A print can look healthy inside while its outer walls resemble a fence with missing boards. That contradiction is useful: it tells you the printer can move filament, but something is starving the visible perimeter. Do not raise global flow yet. In about 15 minutes, you can identify whether the real culprit is outer-wall flow, coasting, wiping, pressure control, seam behavior, or line-width math. This guide gives you a controlled tuning order, practical test values, and clear stop signs so you can fix the shell without turning the infill into an overstuffed plastic mattress.
Read the Gap Pattern Before Changing Flow
“Outer walls have gaps” describes several different failures. Treating all of them as low flow is how a clean print becomes a glossy, dimensionally swollen paperweight.
First, inspect where each gap begins and ends. Use side lighting, a magnifier, or your phone camera. Rotate the part slowly. A continuous groove running around most of the model points toward wall flow, line width, or extrusion consistency. A short opening beside the Z seam points toward wipe, coast, retraction, or pressure advance.
Continuous gaps along the wall
If the wall lines remain separated for several centimeters, check the outer-wall flow ratio, wall line width, wall count, filament diameter entry, and nozzle condition. This is the pattern most likely to improve with a modest outer-wall flow increase.
I once increased retraction for a continuous wall groove because the defect looked “seam-ish” under workshop lighting. It was not. The new setting added stringing, delayed restarts, and supplied a free lesson in changing the wrong variable.
Gaps only at the seam
A gap that appears where each perimeter starts is usually a restart problem. Coasting may be ending extrusion too early. Wiping may be using the last pressurized filament before the new line begins. Retraction may be too long, too fast, or poorly matched to the extruder.
Gaps after corners or acceleration changes
If a wall looks full during straight travel but thins after a corner, inspect pressure advance or linear advance. An aggressive value can remove too much pressure as the toolhead decelerates, then leave the next segment hungry.
Gaps only on tiny features
Small loops amplify every start-and-stop setting. Minimum layer time, cooling changes, short-segment speed, perimeter generation, and restart pressure become more important than average flow.
- Long grooves suggest flow or line-width mismatch.
- Seam gaps suggest coast, wipe, retraction, or restart pressure.
- Corner gaps suggest pressure control or acceleration effects.
Apply in 60 seconds: Mark the seam with a pencil and decide whether the defect follows it.
Who This Is For, and Who Needs a Different Fix
This tuning path is for FDM printers that produce acceptable infill, reasonably solid top surfaces, and predictable extrusion during longer internal moves, yet show missing material or separated lines on visible outer walls.
Diagnostic Eligibility Checklist
This guide is a strong match when:
- Infill lines are continuous and do not click or skip.
- The defect is worse on outer walls than inner walls.
- Gaps repeat near seams, corners, or short perimeter segments.
- First-layer quality is already acceptable.
- The problem appeared after changing a slicer profile, firmware feature, or retraction setting.
Start elsewhere when:
- Every feature is under-extruded, including infill and top layers.
- The extruder clicks during simple straight lines.
- Filament diameter varies visibly or the spool is badly tangled.
- Layer height changes randomly across the print.
- The nozzle temperature is unstable or reported incorrectly.
If the defect affects diagonal walls more than axis-aligned walls, read the related guide on under-extrusion that appears only on diagonal walls. That pattern may involve motion timing, segment handling, or directional mechanics rather than a simple perimeter multiplier.
This article also assumes your model has printable wall thickness. If a CAD wall is thinner than the slicer can represent, flow adjustment cannot create missing geometry without changing dimensions. Plastic is obliging, but it does not negotiate with impossible math.
Safe Tuning Boundaries Around a Hot Printer
Flow testing is low-risk when changes stay inside the slicer. Hardware inspection is different. A nozzle may be above 400°F, heater wiring can remain energized, and loose clothing has a strange talent for meeting moving belts.
Keep hands away from the hot end during extrusion tests. Use pliers or tweezers only when appropriate, and never steady a moving toolhead by hand. Power the printer off and unplug it before opening electronics compartments, tightening heater terminals, or inspecting damaged wiring.
Ventilate the printing area, especially during high-temperature printing or when using materials that produce stronger fumes. Follow the printer and filament manufacturer’s operating guidance. Do not defeat thermal runaway protection, bypass damaged thermistors, or keep running a heater that reports unstable temperatures.
