Your printer may be rated for breathtaking speed, yet Gyroid infill can keep the toolhead sweeping through curves long after a straighter pattern would be finished. Cross Hatch was built for that exact tension: preserve a three-dimensional internal structure while giving a fast printer longer, calmer runs.
The catch is that “faster” does not automatically mean “better.” A bracket, a display model, and a flexible cushion do not ask the infill to do the same job. The right choice depends on the load, part geometry, shell settings, filament flow, and what your own slicer preview reveals.
Cross Hatch
Mostly linear passes, fewer continuous directional changes, and strong potential time savings in broad internal areas.
Gyroid
A continuous three-dimensional structure that is well suited to parts facing mixed, twisting, or hard-to-predict loads.
Gyroid and Cross Hatch do not ask the printer to move the same way
Gyroid builds a repeating three-dimensional surface from smooth, continuous curves. The path does not stack material at same-layer intersections the way Grid can, and its connected geometry spreads support through the part instead of favoring one flat direction.
Cross Hatch takes a different route. It uses mostly straight lines for a range of layers, then transitions their direction so neighboring sections form a three-dimensional crossed structure. The toolhead can spend more time moving along efficient paths and less time constantly steering around curves.
Bambu Lab introduced Cross Hatch in Bambu Studio 1.9.1 and reported that it printed 28% faster than Gyroid in one test cube. That is useful evidence that the pattern can save time, but it is not a universal promise. Model shape, density, printer limits, and profile settings can narrow—or occasionally erase—that advantage. You can review the Bambu Studio 1.9.1 release notes and its infill-pattern guide for the developer’s description.
A 500 mm/s printer does not print every infill line at 500 mm/s
Maximum speed is only one ceiling. A printer also needs enough distance to accelerate, a hotend that can melt the requested volume of filament, and a motion profile that does not sacrifice quality. Short segments, cooling limits, and minimum layer time can all slow the job.
This is why Cross Hatch tends to show its best advantage inside large, open internal areas. Its longer runs give the printer room to accelerate. In a small part dominated by walls, holes, and short toolpaths, changing the sparse infill pattern may move the final estimate by only a few minutes.
Gyroid’s curves are smooth, but the toolhead is continually changing direction. At aggressive acceleration settings, that motion may sound busier or expose resonance in a less-than-ideal profile. Cross Hatch often feels calmer through its mostly linear layers, although its transition layers still require directional changes.
Gyroid vs. Cross Hatch: the practical comparison
| Decision point | Gyroid | Cross Hatch |
|---|---|---|
| Toolpath character | Continuous curves with frequent direction changes | Mostly linear paths with periodic transition layers |
| Print-time potential | Often slower in large infill-heavy parts | Often faster when the printer can accelerate along long runs |
| Load behavior | More uniform three-dimensional structure; a sensible choice for mixed-direction loads | Three-dimensional structure, but speed is its clearest documented advantage |
| Noise and motion | Can sound busier on fast profiles because direction changes are continuous | Often quieter through the predominantly straight layers |
| Same-layer crossings | No Grid-style line crossings | Designed to form the cross structure across adjacent layers rather than Grid-style same-layer intersections |
| Best starting use | Functional parts, twisting loads, flexible structures, and uncertain load direction | Large prototypes, housings, organizers, display models, and time-sensitive prints |
| Main caution | Do not pay a time penalty when the part does not benefit from the geometry | Do not assume a faster pattern can compensate for weak orientation, too few walls, or poor layer adhesion |
For another useful efficiency comparison, see my guide to Gyroid vs. Adaptive Cubic infill. Adaptive Cubic is aimed at reducing unnecessary material inside larger volumes, while Cross Hatch attacks the problem through faster path behavior.
Which infill should you choose?
Choose Cross Hatch when the clock is the real constraint
- The part has a large internal volume with room for long infill passes.
- You are printing a prototype, enclosure, organizer, display model, or fit-check part.
- Your high-speed profile makes Gyroid sound busy or produces a disappointing time estimate.
- The part’s shell carries most of the expected load.
- You can validate the finished part before treating it as production-ready.
Choose Gyroid when the load is harder to predict
- The part may be pushed, twisted, or loaded from several directions.
- You want a continuous internal network rather than a speed-first path.
- You are printing flexible material and want a smooth, connected internal structure.
- Your existing Gyroid profile already prints cleanly and the time estimate is acceptable.
- A failed part would cost more than the time saved by changing patterns.
Do not let infill distract you from the shell
If a functional print keeps breaking, increasing infill is not automatically the best fix. Adding a perimeter, changing the part orientation, improving layer bonding, or reinforcing the exact failure zone can produce a more meaningful gain.
That is especially true around holes, snap fits, thin arms, and screw bosses. Infill cannot repair a weak interface if the load is concentrated in a poorly oriented wall. If the defect looks like gaps, brittle lines, or incomplete infill, use my 3D-printing troubleshooting guide before blaming the pattern.
The fair way to test Cross Hatch against Gyroid
A quick test does not require a lab, but it does require discipline. Duplicate the model and change only one variable at a time.
- Lock the baseline. Use the same model, orientation, layer height, nozzle, filament profile, wall count, top and bottom thickness, infill percentage, and speed settings.
