Relationship between nozzle size, layer/nozzle height, print quality, print speed

Hi,

I must admit, I am new to 3d printing, but you probably got that from reading the topic title. Been reading and watching youtube’s for several weeks now, and I recently build my prusa mk2s printer, even got a few good prints from it.

I am however still learning about all the various print settings and I cant get my head around the relationship between nozzle size, nozzle height, layer height, print quality and print speed.

I understand there’s X/Y resolution and Z resolution (= layer height)
If slicer applications talk about 200 or 100 microns, this is layer height, right? So what defines the X/Y resolution?

And nozzle size, 0.4 is commonly used, but you can print higher resolutions with a 0.4 nozzle, how does that work?

And if you use a print quality of 200 micron, does this mean the height and the x&y is 200 micron? And the ideal nozzle height (from the bed) would be 200 micron?

And print speed, I can imagine the faster you go, the thinner the plastic coming from the extruder? Or is thickness + speed something the printer takes care off?

I hope Tom will do a video on this :slight_smile: Or perhaps he already did, couldn’t find it though. Or if someone else on the forum can enlighten me, that would also make my day :slight_smile:

Ido

Nozzle size affects the size of the line/extrusion width of the print. This is what you would call x/y resolution. With a 0.4mm nozzle you’ll use a line width of 0.4mm to about 0.6mm. Obviously the thicker the lines the quicker of the print. Thicker line width also increases strength of the print. Line width doesn’t have much impact on print quality, except for the fact that if a part of the print is thinner than the line width it won’t be able to print that part.

With a 0.4mm nozzle you can print to about 0.35mm as the max layer height. Layer height affects print strength, but not to the degree of line width. It mostly affects visible print quality and time to print. Larger nozzles can still print w/ smaller layer heights.

Print speed affects quality mostly due to vibrations from the printer frame/chassis as well as heavy X gantry / moving heated beds. A very rigid printer (CoreXY, etc) can print print much faster because there is less vibrations.

It’s important to reorient your thinking regarding “resolution”. Z is the closet thing we have to, say, a computer’s concept of resolution, because the entire layer is printed, then Z is stepped up, and the next layer is printed. On X and Y, your resolution is imperceptibly small because instead of printing line by line like a computer display, lines are drawn continuously, more like a vector display. For example, on my printer my X and Y steps per mm are 160 (pretty standard). So the printer can move to within 1/160 of a mm, or 6.25 microns. This has some slight error, ± 3 microns, let’s say, due to microstepping. There’s also not really a concept of X and Y as independent axes, because they move at the same time. So where on a computer you’d have a diagonal line that would look jagged up close, on your printer the line can be perfectly straight in any direction.

In computer graphics a circle is one of the simplest tests of whether or not you can see resolution, but when you print a circle on a 3d printer it essentially looks perfect. A good test print to demonstrate this is a dome or hemisphere. When you print a dome, each layer will basically be a perfect circle if the model is sufficiently detailed, but the whole dome will have obvious layer lines and therefore Z is a MUCH smaller “resolution” in the traditional sense.

Of course, as lowfat noted, the extrusion width is the limiting factor. Basically you can create any shape to a ludicrously small resolution, but only if you can draw the outline of that shape with, in my case, .45mm wide lines. And the ends of those lines are always round, so for example you couldn’t really print a knife-sharp exterior edge (though you can print sharp indents in negative space). Your settings for number of perimeters lines, etc. will also limit the possibility of fine detail, but as long as there is a bit of room to store your lines, you can print incredibly detailed models. This stuff gets to be important when designing small models to print. I like to think of it like “if I scaled this up a bunch, could I draw this layer with a tube of toothpaste?”

Regarding speed + thickness, yeah that’s calculated when you slice the model, so if you speed up or slow down during the print the extruder will scale accordingly.

Actually, speed and thickness is a response nsibility shared between slicer and the firmware.

The slicer breaks down the model layer by layer, but ultimately down in to moves for the extruder. They are either travel moves (where there is no extrusion, I.e the print head just needs to get to a different location) or extrusion moves, where extrusion must happen.

The slicer has settings for layer height and extrusion width and so on, and with a bit of maths what that actually boils down to is that for a given travel move (from a particular point to another point) the slicer knows the exact volume of plastic that should be extruded. And since it also knows your filament diameter, that can be converted in to a length of filament to extrude.

And that is the basis of the majority of the GCODE commands when you are printing. For example,

G1 X150 Y100 E5

Will cause the printer to move from where it currently is, to the location 150,100 whilst extruding 5mm of filament. It does this at the same rate in each axis such that when it’s half way there it’s extruded half the total amount of plastic etc.

