The Boing Ball Case for the AmigaOne
by Tony Wyatt
This is the story of the Boing Ball case I built for my
AmigaOne. It started out as a joke, but became a serious project that
is now reality.
The sphere is 450 mm diameter and stands on a cylindrical
pedestal 250 mm diameter and 100 mm tall. The top half of the sphere
lifts off for access to the internals. Several of the "checks" are
clear unpainted, so that you can see the AmigaOne inside.
Some of the "cool" features are:
- Space for three HDDs (two are fitted).
- Space for two FDDs (one fitted and one Flash Card-IDE adaptor).
- Space for one CD-ROM/RW/DVD drive (currently a CD-RW fitted, but will be upgraded).
- Top hemisphere lifts off for access to internal components.
- Clear panels allow most of the internal components to be seen from the outside.
- The sphere is smooth, no screws or projections.
- The CD-ROM tray has a cover that forms part of the sphere, so disappears when closed.
- CD-ROM Eject button on pedestal.
- All Power/Reset switches and indicator LEDs on pedestal - blue power LED, of course ;-).
- All connectors extended from the motherboard or PCI cards to similar connectors on the pedestal.
- Room on pedestal for expansion (more connectors).
- Switched mains power outlet.
|
|
|
How it Started
It started on the A1G3Dev Mailing List back in March 2003
- there was a thread going about cases for AmigaOnes, and I jokingly
suggested a spherical case with red and white checks on the outside.
Well, the more I thought about it, the more I thought it might be fun
to build one that way. I was also thinking that if the Amiga Downunder User Group
were likely to be doing demonstrations of the AmigaOne and AOS4, the
sight of the machine in a large Boing Ball would be eye-catching, to
say the least. So I set about sketching the basic essentials. The
design had to support the AmigaOne ATX-sized board (this was before the
MicroAmiga was released), along with a CD, space for several 3.5"
drives, and a floppy drive that had to be accessible, but it was OK if
I had to remove the top cover to get to it. Here's my original sketch:
I reckoned that a 450 mm sphere (18" for backward people)
would fit it all nicely. I could possibly squeeze it into a 400 mm
sphere, but it would be very tight and the saving wasn't worth the
effort.
So I started looking around for something I could use. My
first thought was one of those polycarbonate spheres that you sometimes
see as outside lighting. They are popular in touristy areas like
harbour foreshores. In Sydney we have them all around the Sydney
Harbour foreshore, and around the Opera House. They look great, but
expensive. I finally found the suppliers (Thorn) and their catalogue
details. The Thorn sphere
is 400 or 500 mm diameter, with a large hole in the bottom. It would
have cost nearly A$400 by the time I added tax, and I would have to cut
it into two halves to get the computer inside! Too big, too expensive.
I grumbled to myself and started asking local suppliers. Eventually I
was directed to B & M Plastics
in Sydney. A phone call confirmed that yes, they frequently make this
sort of thing, and 450 mm is a "standard size" in their repertoire.
They would make two 450 mm diameter hemispheres for me, trimmed to fit
together and make a sphere.
The way they make them is interesting. How do you mould a
large plastic dome? Well, you have two pieces of wood - one is a ring,
with the internal diameter that you want. The other is just a flat
surface with a short pipe sticking through. You start with a thick
sheet of plastic (acrylic plastic for instance), clamp it between the
ring and the bottom plate, and heat the whole assembly in an oven. When
it's hot, you pull it out, blow air into the pipe, and the soft plastic
expands upwards in a nice dome! When it has expanded to the right
height, they shut off the air supply and allow the assembly to cool.
Preliminary Design
My two domes/hemispheres cost me A$300, which I thought
was reasonable for something made to order. They were ready for pickup
in less than a week from my order, which was even better. At that time,
I was still anticipating delivery of my A1 board at any time, but in
fact, the A1 was not delivered until the Sydney AOS4 Roadshow in
October, so there was no prospect of my having the Boing Case ready for
the Roadshow.
Meanwhile, I assembled the hemispheres and a wooden salad
bowl as a pedestal for this mockup. The CD shows you how big the sphere
is.
Once I got hold of the A1 board I was able to start modelling. I soon made a couple of important decisions:
- The centre of gravity of the assembly was too high for safety. To
reduce the chance of tipping, I decreased the height of the pedestal as
much as I could, and moved the power supply from inside the sphere to
the base of the pedestal. I also moved the heavy disk drives from above
the CD-ROM to below it.
