Suivez-nous sur X
|
|
|
0,
A,
B,
C,
D,
E,
F,
G,
H,
I,
J,
K,
L,
M,
N,
O,
P,
Q,
R,
S,
T,
U,
V,
W,
X,
Y,
Z,
ALL
|
|
0,
A,
B,
C,
D,
E,
F,
G,
H,
I,
J,
K,
L,
M,
N,
O,
P,
Q,
R,
S,
T,
U,
V,
W,
X,
Y,
Z
|
|
0,
A,
B,
C,
D,
E,
F,
G,
H,
I,
J,
K,
L,
M,
N,
O,
P,
Q,
R,
S,
T,
U,
V,
W,
X,
Y,
Z
|
|
A propos d'Obligement
|
|
David Brunet
|
|
|
|
DIY: Boing Ball Case for the AmigaOne
(Article written by Tony Wyatt and from excerpt Club Amiga Monthly - February 2004)
|
|
This is the story of the Boing Ball case I built for my AmigaOne XE. 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 AmigaOS 4,
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 MicroA1 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!) AmigaOne 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. ;-)
|