Another parameter on the GCODE can set the feedrate, or speed, for the move. The printer will try and move at that speed, if it can - but it might not get that fast due to the fact that it needs to accelerate and decelerate. But the firmware in the printer handles the acceleration and the extrusion so that the plastic is extruded perfectly in sync with the travel move.

I am actually printing a big benchy (300% scale).
Which has a layershift :-/
4 hours with 0.4 layer height and a 0.8 nozzle.




As chumm stated Z is the closest thing we have to resolution and X/Y resolution is practically irrelevant.

When the slicer talks about 0.1/0.2/0.3mm this is indeed your layer height. Again, as chumm explained X/Y resolution depends on the micro-stepping and is theoretically too small to take into account.

A 0.4mm nozzle can print at 0.1mm or 0.2mm layer height because the layers are squished by the extruder, the slicer will adjust flow so that the squished line matches the line width and layer height.

There is a strong relation between nozzle size, layer height and line width. A bigger nozzle will be able to print with larger line width and higher layer heights because it is able to output more plastic than a smaller nozzle. Of course this comes at the cost of print quality because a 0.8mm nozzle will not be able to print details smaller than 0.8mm. All of this also relates to print time, higher layer heights means less layers and quicker prints.

When printing at 0.2mm you don’t want a 0.2mm nozzle because then you couldn’t squish the plastic and the layer adhesion would be really poor. Typically your maximum layer height is 80% of your nozzle size.

Concerning print speed, yes the faster you go, the smaller (more stretched) the line but your slicer will adjust flow-rate depending on the speed to ensure a consistent line width.

Hope that helps

First of all, thank you all for your replies, things are becoming more clear.

So to summarize

  • 3D printer resolution ≠ 2D printer resolution. When advertising a 3D printer, they talk about 100 micron, this is layer height
  • X/Y resolution is defined by X and Y steps per mm where the width is defined by nozzle size
  • A printer can place a line very accurately in the 2D space, but line thickness will be 0.4 (when using a 0.4 nozzle), height can vary, to max of 80% of the nozzle size
  • Z resolution is defined by layer height
  • A printed line doesn’t have the same width all round, common sizes are 0.4 width x 0.2 height

Am I correct by saying this.

Now I also understand why the slicing algorithm is really important. Printing is a complex thing, where a lot of components need to be tuned and work together. Pretty amazing we can actually print stuff :slight_smile:

Yes you are right, but you can have an extrusion width greater than the nozzle width; imagine what happens when you squeeze your toothpaste onto your brush when the tube is close to the brush - it spreads out sideways. It’s just a volume calculation for the slicer. There is a practical limit to how wide you can go though, in the region of 25% to 40% wider.

Oh wait, I think i mixed up the 80% thing, it should be layer height. So a nozzle size of 0.4 can have a max layer height of 0.32 (80%)

mmm, how does that work, when the width is wider, say 25% more, so 0.5mm, will the height still be 0.2 (when selecting this in the slicer) ? or is this something the slicer takes care off, to make sure the width is 0.5 and height is 0.2.

And I also heard single wall thickness should always be a multiple of the nozzle size, so with a 0.4 nozzle, walls should be 0.4, 0.8, 1.2 etc
But how this explain the width thing, you can have any wall thickness…

You’re right about layer height.

Yes the height will still be 0.2 because it is how hight the nozzle is, the slicer will take care of adjusting your extrusion width based on your nozzle size and extrusion multiplier (the more material you push through, the wider the line).

Wall thickness should be a multiple of your nozzle size, making a 1mm wall with a 0.4 nozzle (extrusion width) will leave gaps or introduce artifacts. So 0.8, 1.2, 1.6, … are good wall thicknesses because they are multiples of your nozzle size (extrusion width) and ensure solid walls.

Some slicers will ask how many perimeters you want while others will ask for wall thickness (2 perimeters = 0.8mm wall thickness, 3 perimeters = 1.2mm wall thickness, …)

Yes you can have any wall thickness but usually models designed for 3D printing will have wall thicknesses optimized for 0.4 nozzles (things like 2.4mm or 3.2 mm) if your wall thickness is not a multiple of your nozzle size the slicer will add infill between the perimeters.

But how does this work?
>>the slicer will take care of adjusting your extrusion width based on your nozzle size and extrusion multiplier (the more material you push through, the wider the line)
>>Wall thickness should be a multiple of your nozzle size

If the slicer can define the line width, why is the multiple of nozzle size still important?