- My original plan had been to mount everything on the inside
of the sphere, adding bracing as required. It was obvious, however,
that the sphere was not even rigid enough to support its own shape, let
alone the internals of a computer. So I turned the structure inside out
and made it more along these lines:
Now the assembly in the middle is entirely
self-contained. It is fastened to the pedestal which bears the weight.
The sphere is attached to the pedestal underneath the electronics
assembly and carries no weight at all. This was a much better design.
Originally I had planned on making holes in the sphere for air inlet
and/or outlet, which was undesirable for aethestic reasons. Now, with
the power supply in the pedestal, I could use the power supply's own
fan to circulate air in and out of the pedestal and through the sphere.
I built up a model backplane and webs out of 6 mm
compressed fibre board (we call it MDF for Medium Density Fibre). This
is cheap and allowed me to make mistakes and changes before I bit the
bullet and made it all out of expensive acrylic plastic. The pedestal I
made out of thick corrugated cardboard. You can see the pedestal in the
picture above.
I had to make a sanity test of the airflow design. Would
the path of the airflow through the case be so tortuous that the fan
could not cope? Would there be pockets of stagnant air that got too
hot? I assembled the mockup with the power supply in the cardboard
pedestal and one hemisphere on top. I had not at this stage cut a hole
in the "bottom" hemisphere, and I wasn't going to until the last
minute! I taped newspaper around the bottom of the assembly to make it
airtight and ran the computer as realistically as I could.
The power supply had an automatic temperature sensing
circuit to control the fan speed. While I had known about this before,
I was not sure what the effect would be. My mockup showed that the fan
ran slowly because the hotter air stayed up the top of the sphere,
losing heat to the outside through the plastic rather than by
circulation. The air flow was greatly improved when I shorted out the
control circuit and made the fan run at full speed. I also found out
that the slightest restriction would make a large effect on the air
flow. These little 80 mm fans are about as gutless as they appear to
be.
Building the Internal Assembly
The next stage was to buy the plastic for the internal
assembly and cut it up. It sounds simple, but it took weeks of marking
out, checking, sawing, filing, drilling and checking again. While doing
all the marking out and drilling, I left the protective film on the
plastic sheets to reduce scratching, so these shots show the assembly
still in its "cocoon".
I decided to assemble the plastic backplane and webs not
with metal brackets but with edge-mounted screws. Because drilling into
the edge of the 6 mm plastic was dicey to start with, and a thread in
the plastic was not going to be strong, I had to choose the thread and
screws carefully. I chose the coarsest thread screws that I could
obtain - 6/32 UNC. These are much more coarse than M4, which I used
nearly everywhere else. The screws are 6/32 x 25 mm countersunk, in
stainless steel. I like stainless steel screws, they are expensive and
look cool. Using stainless steel screws indoors is like always having a
towel ready ;-)
The motherboard is spaced off the backplane by 3 mm
plastic spacers, glued on to the backplane. The motherboard is fastened
by M4 screws tapped into the backplane. These are too big for comfort,
but M3 is too small. I would have been happier with 6/32 again, but I
had already drilled and tapped the holes, so it was too late. There are
no nuts anywhere to be dropped or lost. All screw threads are tapped
into the base material.
Looking at the picture above, the motherboard is mounted
on the far side of the vertical backplane. The drives are mounted
between the webs on the side closest to us. The 5.25" CD-ROM drive is
mounted between the outer webs, the 3.5" HDDs and FDD between the
left-hand and middle webs. There is a large slot in the middle web for
the CD-ROM to reach right through to the right-hand web.
The electronics assembly is attached to the pedestal by
three posts, made of 12x12 mm aluminium angle. These posts are secured
to the wall of the pedestal to take the weight of the assembly. Two of
the three posts also secure the power supply to the inside of the
pedestal.
Building the Pedestal
The original concept was to use the pedestal as support
for connectors only. With the removal of the power supply from the
sphere to the pedestal, suddenly things became a little more cramped.
Also the means of fixing the bottom hemisphere to the pedestal became
important. Although the weight of the electronic assembly and pedestal
is entirely removed from the hemisphere, anyone picking up the complete
computer from a table will put his/her arms around the bottom
hemisphere and lift that. Therefore the attachment of the bottom
hemisphere to the pedestal has to be able to bear the entire weight of
the computer.
