gr5

I understand better now. Setting "support horizontal expansion" to a very large number - maybe 20mm might help quite a bit here. Anyway I rarely use support and I'm not sure you need it here. Anyway... I'm pretty sure the problem is the "support x/y distance" parameter. It's trying to stay .875mm from that tilted "wall" (yes it thinks of that as a wall most likely). Of course then you will have the problem that the support is touching other walls and really hard to remove. Do you really need support for this part? It's hard to tell but it looks like it's supporting a flat roof that is bridged between walls. You don't need support when there is bridging like this. I mean it's not as pretty but it works. I recommend you use PVA support. Although I don't remember if PVA works with CPE (just checked the compatibility chart on the ultimaker page and it says it works but not so well with cpe+ evidently). Definitely if you use PVA support set "support horizontal expansion" large enough (20mm? 30mm?) so that the pva reaches from here down to the glass.

So the slicer, when doing one slice, does it by mathematically intersecting a plane with all the triangles (all of them!) to get line segments. Then it stitches these randomly ordered (unfortunately there is no helpful ordering in an STL file - no reference of which triangle is supposed to be connected to another) lines into "islands" or "loops". So let's think of what you are printing as "loops". There are "inner loops" which surround a void inside your part such as a vertical cylindrical hole but most loops are meant to be printed with material on the *inside* of the loop. An island is a loop with material on the *inside*. Now that I have definied "island" and "loop" we can talk about retractions. If the printer is moving within a single island and "combing" is enabled it will not retract. It will move to the other position without leaving the island and without crossing an inner loop. Combing is on by default so the printer SHOULD NOT RETRACT. Maybe you disabled combing? I often recommend to turn on combing except for skin layers. Maybe you did that and you are talking about a skin (top or bottom most in that region of the part) layer? When moving from one island to another it will always retract (well - except for some other rare rules that you seem to have noticed like minimum retraction distance).

When printing through USB the printer can only store a few lines of gcode at a time. The arduino on these printers (I'm pretty sure the A8 has an arduino) stores very little. There is a pause gcode. M something. You could search to see if there are any "M" codes in your gcode file - there shouldn't be any (most likely) after the first 10 or so layers. If the SD card is unreliable then the USB is also likely to be unreliable. They both send signals through the same cable to the arduino and this data path is the most common source of errors. Or you could try blowing compressed air into the slot for the SD card as it can get a single hair in there and then have intermittent (and sometimes complete) reading errors.

@Degs - check to see if you have any plugins or post processing scripts enabled that change temperature. Go to "extensions" "post processing" and "gcode" and make sure the list in there is empty.

@smartavionics - do you have any ideas? There's about 10 or so temperature changes and they aren't mostly in the first 10 layers like I would expect. This seems like some new bug.

@Degs - this is really bizarre. I just looked at your gcode which is helpful as it shows the settings at the end of the file.

@designrtUltimakerS5 - are you near NYC? If so there are some nice people there who can demonstrate the S5. LIke kman. And others. 3D printers have definitely gotten better since the replicator 2X but maybe not as much as you are hoping. The sales volume just isn't there - that's my theory anyway. If they were as popular as cell phones then they would have improved as much as cell phones have since the first motorola cell phone. Instead it's a comparatively small market. Your 300mm requirement makes it tough. Otherwise I might recommend my absolute favorite printer - the UM2go. I have 3 of them (and an S5 and a UM3 and a Um2 and a UMO). On this forum you are likely only going to get one answer: the S5.

There are 5 pins. Two of them go to the temp sensor and that is enough for the printer to know that there is a core loaded. The middle pin goes to the eeprom. If that is an "open" then the printer will think the slot is empty yet turn on the LEDs at the same time I believe. In addition you should loosen the top screw on the BB core (the one through the clear part) and recenter the circuit board left/right so it feels centered. Do not try to loosen the screw on the heater block! You will almost certainly destroy the core instantly if you do that as the heat break takes almost no force to break the heater block from the rest of the core.

So you left out some details. I'm going to guess you have a UM3 or S5 and are printing with PVA support. If so you need to increase horizontal expansion so the support reaches down to the print bed because PVA does not stick well on top of PLA (but the other way around is fine). You need a tower to the glass to hold it still and "horizontal expansion" takes care of that. If not please explain what materials you are using, what kind of printer, and I don't understand the lower picture so - maybe include some more photos? Maybe a photo of the failed part? Maybe?

Is this a dual core print? It cools the idle core so it doesn't leak as much.

Well steps/mm is probably set wrong. Did you update the firmware on the UMO recently? Some versions of Marlin have a "motions settings" menu somewhere in there and you can view and set steps/mm. Fix that value by moving the bed 100mm and measuring how far it actually moves and dividing the difference and correcting the steps/mm of the Z axis by that value. If you did update the firmware maybe you have UMO firmware on a UMO+? They have different Z steps/mm I think. Often you have to do "factory reset" to load all the defaults of the new firmware properly (loading new firmware sometimes messes up all the eeprom saved settings). You have the white circuit board, right? There is a jumper near the center of the board that can affect steps/mm by 2X. If that is loose and jiggling that could cause it to occasionally step 2x but this seems very unlikely.

Double check the gcode file on the USB - maybe you pulled the USB before it finished copying? Maybe the power supply/brick is glitching? Sounds like an overheat issue on the power supply. Does this have a heated bed? If so turn off the bed just to see if it lasts for more than 2 minutes.

