What the nozzle diameter actually constrains
A nozzle has one orifice diameter and it constrains two different things in two different ways, which is why the advice around it is so muddled.
Line width is bounded below by the nozzle, because you cannot reliably lay a bead much narrower than the hole it came out of. It is bounded above by control: past about 150 percent of the nozzle, the extruded plastic spreads sideways further than the tip can shape, and corners round off. So line width sits roughly between 100 and 150 percent of nozzle diameter, and most slicers default near 105 to 120 percent.
Layer height is bounded above by something physically different. The flat face around the nozzle orifice presses each new bead down against the layer below, and that pressure is what fuses them. If the layer is too tall, the tip never reaches down far enough, the bead sits on top rather than being welded into place, and the part delaminates under load. That failure point lands somewhere around three quarters of the nozzle diameter. The lower bound is practical rather than physical: below about a quarter of the nozzle, the bead is so thin that bed flatness, Z accuracy and first-layer squish all become critical, and the layer count explodes.
| Nozzle | Layer range (25-75%) | Line width range (100-150%) | Common working pair |
|---|---|---|---|
| 0.25 mm | 0.06 - 0.19 mm | 0.25 - 0.38 mm | 0.10 x 0.28 |
| 0.40 mm | 0.10 - 0.30 mm | 0.40 - 0.60 mm | 0.20 x 0.42 |
| 0.60 mm | 0.15 - 0.45 mm | 0.60 - 0.90 mm | 0.30 x 0.66 |
| 0.80 mm | 0.20 - 0.60 mm | 0.80 - 1.20 mm | 0.40 x 0.88 |
| 1.00 mm | 0.25 - 0.75 mm | 1.00 - 1.50 mm | 0.50 x 1.10 |
The common pairs are conventions, not recommendations from this page. They exist because the middle of a range is a safe place to start when you have no other information.
The tradeoffs, stated plainly
Halving the layer height doubles the layer count and roughly doubles the print time, and it buys you smoother curved surfaces on anything sloping. It does nothing at all for the accuracy of vertical walls or hole diameters, which are set by line width and flow, not by Z. This is the single most common misunderstanding about layer height: people drop to 0.1 mm hoping for a more accurate part and get a slower one with the same dimensional error.
Going taller improves layer adhesion up to a point, because a thicker bead carries more cross-section across the layer boundary, and it cuts print time hard. It costs you surface finish on sloped faces, fine detail, and the ability to print small features that are only a few layers tall. Going wider on line width improves strength and speed for the same reasons and costs corner definition.
Both of those trades happen through the same physical quantity: line cross-section, which is (width minus layer height) times layer height, plus pi times half the layer height squared. That is the bead modelled as a rectangle with semicircular ends, which is what a squashed extrusion actually looks like in section. Multiply it by print speed and you get volumetric flow, which is the number that determines whether the hotend can keep up.
Flow is the real ceiling
A 0.4 mm nozzle at 0.2 x 0.42 has a line cross-section of 0.0714 mm2. At 60 mm/s that is 4.28 mm3 per second. Push the speed to 150 mm/s and the demand is 10.7 mm3/s, which is at or past what a conventional hotend can melt. The machine will accept the command and then under-extrude, which shows up as thin, gappy walls that people usually blame on the extruder or the filament.
Enter your hotend's flow figure and this page shows the speed at which you run out. It is worth doing before any speed experiment, because it explains a large share of the failures that follow one. The same arithmetic run the other way is the argument for a bigger nozzle: at a fixed flow limit, doubling the line cross-section halves the speed you need for the same throughput, and slower moves are moves the motion system can actually complete.
Layer count, and the top layer that never lands right
Layer count is the part height minus the first layer, divided by the layer height, rounded up, plus the first layer. A 40 mm part with a 0.28 mm first layer at 0.2 mm resolves to 199 layers above the first plus one, so 200 layers, reaching Z 39.88 mm. A 100 mm part under the same settings reaches 100.08 mm, overshooting by 0.08.
Slicers handle that overshoot by either printing a shortened top layer or by rounding the model, and the visible consequence is that a feature you designed to be exactly 40.00 mm tall comes out 39.88 or 40.08. If a height dimension is critical, pick a layer height that divides into it evenly after the first layer, or add the compensation in CAD. That is a different problem from shrinkage, which affects every dimension proportionally rather than only the vertical ones, and the shrinkage compensation calculator handles that side.
The Z-step field is there for the people who will otherwise ask. On a leadscrew machine driven by a modern controller, layer heights that are not exact multiples of the Z step are handled by carrying the remainder forward, and the error averages away. It is a rule left over from hardware that no longer dominates. It is included because checking costs nothing and because knowing why a rule stopped mattering is better than following it.
Questions people ask
Will a smaller layer height make my part more dimensionally accurate?
For horizontal dimensions, no. Hole diameters, wall thicknesses, slot widths and outside dimensions are set by where the toolpath is placed and how much plastic comes out of it, and neither of those changes when you change layer height. What improves is the vertical stepping on sloped and curved surfaces, and the accuracy of the height of features that end partway through a layer. If your parts come out with holes too small or pegs too fat, the fix is flow calibration and hole compensation, not thinner layers. The e-steps and flow calculator and the shrinkage compensation calculator both address that directly.
Can I run a 0.3 mm layer on a 0.4 mm nozzle?
Yes, and it is a normal setting for functional parts. It sits at 75 percent of nozzle diameter, right at the conventional ceiling, so expect layer adhesion to be at the edge of what the nozzle can press together and expect the tuning to be less forgiving than at 0.2. Widen the line to 0.45 or 0.5 mm to help, since a wider bead gives the tip more contact area to work with. If you are printing tall parts that will carry load along the Z axis, test one before committing a long print to the setting.
Why do slicers default line width above the nozzle diameter?
Because a bead that spreads slightly wider than the orifice is more forgiving. At exactly nozzle diameter, any small variation in flow, first layer squish or filament diameter shows up as a gap between adjacent lines. Setting the width 5 to 20 percent over gives adjacent beads a guaranteed overlap, so the surface closes up and the part is watertight. There is also a practical benefit at the first layer, where slicers often go wider still to increase bed contact area.
Does layer height affect part strength?
It affects it in the direction that matters most and least intuitively. Layer bonds are the weakest plane in a printed part, so anything that improves fusion between layers improves Z strength. Thicker layers give a larger bonded cross-section at each interface and fewer interfaces overall, so up to the adhesion limit they are stronger in Z. Beyond that limit they get sharply weaker, because the nozzle stops pressing the bead down. Within a layer, strength is essentially unaffected by layer height and is driven by wall count, line width and how well the lines fuse sideways. If a part is failing along layer lines, look at temperature and cooling first, then at layer height and wall count.
Is there a layer height that hides layer lines?
No. Layer lines are geometric: any sloped surface built from stacked flat layers has visible steps, and the step width equals the layer height divided by the tangent of the surface angle. On a surface only ten degrees off horizontal, a 0.1 mm layer produces a 0.57 mm step, which is plainly visible however small the layer. Halving layer height halves the step and doubles the print time, which is a real but slow-moving trade. Variable layer height, which uses thin layers only on shallow slopes, gets most of the benefit for a fraction of the time cost, and is worth more than a blanket reduction.