A safe tuning envelope
- Change flow in steps of 0.5% to 1%.
- Change temperature in steps of 5°F to 10°F, roughly 3°C to 5°C.
- Change pressure-control values by about 10% to 15% at a time.
- Change one start-and-stop feature per test.
- Stop immediately if the extruder grinds filament or the heater behaves unpredictably.
I have watched a two-minute calibration turn into an hour because a brass nozzle was grabbed with bare fingers “just for a second.” The printer survived. The thumb submitted a strongly worded complaint.
Reset to a Clean Baseline in Five Minutes
Before adding outer-wall flow, remove the settings that can secretly subtract material. The goal is not to build a perfect profile yet. It is to create a profile whose behavior is easy to interpret.
Use this temporary baseline
- Set coasting to zero or disable it.
- Disable extra wipe options that continue travel after extrusion stops.
- Use the filament profile’s normal global flow or extrusion multiplier.
- Set outer-wall flow to 100% of the calibrated global value.
- Use two or three walls.
- Place the seam in one fixed location.
- Use inner-wall-first ordering for the diagnostic print.
- Reduce outer-wall speed to a moderate value, often 25 to 40 mm/s.
Keep retraction enabled if the printer normally needs it, but return to a known conservative value. For many direct-drive systems, that may be roughly 0.2 to 0.8 mm. Bowden setups often use more, commonly around 2 to 5 mm. Those ranges are starting references, not commandments carved into a spool holder.
Print a small two-wall box or a 30 mm square tower. Avoid decorative models. A dragon has too many opinions. You need straight walls, a fixed seam, and enough height to compare several changes.
Visual Guide: Follow the Missing Plastic
Check outer-wall flow, line width, nozzle condition, and wall overlap.
Disable coast and reduce wipe or retraction before increasing flow.
Reduce pressure advance and compare acceleration behavior.
Inspect filament moisture, spool drag, partial clogs, and extruder grip.
- Disable coasting first.
- Fix the seam position.
- Use a simple two-wall test part.
Apply in 60 seconds: Duplicate your profile and name it “Outer Wall Baseline” before changing anything.
Tune Outer-Wall Flow Without Bloated Dimensions
Once coast and aggressive wipe behavior are removed, inspect the baseline wall. If the gap remains along most of the perimeter, increase outer-wall flow in small increments.
Start at 100%. Print a short section at 100.5%, 101%, and 102%. Many slicers allow modifier regions or per-height changes. If yours does not, print three small boxes. Resist the temptation to jump directly to 108%. That may close the groove, but it can also round corners, erase fine lettering, and make mating parts behave like strangers at an awkward dinner.
What a successful flow correction looks like
The outer lines should touch without forming a ridge between them. Corners should remain defined. The measured wall should stay close to the intended wall thickness, allowing for the printer’s normal dimensional tolerance.
| Observed Result | Likely Meaning | Next Change | Stop Sign |
|---|---|---|---|
| Continuous narrow groove | Slight perimeter underfill | Add 0.5% to 1% outer-wall flow | More than about 104% is required |
| Gap only near seam | Restart starvation | Leave flow alone; tune wipe, coast, or retraction | Rest of wall becomes glossy or swollen |
| Ridge between wall lines | Too much flow or excessive overlap | Reduce flow by 0.5% or inspect line width | Corners begin bulging |
| Random thin spots | Intermittent feed or filament issue | Check spool drag, moisture, clogging, and gears | Pattern moves between test prints |
If 101% or 102% closes a continuous gap while dimensions remain sound, keep the correction. If you need 105% or more, look for a deeper cause: incorrect filament diameter, worn drive gears, a partial clog, overly narrow wall line width, or a global flow calibration that was tuned against the wrong feature.
For a separate check of the printer’s actual melt capacity, use the guide to volumetric flow limit testing. Infill can look fine at one speed while the visible perimeter fails under a different combination of temperature, acceleration, and segment length.
Do not calibrate flow from one wall measurement alone
A single-wall cube can be useful, but line-width measurements include bead shape, squish, surface texture, and caliper technique. Judge three things together: wall contact, surface quality, and finished dimensions.
On one machine, my calipers insisted the wall was too thick while the print still showed a visible valley. The bead was oval and the measurement method was lying politely. A two-wall test and a cut cross-section told the better story.