I
tried to find cylindrical acrylic tubing for the pedestal, but it
turned out to be far too expensive in the 2 metre length that I would
have to buy. Therefore I bought a metre of 250 mm PVC stormwater pipe.
This cost me about A$30 from the plumbers' supply shop. The pipe has a
6 mm wall and is a medium grey colour. The ink writing on the pipe
became a problem later.
|
|
|
|
|
I used 12 screws to secure the hemisphere to the pedestal. These are 6/32 x 25 mm, into the edge
of the 6 mm thick pedestal wall.
The connectors are mounted on metal plates screwed to the pedestal, or directly into the pedestal
wall. No connectors are mounted on the sphere, so every motherboard connector has to be
extended with a short cable down to the pedestal.
|
The Power Supply
The 250 W power supply is, like all computer power
supplies, a standard size and shape. It sucks air from the inside of
the computer case and blows it out the back. I wanted to reverse this
flow, so that it sucked in outside air and blew it up over the
motherboard, over the top of the backplane and down over the disk
drives, and finally out through vents in the pedestal. The power
dissipated in a 250 W supply running an AmigaOne is negligible - the
motherboard only takes about 25 W at 800 MHz - so the air coming out of
the power supply is quite cool enough to blow over the motherboard
(only about 1 deg C more than the inlet air). To make the air go where
I wanted it to, I made the backplane come right down to the full width
of the pedestal, with the power supply up hard against it. An aluminium
cowl ducts the air from the inlet port to the fan. I had planned to
include filter material inside this cowl, but I found it interfered too
much with the air flow.
A partition blocks the other side of the power supply so
that the two halves of the pedestal are sealed from each other. The
"exhaust" side has an 80 mm fan-sized hole and wire guard for the
second fan on the exhaust side. The holes in the power supply cover are
all blocked with duct tape except at the top where the air has to
exhaust over the motherboard. Note also that I have had to split up the
cables into two looms, since the motherboard is on one side of the
backplane and the drives are on the other. The power to the drives is
terminated in two 4-way sockets, so that the loom can be connected to
the drives without the power supply. The mains input has been
reterminated, a relay added internally to provide a switched output,
and a 4-way connector is used for mains in and out. There is a screw
terminal for the Earth connection.
The Hemispheres
The two hemispheres are identical (save for manufacturing
tolerances). After studying pictures of the "approved" Boing Ball, I
settled on a 12x12 chequered pattern. That is, 12 checks around the
"equator", with six bands of "latitude". If you think of the latitude
as going from the South pole up through six bands to the North pole,
then down the other side again, you can see why there are 12 checks
around the equator and 12 around the poles. The numbers have to be the
same to make the equatorial rectangles into squares.
I marked out the lines by tracing the circle of each hemisphere onto a
large sheet of plywood and dividing it up into 12. I then placed the
hemisphere over the circle and measured equal arcs from each side to
meet at the "pole". My latitude bands are equal angles of latitude, not
the equal vertical projection that you commonly see when you wrap a
chequered pattern around a sphere in a paint program.
To make the proper tilt (and to make it easier to mount
the CD-ROM drive), the axis of the sphere is tilted 15 degrees from the
vertical. There is a 250 mm hole in the base of the sphere so that it
is open to the pedestal. I marked out a 240 mm hole and cut it very
carefully with a hand-held hacksaw blade.
As mentioned above, I used 12 screws to secure the
hemisphere to the pedestal. To spread the load of each screw head, I
glued small plastic wedges to the inside of the hemisphere, to provide
a flat, level surface for each screw head. The top edge of the pedestal
is tapered to match the slope of the hemisphere where they meet.
The top hemisphere was originally intended to be screwed
to the backplane. However, I decided to have no screws or fasteners at
all on the sphere, so I will use a "bayonet" fitting - the top
hemisphere is lowered onto the bottom one, then turned slightly to lock
it into position.
