After getting the basic dial in done like temperature (usually the higher the better the layer adhesion – up to a point – but the more stringing) and retraction it is time to dial in for accuracy. In many cases it will be a compromise between mechanical strength and over/under extrusion.
This is done by using the Extrusion multiplier and NOT ESTEPs . Esteps are used to calibrate the extruder. I do that usually with no back pressure that means if possible either mathematically or by measuring how much filament is coming out the extruder without having to go through the hotend. As a side note if the amount of filament going through the extruder with lets say the Bowden tube disconnected compared to how much goes through with it connected and going through the nozzle differs then the extruder is slipping and any further tuning will be meaningless until that problem is corrected. That slipping can have many reasons like too low a temperature, too high a speed or the tension on the extruder adjusted improperly or a combination of such.
For the purpose of tuning the print we use a file “multiwall” that is created for
just that purpose and the STL for the file is here https://grabcad.com/library/frc1989-filament-dial-in-1
Here is an image of it sliced in Prusa Slicer
From printing them with different extrusion multipliers and then measuring with a caliper you can get an idea as to the expected accuracy of your part. When you measure with a caliper make sure you measure in a way to not include the first couple of layers in case of elephants foot. So make sure the legs of the caliper end about 2 mm above the part that touched the build plate. For this post I printed the same file with an extrusion multiplier of .9, 1.0 and 1.1. Inspecting those prints you can gleam some additional slicer about your printer/slicer combo. Like for example the 2mm wall used to allow to be split if the perimeter thickness did not add up to a 2mm wall. This stl was designed to test when printing with a .8 so with an external perimeter <1mm there used to be a gap between the 2 lines making that print. You can also measure at the beginning, ending and middle of the wall and find how well you do in regards to things like linear advance issues.
In general the walls for that test should be in % 125, 156.25, 187.5,250 of the nozzle diameter. You can see how your external perimeter settings will influence slicer performance when it comes to splitting the print into perimeters. It will also give you information as to how the slicer places perimeters
We print with the option set to outside perimeters first as that is more consistent when it comes to accuracy as the slicer will place a perimeter of the selected “outside perimeter width” first and you can better predict the outcome when coming up with a slicing plan. This print also will demonstrate to you how your printer/slicer/filament option behaves when it comes to single vs multi perimeter sections of your print. In most cases the single perimeter walls will be wider in relation to the target thickness and the multi perimeter walls thinner in relation to the target width and that most likely has to do with how the filament flows.
Also the higher the flow rate – the more filament – the better the layer adhesion on the plus side – on the negative the most likely over extrusion will cause walls that are thicker than intended and at some point the nozzle will start to “plow” and the next layer will have a rough/bumpy ride and on many parts – especially with a lot of perimeters – a point may come where enough “bumps” build up for the nozzle to get stuck and you suffer a layer shift or knock the print of the print plate
After test printing this in HIPS at 255/105 Bed we came up with a wall thickness as follows
| Wall (target) | .9 | 1 | 1.1 |
| 1 mm | .98 | 1.05 | 1.18 |
| 1.25 mm | 1.14 | 1.29 | 1.42 |
| 1.5 mm | 1.41 | 1.59 | 1.77 |
| 2.0 mm | 1.86 | 1.97 | 2.09 |
| 3.0 mm | 2.81 | 2.95 | 3.20 |
All measurements were performed by taking 5 measurements at the middle in close proximity discarding the high and low and averaging the remaining 3 with a 2 digit accuracy. We can see that the 1 extrusion multiplier came closest to the target width. In practice this filament should extrude well with a multiplier close to 1. For most application we use .98 on this and for some gears as low as .94. At the 1.1 multiplier you can see the phenomen that – as its too high – it starts pushing the walls out due to the pressure from squeezing more filament into the middle. One in theory from the wall placement would expect a wall thickness around 3.1 yet get 3.2
If you examine the prints you will see that the higher the extrusion factor the more pronounced the layer lines.
In summary the higher the extrusion factor – the better the layer adhesion – the lower the extrusion factor the better the visual appearance and the more predictable the dimensional accuracy. Not going to an extend at which prints will fail.
So if within an organization like an FRC team you agree on an acceptable range of parameters. So for example we tune for precision fit parts like gears to 96-98% of target and for low precision high force use to 100-105% target – AND you tune all filaments no matter what make to those ranges. Then you can make a test print of a gear or bracket in lets say HIPS or PLA which is low cost and once it has gone to all the iterations necessary to have a working assembly you can reprint them in Nylon or PETG or whatever other higher cost material the final product will be in and expect it to fit.