Active Leveling Could Not Be Performed Please Check the Nozzle and Bed and Try Again
Week 5 : 3D Scanning & Printing
☛ Grouping Assignment
☐ ane. Test the design rules for your 3D printer(southward)
☛ Individual Assignment
☐ 2. Blueprint and 3D print an object (small-scale, few cm3, express by printer time) that could not be fabricated subtractively
☐ 3. 3D scan an object (and optionally print it)
☛ Testing the design rules of our 3D Printers
This week we are discovering and testing the 3D printers at the Fablab Digiscope, several models from Ultimaker.
☛ Ultimaker 3, Ultimaker iii Extended, Ultimaker S5
At the beginning of the week only 2 of the five Ultimakers installed in the lab were working properly, mainly because in that location is less users at the moment. As we are at present more users for those machines we decided to begin by cleaning the machines before using it : the print cores, the built plates, inside the box itself. This was a very useful first encounter it permitted us to understand how information technology works.
We hands found tutorials to go through these steps. This one, the Hot and Cold Pull is the procedure to make clean the print cores. Nosotros didn't have Cleaning filament and then nosotros took white PLA Filament instead, the beginning four pieces of PLA in the picture below are demonstrating the Hot Pull process. It consists in inserting filamanet through the caput until it resits when touching the hot print cadre. Then pull it of in a motility, and doing it again until the filament becomes articulate. The three adjacent pieces demonstrates the Cold Pull process, the 3D printer heats the impress core, tha filament has to be inserted, but instead of pulling it off right away we have to wait for the nozzle to lose heat. When the print core is maintained at a low temperature nosotros can pull information technology off. The process has to be repeated until the filament is clear.
The next footstep is to calibrate the congenital plate, so the print core can allow out filament in an harmonious mode. Here a several tutorials from Ultimaker to set up their 3D printers properly, the process is slightly different regarding the interface of the model, only the procedure is basically the same, plus in comparing with of the Ender 3, Ultimakers model are very assistive during these steps.
The last link is the error one of the printers shows, we didn't have time this week to wait into it deeper, and so this is a reminder.
- Inserting filament
- Levelling the Ultimaker three build plate
- Levelling the Ultimaker S5 build plate
- ER17
☛ Ultimaker three / Nozzle 0.4mm / Layer summit 0.1mm
To get through his assignment we decided to use the files provided in the grade page. Printing these models immune the states to discover the limits of the printers regarding their capabilities and the parameters nosotros were using. Hither is a tabular array showing the results we obtained :
| Test | Minimum result |
|---|---|
| Angle | 30°-twoscore° |
| Overhang | 1-2mm |
| Bridging | 14mm |
| Wall thickness | 0.4mm |
| Surface stop | Neat |
| Dimension | - |
Ultimaker designers probably thought nearly this item moment in the users practice, taking flick in course and at the end of the press process. This is neat, maybe too much, but this is a satisfying display.
☛ Ultimaker S5 / Nozzle 0.8mm / Layer summit 0.2mm
This printer was longer to put to work, information technology kept showing error messages like "Nozzle probe failed. Please check the nozzles and bed and try again", "Active leveling could not exist performed. Please bank check the nozzles and bed and attempt again.", "Difference betwixt detected summit of both print cores exceeds realistic values.", or "I decided I will non work for the moment, but something check somewhere else and try over again later."
This was frustrating because at first I checked the nozzles and bed the three first times it asked me for, so I decided to endeavor again without checking anything, and it worked. It was a fast task, then I started it once more, and nope again.
The printer became more than compliant and we could 3D print the scale tests. Here are our results :
| Test | Minimum result |
|---|---|
| Angle | 30° |
| Overhang | two-3mm |
| Bridging | 16-18mm |
| Wall thickness | 0.8mm |
| Surface finish | Rough |
| Dimension | 1 side is larger thant the other .8 mm of difference |
☛ Ultimaker iii Extended / Nozzle 0.25mm / Layer height 0.1mm
The Ultimaker iii extended is the Ultimaker iii simply higher. We installed a 0.25mm print core, so we have 3 differents settings nosotros can test. Here are quick experimentations on cubes using Cura Special and Experimental Modes.
- The first on the left is a 10x10mm cube with a brim.
- The second is a 15x15mm cube with "Spiralize Outer Contour" and 0.3mm "Fuzzy Skin" ticked.