Micro-Tune Wipe and Coast Without Starving the Seam
Coasting stops extrusion before the end of a line and relies on residual nozzle pressure to finish the path. Wiping moves the nozzle along a printed path while pressure decays, often during or around retraction. Both can reduce blobs and seam bumps. Both can also remove the exact filament your outer wall needs.
Start with coasting disabled
If disabling coast closes the gap, you have found a primary cause. Do not immediately restore the old value. Begin at zero and add the smallest practical increment only if seam bulges return.
For slicers that express coasting as volume, trial steps around 0.02 to 0.05 mm³ are suitably cautious for a 0.4 mm nozzle. Values around 0.05 to 0.15 mm³ may be useful on some profiles, but the correct amount depends on nozzle size, material viscosity, Bowden length, speed, and pressure-control settings.
Treat wipe distance as a finishing tool
Try 0.2 to 0.5 mm of wipe rather than several millimeters. A long wipe can drag the remaining pressure away from the seam and smear the wall behind the nozzle. If your slicer offers “wipe while retracting” as a percentage, reduce it rather than combining a strong wipe with strong coasting.
One PETG profile I inherited used coasting, a long wipe, high retraction, and pressure advance at the same time. Each setting was defensible alone. Together they formed a tiny committee dedicated to removing filament before every seam.
Use this adjustment order
- Disable coast and wipe.
- Confirm whether the seam gap disappears.
- Correct retraction and pressure control.
- Add a small wipe only if the seam remains raised.
- Add minimal coast only when wipe cannot control the final bulge.
Show me the nerdy details
The molten polymer inside the nozzle behaves as a compressible, delayed-flow system even though the plastic itself is not meaningfully compressed like a gas. Filament elasticity, Bowden-tube flex, gear compliance, melt viscosity, and pressure inside the hot end create stored energy. Coasting assumes that stored pressure will extrude the remaining line after commanded extrusion stops. Pressure advance attempts to model the same delay dynamically. Using strong values for both can double-correct pressure decay. The result is often a clean internal line and a starved perimeter restart because the outer wall contains more visible starts, decelerations, and seam transitions.
- Disable both before raising wall flow.
- Restore wipe in tenths of a millimeter.
- Use coast only when a measurable end-of-line bulge remains.
Apply in 60 seconds: Save your current values, set coast to zero, and print only 10 mm of test height.
Separate Seam, Pressure, and Retraction Problems
A seam gap is rarely solved by adding flow to the entire perimeter. The missing material exists for a few millimeters, so the correction should also target those few millimeters.
Check seam alignment first
Place the seam at the rear or on a fixed corner. Random seams scatter evidence across the model. A fixed seam creates a vertical test record, rather like tree rings for slicer decisions.
If the defect follows the seam exactly, compare three prints: normal settings, retraction disabled for external perimeter transitions, and pressure control reduced by 10% to 15%. You do not need to keep the winning test unchanged. You are identifying which system owns the defect.
Reduce retraction before adding restart compensation
Long retraction can pull molten material too far from the nozzle tip. Fast retraction can grind soft filament or create inconsistent recovery. Reduce distance in small steps, commonly 0.1 mm for direct drive or 0.25 to 0.5 mm for Bowden systems.
Some slicers offer extra prime amount, restart distance, or retraction compensation. Use these carefully. A positive restart value can fill a seam notch, but excessive compensation creates a recurring zit. The related guide on blobs that appear exactly at the layer change explains the opposite side of the same pressure problem.
Pressure advance can be too high
Pressure advance or linear advance reduces extrusion during deceleration and increases it during acceleration. When tuned well, corners sharpen and line starts become consistent. When set too aggressively, the printer may underfeed the end or beginning of short perimeter segments.
Reduce the current value by roughly 10% and compare the seam. Do not copy someone else’s number from a forum. Extruder type, filament stiffness, nozzle geometry, temperature, and Bowden compliance all change the useful value.
Short Story: The Perfect Infill That Framed a Bad Seam
A small electronics enclosure once arrived on my bench with tidy grid infill and an outer seam that opened like a zipper. The owner had already increased global flow, temperature, and wall overlap. The enclosure was now too large for its lid, yet the seam gap remained. We restored global flow, disabled coasting, fixed the seam at the rear, and printed three 12-minute towers. The second tower, with retraction shortened by 0.2 mm and pressure advance reduced slightly, closed the gap. A tiny wipe then softened the remaining bump. The useful lesson was not the final numbers. It was the location of the defect. The printer never lacked filament across the whole wall. It lacked pressure during one transition. By correcting the transition instead of feeding every line more plastic, the enclosure recovered both its surface and its dimensions.
