Adam K Dean

Notes on software, systems, engineering, and things

3.1 A, Sled Tower

Five boards lie flat on trays that slide in from the front, like a tiny server rack. It's 142 × 163 × 277 mm, the second tallest of the four, and the only one where a board comes out without touching the others.

X-ray: the trays, the power hub, the blowers and the power supply

The plinth

The plinth is 77 mm tall and has two levels. Inside it measures 132 × 150 mm, with 5 mm walls.

The x-ray from behind: the inlet above the mains compartment, the plenum beside it

The tower

A 3 mm deck closes the plinth and is the tower's floor. A 72 × 22 mm opening at the back left lets the plenum's air up into the duct, and an 8 mm hole lets the 5 V leads up from the compartment.

Each board sits in its own channel: the tray under it and the next tray over it, 29.5 mm apart. Take away the board and its parts and that's about 2,540 mm² of cross-section, the most of the four concepts. At 2 L/s shared five ways, the air moves at about 0.16 m/s. That's slow, which is the trade for the space.

Perspex

A front door, 197 × 142 mm, held on four 6 × 3 mm magnets. Each side has a 152 × 80 mm window in a 3 mm rebate on the outside of the wall, so the boards show from the side too.

Printing

Part Prints Size (mm) Orientation
Plinth 1 142 × 160 × 77 Upright, open top
Deck 1 142 × 160 × 3 Flat
Tower 1 142 × 160 × 197 Roof on the bed
Tray 5 131 × 119 × 8 Flat, standoffs up

The tower is one print: two sides, the back, the duct wall and the roof. Upside down, the roof is on the bed and the open bottom is up, so it needs no support inside. The rails become 2 mm ledges, which print fine. The slots in the duct wall are the weak spot. Upside down, each slot's top edge is a bridge, and I've kept them to 28 mm so they'll bridge cleanly. That needs a test print before I trust it.

Balance

The estimated mass is 1.74 kg, with the centre of gravity 119 mm up. The nearest edge is 69.3 mm away in plan.

htan⁡10°=119×0.176=21 mm<69.3 mmh \tan 10° = 119 \times 0.176 = 21 \text{ mm} < 69.3 \text{ mm}

So it passes the tilt test easily, and doesn't tip until 30°. The push at the top of the roof (276.5 mm) that tips it is:

F=1.74×9.81×0.06930.2765=4.3 NF = \frac{1.74 \times 9.81 \times 0.0693}{0.2765} = 4.3 \text{ N}

A light push with one finger is 5 to 10 N, so a deliberate shove would still have it over. It's heavy at the bottom, which helps: the power supply is a fifth of the total mass and sits 20 mm off the desk.

Good and bad