I designed a rack in freecad and I try to print it now. It consists of several subparts and although I have done the following steps and iterations the result is not optimal. By far not optimal. Here is a testprint with extruded letters and without:
the build plate is wiped and cleaned with Isopropanol after every print
the AMS has printed housings full with silica gel → 20% moisture inside according measurement
Settings:
0,4mm nozzle
PLA Matte
The rack is mainly printed with 0,16mm high quality settings in bambusstudio. I changed amount of walls to 4, some prints have 6 walls for stability.
25% infill
8 top layers, 1mm total thickness (German: obere Schalendicke)
→ all prints have spots with slight over extrusions
→ the biggest issue is like shown in the picture above.
→ after failing multiple prints I did a complete new calibration. With the standard 0,2mm profile in bambusstudio without any changes the print still looks like the picture above.
The nozzle seems not to be clogged. I have zero idea how to solve this. Can you give me hints / solutions?
The print’s contact surface to the print plate is a mirror image of the print plate (with dry, well calibrated filament of course). So you’ll get a really good surface finish printing upside down on a smooth plate.
As an extreme example, you can even imprint holographic surface texture effects into flat surfaces this way using 3D Effect Sheets.
Also, if you’re using Bambu PLA, I would try to print using the default profile for your Bambu filament, checking the boxes to bed level + flow calibrate, see if that yields a different result.
Thanks for clarification. Ill buy a smooth plate and try it out.
@ printing with default settings:
I did it. Unfortunately the result is the same. And I dont understand it yet…
It is very frustrating.
@ printing up-side-down:
Also here unfortunately it will not work. Please see the attached picture. The string from the first picture was just a small portion, cut out of the bigger model. The bigger model itself has standoff geometry which is higher than the string surface. I hope you understand what I mean. In this case I would use enormous amount of support.
The string over-extrusion was just an extreme example. My surface overall has here and there spots with over-extrusions. I am seeking for support how to solve this. I am printing for ~1month now and my first models were perfect.
Ah, that does explain the desire to print the text as is rather than upside down…
I am not certain what you have tried, yet.
If you really need to print right side up, did you enable “wipe on retract” in the filament settings and for the text object “Arachne” wall generator with a wall count of 999?
You may also have a chance with ironing, but unfortunately, I can not advise so well on that.
Addendum: Printing text well upside down on top of support would require a 0mm z-distance between the support interface and the part. So you would need to use a different material for the support interface. PETG or PVA/BVOH for PLA…
Redo the flow rate and flow dynamicscalibration for each filament.
From looking, it is clear that you are extruding more than needed.
Aftercalibration, try some simple prints with your current settings to assess the surface quality → e.g. three cubes of 30x30x3mm, one for grey and one for orange colours filaments; if both are successful, try with the third cube some multicolour as in your print; with this, avoid filament waste, and make tests faster.
You can keep the bed levelling as a startup routine, but remove the flow calibration, as otherwise you wipe the previous calibration values (if you already do it, ignore);
Infill type: —> you didn’t specify, but if not already using, use gyroid; also, ensure that some void space exists.
That would only mask the issue until it shows up again in another model. There are clear indications of too low PA/K-value here, so printing with the text facing downward does not actually solve the underlying problem.
okay, although I did a calibration with “auto calibration” for each filament I will repeat this process with the manual calibration. Obtaining the correct K value by iteration. This is new to me, cool.
I will retry afterwards the a print and post it here.
The automatic calibration has never really worked for me—not on the X1 Carbon and not on the H2D.
Just last night I tried a fully automatic calibration on PETG-CF, and I have to admit I was quite disappointed. The correct values ended up being off by about 26% compared to the actual sweet spot.
It’s worth remembering that dynamic flow, PA/K-value, and temperature are all interdependent. Dynamic flow should always be established first, but to get reliable results you need a reasonably correct PA set beforehand and avoid changing it during the calibration. Running the calibration in auto mode will only distort the outcome, since the adjustments interfere with each other instead of isolating the parameter you’re trying to tune.
Also, if your PA value is way off when calibrating dynamic flow, you’ll end up chasing a flow value that will never actually be correct.
Okay, I did a print yesterday and still there were small spots of over-extrusion. I don’t speak about the text, I speak about plain and ideally smooth surfaces.
I retried now the manual calibration as follows:
1.) dynamic flow first try
0 to 0,025 in 0,05 steps
→ good first value 0,015 (instead of the automatic calibration one with 0,022?!)
2.) dynamic flow second iteration
0,013 to 0,020 in 0,01 steps
→ better value with 0,013
3.) flow rate fine calibration
There was a predefined value of 0,98 ???
Is this a value from the printer, measured and found due to over extrusion? Is this some kind of correction factor for less extrusion?
Anyway I did the process and the test-bodies look like this:
I’ve succeeded with automatic calibration on the X1C(with the 0.4mm nozzle), yet I always test it after, and repeat if not happy. My tests are mainly for tolerance and accuracy, which differ significantly from aesthetics.
Concerning your tests:
the 0.98 seems to be the default value, e.g. BL PLA Basic has by default a flow rate value of 0.98;
but if you use the BambuStudio Calibration, the result value will override the default one, and get stored there; e.g. on the left is the default BL PETG-CF and on the right is my own after calibration:
In the manual flow rate calibration, you can choose the complete calibration based on the flow rate value. When I run it, I start with the complete calibration and then fine-tune it with a new run using the flow rate from the previous test.
The lines should be visible, but in your case, the distance between the lines seems, in many cases, larger than expected, which is the opposite of what was noticed in your earlier prints. As your goal is an aesthetic surface, I wouldn’t see these results as acceptable, so I would try running a second-stage calibration to fine-tune the results.
The middle region of those probes should be smooth, without visible lines, while at the same time not overlapping lines.
I agree that it isn’t easy to identify the best, but you should expect difficulty from many samples with a smooth middle section and not the opposite.
Edit: check BL wiki images, to see what to expect:
In general, if your best result is at the end of the tested range, you want to extend the range in that direction far enough that you begin to see worse results in that direction. You may have found a better result with something less than 0,013.
In your picture, I think the zero sample is best (but not right) and the rest become increasing worse. The “valleys” between the lines get deeper as flow is reduced. The lines are not squishing tightly against each other, a sign of slight underextrusion.
The predefined values are not wrong, just starting points. They work for well enough to produce a print, but calibrating the filament will nearly always produce improvements. There are always differences in manufacturing batches, colors, printers, printing environment, etc.
To calibrate flow without that additional step, you can print a small hollow cube without a top surface. For example, a 20×20×20 cube with no infill or top layers and X walls. Then calculate the total wall thickness — for instance, three walls should measure 1.32 mm if you keep the default line width. For some filaments, you may also need to use a random seam; otherwise, the wall at the seam can appear inconsistent. Next, take 1.32 ÷ the average measurement of all sides × the current flow ratio to obtain a new value that will be much closer to the correct one. However, this method assumes you already have a reasonably accurate PA value; otherwise, you will end up chasing an incorrect flow ratio here as well.
This 
& 
