GregValiant

Yea, it's installed but I'm not fond of it. When I select "fix model normals" the menu goes away and there's no way to tell if anything actually is being done to the model unless you check the mesh again. I'm stealing the thread here but one poster had a model that was watertight but was composed of some sub-meshes. Cura ignored it but 3D Builder was able to fix that. They are all tools in the toolbox.

Sorry. I misinterpreted "off the bed". Open the Gcode file. Right near the beginning will be a line ";MINX" and one for ";MINY". If either of them is negative then that's the problem. In Cura click on the name of your printer. Choose Manage Printers and then Machine Settings. On the left there is a box "Origin at Center". Make sure that is de-selected. Most printers have the origin at the left front corner. Delta printers have the origin at center. The Home Offsets in the printer, coupled with the Bed Size in Cura also determine where a print will be on the bed. But first make sure Origin at Center is not checked and then slice something. You don't have to print it, just check the MINX and MINY to insure it will print in the positive octant.

Somebody can look at the problem. Post a gcode file. If you have one that works and one that is broken and post them both that would be best. The printer definition files are either provided by a member of the community and submitted to Ultimaker, or the manufacturer of the printer contracts someone to write the definition for them and they submit it to Ultimaker. Until someone gets around to that there won't be a Pro2 definition in Cura.

Your Print Speed (15), Accel (500) and Jerk (7) are really conservative so I don't see it being a belt skipping (even if it was really loose). I read the gcode file into AutoCad (to take Cura out of the loop). Everything looks good in the gcode. All the speeds are right and everything lines up with no layer shift. The models themselves are watertight. Each consists of 3 sub-meshes. That is repairable but that doesn't have anything to do with a layer shift because the resulting gcode is good. I think that somehow someway it's in the printer. I'll think on it but right now nothing is coming to mind. Maybe @gr5 has a thought. @Mari also has an E3Pro and might have seen something like this as well.

I brought one of the bad files we've gotten lately into Prusaslicer. Prusaslicer sliced it without complaining. It does have a "Repair STL" command in the File menu. When selected it brings up an "Open File" dialog and you can pick and repair the file without bringing onto the build plate. It also informs you if the repair has succeeded or not. It would be nice to get an acknowledgement from Cura as to whether a repair was successful or not. I tend to use NetFabb (thanks for that one by the way) and I'm trying to figure out MS 3D Builder.

Like on normal support it controls the amount of support infill inside the trunk or branches. At "0%" you get the outer wall only. At 50% you will see a lot of lines inside the trunk and branches. Below is the definition from the "Cura Settings Explained" webpage. You can also locate it because it is in the sticky "Cura Settings Explained" that is always the first posted thread when you visit the forum. That may have been why there were no responses to this thread. When you go to the page there are example models with different Support Density. Support density This defines the density of support structures. A higher value will lead to stronger support, but these will be more difficult to remove and take more time to print.

In Cura use "File | Save Project" and post the 3mf file here. That will contain the model, the printer, and your settings. You can post the problem gcode file as well. Somebody will take a look to see what is going on.

Yes it was. But even after the repair it wasn't suitable for Vase mode. Now you know what to look for in the model. I've been noticing that lately, both here and on Reddit, there have been some comments like "I opened it in PrusaSlicer and it worked.". What I think has happened is that Cura has gotten fussier about errors in the models and PrusaSlicer (at least some versions) skips over some model errors that Cura catches. I think as PrusaSlicer evolves that it will also "notice" more errors.

Gcode is a simple machine movement language developed at MIT in the 1950's. I started working with it in 1969 when I started on my ME degree. It evolves slowly because there is a huge installed base of machines that use it. When the printers came along it was modified to include their specific needs. The Marlin gcodes are at Marlinfw.org and the RepRap gcodes are HERE. That pretty much covers what you would need and just reading through the descriptions is enlightening. Firmware features can be enabled or disabled. One flavor of Marlin in one printer may accept G60, but it is not enabled in my printer so it doesn't work for me now, but it could if I re-compiled the firmware with the option enabled (if my little printer had the memory space which is doesn't). Gcode is linear in that each line of code is a single command and the lines are read by the printer/planner one at a time. In the case of movement commands like G0/G1, the printer/planner calculates how fast to send steps to each motor so they all arrive at the indicated place at the same time. This makes hand coding some things relatively easy. (Coding a path that describes a big "S" is not one of them. Have you ever tried to hand code a spline? I have. It's a totally useless and time consuming skill.) There are a couple of firmware types that have logic capability ("IF" statements are possible). I don't know their names. Mcodes generally make adjustments to the machine (M104 sets the hot end temperature on a 3d printer but might do something totally different on a CNC milling machine because the firmware interprets it differently) and Gcodes generally are for movement. Open source usually means that anyone can contribute to the development of, or use the software. It does not mean non-profit.

