That’s exactly why it is a partial model/design to rapidly iterate on it: 1g, 10-minute print time. The full version with all the print-in-place parts takes 16 minutes.
Another aspect is collecting designs for a library. From now on I can copy past this subassembly into bigger designs and know it will just work. If I need to modify it I know how and where to change it to get the desired outcome.
If you consider sharing mechanical design concepts as not in line with the spirit it’s fine but others are likely interested in seeing how things work and takes it as inspiration for their designs.
Go and recreate it. Nobody stops you. Could provide the STL but wouldn’t be worth a lot as this is so dialed (tolerances) that it comes down to the specific printer/extrusion system. There are older revisions with huge tolerances (0.4mm) that work but wear down rapidly. To print this exact version it needs to be capable of printing with 0.23mm gap/tolerance between parts.
v10 should still be 0.4mm tolerances (easy to print) on all sides and working. Otherwise not a great design but enough for you to understand that there are dozens of parameters to tune in such a “simple” mechanism and it is (nearly) impossible to nail it on the first try. Have we started talking about optimizing the force required to break it loose? That’s one more thing that needs to be accounted for.
For motors, the limit is the wire insulation so might get away with 100°C ambient. If you can’t move them out there are also water-cooled extruders: www.dold-mechatronik.de/mebs-Hemera-WaCo-Mod-EN
Printing fast/without cooling can also go the other way:
By printing very fast the last layer may still be “hot” when the new layer is added. As the temperature differential is smaller there is less stress within the part once it is cooled down.
Regardless of the scanner use matting spray. Either a commercial that evaporates or baby powder + IPA or baking powder.
OpenScan was already meantioned by somebody else.
The CR scan otter won’t work for small parts and the CR-scan software isn’t great for example there isn’t an undo button to remove the last x seconds of bad scan data.
second failure: Prusaslicer overlapped support and the part. At least my hotend survived that failure.
Called it a day and moved on to a different printer for this print. Also, did I mention that I managed to kill another z-endstop on that printer today? Forgot to remove a finished print before running G28.
The end stop is located on the carriage (toolmount/receiver). Moved sideways into the print and the bed adhesion was stronger than the pin of the switch.
Quick and dirty 5 minutes craft: Draw a rough shape, define the contact surfaces & load, click run, and get the optimized shape. The last step is converting the output to a printable shape and running one more simulation to double-check it is strong enough....
Is there any software that does take toolpaths into account?
Not as far as I know. The next best option is to define anisotropic properties in the z-direction but this doesn’t close the gap between simulation and slicer output.
How good is it? Good enough. Work with safety margins and temper the print if it is close. An important aspect to make Ansys, Nastram, … work with FDM is experience/rule of thumb. Knowing how to read the result and how to set up a simulation/model to get close enough results.
Most valuable is where and how it will fail as this is pretty accurate. For the exact load capacity it is the simulation result decreased by some rule of thumb based on experience.
Inventor has a GUI for it. There are more options.
A slightly higher wall count for prints and the weight of the unoptimized solid is pointless in this instance as it starts with a larger slab and tell it to remove x% of weight. With the simulation result you either increase or decrease the x% removed setting and run it again till the load part strength is correct. Very basic in this regard but this was a quick design. How the spool is relative to the mounting points isn’t optimized. It was just me drawing something.
This particular part is 70g each with the filament being approx. 100mm moved forward (leaver length) and were simulated to withstand roughly 12kg and tested with 5.5kg+spool weights.
In a broader picture: See it as a demonstration of what all those nice tools in the CAD package can do. In this application with a little bit of thought could come up with a similar or better solution but for an I don’t care design approach the output is already good. A proper design approach would be putting thought in in where to place the contact surfaces relative to the spool and then run this software or go a step further and allow a different software to also change that parameter. Keep in mind those simulations are computationally expensive. Complex/advanced questions might take days to solve while a simple question like this is less than 1 minute.
The load was in the circle/groove facing down.
