Clearance per side vs total
Clearance is the gap between the blank and the pocket wall. Two conventions are in use, and mixing them up halves or doubles your gap:
- Per side: the gap on one edge. A 50.8 mm (2 in) blank with 0.2 mm per side needs a 51.2 mm pocket.
- Total: the difference between pocket and blank. The same pocket has 0.4 mm total clearance.
The jig generator takes clearance per side, and so does this guide. Slicer settings often work the same way: Bambu Lab’s wiki states that with XY Hole Compensation “the hole diameter will increase by twice the compensation value.” When you read a number in a forum post, check which one it means.
Too little clearance and blanks jam, chip, or sit tilted, which on a UV printer puts one corner higher than the rest. Too much and blanks shift as you load them, so the artwork lands off-center. The working window is narrow: on a 50 mm blank, the difference between “drops in cleanly” and “rattles” is often 0.1–0.2 mm per side.
What changes the real gap
Four things move the actual pocket away from the number you typed:
| Factor | Effect on the gap | Typical scale |
|---|---|---|
| Laser kerf | Pocket comes out larger than drawn | Roughly half the kerf per side; kerf around 0.1–0.175 mm on the CO2 machines below |
| 3D-print hole shrinkage | Pocket comes out smaller than drawn | Bambu Lab’s example: 0.24 mm undersize on a 6 mm hole |
| Blank size variation | Some blanks need more room than others | Depends on the blank; natural materials vary most |
| Edge condition | Melted or charred edges, burrs, first-layer flare | Small, but it can be the difference on a tight fit |
Laser kerf
LightBurn defines kerf as the material removed by the tool, and its width depends on material, thickness, and settings such as speed and power. When a laser cuts a pocket, the beam runs along the drawn line, so the pocket grows on every side. xTool puts the theoretical kerf compensation at 0.5 × the laser spot diameter, which is the per-side growth.
Kerf is rarely published as a machine spec, but some labs list it for their machines:
| Laser | Kerf listed | Source |
|---|---|---|
| Epilog Fusion M2 40, 75 W CO2 | 0.1–0.175 mm, material dependent | University of Waterloo Rapid Prototyping Centre |
| Trotec Speedy 300, 80 W CO2 | 0.1 mm, material dependent | University of Waterloo Rapid Prototyping Centre |
For diode, fiber and UV galvo lasers, measure rather than assume. LightBurn’s vernier kerf test reads kerf to hundredths of a millimeter (its worked example gives 0.0575 mm) and notes that diode beams can produce different kerf on the horizontal and vertical axes. That last point matters for jigs: on a diode laser, a square pocket can be looser in X than in Y. Test coupons cut in both orientations catch it.
If you use software kerf offset instead: LightBurn’s Kerf Offset moves closed shapes outward with positive values and inward with negative values, and inverts the direction for shapes inside other shapes on the same layer. xTool’s offset accepts −1 to 1 mm with two decimal places. For a jig, it is usually easier to leave offset at zero and let the coupon pick the clearance, since the coupon already includes kerf. The laser engraving jig guide covers cutting the jig in place.
3D printer tolerance
Printed holes tend to come out smaller than designed. Bambu Lab’s wiki lists the usual reasons: filament shrinks as it cools, holes near nozzle size are hard to tune, missing chamfers or lead-ins, and seam placement. In its calibration example, an M6 test hole measured 5.66 mm, 0.24 mm undersize. Entering half of that (0.12 mm) as XY Hole Compensation brought it to 0.04 mm oversize, which shows the per-side vs diameter relationship in practice.
Prusa’s knowledge base says an Original Prusa will be accurate to at least 0.2 mm, that materials can warp or shrink, and that “an initial good measurement for movable parts is at least 0.3 mm.” It does not say whether that is per side or total; treat it as a starting point and test.
For through-hole jigs, watch the first layer. Prusa describes elephant foot as the first layer being squished against the bed so it comes out slightly wider; on a through-hole, the flare makes the bottom of the pocket tighter than the top. PrusaSlicer’s Elephant foot compensation, around 0.2 mm for a 0.4 mm nozzle, is enabled in Prusa’s official profiles. Bambu Studio has its own elephant foot setting. A blind-pocket jig avoids the problem because the pocket walls start above the first layer.
Print the coupon with the same printer, filament, slicer profile and orientation as the jig. Hole compensation set for one filament does not carry over to another.
Blank size variation
Many blanks are cut, cast or quarried, not machined. Suppliers say so: one slate coaster listing gives a “nominal thickness: 3.0mm - subject to variation,” and another notes that “the sizing for slate products is approximate,” with a buyer review reporting pieces from 3 mm to 5 mm thick in one pack. Coatings can add to edge dimensions too, so measure coated blanks rather than relying on the bare part’s listed size.
The fix is to measure, not to guess:
- Measure ten blanks from the batch with calipers, in both directions.
- Design for the largest, not the average.
- If the spread is wide (natural slate, hand-cut wood), use a looser clearance and accept slightly more play, or sort blanks into size groups with one jig each.
The blank sizes page lists nominal sizes for common blanks, but nominal is where you start measuring, not where you stop.
Recommended starting values
These are starting points for a test coupon, not guarantees. They assume measured blanks and pockets at least as large as the biggest blank.
| Jig build | Blank type | Start per side | Coupon range to test |
|---|---|---|---|
| Laser-cut acrylic, MDF or plywood | Machined or molded (acrylic keychains, aluminum tags, cards) | 0.2 mm | 0.0–0.5 mm |
| Laser-cut, any sheet | Coated or painted blanks | 0.25 mm | 0.1–0.5 mm |
| 3D-printed, no hole compensation | Machined or molded | 0.3 mm | 0.1–0.5 mm |
| 3D-printed, calibrated hole compensation | Machined or molded | 0.15–0.2 mm | 0.0–0.4 mm |
| Any | Natural slate, stone, hand-cut wood | 0.5 mm or more | 0.3–1.0 mm |
For blanks you load and unload hundreds of times, lean looser by about 0.05 mm and add finger notches so blanks can be lifted out without prying. On a UV printer, also check pocket depth; the UV printer jig guide explains why the blank must stand above the jig.
Using a test-fit coupon
A coupon is a small strip with the same pocket repeated at stepped clearances, each step engraved with its value (for example 0.0, 0.1, 0.2, 0.3, 0.4, 0.5 mm per side). The generator produces one for any blank shape, as an SVG for the laser or a 3MF/STL for the printer.
Make it count:
- Use the exact sheet, filament and settings you will use for the jig. A coupon cut in 3 mm MDF tells you nothing about 3 mm acrylic.
- On a diode laser, cut it once horizontally and once rotated 90°, because kerf can differ by axis.
- Test with your largest blank and your smallest.
How to read it:
| What you feel | Meaning | Choose |
|---|---|---|
| Blank will not go in, or needs pressing | Too tight, or edge melt or flare | A larger step |
| Blank goes in with a light push, stays put when the coupon is tipped | Snug | Good for machined blanks; loading may be slow |
| Blank drops in under its own weight, barely moves when nudged | Ideal for batch loading | This step |
| Blank drops in and visibly shifts or rotates | Too loose | A smaller step |
Pick the smallest step where the largest blank drops in under its own weight. Then make the full jig with that value. Ready-made layouts such as 4 in square coasters on the E1 let you enter the clearance you found and regenerate the plate, and the jig materials guide covers which sheets and filaments hold their size best.