GregValiant

The end cabinets (that I suppose the drawers go into) might be able to be rotated open end up so they don't need support on the inside. When printing things with large flats on the build plate you can often get by with a skirt instead of using brims. They use less material and don't need to be trimmed off while still getting the flow going. So I would do the legs as a separate print with a brim (so they don't fall over), and the rest of the stuff with just a skirt.

In addition to gr5's suggestions you want the belts to "twang like the strings on a bass guitar". If they're loose they can jump on those pulleys you just tightened and cause that layer shift.

I started printing this Bender head with silky silver and the visor part came out great. When I tried to print the front half of the head the printer struggled and then the layer adhesion jumped up and bit me when I cut off the supports. I gave up and used whatever I grabbed and painted it silver. It looks alright painted, but I wanted that silky finish. .4 nozzle by the way. I haven't tried with the .6 yet.

Urethane needs 100 to 200°C to cure. Hotter cures faster. The ones I was involved with (cutting CNC paths) were aluminum molds and cured on heated tables. Come to think of it some used urethane molds poured around a "hob" (a part machined to pre-shrink dimensions) so a mold printed in ABS might work. Actually, a phone call to a rubber and plastics outfit and "Hey, how much for a thousand of these" might go a long way. You could pick the material specs too. I'm pretty sure you could beat the $3.50/pc. price all to hell since you don't need the brass.

I think the ones I've used were compression molded. Have you considered making a multi-cavity mold and pouring them out of 80-90 durometer urethane?

I think there is a problem with the model. The text appears OK as @fvrmr says, but the pins don't look right. The left model is from your 3mf file and the right is repaired using NetFabb. On your model you can see that there is a counterbore from the backside (it has support inside in the image). It is deep enough to make the pin portion disconnect from the base feature. That won't work because even if they printed (doubtful) they would fall off. The repaired model shows the pins connected to the base feature. The pins are essentially solid on the repaired model. The infill occurs due to the wall count = 3. I'm attaching the repaired file. Take a look because I'm not sure how it's supposed to be. cam_lid_mp04_fixed.stl

My pleasure. "I don't know why but here's how to fix it." never sits well with me. I like to know the "why".

It isn't the model per se (which is watertight), but it might be the way it was modeled. I was using .4 nozzle and line width, .2 layer height. All 3 images are layer 9. My only change was wall thickness. 2 then 3 then 4 loops. This is at 0.8 walls. The infill is into the door area. There are open rectangles throughout the wall. This is at 1.2 walls. The infill has retreated from the door. Note that the holes in the wall have turned into filled rectangles with the lines running at [0,90]. Subsequent layers cover the fancy rectangles. This one is 1.6 walls. It's what I would have expected for any wall thickness. Are there some artifacts in the model that might be leftover from combining/intersecting/cutting operations in the CAD program? Why does changing the wall thickness have an effect on shapes away from the walls? EDIT: I was working with the STL file. I achieved the same results with the 3mf file. GregFacade.3mf

When I opened it I couldn't figure out what was going on. I still don't know but I finally found a couple of parts. After deleting everything but one "TopBackLeft" (there were several) I got it to the plate but it wasn't flat. Using the Rotate Tool "Lay Flat" command and it showed up in a printable position.

I have a nice system for these situations. It is powered by two of the printer stepper motors so they must function. It requires a table the same height as the window sill, and two pulleys (one to open the window and the other to drag the printer over the sill). The system was designed with PencilCAD so no translation software is necessary (though you will appreciate that Email won't work). There is one constraint regarding printer defenestration: A first floor window above a flower garden doesn't work well. A 10th floor window above a concrete parking lot is very satisfying (unless your supervisors Ferrari is parked below).

@gr5. - Here is the x-ray view.

From 6 to 80 eh? That's a better bloat ratio than Microsoft!

I've made several posts here about using Pause at Height. The main thing is that you have to set it to "Layer" and not height. If there are Z-hops in the file then Pause at Height "by Height" can be fooled because the Z keeps jumping up and down. A layer is a layer except that in a gcode file layers start at 0 and in Cura they start at 1. So if you look at the Cura preview and see you want to stop at layer 10, in the plugin you would put layer 9.

I don't know and I truly hope I never have to find out. Try posting a question on the GitHub site.

PLA and ABS PETG. Read up on ABS before jumping in. It's a whole 'nother animal. PETG is much easier to work with, doesn't have noxious fumes, and the mechanical properties aren't that far off from ABS. I don't like to advertise because someone always takes offense. I'm in the US and I tend to stay with MatterHackers Build Series PETG and PLA. It's their bottom end filament line. It's well packaged and they go through a lot so it's always fresh. For what I do it's fine but they do run out and there can be a wait. On Amazon you aren't ever really sure what's going on or how long it's been on a shelf. I ordered a couple of rolls of filament from a vendor there (MatterHacker was out of stock) and it was packaged OK but the moisture content was very high. It was essentially unusable until it had spent 8 hours dehumidifying. The Silky colors are nice but I find that layer adhesion isn't very good and they are better suited for art than for functional parts.

I just came across this site "widgetsupply.com" . I don't know anything about them but they carry a lot of hobby tools and stuff.