The other constrain was the faces contacting the 3030 extrusion being fixed and a keep-out zone was defined around those to ensure no material there was removed.
Otherwise, it was just a flat slab as shape.
What at first surprised me was how this part works: There is a point defined by the lowest/left triangle (tension & compression) on which all the weight rests. The remaining structure is is a cross beam (top mounting point to spool) to support it (tension) and the structure on which the spool rests (compression).
OpenScan is a opensource project for smaller objects. With the PS5 maybe a Creality CR-Scan but it’s quality will be borderline unusable for a shape like a PS5 panel. So after all the entry scanner for a job like this could be a Shining Einstar.
Good news: If you own an iPhone try it’s lidar first. Creality CR-Scan is only slightly better than iPhones. With Android phone, you could try photogrammetry but to scan the PS5 part you would need matting spray and even more tracker (small dots glued to the surface).
Btw. Somebody somewhere at some point in time already scanned or modeled the PS5 side panel. As starting point check GrabCAD and thingiverse for a 3D-model.
How much do you need to turn the knob for 0.3mm? Most people don’t know the answer so they make a guess which is likley wrong.
If the firmware converts 0.3mm to 1/2 rotation it is clear what they should do. This particular printer probs itself after the adjustments and if it isn’t right it shows once more the dialog/instruction with the adjusted rotation value. After 2 rounds, you would already have accounted for the manufacturing tolerances.
This process can be ontop of ABL and z-offset dialog (eyeball it with a shim, testprint, tell the printer which looked the best, the firmware knows what z-offset was each of the 5 prints/lines).
a lot of these printer manufacturers consider it an extra cost since these are subtle things when compared to the manufacturing and mechanical design of the printer.
If there would be a $150 Ender 3 and a $200 one with identical hardware but perfect usability for beginners. I expect more people would buy the expensive machine compared to releasing a noname Ender 3 clone which tries to beat this hypothetical $150 price point.
An example would be the AnkerMake M5 and BambuLab lineup.
I have seen on some budget printers a label with which direction is up and down so that’s a starting point. Also a lot of prints out there to add this information to the bed leveling wheel.
Glue stick also contains PVA. Issues I had is that it stays sticky and cleaning the buildplate can be difficult.
With a pure or close to pure PVA-based slurry/solution it dries up to a film and as such isn’t sticky to dust and finger but still sticky to filament. Further, it can be easily cleaned up with water.
How to apply it? Brush it on. If there isn’t any brush nearby I spread it with my fingers.
Print in place ratchet design (lemmy.ml)
After half a dozen iterations, this was the first reasonably working, acceptable feeling, and good-sounding ratchet mechanism....
[SOLVED] Trouble with ASA prints (sopuli.xyz)
[SOLVED] cause: not enough cooling for overhangs...
Need some advice on scanning
I’m having a bit of a hard time with this, but I also acknowledge this is still relatively niche still....
Shit posting but 3d this time (files.catbox.moe)
Print failed because the layer adhesion is shit on my printer.
Shape optimized spoolholder (lemmy.ml)
Quick and dirty 5 minutes craft: Draw a rough shape, define the contact surfaces & load, click run, and get the optimized shape. The last step is converting the output to a printable shape and running one more simulation to double-check it is strong enough....
Sovol SV08 debuts Apr 18 and <$600 US (lemm.ee)
This looks right in my price range for my next upgrade. I want a core xy, I’ve been on bed slingers, so this looks good....
How do you model complex curves? (kbin.run)
Can someone explain or point me towards a good tutorial that explains how to match complex curves, like the PS5 side panels?...
Why do 3D printer manufacturers not get the details right? e.g. rotation indicator on bed levelling wheels (lemmy.ml)
Building a 3D printer is easy. Getting the details right to build a great 3D printer is hard, as this is where most companies fail. Why?...
Elegoo Neptune Pro 4 - Failed Print Help [SOLVED]
I just sent this email over to elegoo official, so either way I’ll update this post with their response for a community reference point:...