Cleanroom Automation Equipment

Motionwell Automation designs and builds cleanroom automation equipment in Singapore, which in practice means one of two machines: one that runs inside a classified room without spoiling it, or one that carries its own classified zone and stands anywhere. The delivered reference for the second is a series of five automated test equipment variants for electronics manufacturers, each with 4 to 6 fan filter units sealed into the enclosure ceiling holding ISO Class 7/8 over the test fixtures while the machine sits in an ordinary factory bay. The first route covers the GMP filling and sealing platform, with product-contact surfaces in SUS316L electropolished to Ra 0.4 um, welds ground flush and passivated to ASME BPE practice and FDA 21 CFR 177.2600 gaskets, and the 12-station rotary syringe assembly machine, built in stainless steel for ISO Class 7/8 operation at a 15 second cycle. Machines are designed, assembled and tested at our Woodlands Link facility, and the company has delivered more than 150 special purpose machines since 2014 under ISO 9001:2015 and bizSAFE Level 3.

Where we stand, said plainly before you read further. We do not build cleanrooms. We do not design facility air handling, install wall and ceiling systems, manufacture fan filter units or HEPA filters, or classify and certify rooms. What we design and build is the machine, and where the machine carries its own clean zone, the enclosure and the airflow inside it. We are not a notified body and do not issue CE certificates.

This page covers the decision that shapes every other one, which is whether the classified environment is the room or the machine, and what each route costs. Then what integrating the fan filter units changes about the enclosure; which way the air has to move; what you substitute for the mechanisms that shed; what your cleaning regime decides about materials and finish; how the machine physically reaches the room; who has to gown to touch it; how its environment is proved; and when this route is wrong. The class ratings themselves, and the full materials, cable and lubricant treatment, are in our cleanroom machine design guide and are not repeated here. If you already have a class, a cleaning regime and a layout, skip ahead and talk to an engineer.

Should the Classified Environment Be the Room or the Machine?

This is the decision, and it is made on how much of the product’s route genuinely needs classified air rather than on the class number. Three arrangements are available.

The machine lives inside a classified room. The facility holds the class, and the machine’s job is to not spoil it. This is what the GMP filling and sealing platform and the rotary syringe assembly machine are built for.

The machine carries its own classified zone. Fan filter units in the machine’s own ceiling push filtered air down over the process, the enclosure is held positive against the factory, and the classified volume exists only where the product is. This is the delivered test equipment series.

Both, split by class. The room holds a lower class over the whole area and the machine holds a tighter one over the product. That suits a process needing controlled space for people and material flow, with one step needing more than the room economically provides.

Two asymmetries decide more of this than the capital comparison does.

The first is what happens to the second machine. A room is a fixed cost paid once, so the marginal cost of putting another machine into it is low, and a plant already running classified space is usually better off using it. An in-machine envelope does not amortise: each machine buys its own filters, gasketed ceiling and pressure monitoring, so the arithmetic favouring the enclosure on machine one is weaker on machine three.

The second is the product’s route. An in-machine zone protects the product where the machine holds it and nowhere else. If the part is exposed between stations, moves to a second machine, waits in a rack or gets inspected off-line, every one of those handovers sits outside the envelope you paid for. Trace the whole path with a pen before choosing, and count the exposures rather than the machines.

This is a concept-stage decision rather than a purchasing one because it changes the machine itself. The in-machine envelope changes the enclosure design, the path the air takes through it, where the cables run and how maintenance access is arranged. Those four are the frame. By the time the frame is welded the envelope is decided, and revisiting it means a new frame, new cable routing and new actuators, which is most of a new machine.

What Changes When the Fan Filter Units Go Into the Machine Itself?

The enclosure stops being guarding and becomes a pressure vessel with a filter on top, and every hole in it is now a design decision.

On the delivered test equipment, 4 to 6 fan filter units are sealed into the enclosure ceiling with continuous gaskets, air leaves through filtered return at the base, and the enclosure is held positive against the factory bay so that leakage runs outward rather than inward. That last property is what makes the arrangement work, and what makes every opening expensive.

