve built small shelf speakers up to big powerful P.A. systems. Back in the 90’s I also worked at JK Acoustics and improved their JK-01 speaker.
The photo beside this text is a picture of a 4 way system I built when I was around 20 years featuring a 300W 15 inch woofer, 8 inch polypropylene bass-mid, a Dynaudio D-52 AF mid-range and a budget metal dome tweeter (which I blew up – I think over 10 times ;-). Subtility was not the biggest thing in those days. I loved the power and potential of loudspeakers.
Notice the special paint job on the speakers though: back then (1990) already I was drawing drafts of my future speaker project on the skin of my speakers!
When I started the Borque Ball project I was around 35 years old, and I wanted to do my own “final exam” in the speaker building craft. Needless to say, these speakers were originally designed with no compromise properties in mind, although, in the end, the budget turned out not to be infinite (as usual).
Beside the budget, the size of the speaker could not exceed certain measures, as no-one owes a ballroom for a living room. It also happened that I’ve liked ‘blub’ shapes already from my teenage years, drawing them in my school notebooks. Maybe it’s originated by some colorful designs which I’ve seen of the sixties, or maybe it has something to do with smooth female curves, I don’t know.. 😉
Besides my love for this shape, the “blub” shape has many acoustical advantages above the traditional ‘square wooden box’ design. The internal room of the speaker has no parallel walls, so no real standing waves can occur. Of course the self resonance of the air volume in the cabinet is something which is causing a resonance but that one we can calculate exactly to the value which is desired. This results in a very linear frequency response (if the driver provides it too of course). Above that, the outside curves provide excellent radiation of the sound, and enables the speaker to create the seamlessly unaffected sound stage (the position of the speaker sometimes completely dissolves).
The sound of my Borque Balls is open, in your face, fast & forward and therefore very detailed and analytic. Thanks to driver quality and the shape of the balls they feature pin point sound stage and accurate instrument placement. Stereo imbalance is clearly noticable within 1 dB difference.
But note: these speakers are very honest about the signal you offer them, which could sometimes turn out to be a big disappointment when listening to older recordings. (In my opinion especially the masterings done on recordings of the 70’s and 80’s are tormented with the presence of the notorious NS-10(m) as a reference monitor in most popular recording studios. The dull, cheap, and punky sound of these speakers encouraged sound engineers in those days to crank up the 2-3 kHz. range to try and make things more audible) Be aware that the Borque Balls will let you hear audio technical fails like that with painstakingly precision.
The difference between amps and DAC’s used in your signal chain are also clearly audible with these speakers. Please take this in consideration before starting to build them: these speakers are no “O wow! It all sounds so great” -speakers at all, but, I assure you, when you get the signal chain right, the sound provided by these speakers can be astonishing. In the past years I noticed most (real) high end speaker designs actually show similar behavior. It’s like the tires of a car; in a normal budget car one does not experience any difference between good and bad tires (or even hardly a flat one 😉 but in a sports car it is even of utmost importance. It has something to do with honesty and transparency of the whole experience. These speakers tend to make everything bigger you offer them.
Ok now! Off we go to the guide for the construction of these glass-fiber based epoxy ball speakers. Firstly, I actually would like to tell the construction guide as a story, rather than a technical guide, because that gives me the opportunity to put in all the embarrassing trial and error moments I’ve been through, in an attempt of keeping you from the same!
Chapter 2 – The driver selection
The drivers I selected were mainly chosen with 3 things in mind: 1) high performance specs 2) the right kind of material used for the membrane depending on the used frequency working area and 3) a decent and justifiable price tag.
First of all, the big and obvious question surfaced: how many drivers shall I use? One, Two or Three?
One full range driver can be shockingly honest and pure in sound, I know, but has one big disadvantage: maximum membrane surface. To be able to produce 20 kHz the diameter of the driver cannot surpass a certain physical border. The maximum sound pressure is quite limited and besides that these drivers features insane price tags, therefore the full range single driver was no option for me.
