Design-check catalogue
Every check in this catalogue is advisory. None of them disables an export, greys out a button, or stops you doing anything. The panel has no way to express “blocked” — you are the one holding the printer, and you know things the geometry does not.
What each check does is tell you what it measured, what it compared that against, and what tends to go wrong. The rest is your call.
Each row in the Design checks panel carries a Learn more link that lands on the matching section below.
How to read a warning
Section titled “How to read a warning”A warning has four parts.
| Part | What it is |
|---|---|
| Severity | risk, caution or info — how likely you are to regret ignoring it. Nothing more. |
| Title | The problem, with the measured value in it. |
| Detail | The consequence, and the usual fix. |
| Measurement | value · limit unit — what was measured and what it was compared against. |
Severity ranks regret, not permission. A risk on a chip you are deliberately
pushing past its printer’s rating is information, not an error.
Clicking a warning selects the elements it names and centres the view on them, in whichever tab you are in. Switch tabs and the same elements stay highlighted.
Where the thresholds live
Section titled “Where the thresholds live”Three of the numbers below are yours to change, in the Thresholds panel beside the warnings list:
| Setting | Default | What it governs |
|---|---|---|
| Taper ratio | 2× |
abrupt-taper |
| Minimum printable wall | 0.4 mm |
channels-too-close |
| Overhang limit | 45° |
port-overhang |
The rest come from your printer profile (minFeatureSizeMm, buildVolumeMm) or
from the chip itself (wallThickness, minEdgeSpacing), and are edited where
those live.
Manufacturability
Section titled “Manufacturability”These ask whether the printer in your profile can make this part.
channel-below-min-feature
Section titled “channel-below-min-feature”Channel below minimum feature size
Severity risk · Measured against your printer profile’s
minFeatureSizeMm
Triggers when the narrowest point along a channel falls below the smallest feature your printer profile claims to resolve. Measured at the narrowest sample rather than an average: a channel that tapers to 0.05 mm at one end fails there, however generous the rest of it is. For a rectangular or trapezoid channel, height counts as well as width — either can be the limiting dimension.
What goes wrong. Below the resolution limit a channel does not print small, it prints closed. Resin cures across the gap, the void never forms, and you get a solid block with a decorative line in it. Just above the limit you get the worse failure: a channel that opens in some places and not others, so the chip appears fine until you push fluid through it.
How to fix it. Widen the channel, or move to a finer printer profile. When several segments share a dimension marker they are reported as one warning — “3 segments are only 0.05 mm across” — because widening that one marker fixes all of them.
channel-near-boundary
Section titled “channel-near-boundary”Channel or port near the chip boundary
Severity caution, or risk when the element is outside the outline ·
Measured against chipParams.minEdgeSpacing
Triggers when a node or any point along a swept channel comes closer to the chip edge than your minimum edge spacing. Ports are measured by their outer footprint, not their bore: a luer socket is about 4.3 mm across where the bore is 1.6 mm, and it is the boss that runs out of material at the edge.
The check walks the swept channel as well as its endpoints, because a bowed bezier can stray outside the boundary between two perfectly placed nodes.
What goes wrong. A thin rim is fragile. It cracks when you lever the part off the build plate, and it is the first thing to fail when you tighten a fitting. If the clearance is negative the element is simply outside the chip and will be cut off — a port becomes a notch in the side wall, and a channel becomes a groove open to the outside.
How to fix it. Move the element inward, or enlarge the chip. If your resin
and geometry genuinely tolerate a thinner rim, lower minEdgeSpacing — it is
your design rule, not a printer limit.
channels-too-close
Section titled “channels-too-close”Channels too close together
Severity caution · Measured against the larger of
chipParams.wallThickness and the minimum printable wall setting
(0.4 mm by default)
Triggers when the wall between two channels that do not share a node falls below that limit. Segments that meet at a node are skipped — they are supposed to touch, and measuring them would flag every corner.
