Solve defects in 3D printed wire mesh pillars with an accuracy rate over 94%
This page is for engineers whose lattice pillars come out slouched, fused, or out of tolerance. We map each visible defect to a process cause and a concrete fix, using metal and polymer lattice work on our own machines. Read it before you scrap a second build.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
3D printed wire mesh pillar defects: symptom, cause, action
Use this table to pick the right corrective route before you rebuild the file.
| Symptom | Likely cause | Action |
|---|---|---|
| Struts sag or bow mid-span | Layer cooling too slow for thin struts | Add local fans, drop layer height to 50–80 μm |
| Nodes fuse into solid blocks | Over-exposure or excess overlap at junctions | Reduce node overlap to 0.05–0.10 mm, recheck exposure |
| Pillar axis drifts off center | Part shifted on build plate or recoater drag | Re-level plate, add sacrificial skirt, slow recoater |
| Strut diameter runs oversized | Thermal bloom on thin walls | Offset contour inward 0.03–0.08 mm, verify with CT |
| Micro-cracks at node roots | Residual stress from fast cooling | Stress-relieve, slow final layers, inspect at ×10 |
| Pillar bows after heat treat | Support removed before stress relief | Keep supports through heat treat, cut after cooling |
| Porosity between struts | Shielding gas flow too low | Raise flow, recheck chamber O2 below 100 ppm |
| End face not flat | Built without a machined datum | Add 0.3 mm stock, face on CNC to ±0.005 mm |
Fix the process, not the drawing
Thin lattice pillars fail for three reasons: cooling, node overlap, and an unmachined datum. Correct those and the defects stop repeating. If your pillar still drifts after all three, send the file and we will run a DFM review within 12 hours.
What makes a 3D printed wire mesh pillar hard to print
A wire mesh pillar is a truss. Thin struts meet at nodes, and the whole thing carries load through geometry rather than mass. That is why the structure wins on stiffness per kilogram, and also why it prints badly. Every strut is a small thermal mass sitting next to a large one. Heat leaves the strut faster than the node, so the two regions shrink at different rates.
The second problem is support. A 45° overhang in a solid part has material behind it. A strut in a lattice has nothing behind it. As soon as the strut angle falls below roughly 40° from vertical, the next layer has less than half of the previous layer under it. On polymer machines that means droop. On metal powder beds it means dross and a rough underside.
The third problem is inspection. You cannot confirm a 0.6 mm strut with calipers. We use CT scans and cut-and-mount cross sections. That costs time, so the fix has to be right the first time. The rest of this page works through the defects we see most and what actually stops them.
- 1Thin struts cool faster than nodesDifferential shrinkage pulls the node out of position.
- 2Lattice struts have no supportBelow about 40° from vertical, droop starts.
- 3Calipers cannot verify a 0.6 mm strutUse CT or cross sections instead.
Strut sag and bow in 3D printed wire mesh pillars
Sag shows up as a strut that leaves the build straight and arrives at the node late, usually with a rough underside. On polymer machines the cause is almost always layer cooling. A 0.5 mm strut has very little heat capacity, so if the next layer lands on semi-molten material, it slides. Drop the layer height into the 50–80 μm band and add a ducted fan aimed at the pillar. On our polymer builds, sag below 30 μm over a 20 mm strut length is acceptable.
On metal powder beds, sag reads differently. The strut is not molten, it is sintered at the edges. Low laser power or high scan speed leaves unmelted powder clinging to the downskin. The fix is usually a downskin parameter set at 60–70% of the core energy density, plus a contour pass before the infill.
Do not fix sag by thickening every strut. That adds mass at the nodes and can push the pillar outside the envelope. Fix the cooling or the downskin first, then re-measure.
- 1Polymer: cool the layerLayer height 50–80 μm, ducted fan on the pillar.
- 2Metal: set downskin energy60–70% of core density, contour pass first.
- 3Do not thicken all strutsIt adds mass where the part is already stiff.
Node fusion and blocked openings in 3D printed wire mesh
A node is supposed to be a joint, not a lump. When six struts meet at one point, the exposure from each strut stacks. If the overlap radius is set too large, the node grows into a ball and closes the adjacent openings. On a 1.0 mm strut, keep node overlap between 0.05 and 0.10 mm. Measure a few nodes under a toolmaker's microscope before you commit to the full build.
In metal, a blocked opening is often not a geometry problem at all. It is unmelted powder trapped in the pore. If the pore throat is smaller than about 3× the powder particle diameter, powder will bridge. With 20–45 μm powder, that means any opening under roughly 120 μm will clog. Redesign the pore or switch to a coarser powder grade.
Blasting removes loose powder but not bridged powder. We use ultrasonic cleaning in a compatible bath, then verify by CT. If the pore is still blocked, the geometry needs to change, not the cleaning recipe.
- 1Node overlap 0.05–0.10 mmMeasured on a 1.0 mm strut.
- 2Pore throat above 3× particle sizeAbout 120 μm minimum with 20–45 μm powder.
- 3Blasting will not clear a bridgeUltrasonic clean, then CT to confirm.
Dimensional drift in 3D printed wire mesh pillars
Drift is a slow error. The first 10 mm of pillar is correct, the last 10 mm is 0.15 mm off axis. Look for a plate that moved, a recoater blade dragging on a tall thin part, or thermal growth in the chamber. Tall lattice parts are light, so a small nudge moves the whole top. Add a sacrificial skirt tied to the base and slow the recoater to 80–120 mm/s.