I designed some two-way speakers which sounded right. In the studio I even use two way speakers, but I think the ultimate design is a three-way system, with a HQ driver devoted to the delicate mids, and two speakers accompanying it in the lowest and highest frequencies.
Starting off with the woofer; this driver should at least feature a very rigid membrane, which should be also as light as possible, together with a strong magnet and powerful voice coil. The maximum linear excursion together with the membrane surface should house enough potential energy to really shake up a medium-large living room.
A sturdy built chassis was of course also an important pro, as well as a low self resonance frequency and small cabinet volume suitability (Qts-Vas). Searching for this, my eyes were drawn to the Alcone 10 HE, which embodied all of these demands with a reasonable price tag of about 139 Euro / piece (2016).
Also, this driver is featured in the Reference Speaker developed by speaker developer guru Joe D’ Appolito (even approved of by Ralph Mantel, founder of the firm Alcone itself). This speaker features the AC-10 HE driver in the main speakers of The Alcone Reference System and from which I derived later on a lot of valuable data for use in my speakers. This very well designed driver features a rigid aluminium membrane. The only thing with this driver (probably due to the use of the sturdy aluminium) is the fairly extreme (+15 dB) breakup point around 2,5 kHz.
At the time I was starting the calculation of my balls, I was still thinking of using my Dynaudio D-52-AF (discontinued) mid-range driver. I’ve been listening to this mid-range for over 10 years with great pleasure. Smooth and detailed mid with excellent sound-stage. But these mid-range drivers can hardly handle frequencies below 700 Hz. Thinking of an 12 dB/oct. Linkwitz filter, these specs did not match.
There I was, facing a big dilemma, use the fine Dynaudio D-52 mid-ranges and choose another woofer, or, the almost impossible task of finding a mid-range excelling the Dynaudio’s and being able to reproduce lower frequencies until about 300 Hz. Still the last option was the one I chose, being convinced of the superb low reproduction capabilities of the Alcones thanks to Joe’s Reference System…
I searched and searched… and found… The Best Mid-range Driver In The World! …Anyway, I recommend not to be too griping when it comes to choosing your mid-ranges 😉 The mid-range reproduces the frequencies a human ear is utmost sensitive for. In this frequency range the human ear is capable of registering things in the audio far beyond specs and measurements. This is also the base for the fact that I designed a 3-way system (as being ultimate), in spite of all the popular 2 way systems out there, even in spite of the renowned filtering disadvantages. A crossover frequency in the upper mid-range area (1-3 kHz) is a very dangerous thing to do in my opinion. It’s safer and more audiophile in my opinion to design a three way system. One just has to design the filter right.
Anyway, I found the Seas Excel M15CH002 mid-range with impressive specs, accompanied with an impressive price tag of about 180 Euro per unit… (Oh noes! The mid-range driver was more expensive than the woofer!) I Never encountered such a thing before in my life, being a bass freak in my early days, and I spent a lot of money on woofers, but in a strange way it felt as something being very true that my focus over the years changed from bass to mid. As a sound engineer I’ve learned the mids are always the biggest challenge to get just right. The human ear features the highest sensitivity and resolution in this range.
This Seas mid range features a very lightweight (6,75 grams) specially coated membrane, and a revolutionary open Neodymium magnet construction, which gives this driver superb capabilities. Also it has no issues with reproduction reaching as low as 150 Hz. In fact, I really recommend this mid-range as an “affordable” reference driver for all frequencies between 150 Hz – 4 kHz.
The tweeter I had used before in another project I chose to use again, the Vifa XT-25 (discontinued as wel, and now replaced by the Scan-Speak R2604/832000 ‘Discovery’ which still only costs a mere 39 Euro/ piece). It features an impressive flat frequency response, going well beyond 40 kHz and excellent transient response. The concept it is based on what is called a ‘dual ring-radiator’ tweeter. I remember Sonus Faber used tweeters from this series by Vifa in several models a few years ago. It has excellent specs and gives a clear, musically detailed high.
Even though, I have to be honest, in the future I just might try another tweeter. In the end I find the Vifa’s to be a bit too bright for my ears and also the radiation pattern is too bundled for my taste.