What goes wrong. This is the classic resin failure. Light bleeds through a thin wall, the resin between two channels cures, and the two channels become one. The chip looks right, passes a visual inspection, and then mixes two streams that were supposed to stay separate — which on a gradient generator or a droplet device means every number coming off it is wrong.
0.4 mm is a conservative floor for a typical desktop mSLA machine and a pigmented resin. A clear resin bleeds further; a fast, well-characterised resin may hold less.
How to fix it. Move one channel, or raise the chip’s wallThickness and
redesign around it. If you have measured your own resin’s real limit, set
Minimum printable wall to that number.
channels-overlap
Section titled “channels-overlap”Channels overlap
Severity risk · Measured against the same wall limit, with a negative
result
Triggers when two channels that do not share a node intersect — the gap between them is less than zero.
What goes wrong. They print as one merged cavity, not two channels. Whatever the design was doing, it is not doing it.
How to fix it. If the crossing is accidental, move one of them. If the two are meant to join, connect them at a node rather than letting the geometry overlap — a real junction is what the flow solver, the drainage planner and the CSG sweep all read.
footprint-exceeds-build-volume
Section titled “footprint-exceeds-build-volume”Chip larger than the build volume
Severity risk · Measured against your printer profile’s
buildVolumeMm
Two separate cases:
- The chip’s footprint does not fit on the plate. The check tries the chip both ways round before complaining, so a 40 × 143 mm chip on a 143 × 89 mm plate is reported only if it fits in neither orientation.
- The chip is taller than the machine’s Z travel. For a two-piece split the measurement is the taller of the two pieces, not the assembled block — you print them separately.
What goes wrong. The slicer will either refuse the file or silently crop it.
How to fix it. Shrink the chip, split it, or change printer profile. A two-piece split is often the answer for the Z case on its own.
port-overhang
Section titled “port-overhang”Port socket overhang
Severity caution · Measured against the overhang limit setting
(45° by default)
Triggers when the wall of a port socket flares further from vertical than the
limit. The angle is computed from the socket’s own dimensions —
atan((wide − narrow) / 2 / depth) — using the taper diameters and length for a
luer or luer-lock port, and the outer diameter, bore and depth for a custom one.
What goes wrong. Past roughly 45° the roof of the socket is unsupported. In mSLA that means it sags into the cavity as it cures, and the seat a fitting needs to press against is no longer flat or round. You feel this as a fitting that will not seal no matter how hard you push it.
How to fix it. Deepen the socket, or reduce the diameter change across it. Both reduce the angle. Printing the chip flat rather than tilted also helps, but does not change the geometry.
thread-support
Section titled “thread-support”Luer-lock thread support
Severity info · No threshold — this fires for every luer-lock port
A luer-lock thread is a helical rib, so its underside is unsupported at every turn no matter how the cone angle works out. This is a note, not a measurement.
What goes wrong. Printed tilted, the thread’s lower flank sags and a real collar binds or cross-threads.
How to fix it. Print the chip flat. Expect to test-fit before committing a chip to a run: OpenChip lays a single-start helix, which is a deliberate approximation of the ISO 80369-7 form and not the certified profile. See Ports and fittings for what that does and does not guarantee.
Flow feasibility
Section titled “Flow feasibility”These ask whether fluid can actually move through what you drew.
dead-end-channel
Section titled “dead-end-channel”Dead-end channel
Severity caution
Triggers on a node that is not a port and has exactly one channel attached — a blind end. Also fires when a whole connected section has only one port, which is the same problem at a larger scale: with a single opening there is no through-path.
What goes wrong. Fluid entering has nowhere to go. On the bench the branch traps air when you prime the chip and never fully wets. On the printer it traps uncured resin that cannot be flushed, because pushing solvent in has nowhere to push the resin out to.
How to fix it. Add a port at the blind end, or delete the branch. If the dead end is deliberate — a pressure tap, a reservoir — add a drainage vent so it can at least be cleared after printing.
disconnected-network
Section titled “disconnected-network”Disconnected network
Severity caution for a chip in several pieces; risk for a section with no
port at all
Triggers when the channel graph has more than one connected component, and separately when a component contains no port.