Here is the trade: every service panel is a leak path, and every sealed panel is a maintenance problem. You cannot have neither, so the design question is which openings you keep and where they go.

Where an axis must cross that boundary, it crosses through a sealed bellows or a glanded penetration rather than an open slot.

Reducing the number of openings is a layout decision rather than a sealing decision, and the delivered machine shows what that looks like. The test equipment uses a circular pallet transfer that returns every specimen to a single load and unload station, so one collaborative robot serves the whole machine and only one opening has to be made in the enclosure. The station count is then fixed by the loop geometry, which is what that choice cost. The rest of that architecture is on our automated test equipment page.

Two consequences follow the units into the specification and are routinely left out of it. Overall height grows, because they sit above the working level and the room has to clear the machine plus its ceiling. And they draw power and make noise continuously, in a space where both are budgeted.

Because the same panels are safety guards and contamination boundaries at once, the guarding layout and the airflow layout have to be resolved together rather than in sequence, alongside the risk assessment described on our machine safety and compliance page.

Which Way Does the Air Move, and What Does That Do to the Station Order?

Down over the product, then out at the base. The rule that follows reshapes layouts: the product path has to be upstream of anything that sheds, and upstream in a downflow machine means higher, and earlier in the flow.

That single rule sorts the machine vertically. The open product, the nest and the tooling that touches it sit in clean air at the top of the flow. Guides, screws, belts, gearboxes, cable carriers and pneumatic exhaust sit below the product plane or behind the boundary, so what they shed is carried away from the part. A mechanism that must work above the product is the expensive case, and it gets a sealed housing with local extraction rather than an apology.

One ordering rule comes out of the same physics and is worth applying to the station sequence: put the step that exposes the product as early in the air path as the process allows, so it is not working in air that has already passed over four other stations.

Where the process forces the wrong order, the answer is a barrier and local extraction rather than more filtration. Adding fan filter units to overwhelm a source you could have moved is a running cost paid forever to dodge a layout decision that was free at concept.

Where Do the Particles Come From, and What Do You Substitute?

Anything that slides, rolls, meshes, flexes or exhausts is a source. The substitutions matter more than the list, so here they are with their costs.

The reasoning reaches into parts nobody thinks of as particle sources. On the test equipment the specimen nests are machined from PA66 nylon and POM rather than aluminium, partly because surface resistivity above 10^12 ohm keeps the fixture from loading the device under test, and partly because a non-metallic nest does not shed conductive particles into a Class 7/8 envelope. The cost comes in the same breath: engineering plastics creep under sustained clamping load, so those nests are consumables with a defined replacement interval rather than permanent tooling. The tooling side of that argument is on our end of arm tooling page.

What Does Your Cleaning Regime Decide About Materials and Finish?

More than the class number does. A surface specification answers two questions: how is this surface cleaned, and how much residue may survive that cleaning. The ISO class tells you neither.

Two consequences are worth acting on. Product-contact parts that come off without tools get cleaned properly, because cleaning stops queuing behind a technician looking for a spanner. And the chemistry belongs in the user requirement specification as a named compound rather than as the word “cleanable”, one of the lines discussed on our writing an automation URS page. The delivered platform is on the filling machines page and in the GMP filling and sealing case study.

None of it works without the geometry behind it, drawn rather than cleaned in: no blind holes, no lap joints, no exposed bolt heads.

How Does the Machine Get Into the Room?

This is the constraint discovered late, and it is a design input rather than a delivery problem. The machine has to arrive through the doors and around the corners that exist, in the sections it ships in, and those sections are decided in the same weeks as everything else here.

Scale is why it bites. The filling and sealing platform occupies roughly 3 m by 1 m by 2 m, and that footprint was an achievement in itself: the filling mechanism is stacked above the container transport level and the pneumatic and electrical cabinets are side-mounted to keep the plan area down. A machine deliberately made compact is still a two-metre object crossing a classified boundary, and fan filter units on the roof make it taller.