I have two (presumably) ELAC tweeters (from a pair of old Musical Fidelity MC-2’s) which sound more relaxed and honest, together with excellent pin-point sound-stage. True candidates and match for the rest of the system! If I try these Elacs, I will update this part of course. I hope to obtain a set of those tweeters second hand, as the MF MC-2’s sound very nice and actually have sentimental value for me, it would break my heart just to take them apart for my ball speakers.
All together I am talking about a fairly modest 350 Euro (2016) spent on the drivers per channel.
Chapter 3 – The cabinet design
Enjoying the highest priority with speaker design is of course always acoustical behavior of the cabinet. Having built enough wooden cabinets, I learned that the sturdiness of the whole cabinet and especially the front panel is very important. Standing wave forms between driver and back panel will also result in coloring of the reproduction, even when damped with the best damping materials.
Killing those two birds with one stone, a bowl-shaped cabinet design was clearly the best choice. The sturdiness of a ball isn’t as much in wall thickness, it’s in it’s shape, a ball just cannot easily deform at all by its form. As mentioned before, the ball shape even features a third unique quality on the outside; excellent sound radiation behavior.
Of course one cannot mount three speakers in one ball, having tweeter and mid-range pointing at different directions, so three balls “stitched” together was the logic outcome. Here you have a couple of early drawings of my ball speakers:
The ‘necks’ between the two smaller balls and the woofer ball are a personal preference, I thought of the neckless design (drawings above) that it was less appealing. Being a bit worried about the distance between woofer and midrange/tweeter, I tilted the design a bit until it became this (see drawings below).
In a later stage I ditched the idea of a ball for the Bass Reflex port, also because the pipe itself was too long to fit in a ball, and I thought it looked even better. The Bass Reflex port moved to the back of the biggest ball. Joe D’ Appolito calculated for the woofer 42 Litres volume with a bass-reflex pipe of 10 cm diameter and 16 cm length to be ideal.
It was possible to reach that with the volumes of the three spheres: Woofer-ball (diameter of 42 cm)=>r=21 -> vol.=38,8 liter + Tweeter-ball (diameter of 15 cm)=>r=7,5 -> vol.=1,8 liter + rings (necks)= 2x (about) 0,5= 1 liters = 41,6 liters. I think the 0,4 liters remaining is easily filled in by lower part of the mid-range ball, which is actually shut a little above it’s neck as I mentioned before.
I actually used a smaller Bass Reflex pipe, which of course affects the length of the pipe, I used a 7 cm diameter drain pipe (inside = 6,5 cm.) and calculated a length of 18 cm. I believe in more effectiveness of a BR port when not being too big.
The material I wanted to use for my balls was still not even clear back then…
I was still thinking of buying 2 large blocks of wood from a big tree and carve the whole design out of it… I estimated it would cost me at least 5 years to finish it, and I probably would have needed twice that (in most cases every job I have does). An obvious alternative was glass-fiber reinforced epoxy. I had no experience whatsoever with it, though. But -nothing ventured nothing gained- I decided to give it a try.
I thought I needed a mold first, on which I could apply the epoxy. I found Styrofoam balls in the creative DIY shop in Utrecht (here is an online shop which also sells them).
Diameters: 1 x 42cm, 1 x 25cm, and 1 x 12 cm. The necks were made of 2 Styrofoam rings of 12 cm (for the mid-range ball) and 2 x 10 cm diameter (for the tweeter ball). I made an mold out of them using normal hobby adhesive and paper tape. (Of course I tried the acoustic behavior of these molds right away, and was not even disappointed! (notice the different woofer hole in the Styrofoam; I was firstly designing the whole thing for another woofer, similar to the AC-10HE, but smaller, and in need of even more than the same equivalent volume.)
