BRICKAID

Inspiration

To place a brick in an interlocking-brick model, you need both hands. One hand pinches and positions a small part, while the other holds the model steady so the brick can be pushed into place. For most builders the process is invisible. For people with one working hand, arthritis, tremor, or hemiparesis, it is three separate barriers stacked on top of each other.

We looked at the existing solutions and they all changed the brick rather than the builder. Patents for adaptive bricks (US 6,050,873; US 5,281,185; US 8,961,259) all make the element larger and easier to grab. The manufacturer's accessibility work has focused on vision — braille elements, audio instructions — and the tools available to buy are separators, hammers, pliers and tweezers, every one of which still requires a pinch.

Builders describe the barrier in simple terms. A stroke survivor on a fan forum wrote that he used to hold the model in one hand and add parts with the other, and can no longer do so. A newspaper profiled a builder with more than 60,000 bricks whose right side was weakened by a stroke, who continues to build as part of his rehabilitation. A writer on Medium described stopping because of hand pain.

No one had built a tool for the hand.

What it does

BRICKAID grips a brick by its own stud, so you never pinch it, and seats it with a force-limited plunger, so you never press.

1. Grip. The head is an inverted stud socket. A split collet slides over a stud and holds it through light interference — the same principle that holds two bricks together.
2. Position. The tool carries the element. Two printed clips lock the baseplate down, so no hand is needed to steady the model.
3. Press. Bear down on the palm pad with the heel of your hand or your forearm. A disc-spring limiter caps the force.
4. Release. A stripper blade pushes the element off the head as the plunger reaches the bottom of its stroke.

Nine swappable heads cover studs, tiles, slopes, rounds, pin-hole beams and smooth curved pieces. A clip-on module with a linear actuator and a momentary button drives the same stroke for users who cannot press at all.

Engineering

The whole design reduces to one ordering that has to hold:

$$F_\text{retention} < F_\text{seating} < F_\text{limiter} < F_\text{ceiling}$$

If retention exceeded the seating force, the brick would lift back off the model with the tool. It doesn't, by a factor of six.

Retention treats each collet finger as a cantilever:

$$P = \frac{3EI\delta}{L^3}, \qquad F_\text{retention} = \mu \sum P$$

where $E = 2000\ \mathrm{N/mm^2}$ for PETG, four fingers and $\mu = 0.35$ against ABS.

That formula produced the most useful result of the build. The first version specified a 5.0 mm collet bore over a 4.8 mm stud. Substitute those values: the interference $\delta$ is negative, retention is zero, and the brick falls straight off, because the fingers never spread around the stud. The socket has to be smaller than the stud. Changing bore_d from 5.00 to 4.60 took retention from 0 N to 8.7 N. One number made the difference between working and not, and it survived several reviews because 5.0 and 4.8 look similar.

How we built it

The assembly is a native parametric Fusion 360 model generated by a Python script: 47 parts, 45 solids, 50 occurrences, 9 joints and 7 rigid groups.

Building it as code meant a check harness could run alongside it. Fifteen rules were re-evaluated on every rebuild: plunger travel against spring solid height, return preload, foot travel, three feet coplanar, press axis inside the support triangle, grip clearance above the foot tips, leg deploy sweep 0–90°, collar seated on a continuous ring, envelope, mass, no brand name, and zero unexpected interference across four poses — rest, plunger at −14, feet compressed, leg deployed.

All fifteen pass. The tool's footprint is 273 × 66 mm, working height 104 mm, mass 696 g of which 504 g is printed PETG.

Hestus Sketch Helper was used on the head-face sketches. Each face is a constrained 2D sketch — equal circles on 8.0 mm pitch, symmetric margins, equal relief slots — which is exactly what a constraint assistant is good at. We drew the collets freehand, accepted the constraints it proposed, then dimensioned with parameter names rather than numbers. Every suggestion is logged, including the ones we turned down: it repeatedly proposed "All Dimensions," which we declined because typed dimensions would have broken the link to the parameter table.

Honest scope on that: boss_od, bore_d and bore_depth are live in the document. slot_w and pitch are read from the parameter table at build time and written into the sketch as coordinates. Partially parametric, not fully.

Challenges

The tool couldn't stand up. The check file reported foot-tip-to-pad at 104 mm and overall height at 159 mm. That 55 mm gap was the grip hanging below the feet — the three-point stance was real and useless, because the grip hit the table first. Fixing it meant solving grip angle and length together, since swinging the grip back tilts the trim plane through it and eats internal length by $10.3\tan\alpha$. It resolved at 74° and 116 mm.

Thin walls that looked fine on screen. An audit found a 0.4 mm web between two lithium cells and a 0.8 mm wall on the collar spigot that the heads hang from. Both would have printed as wisps and snapped. Bringing every printed wall to ≥ 1.6 mm cascaded through the design: the plunger grew from Ø18 to Ø20, which moved the rack, seat ring, return spring and pawl.

Automation failed quietly, three different ways. A rebuild flag did not take effect and the script built a second copy of the tool into the same document. A JOIN on ribs that touched nothing did not error; it silently produced floating bodies. And the worst one: eight of nine head STEP files exported at ~4 KB with zero solids. The obvious reading was that the faces had never been built. They had. The exporter hides the spare heads to shoot the assembly and never unhides them, and Fusion writes a hidden occurrence as an empty file. The one correct export was the mounted head, the only one not in the spare list. The model was right and the deliverable was wrong.

We caught it by counting MANIFOLD_SOLID_BREP entries in the exported STEP files rather than trusting the build report. Two lines fixed it.

What we learned

• Check the sign before the magnitude. A collet bore larger than the stud is not a weak grip, it is no grip at all.
• A pressed slope skids. A vertical press on a bare 45° face produces an equal sideways force. The slope heads had to capture studs on the flat strip; the angled pad only steadies.
• A sideways grip can't take a downward load. The pin head needed a rim landing on the top face of the element.
• Verify the artifact, not the process. Every check passed while eight of the exported files were empty.
• Automated review finds real defects and invents fake ones. Multi-reviewer passes surfaced genuine problems, including a 0.21 mm feather wall and an edit that had never been applied, alongside confident findings that turned out to be measurement errors in a rotated coordinate frame. Every finding had to be checked by hand.

Honest limits

• Nothing has been physically tested. Geometry is closed and interference-checked; mechanical performance is unvalidated. The 6.4 N clutch force and 8.7 N retention are hand calculations, not model outputs.
• Springs are modelled as rigid coils at installed length, so the palm plate passes through them at full travel in the STEP. Expected, and listed in the check file.
• Rear-foot reaction at the 110 N limiter load is 24 N against a 20 N target. Reaching 20 N needs the rear foot at 96 mm, not 80 mm.
• The clips fix the plate to the clip; the clip is held to the table by friction or the user's forearm.
• Scope: bricks, plates, tiles, slopes, rounds and pin-hole beams. Not minifigures, panels or flex elements.

What's next

Print it and measure it. Three numbers matter: the collet retention, the real seating force on a real plate, and whether the limiter caps where the disc stack says it does. Then a session with occupational therapists to find out whether the palm-press geometry works for the hands it is meant for.

The files are printable and free to any OT, school or library that wants them.

Compatible with standard interlocking bricks. Not affiliated with, sponsored by, or endorsed by any brand.

Built With

• autodesk-fusion
• autodesk-fusion-360
• cad
• claude
• fdm
• fusion-api
• hestus
• iso-ts-15066
• python