Wow! That's some serious splitting. Okay first you need to figure out what kind of plastic this is. I'm guessing it's ABS and you are printing with PLA settings. ABS prints best with very little or zero fan. Try setting the fan to the minimum where the fans still rotate (assuming this is ABS) also make sure Cura is doing ABS settings. Also cover the front and top of the printer. You need to get the air up to 35C and the bed up to 110C and you can't do eithe without covering the front and top. Really you should avoid ABS as there are much better materials out there. Like PLA which is much easier to print and doesn't need any of the above advice.

As far as wear and tear - I wouldn't worry about that. I've seen prints that do a kilometer (a kilometer!) of retractions in just one print (voronai vase). If you don't retract you will get stringing, right? I'm confused - how can retractions be bad when going from one gear to another? In cura it shows light blue and dark blue lines. One is retraction moves and one is non-retraction (and non extruding moves). I assume you know this.

When not printing it leaves it "unlocked" so you can swap cores without doing anything. when it starts printing with the left core it "locks" it (aka moves the core up). If you look at the printer when it's not printing the right core is always down. If you look at it while printing with the left core the right core is always up. When starting up a print you can leave the lever in either position and it will "correct" it.

This is a complicated subject. I would have to type for 4 hours to answer your questions properly. Can you pick just one thing please? The thing that is the largest problem for you personally. If you could only fix one of these issues this week and if you could narrow it down to something that is very subjective e.g. "the part is wider at the bottom then further up yet the model has vertical walls - as seen in this photo - I care most about the bottom layer though". Then I can help you out. The AA 0.4 profile with 0.15 layers for Tough PLA is probably the best starting profile but there are so many profiles I don't know which ones I've tested. But this is probably the most tested and best one to start with.

Making one up also assures it isn't https as most sites these days are https and you can't get to the UM wifi setup web page if you choose an https site. Such as google.com. So making up a new site is best like: a.a

No need to remove the thin layer of PVA off the glass. One coating will last me a dozen or so prints. The way kman describes it is the best and easiest way. Alternatives: 1) Use glue stick then use a wet tissue and spread around and at the same time removing 90% of the pba (you want a very very thin layer). 2) Use hair spray (also contains pva glue). but remove the glass first. Because of this extra step I prefer using a pre-mixed slurry of elmers wood glue and water. kmans addition of alcohol sounds great and I will try that next time I make a batch. Oh and you can see me doing all 3 of these methods in a video:

I'm not sure if you understood IRobertI's question but if this model is truly a shell and the model has an empty inside then as he suggests your walls are probably too thin. A quick test would be to set the line width to 0.2 in cura just to see what happens. A 0.4mm nozzle can do line width down to about 0.3 and still have reasonable quality. Or maybe you have coasting turned on. Turn that off just to see if that makes a difference in layer view. Or maybe your model has errors in it. Try viewing the part in "xray view" in cura and if you see any red then the model is a problem. Whatever your issue is there are multiple solutions but first you have to narrow it down a little.

I love heated glass. I've never tried build tak. I know lots of people who have. I don't know how well PVA sticks to it - maybe that's your entire problem? I don't know. I tend to go overboard and make my prints stick too well. I would have to back off quite a bit not to damage build tak. I have literally picked up my printer by picking up a just-printed Ultimaker Robot (default size). I can get anything to stick to heated glass.

ABS is quite difficult to print. Have you printed it before? Do you have a heated bed? This video will tell you the tricks to get it to stick well. Very well. But without a heated bed it's going to be very hard if not impossible. I guess my advice is: don't print ABS. It was the first popular material back in the day for 3d printing but we moved on a long time ago. PLA is the easiest to print. If you need something that can withstand higher temps then try nGen/PET/CPE.

Underextrusion can have many causes but basically it's printing too much volume in too little time at too cold a temperature. Even a partial nozzle clog can be overcome by slowing down a lot. Thicker layers print with more volume per second. Did you look at this in cura layer view? It may be skipping your second layer in cura and jumping to the 3rd or 4th layer. Or any of the things I mentioned before. ALL of these things are very easy to test. You can raise temp from tune menu by 25C to see what happens. You can cut the speed in half to see what happens. You can turn off the fans to see what happens. I'd do all 3 and if it gets better than undo them one at a time. You can do all of these experiments while it's printing that second layer. If doing all 3 things fixes it then it's underextrusion related. If they don't then the z axis moved too far (either slicing issue or z screw issue). But I'd check cura layer view first.

Oh - and another quick experiment is to push up on the bed gently when the PVA isn't sticking. See if that helps. Personally I just reach for the 3 screws and turn them the same amount. The first 5 times I did that I had to think "um want nozzle to go down - oops no I want the bed to go up -- so loosen. - so counter clockwise from below - okay do it". After hundreds of prints I don't have to think anymore -- I just do it.

@Ryan8696 - Yes but the bottom layer is typically .27mm by default and squishing by an additional. 0.1mm isn't going to change things too much on the left core. Look this is mostly an experiment to see if the problem is: 1) too much moisture 2) not enough squish 3) oil/wax/grease on the bed 4) Something else? Bed is cold? When I see your photo I instantly think "not enough squish". I've seen it hundreds of times. It's happened to me many times. Personally I only do manual leveling. When you do the last step of manual leveling on the UM3 you need to set the second core height to be the EXACT same distance from the bed as when you do the left core. You want to use the calibration card and you want to feel the same amount of friction. Once that is done and calibration is done, then it's okay to mess with the 3 screws (in equal amounts) to get the bottom layer good for either core. It will be good for the other core as well since you got the two heads exactly the same in manual calibration.

@AndersK Yes. I usually do nylon at 100C overnight but my spool did melt slightly (must be ABS) so next time I think 95C.