Where the drivers would be tightened to the cabinet I placed MDF rings, where I could use T- nuts in, a fixation for drivers which prefer well above normal screws. Don’t forget these screws hold the weight of a whole driver. The positioning of the MDF rings gave me the possibility to position the 3 speakers exactly in phase (one needs to position the voice coils of the 3 speakers in line with each other.) Depth voice-coil of woofer: 3 cm, Mid-range 1,2 cm (to back of the frame front-plate) Having constructed a rough mold for my balls speakers, I painted them excessively with latex:
Be aware: Styrofoam and epoxy are natural enemies. The Styrofoam tends to ‘melt’ even if only exposed to the fumes coming from the hardening epoxy. Important notice: take the word ‘excessive’ serious when applying the latex to the molds. I mean it. Do the inside too, and apply at least 3 dripping thick layers and let them dry for a week. It will save you an enormous amount of work later on. Trust me, I’ve been there… tell you all about it later on.
I found a wholesaler in Amsterdam who could deliver epoxy and glass-fiber mats at fair prices. I bought a mat of 20 m2 fiberglass, 2 cans of 5 kG resin with the the appurtenant hardener (just take advice of the salesman, he knows exactly the ratio of his epoxy). I also bought 5 kG of epoxy putty and a tool to apply it with and spent all together around 200 Euro.
Happily I could use the garage of my parents, and yes, I really recommend finding a space which is roofed in, but has as much ventilation as possible (unless you want to pass out because of the highly suffocating fumes coming from hardening epoxy). Don’t worry about long drying times: I noticed that within an hour even the thickest layer of the epoxy was already as hard as stone.
Chapter 4 – The application of the epoxy
Ok, I have not been totally honest with you: I did not drench my molds completely in latex. I just applied 2 layers of normal thickness latex, and I did not even think of doing the inside. As soon as I mixed the resin with the hardener and applied a first layer of it on the mold, it started to deform a little. I just worked on, actually not even taking notice of the phenomenon, which became even worse. I went on applying the first layer of glass fiber mats on the balls.
Very important tip when using the glass fiber: cut it in small strips because glass fiber mats tend to keep their original form, and do not want to bend. If you cut it in small strips, you can take them apart a little, and it will be easier to form. Of course I didn’t think of this and found it out during application. Can you imagine: cutting with a pair of scissors in bulging pieces of glass-fiber in already hardening resin, with a mold deforming beneath it… I could throw away those scissors and also the gloves (I used thin rubber surgical gloves), they were ripped apart by the glass fibers. In the end I could also throw away the skin of my hands, they blackened out within a few minutes… I recommend using thicker gloves with still some “feel” in them, my advise would be dish-washing gloves.
Ok, I ended up with ball speakers that were hairy and cockled, rock-hard and very remotely the shape of what I desired, and they had no smooth outside whatsoever (see photo). I had to pinch my eyes firmly to see the vague shadow of my original idea. I already applied some putty on the cockled surface, just to cover up the worst dents and bulges. I removed the inner Styrofoam, and found lots of holes and thin spots in the epoxy skin.
Luckily I had the 5 kG epoxy putty which I now needed more than ever. Looking backwards, I think it is better to take a smaller amount of resin with hardener (also less glassfiber mats) and about the same amount of putty I had, 5-10 kG. Apply it evenly on the surface until you get a skin thickness of about 5-7 mm.
The only benefit I had from the bulgy, dented and hairy surface was that I was forced to fill it up with a lot of putty, which turned out later to be just fine for the rigidness of the balls. I used about 7-8 kG of epoxy material per speaker, and I recommend you do the same. It will yield a sturdy and acoustically solid cabinet, and indeed an average skin thickness of about 7 mm. (See picture of the cut out bass reflex hole)
The application of the putty made the speakers a lot more credible, and they looked more and more like they should. I used a wooden pole where I could put the speakers on, while applying the putty. This way the speakers would never rest on a spot which still had to dry.
Take about 1 hour per layer (of about 5 mm.) to dry. In all, you could do the whole job in 1 day, but it took me 2 days of course. After I applied all the putty I had to the cabinets, it was time to sand down the speakers.
It cost me about 2 weeks to get the speakers like this (see picture). Never underestimate the strength of hardened epoxy! It’s almost the same as sanding down hard baked stone. Also take good precautions about the enormous amount of dust generated by the sanding of epoxy.