Dont get me wrong, I am not in doubt, there are plenty of posts about wall thickness and nozzle size, but don’t understand the relationship

The nozzle width will define the minimum width of the line, i.e. there’s no way to create a .3mm wide line with a .4mm nozzle. You can create a wider line, though not by much (up to 25% more than the nozzle width generally). Think of it like squeezing out a line of toothpaste, you can squeeze harder and increase the flow and it will push out to the sides. So, the above is true, but only in a really limited range.

If you want to make very thin lines you need a very thin nozzle, and a large nozzle can help you push through a ton of material, which can greatly decrease the time a print takes, at the expense of having to use thicker walls. This is also potentially limited by the hotend’s ability to keep the filament hot when extruding at a high rate. .4mm nozzle and .45mm extrusion width is the standard and should serve almost all applications.

You’re right that the multiplier point was slightly misspoken, wall thickness should be a multiplier of your extrusion width, not necessarily your nozzle width. This also only applies to thin walls: if your perimeter layers are 3, and you have a .45mm extrusion width, anything larger than (.45 x 6) 2.7mm could be any size you want, since the space between the walls can be any size.

Thanks. Things have become clear concerning the Z en X/Y resolution.

Funny how Z resolution is not dependent on X/Y resolution. You’d expect when printing in a 0.20 layer height, you would also want to use the 0.2 nozzle, so height and width are the same. But in practice this doesn’t seem to matter and its pretty handy you can vary Z resolution quite easily.

One thing I am still pondering about, say you sliced with 0.20 layer height.
If the printer is busy with layer number 2, does the printer always print this at 0.20 layer height (does the printer even know how high it is?), or does it print on top of layer 1?
With my Prusa mk2s I can adjust my Z value for the first layer while printing. Sometimes the layer height after printing is 0.2 mm, some prints 0.24 (wrong Z value).
Will the second layer print at 0.24 + 0.20 = 0.44, or will all layers be 0.24, so second layer will be 0.24 + 0.24 + 0.48?

Each subsequent layer will just be .20 above the prior, so .24, .44, .64 etc. Adjusting Z on the first layer isn’t about changing the layer height, it’s about correcting for Z = 0 being inaccurate. So ideally if you made that correction appropriately on say a 30mm tall print, it would measure exactly 30mm, not 30.04, because the adjustment was to ensure that the first layer printed at the right level.

Just to add something for a new person, there are layer hights that will work better or more consistent on your specific printer due it’s mechanical properties. I mean by that that it’s possible that your print result could be better with a layer hight of 200 micron instead of 150 micron for example, even though you should normally have better details on the 150 micron you can have gabs, inconsistent lines or bad layer adhesion or over all a bad looking print in the end.
This can happen due the manner how stepper motors and micro stepping_ _ (Tom has a video about them) works and which threaded rods you use, more pricise which pitch they have. The pitch size is the travel in (normally) mm that the threaded rod is able to transport with one rotation.

You may know this behavior when you rotate a stepper moter that it’s snaps every few degrees (when it’s not on) in a specific position so to a specific step. Your electronics normally calculates the needed amount of steps for your desired movement. For the z-axis every X micron setted in layer hight per each layer. But if you set in a layer hight which isn’t or isn’t nearly matching a multiple of the travel amount per each step of your z-axis mechanic it can only move to the nearest possible. The whole micro stepping is actually a thing to emulate parts of a full step but aren’t as accurate and not effective if the motor has to hold it’s position. It is also a problem that the slicer calculates the rest of the model with the dialed in layer hight instead of the “right layer hight for your printer”, so if the last layer is a few microns less or to tall than expected with the _wrong _layer hight the next layer has to neglit that which reduces in bad layer adhesion.

But what is the right layer hight for your printer you may ask? Well as I said the effect is caused by the stepper motor and threaded rod, you can increase the amount of “perfect working” hights by a factor of two for example by picking a stepper motor with 400steps/revolution one instead of the common 200step motors or you can pick a threaded rod with a super low pitch size but this way your printer has to step several thousand steps more than necessary for homing (=slower movement, for z-axis only not a big deal) or fast movement which stresses the controller.
Another easy way which also doesn’t lead into the need of new parts is to actually know which layer hight are ideal for your printer. The super handy calculator from Prusa (http://www.prusaprinters.org/calculator/) can help you with that, just dial in under Optimal layer hight for Z-axis which pitch your rod has (and layer hight if you look for a specific one) and it will show you if the mechanics of your printer is capable to replicate these accurately.

For this example of a printer with 200step motor and a threaded rod with a 8 mm pitch, are increments of 40 micron when choosing the layer hight ideal.

EDIT: corrected a bit of spelling and the image