Custom Animatronic Design Process: 6 Client Collaboration Stages

Ask ten people what a custom animatronic build is like and you'll get ten versions of "a lot of back and forth." The sculpting is the fun part. The part that decides whether the thing actually works is the six conversations that happen before anyone touches foam.

We've been building these since 2015, and the process has settled into a fairly repeatable shape: six stages, each with its own approval gate, each answering one question. Below is how they run in our shop, what we need from you at each one, and the numbers that actually drive decisions — joint counts, skin thickness, cable runs, lead time.

Design and prototype together usually land in the six-to-eight-week range. Production after sign-off runs roughly 30–45 days depending on size, finish and control complexity. Freight and on-site installation sit outside both numbers.

Stage 1 — The First Concept Conversation

Almost nobody shows up with a drawing. What shows up is a phone: museum photos, a game screenshot, three frames from a documentary, and a rough sense of where the piece is going to stand. That's plenty to start with. The images give us the silhouette; the conversation fills in everything the silhouette doesn't say.

What we actually need from you

Four things, in roughly this order of importance: where it's going, who's going to stand next to it, how it should feel, and which parts need to move.

Where it goes decides material and frame long before it decides anything else. An atrium with a glass roof and a seaside plaza are two different engineering problems. Who stands next to it decides height and reach — a piece meant for a photo queue needs a different pose and a different clearance than one meant to be viewed from behind a rope. How it should feel decides sculpture style: documentary-real, cartoon-friendly, theatrical. And the movement list is usually shorter than people expect once we get into it.

Bring photos of what you like, and one or two of what you don't. "Not like this" is faster than ten adjectives.

The requests that get reworked

A head that turns 180° on a six-metre neck. A jaw that snaps shut in under a quarter of a second. A full-body undulation on a piece that has to pass through a 1.2 m door. None of these are impossible in principle. They're expensive in torque, frame weight and safety clearance, and the cost usually lands well past the visual payoff.

We'd rather say that in week one than after the frame is welded. Same goes for the reverse: sometimes a request we'd normally push back on turns out to be the whole point of the project, and then we design around it properly instead of talking you out of it.

What comes back to you

A one-page concept brief: a proportion sketch, a motion list with the joints we're proposing, and the assumptions we're working from. You sign off on the brief, not on a finished design. It's cheap to change at this point, and it becomes the document the rest of the project is measured against.

Stage 2 — Size and Movement Requirements

Size is the number that quietly drives everything else: joint count, motor selection, frame section, crate dimensions, and whether the thing fits through the door it has to fit through.

Treat the table as a starting point, not a rulebook. A 4 m piece built for a museum gets more joints than a 4 m piece built for a photo spot, because one is meant to be watched up close and the other is meant to be photographed from three metres away.

The tail is where people get surprised

Torque at the base of a tail doesn't scale with length alone. It scales with length and arc together. A 2 m tail swinging through 120° loads the mount considerably harder than a 1.5 m tail doing the same swing, and the mount is usually where that shows up first — a bit of play after a season, a rattle at the end of the stroke.

That's not a motor problem. It's a moment arm that got specified for a shorter tail. Worth half an hour of arithmetic in stage 2.

Weight, headroom and weather

Every extra 100 kg changes the frame section and the motor selection, and it changes them together — you can't stiffen one without revisiting the other. For headroom, a 5 m piece standing at full neck extension wants somewhere around 6.5 m of clear space. Measure the actual ceiling, not the drawing; plant rooms and sprinkler runs have a way of eating 40 cm.

Outdoors, sustained wind above roughly 30 km/h usually means a sturdier frame and a reduced motion envelope. We generally do both, because the frame is a one-time cost and the motion envelope is a software setting. Cold matters too: below about 5°C grease thickens and everything slows down. If it's a winter installation, mention it early and we'll spec the lubricant accordingly.

Then we put it on a drawing

We issue a dimensioned drawing with tolerances: ±2 cm on limb lengths is workable, gear alignment wants closer to ±0.5 mm. Two rounds of review, then it's frozen. Frozen doesn't mean locked forever — it means a change later arrives as a change order with a date attached, which is a much calmer conversation than discovering the discrepancy on the shop floor.

Stage 3 — Material Selection

Material choices show up later as maintenance bills, so this meeting is less "what looks nice" and more "what does year three look like."

Frame

Aluminium keeps a small piece light and easy to reposition; past roughly 300 kg, steel starts to make more sense. Steel is heavier and slower to weld, but it carries a load without arguing about it. On bigger builds we usually mix the two — steel where the load path runs, aluminium where the structure is only holding skin.

Skin: the part everyone asks about

Most of our skins are platinum-cure silicone, 5–7 mm thick, sprayed and painted in layers. That thickness is a balance point. Thin enough to wrinkle naturally at the joints, thick enough to survive a kid leaning on it and a handler grabbing it during a show.

Outdoors, with a wash and a re-spray every couple of years, five years is a realistic service life. Urethane costs less per square metre and lives shorter — a reasonable trade for an indoor piece that nobody touches.

One cost note that surprises people: on a skin, the silicone is where the money sits. The foam underneath is comparatively cheap. When a quote reads higher than expected, it's almost always skin area and paint hours, not the sculpt.

Climate, and the small hardware

Coastal sites want stainless fasteners and a marine-grade coating on exposed metal, because salt reaches standard hardware first and quietly. Desert and high-UV sites want UV-stabilised pigments; ordinary pigments lose saturation noticeably within a couple of seasons. Indoor museum pieces sometimes go latex-free where visitor allergies are a concern, accepting a shorter skin life in exchange.

Foam density is the other quiet choice. Around 20 kg/m³ reads soft and expressive at a jaw or an eyelid; 45 kg/m³ holds shape better at a shoulder. Most builds use both, in different places.