After the sanding down I applied 2 thick layers of brush filler to cover up the smaller unevenness es in the surface. Of course all of this would have been a lot less work if I latexed the molds better, my balls would have been a lot smoother to start with. In the picture on the left you can see how I had to sand it down (dark spots) to get the balls back in the shape I intended. After sanding the brush filler, it started to look like the balls I had in mind.
Chapter 5 – The cabinet construction
In the MDF rings I drilled holes (using the drivers themselves as a mold ), and inserted the T-nuts in the MDF. (A tip when you do this is to use a ring on the outside of the cabinet and a bolt, and tighten the bolt (with an electric screwdriver) so that the T-nut is pulled into the MDF.)
I applied a little 2 component glue on the outsides of the T-nuts and the MDF so they won’t break loose if one accidentally presses too hard on the bolts when tightening the driver. (Mind you; that can easily turn out to be a nightmare, ending up with ripped up cones, sweat, blood, and tears…
After the T-nuts it was time to separate the mid-range ball volume of the rest of the cabinet volume. In the start of this project I experimented with a pack of 5 CD’s glued and bolted together (also for testing and first measurements taped them to the bottom neck of the mid range ball). I found that a disc a bit larger than a CD (about 15 cm diameter) would be ideal, adding a bit of the mid-range volume to the woofer cabinet.
The fairly large volume of the mid range balls, about 8 liters, (which actually is more than enough for the M15CH002’s, who can suffice with around 2 – 3 liters) made me decide to put the crossover components on this MDF wooden disc as well. Of course the dimensions of the crossover filter took a part in this; it should fit on this separation disc, and some of the components were quite bulgy. I think the total net volume of the midrange ball ended up around 5 liter. The thickness of the MDF disc was 15 mm, which I sawed tapered to fit the ball more easy and to increase the contact surface between disc and ball. Glue the disc with construction foam glue. Please notice: one should first secure the crossover filter to the MDF circle, before glueing it into place……
Chapter 6 – The crossover filter
I first tried some prefab and home made filters I had lying around, and changed some components on them, but it soon turned out that was a dead end: I would have spent hundreds of Euro’s on capacitors and coils to get it right, way more than an active filter would cost (which I even could sell afterwards.) This proved to be a crucial step for the precise tuning of the speakers.
That’s why during testing and designing the Borque Balls I bought a DBX 234XS 3-way active filter in order to exactly determine the crossover frequencies and damping, and adjust them as much as I wanted, especially based on listening sessions. The DBX featured 24 dB/oct. Linkwitz/Riley characteristics.
Luckily back then I was also the proud owner of 2 NAD 2200 PE power amps, and therefore I had 4 channels of nice linear power amp for measuring and tweaking (For stereo listening sessions I also used a Sony amp (with the gain fully open) for the tweeters). After extensive listening and measuring I came to the conclusion that the mid-low separation frequency should be around 300 Hz.
The fact that the separation frequency of the low-mid will cost you a lot more money if you go much lower and because of the big coils and capacitors have to be used, the total volume of the filter would start to give consequences on the mid range ball volume, especially in the low end. At 300 Hz the Alcone woofer stayed nicely warm and still punchy, with no trace of the harshness it has in the upper midrange, using the 12 dB /oct. Linkwitz filtering. On the right you can see my measurement setup with my Spectralab software and ECM-999 mic.
The mid-high separation turned out to be sounding at best around 3,5 kHz. I think it is caused by the slight attenuation of the M15CH002’s around 6 kHz. I think especially this crossover frequency is the most delicate, because in this frequency range a lot of audio information is stuffed (2-5 kHz.) in most popular music. It took me several listening sessions with “clean ears” to determine it for sure, because on the screen of the measurement tools there was almost no visible difference between 2,5 and 5 kHz. It was only audible.
As you can see on the measuring graph here, the system with the DBX filter featured an odd -5 dB downwards spike of 0,1 Oct. on 750 Hz, but I ignored it, because of very nice listening sessions, which proved the filtering of being just right. Maybe it had something to do with the 24dB/oct filtering of the DBX. So in the end I turned up with a 2nd order 12 dB/oct Linkwitz/Riley filter on 300 Hz and 3,5 kHz.