I think this will work. In the Cura Pause At Height dialog box there is an optional "Gcode after pause". If you enter G4 S180 in the box then the printer will do the correct pre-pause routine, skip over the M0, and then "dwell" for 3 minutes. In the G4 line the S parameter is in seconds so you can adjust it to suit yourself. Once the pause happens the dwell cannot be shorted - the printer will be paused for the amount in the S parameter. Make sure you have the "Disarm Timeout" set to 1800 so you don't loose the stepper locations. In any pause routine the re-start cannot involve a G28 (Auto-Home) as that slightly changes the zero of all the axes (switch error) and results in a layer shift. That doesn't always happen, but it happens enough.

Those are the same people who regularly print at 1000mm/sec. Are your E-steps calibrated? That would be the first thing to check. You can tune the "Flow" on the fly. You might try that to work towards an acceptable finish. I've attached a little ditty I wrote. It's a Windows app (I don't know what OpSys you are using) that I call "Greg's Accel and Jerk Tool". If nothing else it's an interesting toy. Start out slow as high speeds with high Accel and Jerk numbers can cause missed steps or even belt skipping (we are talking about Creality printers after all). It's an unsigned app and if you decide to try and install it you might have to convince your anti-virus to let it install. It requires a USB data connection to the printer. Greg's Accel and Jerk Tool.zip

"But an Ender is a whole 'nother level of cheap..." Amen brother. Every body who owns one ends up replacing the fans at one point or another as the sleeve bearings wear out or the fans just quit. This is not to mention how noisy they are. And the layer cooling blower is a proprietary design so the only direct replacement for a stock piece of crap is another stock piece of crap. It's one of the main reasons that so many owners move to a 5015. (I don't get a lot of the fan contraptions I see on Enders though. They are borderline ridiculous.) And it is my (never humble) opinion that the switches are every bit as good of quality as the fans.

@gr5 10 years ago they were probably "high quality" Taiwanese switches. When I have to restart a print and I'm forced to include a G28 there is almost always a layer shift of up to .5mm. It's the switches on the Ender. My Z switch is actually pretty good but the X switch in particular is terrible. The 1 out of 3 gives me an Acceptability Rating of 33%. I think the most I get out of a "leveling" session is about 5 prints. Part of that is removing the glass to get the print off, but 30 days on this thing? Ain't no way.

In order to use "Spiralize", the model really needs to be a solid. Spiralize then figures out where the outer walls should be, gets rid of the top, and generates the spiral path up the wall. If it isn't solid, then Spiralize can't quite figure out a path that works and you get things like below (your model) where you can see the inside path is over air and would just make a mess. Below is also a vase model from Thingiverse. You can see that it is solid throughout. It slices as we would expect but because the outer "gear" features are so close together a 1.0 nozzle wouldn't do well with this one as a lot of definition would be lost. You will come to find that the stock hot end on the Enders (and CR-10's) aren't very capable. Consider that a 0.4 nozzle at 0.2 layer height results in a .08mm² area to fill while a 1.0 nozzle at .5 layer height is .5mm² to fill. If you print with a .4 nozzle at 60mm/sec then the flow through the nozzle is 4.8mm³/sec but the 1.0 requires 6 times the amount of plastic. To get the same flow through the nozzle (that we know it can maintain) with a 1.0 nozzle you would have to slow the Print Speed to about 10mm/sec. You can print hotter to make up for the excessive cooling that the hot end gets at high flow rates (the plastic is removing heat), but at some point it just doesn't work right. It's one of the reasons people move to Micro-Swiss or Volcano hot ends on their Creality printers. They allow for big nozzles and higher speeds.

If you will post the STL somebody will take a look. Guessing from the image it looks like a problem with the model. Is that model solid through and through or is it hollow with a thin wall?