- The third is a 20x20mm cube with 0.3mm "Fuzzy Skin" and "Wire Printing" ticked
☛ Creality Ender 3 / Nozzle 0.4mm / Layer tiptop 0.2mm
I made the acquisition of a Creality Ender 3 almost a year agone, it is 18 times cheaper than an Ultimaker 3, and and so far I recollect that it is a more than comfy machine to use because you can make errors on it that volition clearly show yous what'southward is mechanically wrong. I found that this is express on an Ultimaker because it doesn't permit you that power, information technology shows error messages before. Simply yes, okay, you tin can simply plug and unplug a new nozzle, basic tasks equally the calibration travel and heating/cooling when it is fourth dimension to change the filament or cleaning the print cores are mostly automated and supported past sensors. And, yes likewise, the Ender three is a very noisy machine (regarding this, upgrading to the silent motherboard V4.ii.7 has been a real comeback regarding noise emitted. The noise remaining is mostly caused by the fans cooling the extruder and the nozzle.), it takes a large corporeality of time to calibrate the build plate the start times, and changing a nozzle is much more challenging than with an Ultimaker. Fifty-fifty with that, my preference goes to Creality.
I reproduce the calibration tests using the Ender three, and the results are quite satisfying, particularly regarding the angle of printing and the bridging tests.
| Test | Minimum outcome |
|---|---|
| Angle | x-20° |
| Overhang | 2mm |
| Bridging | 20mm |
| Wall thickness | 0.5mm |
| Surface terminate | Cool |
| Dimension | Ane side is larger thant the other . mm of difference |
| Clearance | 0.2mm is tight just moves, 0.1mm is stuck |
☛ Blueprint something that could not be fabricated subtractively
For this part I tried to brand sort of airless tires for the "Drawing Car" I designed during Calendar week two. I exported a STL file from my previous work and imported it in MeshMixer Autodesk. Hither are the steps I went through :
- Select the model and display meshes, then
- Click Edit > Reduce fifty%, and repeat this footstep : information technology works ameliorate if there is less triangles
- Clear the Pick
- In the left panel, select Edit > Create Pattern
- In the window appearing select Dual Edges pattern and alter the parameters until it fits, and click Have
- You tin then export the file
Hither is the 3D model of this object on Sketchfab
Here is the 3D printing endeavor with the Ender 3. I reduce the size of the tire from 80mm of diameter to 50, so information technology would take less time to impress. Thinking twice, this was not a good idea as the meshes were already 1mm of width. The result shows a lot of stringing on the printed model, it probably means that the retractation of the filament is too ho-hum in that endeavour.
I made another try with the Ultimaker 3 Extended, using a nozzle of 0.25mm. I kept the reduced size of the model for this test as well, there is less of the stringing result, but at that place is still. I had to use an adhesion support, between the model and the build plate only, because it would not attach to it without. This is non ideal equally it is difficult to pull it off of the model without dissentious it.
Here is some other effort, still using the Ultimaker 3 Extended, with 0.1mm of layer top, 20mm/s of travelling speed, 28mm/south of retractation speed, and "Enable retractation at layer speed" ticked, in Cura slicer. Stringing is less present only the brim is all the same hard to get rid of, and the lesser layers (the first layers printed) are not very smooth. At this signal I could try to lower the heat of the print core, or merely redraw this model and then the angle is easier to produce with this machine.
☛ 3D Scanning
Sample text
☛ 3D Scanning with the Sense (failed attempt and so far)
I tried a long fourth dimension to understand why it would not work, merely it doesn't work, I installed both versions of the software, in case I was wrong from the beggining. I probably missed a pace and information technology took me hours non to succeed in making it work with my calculator.
3D Systems Sense 5.1 Sense Scanner softwares Intel RealSense SDK 2.0
☛ 3D Scanning with a Kinect
Fortunately, we take a Kinect at the Fablab, I had issues to find a software that would piece of work with it. I first followed a tutorial to use it with Skanect software but it did not detect the Kinect and I all the same don't know why. I installed Kinect Studio only it asked for a Kinect Enabled App, and then I went through installing different softwares that are "Kinect Enabled", none of them worked. I tried 3D Scan, whiwh was suggested in the Microsoft app store, it detected the Kinect at outset opening so I went with it.
Kinect for Windows SDK v1.eight and the Kinect for Windows Developer Toolkit v1.8
"4 Kinect enabled apps" 3D Builder 3D Browse
The process is very unproblematic, as minimalistic equally the interface. In the 3D Scan left panel it possible to
- Adjust the depth
- Suit the width
- Suit the height
- Rotate the Browse zone
- Add a timer
In the main body you can Scan and Abolish the scan.
The starting time test I fabricated were using my desk chair, which is the only rotative thing I could recall of and find at my place. I installed the Kinect on a tripod and adjusted the parameters to scan properly.
Once the browse is done, 3D browse constructs the meshes that can be open up in 3D Architect. This was my showtime time using a 3D scanner, so the issue is quite pleasant. I tried to makke a closer scan, and modified the lighting "set upwards" I have (lamps). In that location is less details when the scene is nether-lighted, but too much calorie-free seems to distort the subject. (Or it could be me).
I tried to scan a toy from my nephew (I tried to browse my nephew first, but the "not moving" part was complicated to achieve). And I think the light set-up is in crusade in that result, in that location wasn't enough low-cal.