- Slice both patterns. Record the estimated print time, total filament in grams, infill time if your slicer reports it, and maximum volumetric flow shown in preview.
- Inspect the toolpaths. Look for unsupported roof spans, tiny isolated infill segments, very short direction changes, and areas where the shell—not the infill—dominates the job.
- Print the meaningful candidate. Listen for resonance or scraping, then check top-surface quality, wall finish, dimensional fit, and the part’s behavior under its intended load.
- Change one setting next. If strength is inadequate, test an extra wall before making a large jump in infill density.
Beginner-friendly starting settings
These are test baselines, not engineering guarantees. Your part geometry, material, nozzle, layer height, and printer profile still control the outcome.
| Print goal | Pattern to test first | Starting structure | What to inspect |
|---|---|---|---|
| Display model or fit check | Cross Hatch | 8–12% infill, 2–3 walls | Top-surface support and total time |
| Everyday enclosure or organizer | Cross Hatch | 12–18% infill, 3–4 walls | Screw bosses, corners, and roof spans |
| Bracket or mount | Gyroid | 15–25% infill, 4–5 walls | Layer orientation and the actual load path |
| Flexible or twisting part | Gyroid | Low-to-moderate density, then tune by feel | Compression response and wall-to-infill bonding |
Clean testing starts with consistent material. Damp filament, unstable extrusion, or a profile that exceeds the hotend’s flow capacity can make one pattern look guilty when the real problem is elsewhere. I partner with COEX 3D; readers can use code 3DPRINTINGBYKEVIN for 15% off eligible filament. Availability and terms can change, so verify the offer before ordering.
Use the P.R.I.N.T. Method to make the decision
P — Problem: What is the part supposed to do, and what would failure look like?
R — Requirements: Rank print time, weight, stiffness, surface finish, noise, and durability.
I — Interfaces: Mark screw holes, clips, mating faces, load points, and directions where force enters the part.
N — Next-Best Materials & Methods: Decide whether a different filament, orientation, extra wall, modifier mesh, or manufacturing method matters more than the infill pattern.
T — Test & Tune: Slice both patterns, compare time and grams, print a representative sample, and change only one variable in the next round.
If you are still building confidence with slicer terms and print planning, my step-by-step guide for absolute beginners provides the broader foundation. The expanded P.R.I.N.T. It Practical guide carries the same decision-first approach through an entire project.
Four-question knowledge check
Choose your answer, then open the panel to check your reasoning.
1. Which pattern should you test first for a large display model when print time is the main concern?
Answer: Cross Hatch. Its mostly linear path behavior can make better use of a fast printer across broad internal areas.
2. Does the same 15% infill setting guarantee the same material use for both patterns?
Answer: No. Pattern geometry and slicer implementation can change the estimated grams. Check both percentage and material use.
3. What should you consider before increasing infill to fix a broken bracket?
Answer: Orientation, walls, and layer bonding. Those factors may affect the failure more directly than a large density increase.
4. Does Bambu Lab’s 28% test-cube result mean every Cross Hatch print will be 28% faster?
Answer: No. It demonstrates potential under one test condition. Your geometry, density, motion profile, and flow limit decide the real difference.
Frequently asked questions
Is Gyroid the strongest infill?
No infill pattern is universally strongest. Gyroid is valued for its continuous three-dimensional structure and more balanced behavior across directions, but FDM parts remain sensitive to layer orientation. Material, walls, temperature, geometry, and load direction can change the winner.
Is Cross Hatch always faster than Gyroid?
No. Cross Hatch has a clear path-efficiency advantage in many infill-heavy models, and Bambu Lab documented a 28% improvement in its test cube. Small parts, low-density prints, or models dominated by walls may show a much smaller difference.
Does Cross Hatch use less filament?
It can, but do not assume it from the percentage alone. Slice both patterns and compare the estimated grams. If you are testing strength per gram, adjust density until the material estimates are close.
Which pattern is less likely to scrape the nozzle?
Gyroid and Cross Hatch both avoid the stacked same-layer intersections associated with Grid. However, scraping can also come from over-extrusion, warping, poor bed adhesion, loose motion components, or an unsuitable travel profile. A pattern change should not replace troubleshooting.
Should I use Cross Hatch for production parts?
Use it only after validating the actual part. A prototype that looks clean is not proof that a mount, bracket, or safety-relevant component will survive repeated loads. Document the slicer version and settings so the approved result can be reproduced.
The bottom line
Cross Hatch is the more compelling high-speed choice when the part gives it room to run. It can reduce time and calm toolhead motion without falling back to a simple two-dimensional line pattern.
Gyroid still earns its place when load direction is uncertain or the internal structure is part of the design decision. Just do not let its reputation make the choice for you. Preview both, compare grams as well as minutes, and test the finished part under the load that matters.
Share the model type, slicer, infill percentage, estimated time, and actual result in the comments. A useful comparison gives the next maker more than a pattern name—it gives them context.
Reference note: Bambu Lab documents Cross Hatch’s path design and its 1.9.1 test-cube result. For a broader explanation of Gyroid, Grid, Rectilinear, and other patterns, see the Prusa Knowledge Base infill guide. Slicer behavior can change with software updates; verify the preview in the version you use.