When you get a splice like that horn where you see a layer all by itself then that usually indicates either an error in the model, or a feature that is too fine for your nozzle and line width. When that is the case, Cura doesn't recognize that the area needs support because it can't interpret the model. First things first - go to Service.Netfabb.com and create a free account. Upload the model. The model will be repaired and then you can download the fixed version. If you haven't done so, enable "Shell / Print Thin Walls" (it can help with fine details) and re-slice the fixed model and see how it goes. Those teeth are going to be tough to get right. They will need support but they are so small that removing the support without breaking a tooth will be difficult. If you can post the STL file here then maybe @kmanstudios can take a look. I think you'll need an artist's advice more than a guy who prints plastic angle iron parts.

If you use "File|Save Project" and post the 3mf file here someone will take a look. The support blocker in 4.8 works for me but I may be trying something different than what you want.

That's an interesting question. These are some random thoughts. The G1 and G0 lines within a Gcode file will tell the printer what point to go to in space. That often looks like the axes are tied together somehow, but they aren't. Consider these two lines... G1 F500 X12 Y12 Z25 G1 X25 The first line tells the printer to move from where ever it is to the 12,12,25 point at 500mm/min. Within the printer/planner the calculations are made to have all three motors arrive at the destination at the same time...a linear move from point A to point B. Within the second line there is no mention of F, Y, or Z so there is no movement on the Y or Z axes and no change in the feed rate. The head simply moves in the "X only" to 25,12,25 at the previous speed of 500mm/min. So the X axis is independent of the Y and Z. That independence is also true for the Y and Z (and E). Your problem might end up being the feed rate of 1-10mm/min. I don't know the Ultimaker printers. Are their stepper motors and drivers capable of going that slow? I just checked my Ender by sending G1 X100 F1 and the actual speed was 24mm/min. That "minimum speed" is between 2.4 and 24 times faster than you want. I'd need a heckuva gearbox to get it down to 1mm/min - and I'd need one on each axis. There are finer steppers available, and there is micro-stepping but I'm thinking you'll still need a gear reduction unit to go that slow. You can make a printer go in an approximate circle with a single line of Gcode using G2 or G3. That places the calculations for the circle on the printer/planner whose "discrimination" is dependent on the firmware. The circle is going to be a collection of short lines. Referring again to my printer running Marlin firmware, the shortest line for a G2 / G3 circle is 1mm. A circle of radius 1 would end up as a six sided polygon rather than a real circle. Pushing plastic, that doesn't matter, but in your case it might.

The black goo is the PLA breaking down from the heat and it causes high friction at the end of the bowden tube. You can cut 3 -5mm off the end of the bowden and re-install the bowden. That is normal maintenance and it's why the bowden tubes are consumable parts of the machine. With the nozzle out, shove the bowden tube down through the hot end and push out any plug of plastic that might be in there. I do full machine maintenance once per month whether it needs it or not. The Y bed travels fore and aft on a 40 x 40 beam (the yellow beam in the image). That beam is connected to the bottom 40 x 40 cross beam (green in the image). There are 4 screws on the left and 4 screws on the right that are directly below the 40 x 20 vertical Z beams that hold the green beam in place. So if the green beam isn't exactly square to the machine, the gold beam is at an angle and so the Y bed is at an angle. Many Ender's are assembled like that causing the Y beds (the trolley) to run up and down hill. Leveling isn't inconsistent, just wrong. These are at the same time one of the simplest and the most complicated machines I've worked on. Everything affects everything and so each machine part must be adjusted correctly to take it out of the problem equation. As you've learned, troubleshooting is a big part of owning a Creality printer. The "Tune" is on a knife edge and once it falls out of tune it can be damn hard to chase the problem down. That's why I'm so fussy about maintenance. When there are several problems at once, and they are over-lapping, and any of them could cause the problem we are having, then frustration creeps in. One thing at a time is the way to approach it. I was over-thinking a problem I was having with the Z location and finally looked closely at the Z switch. There was some strings of PLA that had gotten caught in it. It was one of those "DUUHH" moments. Highly annoying.