What goes wrong. Independent networks on one chip are a legitimate design — two experiments on one plate. A break you did not intend is a channel the pump can never reach, and you find out after printing. A section with no port at all can never be filled or drained, and holds resin permanently.
How to fix it. If it is unintentional, connect the pieces at a node. If it is deliberate, give every piece at least two ports.
no-ports
Section titled “no-ports”The chip has no ports
Severity risk
Triggers when the chip has channels and no port nodes at all.
What goes wrong. There is no way to get fluid in or out, and no opening for resin to drain through when printing. The chip is a sealed void.
How to fix it. Place a port at each end of the network. Select a node and set its type to Port in the inspector, or draw with the port tool.
abrupt-taper
Section titled “abrupt-taper”Abrupt taper
Severity info, rising to caution at twice the threshold ·
Measured against the taper ratio setting (2× by default)
Triggers when the channel width between two consecutive dimension markers changes by more than the ratio. Measured along a whole connected chain — the same axis the loft interpolates on — so the ratio reported is the ratio the solid actually has.
The distance is reported alongside, because a 3× change over 20 mm is a gentle taper and the same change over 0.5 mm is a step. The detail line gives the half-angle.
What goes wrong. A sharp expansion sheds vortices and traps bubbles at the
step; a sharp contraction raises pressure and wall shear, which matters if you
are flowing cells or anything else shear-sensitive. Neither is automatically
wrong — a flow-focusing droplet generator needs a sharp contraction, which is
why this defaults to info.
How to fix it. Spread the change over a longer run by moving a dimension marker, or accept it. If your work routinely uses sharper transitions, raise the threshold.
Drainage
Section titled “Drainage”These only run on a chip that is fully enclosed. A two-piece split is open by construction — the channel ceiling sits on the cut plane, so the groove is exposed until the lid goes on — and none of these fire.
enclosed-cavity
Section titled “enclosed-cavity”Enclosed cavity
Severity caution while the chip has no vents; info once vents are placed
or there is nothing left to reach
Triggers when the chip prints as a sealed solid. The detail line reports roughly how much uncured resin sits inside, computed from the swept channel volume.
What goes wrong. Uncured resin left in a sealed channel does not stay liquid. Ambient light and time set it solid, and at that point the channel is gone and the chip is scrap. This is the failure that ruins most first chips.
How to fix it. Add drainage holes, or switch to a two-piece split so the channels are open until you bond the lid on. The app can plan vents for you — see Draining uncured resin from channels.
A monolithic chip is worth catching in one specific accidental case: a split height set outside the block means no cut happens and the chip prints sealed even though the split type says two-piece.
undrained-region
Section titled “undrained-region”Undrained region
Severity caution when the region cannot reach an opening at all; info
otherwise · Reference distance 15 mm of along-channel travel
Triggers per region, with three distinct reasons:
| Reason | What it means |
|---|---|
| Unreachable | No path to any opening. A vent is the only way to clear it. |
| Dead end | A branch with one closed end. Solvent pushed in has nowhere to push resin out to. |
| Far from an opening | The furthest point is a long way along the channel from any port. |
What goes wrong. Resin that has to travel a long way to escape often does not. Surface tension in a sub-millimetre channel is strong, and without a second opening to let air in behind it, the column does not move at all.
How to fix it. Add a vent at or near the reported node. A vent shortens the escape path and gives air somewhere to come in, which is usually the more important half.
What these checks do not do
Section titled “What these checks do not do”Stated plainly, because the gap matters:
- They are geometric rules of thumb against a printer profile you entered. They know nothing about your resin’s actual cure-through depth, your exposure settings, or how your machine is behaving today.
- They do not simulate printing. There is no light transport model here and no cure kinetics.
- They do not check whether the design works — only whether it can be made and whether fluid can move through it. A chip that passes every check can still be the wrong chip.
- A clean panel is not a guarantee. A test print is the real test.
See also: How to read the Design checks panel for working with the panel itself, and The limits of analytical estimates for the equivalent statement about the flow numbers.