Balance Speed, Temperature, and Line Width
Outer-wall flow is a ratio, but the nozzle must still melt and place the material at the requested rate. A wall can receive the correct calculated amount and still appear thin if the polymer is too cool, the line is too narrow, or motion changes are too abrupt.
Use moderate outer-wall speed during diagnosis
Set outer walls to roughly 25 to 40 mm/s for a 0.4 mm nozzle and ordinary layer heights. High-flow printers may run much faster, but a conservative diagnostic speed removes one variable. Once the wall is stable, raise speed in 5 to 10 mm/s steps.
Temperature affects pressure recovery
A slightly warmer melt flows more readily and may recover after retraction with less delay. Increase temperature by about 5°C only when the filament manufacturer’s range allows it. Watch for stringing, loss of detail, weak overhangs, or a surface that changes from matte to glossy.
If your print shows alternating sheen rather than true missing material, see the guide on walls that turn matte and glossy in bands. Speed and temperature transitions can imitate flow inconsistency under angled light.
Outer-wall line width matters
For a 0.4 mm nozzle, an outer-wall width near 0.40 to 0.44 mm is a practical starting region. A very narrow width may improve fine detail, but it also reduces contact area and makes line placement errors more visible. A width near 105% of nozzle diameter is often forgiving without becoming clumsy.
Mini Volumetric-Flow Calculator
This estimate uses line width at 105% of nozzle diameter. It helps reveal whether a “slow” wall is quietly asking for substantial melt flow.
Estimated volumetric flow: 2.94 mm³/s
The calculated number is not a pass-or-fail score. Compare it with a known volumetric-flow test for that filament, nozzle, and temperature. Keep routine outer-wall demand comfortably below the point where extrusion becomes inconsistent.
- Diagnose at moderate wall speed.
- Use temperature changes only within the material’s safe range.
- Keep outer-wall width close to the nozzle’s practical range.
Apply in 60 seconds: Calculate your wall’s volumetric demand and compare it with your tested filament limit.
Check Slicer Geometry Before Blaming the Extruder
Sometimes the printer follows the toolpath perfectly and the toolpath is the problem. Preview the sliced model line by line. Look for actual empty spaces, abrupt line-width changes, missing gap fill, or wall segments that begin and end repeatedly.
Compare perimeter generators
Variable-width perimeter systems can fill thin geometry more gracefully, but they may create frequent width and speed changes. Traditional fixed-width generators are easier to predict, yet may leave narrow spaces unfilled.
Compare one slice using each system. The guide to Arachne versus Classic perimeter generation explains when variable-width walls help and when they complicate calibration.
Inspect wall thickness against line width
A nominal 0.8 mm CAD wall does not always become two perfect 0.4 mm extrusion lines. The slicer may account for overlap, bead geometry, variable width, and dimensional compensation. Use widths that the slicer can represent without repeated micro-gaps.
Check wall overlap and gap fill
Wall-to-wall and wall-to-infill overlap settings affect internal bonding, but excessive overlap can push material outward. Start with the slicer’s proven defaults. Increase overlap only when the preview shows a consistent internal separation, not merely because the outer surface looks thin.
Look for excessive short segments
High-resolution meshes can contain thousands of tiny line segments. The printer repeatedly accelerates and decelerates, which makes pressure-control errors more visible. Simplifying the mesh or using sensible resolution limits can produce a cleaner wall without changing flow.
I once received two visually identical cylinders. One sliced into smooth arcs; the other became a glittering necklace of microscopic segments. The second print had seam-like pores everywhere. The filament was innocent. The mesh had written a percussion solo for the motion planner.
Check wall order
Outer-wall-first printing can improve dimensional control in some parts, but it exposes the visible wall to unsupported pressure transitions. Inner-wall-first often gives the outer line a stable neighbor and a more consistent thermal environment. Test both before making large flow changes.
Common Mistakes That Create New Defects
Raising global flow because one feature is thin
Global flow affects infill, top layers, walls, bridges, supports, and sometimes first-layer behavior. If those features already look correct, a global increase spreads the correction far beyond the defect.