Hi gr5. Looking at the picture of the white model the skirt looks pretty good and so do the loops of the model itself. They look like they welded together so the height couldn't be too far off. On the other hand, the first layer skin is a mess. Scouty, gr5 made an in-depth video on bed adhesion that covers most issues. For me, I clean the glass with IPA, and then give it a dusting with Aquanet Super Hold hairspray, and I level with a piece of parchment paper. I get consistently good starts to prints. Since gr5's video will cover adhesion, I'll go over some things about leveling (setting the gap) that you might be running into. These are a few mechanical things that can mess up leveling and make it so inconsistent it drives you nuts. The wheels on the Y trolley can loosen up / wear in. That causes wobble and leveling is very inconsistent and can change by a lot from one print to the next. You can adjust the eccentrics on the wheels on the Y trolley and take any wobble out. Like the other wheels, you should just be able to turn them by hand. you might also want to make sure that the beam that the trolley runs on is square to the rest of the frame. Mine was off by 3° and it was near impossible to level. The screws that mount it are the 4 below the Z uprights on both the right and left sides of the machine. The stop switches on these printers aren't high quality and so the repeatability of the stop position varies. The stop position can vary by .25mm over to .25mm under. That isn't so important in the X and Y but for the Z it's what sets the "0" for everything. Example: You're going to start a print and you Auto-Home and that gives a Z=0, then you level the bed however you want (BLT or paper) and then the print starts and there is a G28 so the printer Auto-Homes again and this time the Z is .25mm above (or below) where it was during your leveling. Now the first layer height is totally wrong and you either scrape or the plastic doesn't want to stick. You might be able to swap the Z switch with the X or Y switches. One of them might be more accurate (if this is the problem). If the right end of the X beam has movement up and down (wiggle it), and when you are leveling, you move the Z up, then go to the next leveling position, and then move the Z down, on the right side the movement is different than on the left side. That causes inconsistent leveling. You know that the first layer has to go down right. Get your tools out and go over the mechanicals of each axis and remove any slop that might be there. After that, if it is still inconsistent, take a good hard look at what is going on with the Z stop switch and check the set screws on the Z coupler as well. Take a break by sitting down with a bag of popcorn and watch gr5's video. He covers a lot of material in it. Now about those zits...I don't think I've ever seen that before. I print regular PLA at 210 but white at 195-200. There is just so much colorant in it (titanium oxide?) that it's tough to get right. The Gcode file I've attached is a leveling routine. The print head will home and then move to a corner and wait. When you're ready for the next position click the button on the LCD. Some Ender 3's with the 4.2.X boards don't pause correctly but for most the M0 pause command does work. The bed and hot end temperature lines are commented out. If you remove the semi-colons from in front of the lines then the bed and hot end will heat up. Level.gcode

If you can - post an STL / 3mf file. I've never seen those jittery moves before and I don't think it's any Cura settings that do that. Cura may be confused over something in the model.

Bring in a support blocker. Select the Support Blocker. Using the SCALE tool, de-select "Uniform Scaling" Set the X and Y sizes sufficient to cover the area you want to change. Set the Z = {Layer Height} * {number of layers you want to change}. You only want the top layer, then the Z dimension would be 1 layer thickness. Select the Move tool and move the Support Blocker into position over the model. Select the tool above the Support Blocker tool (it has 4 small shapes on it). Select "Modify Settings for Overlaps". Select "Infill Mesh Only" and change it to "Cutting Mesh" Select "Select Settings" and under Shell select Wall Line Count and Extra Skin Wall Count. Close the dialog. Set the Wall Line Count to 1. Set the Extra Skin Wall Count to 0. Move what was a Support Blocker and is now a Cutting Mesh into the correct Z position so the top of the mesh is the same as the top of the model. Here is a block with a single wall for the top 4 layers. If you select the "Eraser" you can change the Wall Line Count to 0 and there will be no walls for the top 4 layers. GregV_cubetestv6.3mf

When you get a moment put up a screen shot of what you are seeing. Make it with the preview slider on the topmost layer. And yes, it's Mr. Natural. You're the first to comment. I think because nobody else wanted to admit they were old enough to remember. Here is a shot of the test model. The top is layer 83 of 83 and the six below it are the same. 1 outer wall, 2 inner walls, 1 extra skin wall, then the skin at 45°.

I have to be because you get up 5 hours ahead of me🙃. Within each of the last 7 layers at the end of the Gcode file there should be a comment line ";TYPE:SKIN" that indicates the upcoming code will do a top layer. Check the file on the SD card for that. Some times the card can get funky and the end of a gcode file gets cut off as it gets written into a bad sector. You can try re-formatting the SD card, copying the file back to it and printing again.

When I slice that 3mf I get 7 layers on top and 7 layers on the bottom. The total layers is 83. Using the Cura mesh checker the model is fine. So from here I don't see a problem either with the model or with the Cura settings.