I bought Intertechnik components for the filter, having good experiences with this brand, using the highest affordable quality in the signal route (components serial to the drivers) and “normal” quality for the components parallel to the drivers and signal route.
I chose for the serial capacitors the Audyn Plus series, and for the serial coils the Corobar series, and for the (parallel) capacitors Audyn “normal” and MKT and MKP, and the parallel coils of the “Bell core” type (see photo). R1 and R2 are implemented to dim the mid-range a bit (-3dB) in order to create the most linear frequency response. Here you have the final passive filtering design after some tweaking:
I did not use a circuit board or whatever, because I believe it is not in favor of the signal quality (being only microns thick). I just used speaker cable (6 mm2) for every connection (see picture of the build up filter) and glued and tightened all components with tie-wraps, screws and hot melt to a small multiplex board. Just make sure everything is tightened firmly to prevent the microphony effect as much as possible. Make sure NOT to screw the air coils with a metal screw, but use tie/wraps instead, as the screw will influence the inductivity value of the coil.
The final frequency response even turned out better than the active version:
The frequency response @ 1 Watt / 1 meter distance.
Chapter 7 – The Bass-reflex port
The Bass-reflex port actually is no more than a draining pipe of the right length and diameter. I used a draining pipe of 70 mm (outside diameter) of a length of 18 cm. Make sure you drill a hole just a tiny little bit smaller, so you can tighten the bass-reflex port only by driving it in with a rubber hammer, and securing it with construction foam glue from the inside. Well, that’s how I did it, but as long as you use some sort of glue or kit to seal the pipe in its hole everything is fine.
The placement of the BR pipe should be not to close to the membrane of the woofer, as the Bass reflex effect will be influenced negatively then. I made my BR port at the back of the cabinet, on the lower part of the sphere. See picture below. If you find the BR pipe too noisy (what can happen if the port faces more to the front of the speaker) you can fill the hole with straws to reduce the noise.
Chapter 8 – Finishing
For a really smooth surface I recommend Brush filler as mentioned before. I used several layers of it to smoothen out the small dents and scratches of the surface. After that I used normal lacquer (aerosol). I filled up the cabinet with damping material, mainly glass wool. In the midrange ball I used Dr. Bailey’s longhair wool. Not too tight, not to loose. I cannot describe it differently, maybe you can tell how much to use from the picture on the right. Make sure there is no damping between the membrane of the woofer and the bass reflex pipe. The functionality of the port is decreased and even more irritating: glass wool splinters will whirl around your living room when you turn up the volume. To make my balls stand up right I used 3 threaded rods underneath each cabinet, which I drilled through three spots on the bottom of the balls. That way they can never wobble (so don’t use 4!).
I recommend you the use of fairly large metal rings inside the balls to level the force of the weight over a bigger surface. On the end of the bolts I placed small plastic balls (closet grips from Ikea) to prevent scratching. Also the contact energy loss to the surface below is reduced to a minimum with the use of those. On the back I placed two connection terminals in two small drilled holes (see picture of bass-reflex hole). The type I used is similar to what is offered on this page.
Finally they are looking like this:
As you can see in the picture today I have also placed metal shielding in front of all the drivers for to make them child proof, which I can recommend if you have children or cats 🙂 An easy and cheap way to do so is cut the circles of the desired size of the grills out of a metal front of an old stage speaker for instance. Make holes on the exact spots where the drivers have their bolts aswel and put small metal busses between the driver and the grille (so that the driver and the grille are tightened by the same bolt.)
Allright, that’s all folks! I enjoyed building these speakers even though it was much more work than I expected, mainly because of the number of first time errors which I made, which you hopefully won’t make if you read this carefully. I can assure you, it is really worth the effort.
Don’t hesitate to contact me (on the contact form on the left side of this page) if I can help you if you want to build your own pair of these extraordinary speakers! Of course you can contact me for any other question about my Borque Balls too 🙂
Enjoy music! Bernie
Copyright: Bernie Borkent (2016)