Every split line has a price on both sides of the wall. A joint in the frame is re-aligned, re-levelled and re-commissioned on site, and on a machine carrying its own envelope it is also a gasketed joint to be re-made and proved inside the room. Assembling in a classified space brings people, tools and packaging with it, so the crating stops at the boundary and the machine is cleaned before it crosses. That belongs in the programme rather than in the goodwill of the installation week.

None of this is unique to a locally built machine. Rigging equipment into a unit with a tight door and a lift with a weight limit costs what it costs wherever the machine came from, one of the rows in the comparison on our local versus overseas machine builder page. What is avoidable is learning the door dimension at the end of the build rather than the start, and a photograph of the route at enquiry stage prevents it.

Who Has to Gown to Touch the Machine?

Whoever the room’s protocol says, and that answer carries a cost per event landing in maintenance rather than capital.

The arithmetic is simple and usually left out of a comparison. Every intervention inside a classified space carries the gowning routine before it and the re-clean after it, and so do the tools and the spare part, which come in through a pass-through or an airlock. A five-minute fault does not cost five minutes. It costs five minutes plus the boundary, twice.

Three design consequences follow, and all of them are cheap at concept.

Split the machine at the boundary. Everything touched often belongs outside it: the drives, the air preparation, the pneumatic manifold where the schematic allows. The delivered test machines run SMC and Festo valve banks with regulated air preparation at each bank, and where the boundary permits that hardware and its adjustments sit on the non-clean side.

Make the consumables the shortest visits. Filters, nests, seals and grease points are the recurring reasons anyone enters. Ask for the consumable list, the interval and the access route together, because an interval that reads reasonably on paper reads differently once each occurrence is a gowned entry.

Write the recovery step before it is needed. Any panel opening into the clean zone gets a documented re-clean, and on a qualified machine a documented return to service. Without that procedure the machine leaves its qualified state the first time it is serviced, and nobody notices until an audit.

How Is the Machine’s Own Environment Proved, and Who Owns the Number?

By measuring it while the machine works, and by agreeing who owns the number before anyone measures.

At rest, almost any machine passes. The meaningful measurement is taken with the axes moving at production speed and the cylinders firing, which is why particle verification belongs in performance qualification rather than at handover. A machine at rest, one running empty and one running product with people present are three different results, and the specification has to say which the acceptance criterion refers to.

Ownership splits cleanly if it is written down. The room’s classification belongs to whoever built and maintains the room. The environment inside our enclosure is ours, measured against the factory air outside it. The interface is the pressure differential and the openings, and the interface is what gets argued about when a number disappoints, so name it in the contract.

One useful consequence follows for a machine carrying its own zone: the envelope can be proved at factory acceptance testing at Woodlands Link, because the enclosure brings its own air with it. A machine destined for a room cannot be fully proved before it is in the room, so its acceptance splits across factory and site testing, and the sampling points and criteria have to be agreed early enough to decide where the sampling ports and access panels go.

What the evidence pack contains depends on the regime the machine is bought under. The rotary syringe assembly machine shipped with an IQ/OQ/PQ package prepared under GAMP 5, including executed test records and a validation summary report, and the GMP filling platform ships the same way, with material certificates and surface finish records for product-contact parts. The mechanics of that pack are on our computer system validation page. Where the resulting records are retained as GMP electronic records, 21 CFR Part 11 governs how they are stored, who may alter them and what audit trail follows, which is a documentation project running alongside the build.

When Is a Machine-Carried Clean Zone the Wrong Answer?

We build these enclosures, so read this as the argument against our own preference.

The product is exposed outside the enclosure. Covered above under the second asymmetry, and repeated here because it is the failure that shows up after installation rather than at design review.

The process is bigger than an enclosure can sensibly be. An enclosure a person has to walk into is a small room with worse airflow and no gowning discipline. At that size, build the room.