First Assembly
Having built up the electronics assembly and constructed
all the fastening devices (brackets, screws, etc), I had to put it all
together to make sure it went together. I also wanted to see
how the power supply fan coped with all the internal cables in the way,
a hot day, etc, etc. So I assembled the whole machine, complete except
that the hot air was exhausting out the base of the pedestal, rather
than the side. I had to make a fresh air duct to prevent the air from
being sucked back into the inlet. I was also suspicious that the CPU
fan was recirculating some air, although I had never suspected it
before. In previous tests, the motherboard had always been lying
horizontally on the table, now it was sitting vertically with a
definite airflow over the board. Certainly the CPU seemed to be running
hotter, but the ambient temperature was 30 degrees instead of the
cooler temperatures of earlier tests. The temperature rise between the
inlet air and the exhaust air was about 7 degrees, the CPU indicated
(by its "uncalibrated" sensor) temperatures of 50-54 degrees.
Although the temperature rise was quite low and the CPU
temperature well within reasonable limits, I still was not entirely
happy with the amount of air being blown through the machine, so I
decided to look for a more powerful fan or some other means of
increasing the amount of air flow. I bought a new fan with ball
bearings and "high flow rate". It certainly moved a lot of air, but the
noise was excruciating. I could never use a machine that made such an
awful whine. That fan stayed in its box, and I decided to try a second
quiet fan in the base of the pedestal as an exhaust fan. It increases
the flow rate by helping the supply fan.
Painting
The sphere has to be painted in red and white checks. You also have to be able to see
through the top hemisphere to see the (Ooh! Ahh!) A1 XE inside. So some checks have to
remain unpainted, both red and white ones. I bought some spray paint and masked off
the hemispheres.
I was hoping to have some of the squares with just an outline of the square's colour. In other
words, it would look like a dirty window with a clear spot in the middle where someone has
cleaned it. I would have preferred to spray the outside of the square with a "vignetting" effect.
However, I would only get one attempt at it, and it had to work first time, so I used a rectangular
"frame" around each window. Only windows that abut each other have this frame.
|
|
|
|
|
I was originally going to spray the red squares first, then mask over them and spray the white
squares. However, after trying a "test spray" on the pedestal, I found that the red paint did not
cover the ink markings on the grey PVC pipe. Also, the red was just too dark - I wanted a
bright red.
|
Maybe if the red was sprayed over a white surface first?
To test the idea, I sprayed a piece of scrap plastic white, then later
sprayed it red over the white. The red turned out much brighter with a
white backing.
So I decided that the sphere and the pedestal would be painted in white
first, then in red checks where appropriate. Finally the red checks
would be taped over and the white areas re-sprayed so that every part
had two full coats of paint.
So, I sanded all the red paint off the pedestal (that
took a couple of hours) and started again. The second attempt was much
better, the red was good and bright.
Next I tore all the masking tape off the sphere (except
where the checks were intended to be clear transparent) and sprayed it
all a thin white coat. I left that for two days to harden, then taped
it over again and sprayed the red checks. That took a day of several
minor coats, leaving it to dry for an hour or two after each.
Next I had to carefully remove the tape from the white
parts, tape over the red parts, and repeat. Each time it took about
three hours to tape up the whole sphere. I also made a small window
that slides out with the tray of the CD-ROM drive.
The whole painting job took about a week!
Final Assembly
At last the time came for the final assembly (well, when
I say final, maybe I mean the "first" final). I'm glad I didn't try to
squash it all into a 400 mm sphere. It takes some squeezing to get all
the cables into place, without getting in the way of the airflow.
The final assembly can be seen in the pictures. I
expected it to occupy a large footprint on the desk, but because of the
height of the pedestal (100 mm), things like keyboard, mouse/trackball
and cables can slide underneath the sphere so that it takes up less
room than the tower of my A4000.
Future Work
- I originally intended to light up the inside with cold
cathode lamps or some such. However, the paint on the sphere is not a
uniform thickness, so it looks "blotchy" when illuminated from the
inside. I still have some work to do on that. I also thought of putting
LEDs in the edges of the acrylic partitions, so that the backplane and
webs are illuminated. I haven't spent much time on that yet.
- I bought a CD-RW drive for the machine about nine months
ago. Since then, DVD-RW drives have become cheaper and now the CD-RW
drive is old hat. So that will also have to be upgraded.
- I have not extended all the cables from the PCI cards - only those that I need. In time, some more will have to be brought out.
- I need a dust filter in the intake, to keep the inside of the computer clean. We can't have a dirty AmigaOne, now, can we?
- I
also need a big TFT screen like 17", 1280x1024 at least. A 16x9
(1920x1080) widescreen version would be even better. But that's not
really part of the Boing Ball project ;-)
|