I selected the more detailed scan I had fabricated so far and exported it top open it with Cura Slicer without whatsoever modification.
Information technology was printed with support, as the model is not "plane", and with a brim considering it would not adhere - adhesion bug on my buildplate ? After two hours and ten minutes :
☛ Photogrammetry setup
I really wanted to endeavour photgrammetry with a turntable and then I made one. I took a part of an older project, a big laser cutted gear and a continuous servo. The continuous servo was installed in a carte board box, the gear was slightly modified so the servo can make it turn. And then I double taped a white piece of newspaper as a clear background.
☛ 3D Scanning using ReCap from Autodesk
Recap Photograph is an Autodesk scanning software.
Once I had took all the pictures I needed, I tried to import them in Epitomize. The amount of images you tin import is limited, 100 is the maximum pictures y'all tin can process at a fourth dimension.
Here are three reconstructed mesh from my pictures. The first image is a little blackness elephant, but equally you tin can run into on the movie below it, the cameras are not placed all effectually the scene. Just it was a dark and cogitating material.
the second image shows the reconstruction attempt of a little pinecone I brought from Lahti in Republic of finland (very important detail), as it is a complex shapethe result is not totally surprising, plus the cameras aren't tracking efficiently once again.
The last i is a seashell, as information technology is a uncomplicated shape with a poor reflective capability, the result is more accurate, there is yet an outcome regarding tracking.
☛ 3D Scanning using MeshRoom
The link to Meshroom
I tried to import my pictures in MeshRoom.
☛ 3D Scanning using 3DF Zephyr
3DF Zephyr I downloaded 3DF Zephyr with the 14 days gratuitous trial license. I imported my pictures in it. I found the interface easier to understand than MeshRoom, and it is possible to edit the mesh afterward the reconstruction. This first row of images shows the reconstruction endeavor of the pinecone, only the tracking is a real upshot that I missed then far so I cannot do anything with that consequence.
I tried once more with the seashell, the editing features immune me to select a colour range an delete it. Merely I couldn't export a model anymore after this. This result could probably be resolved with more time.
☛ Photogrammetry setup update
Afterwards realising several test it occured that the fact that I was using clear white newspaper was making it hard for the software to runway the object, then the result would not exist useful as a scanned object. I took a grided paper and added dots on it. Ambroise told me that the result would exist more efficient if the grided pattern is not regular, a page of newspaper would be ameliorate because with all the block of text of different sizes the sofwtare can differentiate where it is effectually the objets more easiy.
A tutorial for using the masking tool on 3DF Zephyr and another tutorial to practice it with a turntable
I added my grided background and the consequence was improve, but the grided background is reconstructed as well. Ambroise advised me to await at those tutorials. The masking tool in 3DF Masquerade allowed to show manually what is the foreground (red strokes) and what is the background (dejection strokes) to the sofware. This result is cleaner than the previous attemps !
I tried the Confidence Tool in the editing section information technology makes a sort of topography of the model, this feature allowed me to take out the role that was more than than the seashell - only again, I couldn't export a 3D model after this step.
Hither is the 3D reconstruction of the seashell :
I imported the model in Cura and printed information technology. The result is missing details, simply information technology is already satisfying regarding the previous results.
I thought I would do it over again with the pinecone and using the masking feature, and I thought it would work this time. The offset prototype I saw on the software was already much more than detailed than my previous atttempts, simply I turned the model around, and only the half of it had been reconstructed. five images out of 24 were used by the software. It demonstrates that the quality of the browse does not rely on using an high amount of images of the object. And then far I learned that it does rely on the quality of the images, of the light exposure, on the fact that these expert images has to overlap on each other. A good event also depends on the abilty of the software to detect the tracking around the object using images, a blueprint that is irregular as a background is useful, its irregularity will be easier to track. The framing is also very important and wil affect the effect, when using the photogrammetry procedure with a turntable information technology is best to avert framing out of the turntable zone : the object is moving, but the groundwork is not, the tracking is therefore more complicated to process, and it creates a lot of artefacts. Using the mask tool is also helping to prevent this.
A 3DF Zephy capability is to give it video and it exports images out of it. I thought it could be useful to try, as the photographic camera is non moving peradventure the tracking will exist more than accurate.
And yes it was, it is the best result I got using these softwares so far - I'll have to find how to edit and export mesh using 3DF Zephyr (in the 12 remaining days of my gratis trial).
☛ Textures
This part is a Bonus, what textures from different 3D sofwares look like.
AutoDesk Recap Textures
AliceVision MeshRoom Textures
Microsoft 3D Builder Textures
‹ Previous : Electronics Production • Adjacent : Electronics Design ›
Source: http://fabacademy.org/2021/labs/digiscope/students/selena-pere/Week05.html