The product has to be reached constantly. Every opening cycle breaks the boundary, and a machine whose door is open half the shift is not holding anything.

The class is being specified upward for comfort. If the product needs Class 8, buying Class 5 sealing produces a machine that is harder to maintain and no better yielding. Specify against the product’s requirement and the cleaning regime, not against the tightest number in the room.

The machine already exists. Making a built machine cleanroom-compatible rarely works, because the frame material and finish, the fastener style, the cable routing and the actuator choice are frame-level decisions, and changing them is most of a new machine. Where a retrofit is genuinely the right route, the honest scope is on our retrofit versus replace page.

One more exclusion, said directly. Where a standard machine with a cleanroom option covers your process at a lower price than anything we would build, the useful answer is to say so, and that comparison is on our custom machine versus standard equipment page.

What Has to Be Settled Before the Frame Is Welded?

Run these in order. The first hard answer usually settles the route.

1. Where on the product’s path is it actually exposed? Mark the exposure points on a route drawing. One points at an enclosure. Five across three machines points at a room.
2. What class does the product require, and on what evidence? A requirement from a customer specification, a regulator or a yield study is a number. A preference is not.
3. What is the cleaning regime, named as a chemistry and a frequency? This sets the surfaces, the seals and the labels more directly than the class does.
4. What has to happen above the product? Overhead tooling and overhead axes are the expensive case, and knowing early is what keeps them affordable.
5. Which interventions recur, and can they happen outside the boundary? Every one moved outside is a gowned entry you never pay for.
6. What is the route in, in millimetres? The narrowest door, the lift, the corner, the slab, and the height under the services.
7. Who owns which particle number, and when is each one measured? Room, enclosure, factory acceptance, site acceptance, performance qualification.

Lead time runs 16 to 24 weeks from concept approval to factory acceptance testing on a standard build, and 24 to 32 weeks where cleanroom compatibility or full GMP validation applies. Design, fabrication, assembly and testing happen at Woodlands Link with an in-house design team of eight, which is why a Singapore buyer attends the factory acceptance test instead of flying to it. That matters here because the acceptance test is where a particle count under motion either holds or names the axis that does not. The delivered work sits in context on our electronics and semiconductor automation page, our medical device automation page and our pharmaceutical packaging page, and the enclosure is documented in the cleanroom ATE case study.

Frequently Asked Questions

Does automating a cleanroom process reduce the number of people who have to gown up?

Usually per shift, and not to zero. Automated handling removes the operators who were moving product between steps, which is one of the reasons this equipment gets bought. What it does not remove is entry: somebody still loads consumables, clears faults, changes filters and replaces wear parts, and those entries are less predictable than a production shift. The honest version is that you trade a known number of gowned production hours for a smaller and less predictable number of gowned maintenance hours. Design the machine so the frequent interventions happen from outside the classified boundary and the trade goes your way.

Who maintains the fan filter units, and how often?

Your maintenance team, on an interval driven by loading rather than by the calendar. A filter loads as it works, the pressure drop across it rises, and the fan compensates until it cannot. That is why differential pressure monitoring belongs on the enclosure as an instrument somebody reads, not as a commissioning check: it is the signal that a filter change is due before the class drifts. Ask any supplier for the filter part number, the differential pressure that triggers a change, whether the filters come out from above the ceiling panel or from inside the enclosure, and what the re-clean step is afterwards.

Should the control cabinet sit inside the cleanroom or outside it?

Outside, wherever the cable lengths allow, because the cabinet is where the frequent interventions happen. Drive faults, resets, parameter changes and program work are the interventions that recur, and every one of them costs a gowning cycle if the cabinet is inside. What moving it out costs is real and should be priced rather than assumed: longer shielded motor cables, a wall penetration that has to be sealed and stay sealed, and a machine split across a boundary at installation. Settle it while the layout is being drawn, because it decides where the cable route runs.