That's why it's orange then. The recommended Max layer height for a .4 nozzle is .32.

It depends a lot on the printer and firmware. Sometimes allowing the build plate to cool off will pop a print off. Once its location is lost, all is lost. It's the same with the stepper motors losing position. You didn't mention your printer and I don't know about the Ultimaker models, but on my Ender, the maximum time I can enter for the stepper motor disarm is 3 hours. After that the steppers lose their position. If that happens then a homing move is required. Sometimes the print is in the way and it can't be done. Even if the print isn't in the way, the end-stop switches on my printer aren't consistent regarding the exact stop point so a small layer shift (up to .5mm) can occur. If I pause a print to change filament I really don't like to have to Home the print head. With the disclaimers out of the way...depending on your printer and how the firmware is configured, you can pause the print and then adjust the bed temp back to what it was during the print. That at least eliminates one possible failure mode. The whole process really has to be practiced so you know what you are doing and how to overcome any hardware/firmware constraints. On some printers all you have to do is shut the machine off and sleep with your fingers crossed. When you get up in the morning turn the printer on and it asks you to resume the print. If the print didn't move on the build plate it should just go back to work. Shutting off while it's printing support is the best situation with infill being second. Print a calibration cube and after the first few layers are down practice different methods of pausing and restarting. No matter how you do it - if the nozzle stays in contact with the print it will leave a divot and may glue itself to the print. Re-starting like that would be a bad thing.

Having an Exacto modeling knife, small putty knife/scraper, micro-files (sometimes called needle files), a pick set (like dental tools), sandpaper of various grits, long-nose tweezers, a small pair of side-cutters (nippers), needle nose pliers, etc., can make life a lot easier when it's time to remove supports and clean up a print.

Red means illegal. I think orange is legal but "not recommended" or "beyond calculated value" or "do you really want to do this?". If you are running a 0.6 nozzle and want that initial layer height at .666, then yeah, it's going to turn orange. If you make the layer height 0 it will turn red.

Glad to have been some help. By the way, the "speed factor" in the plugin refers to the time difference that might exist between what Cura says a print will take, and how long it actually takes. It's a percentage based calculation. If the Time line in a Cura gcode file (second line) says 34222 (seconds), and it actually takes exactly 10 hours to print (36000 seconds), then 34222/36000 = 0.95. That is the speed factor for that print. You would want a fair sampling of print times to come up with a good average for the "offage" and enter that average as the Speed Factor. My thanks to Louis Wooters for writing ShowProgress.Py.

For more joy... The next thing is a problem between the 4.2.2, 4.2.7 mainboards the the TFT screens. M0 (used in the Pause at Height plugin), M1, and M117 don't work (on some) because they send text messages to the screen. Once again it's a Creality issue. There was a post here and the OP said he installed a different firmware flavor and the commands started working. Then he disappeared without telling anyone what flavor it was. The folks at Marlin aren't happy with Creality and won't discuss it. I hit a dead end there and lost interest. I was considering a 2-in-1-out shared heater hot end. I'm going to pass. I would have to change mainboards and firmware and it just sounds like a headache.

If the firmware doesn't understand arc moves then any G2 or G3 commands are ignored by the printer. The printer then cycles through the gcode file until it finds a line that it understands. Extrusion jumps to a new value and causes massive over-extrusion along a line that shouldn't have been connected in the first place, and in fact may be at a significantly different Z elevation. My Ender 3 Pro with Marlin 1.1.8 understands G2 and G3. It sees this Arc Welder gcode file like this. A printer that doesn't understand G2/G3 sees the same gcode file like this. It would be a mess. If you don't understand your printer/firmware capabilities you will often be disappointed, or even worse, surprised.

In the year or so that I've been hanging around here I've developed some feel for the people involved in the work on Cura. I'm sure that some would be able to give you guidance into what the thoughts were behind the software design. Whether or not they have the time (or inclination) is another thing. The Cura Team is heavily involved in the new Arachne version along with maintenance and support of the current and past versions. Additionally I believe most are Ultimaker employees which will no doubt have some political implications regarding how they spend their time and what they give away for free. Firmware design, mainboard design, and the physical design of your machine, are other things you have no doubt considered. It would appear to be a fusion between FFF/FDM gcode generation and translation and CNC gcode generation and translation. I would call it an ambitious endeavor. I know some of these are directly involved in the Cura Team at Ultimaker. In no particular order - @ghostkeeper, @nallath, @fvrmr, @rburema, @SandervG and @ahoeben at FieldOfView. In addition @burtoogle (@smartavionics on GitHub) has his own thing going with Cura itself. I hope none take offense that I've included their handles here (and that none I've left out take offense). As for me...I know just enough about this stuff to hurt myself.

When designing scale models you must keep in mind that detail will be lost. When designing scale models for FDM more detail is lost as the printer isn't scaling smaller. There is a point at which a detail disappears from the preview screen because the printer is physically incapable of printing it. I mention all of that because a primary design constraint must always be the fact that the part will be FDM printed and the printers are good, but they don't scale downward beyond going to a smaller nozzle. Keep it simple and your success rate will be high.