Thermal bloom is the other driver. Thin walls lose heat to the powder and to the plate, so the effective melt pool is wider than the slice. If the contour runs oversized, offset the contour inward by 0.03–0.08 mm and rebuild. Verify with CT rather than calipers, since the error changes along the height.
Once the printed pillar is within about 0.2 mm, we face the end datum on a 3-axis mill. That gives a flat mounting face at ±0.005 mm and a clean reference for the rest of the assembly. Printing to final tolerance on a lattice is not realistic. Machining one datum is.
- 1Check the plate firstA shifted plate shows as a tilted pillar, not a bent one.
- 2Offset contour 0.03–0.08 mmCompensates for thermal bloom on thin walls.
- 3Face the datum on CNC±0.005 mm flat face after printing within 0.2 mm.
Cracks and early failure at pillar nodes
A crack at the node root after heat treatment is a stress problem, not a print problem. The strut cools fast, the node cools slow, and the joint carries the difference. On titanium and 17-4PH, we stress-relieve before cutting supports. Cutting supports on a warped part releases the stress and the node splits.
For polymer lattices, cracks usually start at a layer seam where the fan was too aggressive. If the strut surface looks chalky, the material cooled below its glass transition before the next layer bonded. Reduce fan speed to 40–60% and raise the nozzle 0.1 mm. Bond strength recovers.
Test the result rather than trusting the print. A room-temperature pull test on a single pillar, loaded along the axis, tells you whether the nodes are sound. If the failure is at the strut and not the node, the joint is fine and the strut is the weak link.
- 1Stress-relieve before cutting supportsEspecially on titanium and 17-4PH.
- 2Chalky surface means cold bondFan at 40–60%, nozzle up 0.1 mm.
- 3Pull test one pillarFailure at the strut means the node is sound.
Step by step: fixing a bad wire mesh pillar build
Work in this order. Skipping to the last step usually wastes a build.
- 1Inspect and classify the defectCut one pillar out of the build. Photograph the node and the strut at ×10. Label the failure as sag, fusion, drift, or crack. Do not start a rebuild until the class is clear.
- 2Measure the real strut diameterUse CT or a mounted cross section, not calipers. Record diameter at three heights: base, mid-span, top. A 0.6 mm nominal strut that reads 0.72 mm at mid-span is a thermal bloom problem.
- 3Check plate level and recoaterRun a plate map. Anything above 0.05 mm deviation gets corrected. Slow the recoater to 80–120 mm/s on tall lattice parts and add a sacrificial skirt to the base.
- 4Reset cooling or downskin parametersPolymer: layer height 50–80 μm, ducted fan on the pillar. Metal: downskin energy at 60–70% of core, contour pass before infill.
- 5Adjust node overlapSet overlap to 0.05–0.10 mm on a 1.0 mm strut. Print a coupon with five nodes and measure before running the full pillar.
- 6Stress-relieve before support removalRun the relief cycle with supports still attached. Let the part cool in the furnace. Cut supports only after the part reaches room temperature.
- 7Machine the end datumFace the mounting end on a 3-axis mill to ±0.005 mm. Add 0.3 mm of stock in the print for this operation so the finished pillar meets the drawing.
- 8Verify and releaseRe-scan the pillar, run one axial pull test, and record the result. Release only when the node holds and the axis is within tolerance.
Questions engineers ask about 3D printed wire mesh pillars
What strut diameter is practical for a wire mesh pillar?
Below 0.4 mm the strut becomes sensitive to every drift in the machine, and yield drops fast. Between 0.6 and 1.5 mm you get a stable process on both polymer and metal platforms. Above 2.0 mm the lattice stops behaving like a lattice and starts behaving like a drilled block.
If the design calls for 0.3 mm struts, expect to lose a percentage of the build to handling damage even when the print itself is clean.
Can you print the pillar net shape and skip machining?
For a lattice, no. The struts can print within about 0.2 mm, but the mounting face needs a flat datum. We add 0.3 mm of stock to the end face and machine it to ±0.005 mm on a 3-axis mill.
That one operation fixes the interface to the rest of the assembly without touching the lattice itself.
Why does the same file print well on one machine and badly on another?
Lattice parts are sensitive to gas flow, recoater speed, and plate flatness. Two machines with the same laser power can differ by 10% in downskin quality if the shielding flow is set differently.
Match the flow and the recoater speed across machines before you compare results. Otherwise you are comparing atmospheres, not builds.
How do you inspect a lattice without destroying it?
CT scanning gives you strut diameter and node position in one pass. We use it for first articles and for any change in parameter set. Destructive cross sections are faster and cheaper for a single suspect node.
For production, we combine CT on the first part with dimensional checks on the machined datum for every part.
Does heat treatment always distort a metal lattice?
It distorts if you cut the supports first. The supports hold the geometry while the part is soft at temperature. Run the relief cycle with supports attached, cool in the furnace, then cut.
On 17-4PH and titanium we also slow the final print layers to reduce the residual stress that the relief cycle has to remove.
What accuracy rate should I expect after correction?
Once cooling, node overlap, and the machined datum are under control, we hold the pillar axis within tolerance and the node geometry consistently. The 94% figure quoted for this class of part comes from builds where one of those three was left unchecked.
Fix all three and the pass rate on the machined end face is governed by the CNC tolerance, not the print.
Send us the pillar that keeps failing
Upload the model and the defect photos. We review printability, node geometry, and the datum plan, then quote the print plus the finish machining. No minimum order quantity, from one prototype upward.
12-hour quoteFree DFM analysis100% inspectionNDA on request
More process notes from the shop
We publish setup notes, tooling trials and inspection data from the factory floor.
