BTW, if you're measuring wall thickness, and it's off, then most likely the issue is with backlash on your x and y axes, so that theres some variability in your head position on each layer resulting in walls that are too wide when you measure at the extremes. You aren't really measuring bead width, but rather positional accuracy.
Also, bear in mind that because of acceleration and deceleration of the head there walls are likely to be thicker at the corners - the head slows, but the extrusion lags a bit. Plus filament width varies, adding or removing another 5%-10% to the extrusion amount. And since the bead has somewhat rounded corners, measuring the width accurately is pretty hard - the bead will tend to deform slightly as you try to measure it.
Recommended Posts
illuminarti 18
Filament density is nothing to do with slicing. It's just so slicer can tell you the weight of the part.
If the flow rate is down to 40% and you're still getting lines that are 'too thick' then you're not doing the test right, or your margin of error on the measuring is obscuring what's really happening (or your default settings are way off).
As we've pointed out before, that calibration routine is really nothing to do with modern slicers or printers. It's for calibrating Skeinforge, and I think mostly for dealing with constant extrusion-rate printers.
On an Ultimaker, the thickness of the lines that you get is a function of the declared filament width, the steps-per-mm setting for the extruder, and - for the first layer - the homed z-height, since that determines how widely squashed the plastic is. The extrusion volume can also play a role as it reduces the extruder drive efficiency. But, in general, the slicer and firmware make sure that the right amount of plastic gets fed in to exactly fill the width x height x length of the line that you are printing.
Furthermore, for actual printed parts, the finished size is determined by a whole bunch of things, including thermal shrinkage, nozzle pressure, temperature, part geometry, etc. One of the biggest challenges is going to be that the actual volume of extrusion is going to vary with the head pressure - and therefore the temperature, the print speed, and also the intricacy of the geometry, and hence the actual head move speed on each part of the print. The variation in extrusion amount can easily be 10%-20% as a result, which will swamp any fine calibration you try to do.
Trying to calibrate all those things away as you are just isn't worth it, in my opinion, and at best all you'll do is be able to print that test really well. As soon as you print something else, none of it will apply.
Far better to focus on simply getting the input variables measured right - that's primarily the filament width, steps-per-mm, and z=0 position. Then trust the firmware to generally do the right thing, and compensate flow rates or object sizes as needed on a piece-by-piece basis. There's nothing 'impure' about doing that - traditional manufacturing also has these kinds of compensations and post-processing steps built in. Each process has its limitations and tolerances; it's just not practical to try to eliminate all of the variables.
Link to post
Share on other sites