Tuning flow, temperature, retraction, and speed together
A successful print after four simultaneous changes does not tell you which setting helped. It also hides which setting caused the next problem. Change one category at a time and label every test.
Using coasting and aggressive pressure advance together
Both reduce pressure near the end of a move. Strong values can stack into a seam notch. Choose a primary pressure-control method, then use only gentle cleanup settings.
Increasing wall flow to compensate for wet filament
Moist filament can produce bubbles, uneven extrusion, roughness, and random voids. The guide on filament foaming caused by moisture can help distinguish pressure starvation from contaminated melt flow.
Measuring only the most flattering wall
Check all four sides of a test box. Cooling, seam placement, belt direction, and airflow can create side-specific differences. One beautiful face does not acquit the profile.
Copying retraction values from another printer
Two machines with the same extruder name may have different hot ends, tube lengths, firmware behavior, nozzle wear, and filament paths. Borrow the method, not the final number.
Chasing a perfectly invisible seam
FDM perimeter lines must begin and end somewhere. The practical target is a small, consistent seam that does not weaken the wall or disrupt fit. Trying to erase it entirely can trade one faint mark for a recurring hole.
Risk Scorecard: Is Your Change Too Aggressive?
- Low risk: 0.5% flow change, fixed seam, short test print.
- Moderate risk: 5°C temperature change or 10% pressure-control change.
- High risk: More than 4% extra wall flow combined with coast, wipe, and restart compensation.
- Stop and reset: Dimensions worsen while the original seam gap remains.
Know When It Is a Hardware or Filament Problem
Slicer tuning should produce repeatable changes. If identical files produce different gaps each time, inspect the filament path and hardware.
Signs of a partial nozzle clog
- The extruded strand curls sharply as it leaves the nozzle.
- Flow changes after a retraction-heavy region.
- The extruder clicks only at certain temperatures or speeds.
- Manual extrusion feels inconsistent.
Follow the printer manufacturer’s cleaning procedure. Replace a worn or damaged nozzle rather than applying increasingly heroic flow multipliers.
Check extruder grip and spool resistance
Mark the filament above the extruder and watch for hesitation. Inspect the drive gear for packed filament dust. Confirm that the idler is firm enough to grip without crushing the filament. Make sure the spool can rotate without tugging the filament path.
A nearly empty cardboard spool once caused perfect calibration towers and defective full-size parts. The lightweight spool shifted sideways on its holder, caught an edge, and briefly starved every long perimeter. The slicer spent several hours being falsely accused.
Inspect temperature stability
A failing thermistor, loose heater connection, or poor PID behavior can change melt viscosity during a wall. Stop using the printer if temperature readings jump unexpectedly, wiring is damaged, or the heater continues rising without control.
When to seek experienced help
Contact the printer manufacturer, a qualified repair technician, or an experienced local service provider when the hot end leaks above the heater block, electrical connectors show heat damage, the heater cannot hold temperature, firmware protection triggers repeatedly, or extrusion remains erratic after a known-good nozzle and filament are installed.
Decision Card: Keep Tuning or Inspect Hardware?
Keep tuning when the defect appears in the same location and responds predictably to coast, wipe, retraction, or flow changes.
Inspect hardware when the defect moves, changes between identical prints, or arrives with clicking, grinding, temperature instability, or random surface bubbles.
Seek help when heater wiring, thermal protection, melted connectors, or hot-end leakage is involved.
Use a Repeatable Test Protocol
A repeatable test is faster than a dramatic guess. Use the same model, filament, temperature, wall speed, seam position, and camera angle. Change only one setting family per print.
Recommended four-print sequence
- Print A, baseline: Outer flow 100%, coast off, wipe off, known retraction, fixed seam.
- Print B, flow test: Increase only outer-wall flow by 1%.
- Print C, restart test: Return flow to baseline and reduce retraction or pressure advance slightly.
- Print D, finish test: Keep the best restart settings and add a tiny wipe only if a seam bump remains.
Record four observations
- Gap length in millimeters.
- Whether the gap touches the seam.
- Wall thickness at three heights.
- Corner sharpness and surface gloss.