Samples, not photographs

We courier physical swatches — a skin square, a foam block, a painted panel — and let you stretch and bend them. A colour on a monitor isn't a colour match; a swatch in your hand, under your own lighting, is.

Stage 4 — Control System Preferences

The controller decides whether the movement reads as an animal or as a machine with a motor in it. It's the least visual decision in the whole project and the one visitors feel most directly.

How many axes, honestly

Twelve to eighteen servos covers a convincing small-to-mid piece: head, neck, jaw, eyes, tail base, four limbs. Past roughly 24 axes the wiring and the timing get awkward, and we move to a PLC-based system with distributed I/O. That isn't an upsell — it's what keeps a 30-axis show in sync instead of half a beat behind itself.

The numbers behind "lifelike"

A neck rotating at 15–25° per second reads as natural. Push past that and the audience starts reading "animatronic" in the wrong way. Servo positioning around 0.1° is fine for limbs; eyes and jaw benefit from something closer to 0.05°, because small parts magnify small errors. Input-to-motion delay under about 50 ms feels real; past 100 ms people notice without being able to say why.

Power is the unglamorous part. Each servo pulls roughly 0.5–3 A, so a 24-axis piece wants 50–60 A of peak headroom. Cable runs beyond about 5 m start to sag in voltage, which is why we drop power nodes every 3 m or so and hold the supply within ±5%.

The interface your staff will actually use

Operators don't read manuals at nine in the morning with a queue at the gate. So controls get grouped by body part on a single screen, show sequences go onto a timeline, and position memory means a power cut doesn't turn into a 45-minute recalibration. Emergency stop within half a second, tested in front of your team rather than described on a spec sheet.

Outdoor cabinets get an IP65 housing, a −20 to 50°C rating, and a door that opens without a ladder. Small detail. You'll appreciate it in February.

Stage 5 — Prototype Testing

By this point there's a partial build — usually head, neck and tail on a test frame, with the control system wired in. It looks rough. It's also the cheapest place in the entire project to find problems.

Skin and paint get abused on purpose

Silicone samples go through a few hundred stretch cycles at the fold lines, because that's where creases start. Painted panels sit in a humidity chamber cycling between 30% and 90% RH. Metal coupons go through salt spray. None of it is glamorous, and it's the reason a piece still looks right in year three instead of going chalky in year one.

Your team drives it before it's pretty

We hand over the controller while the cabinet is still bare metal. First-time operators usually have a basic sequence running in about fifteen minutes; where they hesitate is where the interface needs work. Better to find that now than after the paint is on.

Stage 6 — Final Adjustments Before Production

Last gate. Everything gets signed, measured and photographed, because after this point the drawings go to the workshop and changes get slow.

The calibration checklist

Five items, run in order, every time:

1. Limit switches trip within about 1 mm of the intended position.
2. Three or more moving parts stay in sync — a 50 ms stagger is visible to a viewer.
3. Peak current stays under roughly 90% of rated load through the full sequence.
4. Components hold below about 70°C across 30 minutes of continuous operation.
5. Emergency stop halts everything in under 0.3 s without a current spike.

Paint, seams and the things visitors notice

Colour is matched to the approved sample at ΔE below 2.0 — past that point the eye picks it up even when it can't name the problem. Skin seams stay under about 1 mm at the joint; wider gaps collect dust and read as a hard line under gallery lighting. Texture is checked across panels at a 45° angle, because that's the angle a visitor's eye actually catches a patch.

Paperwork that saves an argument later

Production drawings match the prototype dimension for dimension. The spare parts list covers the wear items — servos, bearings, skin patches — rather than just the big assemblies. Crate design protects the jaw and the tail tip, which are what arrive damaged when nobody thinks about it. Pass/fail criteria are written down so a QC check isn't a matter of opinion. And the operator manual uses photographs instead of paragraphs.

What you walk away with

An approval checklist, material certificates, a duty cycle agreed in writing (most park pieces are specced at 8 hours a day), and a maintenance schedule tied to measured thresholds rather than a calendar. Motor resistance up 20% means bearings need attention — "service in June" means nothing if the piece runs all year.

Frequently Asked Questions

How long does a custom animatronic take from first call to delivery?

Design and prototype usually run six to eight weeks together. Production after sign-off runs about 30–45 days depending on size, finish and control complexity. Freight and on-site installation sit outside that window, and for overseas projects the shipping leg is often the longest single item on the calendar.

How much does a custom animatronic cost?

Four things drive it: length, joint count, skin area and control complexity. We quote after stage 2, when there's a drawing with real dimensions. A range quoted before that is a guess wearing a number, and it tends to cause more trouble than it saves.

Can you build a specific dinosaur from a film or a game?

We can build to any design you hold the rights to, and we can build an anatomically referenced or stylised version of any real species. Copying a protected character design isn't something we take on — that's a rights question, not a manufacturing one.

Can it be repaired on site?

Most of it, yes. Skins patch, servos swap, and the control cabinet is laid out to be serviced with hand tools. Structural welding stays a workshop job, which is one reason we send a spare parts list that covers the wear items rather than just the major assemblies.

Do you handle shipping and installation?

We can arrange both, and they're quoted as separate lines from the product. Freight, duties, certification and installation labour vary too much by destination to fold into a unit price — better that you see them clearly and can compare them against your own local options.

Do we get the drawings afterwards?

Yes: dimensioned drawings, the motion list and the I/O map. Those are the documents a maintenance team actually needs in year four, when the people who commissioned the piece have moved on.

YESDINO — Zigong Dino Culture Art Co., Ltd. Building animatronic dinosaurs and custom creature effects since 2015.