Whether an appearance design can actually ship comes down to something plain: can the factory make that line, does it still look good once made, and does it feel right in the hand. If any one of the three was not settled before the model was cut, what follows is tooling changes, rework and redrawing. We did a handheld inspection device once. The side had a beautiful gradient edge in the render and the client approved it on the spot. After the first production batch, that edge turned into a dirty band that flickered as the light moved, because the edge radius was 0.3 mm and mold polishing simply cannot reach the finish the render promised. The appearance was redrawn and the tool was repaired twice.
The short version: four boundaries decide whether an appearance design ships. Surfaces need a manufacturable transition radius — at least R0.5 mm generally, and R1 or more on visible cosmetic surfaces. Parting lines, gates and draft directions have to be marked at the render stage. CMF has to be derived backwards from real process tolerances, not tuned on a screen. And ergonomics has to be verified against real hands. Miss any one of these on the drawing and you will pay for it in tooling.
Surface boundary: the transition you drew, the mold cannot polish
The most common failure in appearance design is the surface transition in the render. The renderer will not tell you the radius is impossible; it will simply make it look lovely.
The self-check line I give the team: visible cosmetic transitions no smaller than R1, structural transitions no smaller than R0.5. Below that it is not that the part cannot be made — it is that it will not look good. Polishing becomes dramatically harder, tool marks show up in the texture right at that transition, and melt flow through that section becomes unstable. The inspection device died on 0.3 mm.
There is a subtler problem too: curvature that changes too fast. Some forms chase a "blade edge" feel by going from nearly flat to a tiny fillet within a dozen millimetres. It reads as crisp in the render, but on the real part you get a visible highlight band that shifts and flashes as the light moves. Our current approach is to keep the rate of curvature change below what the eye notices, replacing abrupt transitions with continuous ones. A blade-edge look should come from colour breaks and tangent chamfers, not from forcing a curvature jump.
Three checks at the render stage: mark the minimum radius on every visible transition and justify anything below R1; check the curvature comb and either soften the jump or break it into a deliberate separate part; and put the model under a single strong light and rotate it, simulating how it will read under spotlights.
CMF boundary: the colour on screen is not the colour of the product
We have been burned by this more than once. One batch of housings, same colour code, came out half yellow-shifted and half cool. The client assumed we had changed the colour. The cause was different machines, different material lots, and a small difference in texture depth — the diffuse reflection came out different.
Three things have to be nailed down early for CMF to be manufacturable: physical colour chips, texture specified by depth and demoulding direction, and gloss given as a process range rather than a single number. There is no conversion formula between an RGB value on screen and a plastic chip on the table. You can only match by making boards.
| Element | Risk if confirmed only on screen | How to lock it early |
|---|---|---|
| Colour | Batch-to-batch drift, hue shift between materials | Make a physical chip, bind it to the material code |
| Texture | Too shallow looks cheap, too deep traps dust | Specify texture number and depth, note demoulding direction |
| Gloss | Glossy surfaces amplify defects on the real part | Give a gloss range by process, leave room for tuning |
| Colour break | Break line drifts, visual weight shifts | Put the line on the parting surface or a deliberate separate part |
The last row is the one most often missed. I have seen a scheme that used an extremely fine colour-break line to visually cut the body into upper and lower halves; it looked sharp in the render. But that line sat on the joint between two parts, and once assembly tolerances stacked up the line was no longer straight — the whole sense of precision collapsed. A colour-break line either lives on a single part (paint or two-shot moulding), or you accept that it is an assembly gap and give it enough visual tolerance.
Ergonomics boundary: a good feel starts with measuring hands
Ergonomics is the first thing sacrificed in appearance design, because it does not show in a render.
For handheld products we now measure hands before drawing forms. Adult hand breadth is roughly 75-90 mm, and a grip diameter of 30-40 mm feels natural to most people; above 45 mm the palm has to open noticeably wider and long holds become uncomfortable. The numbers do not need to be precise, but the range has to frame the form before the surfaces are drawn — not after you discover the thing cannot be held.
One more thing often ignored: visual weight versus actual weight. Widen the lower half and narrow the upper half and the product reads as more stable; lift the visual centre of gravity and it reads as lighter. That is a modelling-language problem, but it only works if you know where the grip support sits. On handheld devices we mark the thumb web position and the four-finger contact area on the side view, and keep buttons, indicator lights and decorative lines out of that zone.
The same applies to buttons. If the gap between two adjacent buttons is smaller than a fingertip, users will press the wrong one without even looking. None of this is visible in a render, and all of it matters on the real object.
The check for the ergonomics boundary: make a 1:1 foam or 3D-printed white model and have a colleague who is not on the project close their eyes, grip it, then open their eyes and press the buttons. If the grip feels wrong or the buttons have to be hunted for, the form goes back for changes. One or two days here saves an enormous amount after tooling.
Process boundary: separate parts versus visual integrity
The designer wants one continuous volume; the mold demands the object be split. The balance has to be allocated while drawing.
Our priority: keep the primary face the user looks at intact, and push part joints to the sides, the bottom and hidden corners. If a parting line must be visible, make it part of the design language with a defined width and consistent direction. The worst outcome is a parting line in an arbitrary position in the middle of a curved surface — neither symmetric nor linear, so it reads as a mistake.
Then there is draft. Every vertical wall in the form will need a draft angle in reality. The closer a surface is to vertical, the more the draft angle distorts the visual proportion. Mark the draft direction on a section view at the render stage. On forms with a large flat face and a narrow side wall, adding draft visually thickens the wall and changes the proportions.
A failure from a few years ago: a small countertop appliance whose form was a single forward-leaning plane, which I thought looked very clean. After tooling we discovered that the junction between the front face and the sides had become a transition band with an angle, because it had to demould. Under frontal light that band cast a shadow and cut the originally continuous plane in two. The process was not wrong — I had simply never drawn that band in the render. We changed the transition so the shadow fell inside the corner, and the form read as continuous again.
The five-item checklist before you send drawings
- Every visible transition has a minimum radius marked (R1 or more on cosmetic surfaces).
- Parting lines avoid the primary face.
- Draft direction is marked on section views.
- Colour chips, texture and gloss have been matched against physical boards.
- A 1:1 white model has been gripped and pressed.
Do these five at drawing stage and post-launch appearance rework stays low.
These boundaries sound like restrictions on form. After enough projects you find the opposite: once the boundaries are clear, drawing goes faster, because you know what not to touch and you can push hard inside the space that remains.
Written by the Hezi Industrial Design team in Dongguan. We do product appearance design, structural design and CMF as one service, with 320+ products in mass production.
Top comments (0)