Use a permanent marker to label each tower. Slicer filenames such as “final-v7-real-final-fixed” are archaeological artifacts, not records.
| Test | Outer Flow | Coast/Wipe | Pressure/Retraction | Result |
|---|---|---|---|---|
| A | 100% | Off | Known baseline | Record gap location |
| B | 101% | Off | Unchanged | Judge continuous wall contact |
| C | Best known | Off | Reduced slightly | Judge seam restart |
| D | Best known | Tiny wipe | Best known | Judge final seam finish |
- Use one diagnostic model.
- Change one setting family per print.
- Measure the seam and the full wall separately.
Apply in 60 seconds: Create four filenames labeled A through D before starting the first print.
FAQ
Why does my 3D printer have gaps in outer walls but not infill?
Outer walls contain different speeds, accelerations, line widths, seam transitions, and pressure changes than infill. Coasting, wiping, retraction, pressure advance, or a separate outer-wall flow setting may therefore starve the perimeter while longer infill lines remain continuous.
Should I increase flow if only the outer wall is under-extruded?
Increase only outer-wall flow when the gap continues along much of the wall and remains after coasting and excessive wipe behavior are disabled. Use steps of 0.5% to 1%. Do not raise global flow when infill, top layers, and internal walls already look correct.
Can coasting cause holes at the Z seam?
Yes. Coasting intentionally stops commanded extrusion before the line ends. If the chosen volume is too large, residual nozzle pressure cannot complete the perimeter, leaving a notch or hole near the seam.
What is a good starting value for outer-wall flow?
Start at 100% of a properly calibrated global extrusion multiplier. Test 100.5%, 101%, and 102% when a continuous wall groove remains. Needing more than roughly 104% suggests that line width, filament calibration, nozzle condition, or feeding hardware deserves inspection.
How much wipe distance should I use?
Begin with wipe disabled. If a seam bump remains after pressure and retraction are stable, test a small distance such as 0.2 to 0.5 mm. Long wipe distances can consume residual pressure and create a starved restart.
Can pressure advance cause outer-wall gaps?
Yes. An overly aggressive pressure-advance or linear-advance value can remove too much pressure during deceleration or fail to supply enough material at short line starts. Reduce the value by about 10% to 15% and compare a fixed-seam test.
Why do gaps appear only on small circles or lettering?
Small features create repeated acceleration, deceleration, retraction, and restart events. Variable-width perimeter generation, minimum feature speed, cooling, pressure advance, and mesh resolution may matter more than average flow on these shapes.
Does increasing nozzle temperature fix outer-wall under-extrusion?
A modest temperature increase can improve melt flow and pressure recovery, but it is not a universal fix. Stay within the filament manufacturer’s range and watch for stringing, soft details, glossy bands, or weak overhangs.
Should outer walls print before or after inner walls?
Inner-wall-first is often easier for diagnosis because it gives the outer perimeter physical support and a stable neighboring line. Outer-wall-first may improve some dimensions, but it can expose pressure and placement errors more clearly.
How can I tell whether the nozzle is partially clogged?
Suspect a partial clog when identical files produce different defects, manual extrusion is inconsistent, the strand curls sharply, or the extruder clicks at ordinary speeds. Random gaps are less likely to be solved by a precise outer-wall multiplier.
Can wet filament look like low outer-wall flow?
Yes. Moisture may create bubbles, foaming, roughness, popping sounds, and intermittent voids. Those defects often move between prints rather than appearing at the same seam or corner every time.
What test model is best for tuning outer-wall gaps?
Use a plain 25 to 35 mm square tower with two or three walls, no decorative texture, a fixed seam, and enough height for several setting changes. A simple model makes each defect easier to locate and measure.
Fix the Shell, Not Everything at Once
When infill looks healthy but outer walls have gaps, the printer is giving you a useful clue rather than a contradiction. Continuous grooves may justify a small outer-wall flow adjustment. Seam notches usually belong to coast, wipe, retraction, or pressure control. Random voids point toward hardware, moisture, or inconsistent feeding.
Your next step can fit inside 15 minutes: duplicate the profile, fix the seam position, disable coasting and extra wiping, and print a short two-wall tower at moderate speed. Inspect whether the gap follows the whole wall or only the seam. That single distinction prevents most unnecessary flow changes.
Keep the correction local, measurable, and reversible. The goal is not the largest number that closes a gap. It is the smallest change that restores wall contact while preserving dimensions, corners, and repeatability.
Last reviewed: 2026-07
