gr5

I recommend you order the cable in my above link - it's not expensive - I think around 10 euros or $10. Then you don't have to be afraid of bricking the printer - the cable is a serial cable that connects to any computer and you have a terminal window directly into the linux machine and you can see it booting and when it's done booting you can log into root (password ultimaker) and do whatever you want. The "bricking" that can happen with bad json files still allows the computer to boot - it just blocks the ssh daemon and never tries to connect to the arduino computer. However maybe @Daid can answer your question about flask hooks and the "mock" printer in the griffin folder. Please publish whatever you come up with on github and link to it here in this subject as many other people can use this kind of 3rd party improvement to the printer.

Try two things - try setting line width extra small briefly just to see if that makes a difference - try 0.1mm line width. Also look at the model in "xray" view mode. If you see any red then that's the problem. Red indicates that the model isn't manifold (doesn't have an obvious (to a computer) inside and outside due to too many walls or holes.

What material are you printing? If you clean the glass well and you are printing PLA then the initial priming extrusion should be fine. At least it works fine for me with pla - they always stay where they belong. There is a printing kit you can get from UM that has adhesion sheets. If you cut out a small square of adhesion sheet and place it where the prime poops appear that works for many people. I'm pretty sure @lrodriguez does this. There are alternate materials out there like buildtak.

1/2 degree temperature should not be a problem. Any z banding is probably imperfections in the Z screw which I'm guessing are only fixable by getting higher tolerance parts. I'd not worry about it too much. Maybe you should just prime and paint your parts to hide the banding?

If you post some GrabCAD gcode output - the first 500 lines of gcdoe, someone could probably tell you how much you will need to mess with it to get it to work on a UM3. For single color prints it should mostly just work - maybe change out the first 30 lines of gcode. But for dual extruder printing I dont' think you'll get it to work. A huge amount of the magic for doing PVA support on the UM3 involves careful temperature control of both nozzle so one doesn't drip while using the other nozzle and so changing to the other nozzle is quick. So there is pre-heating of the idle nozzle. There is pre-cooling of the active nozzle seconds before it's done printing, there are standby temperatures, and I think more. The gcodes for changing to the other nozzle vary by printer so the odds it will just work are also low. Normally you use T0 (tool zero) for one nozzle and T1 for the other. But some printers (like the UM3) allow you to control the idle nozzle temperature without having to issue T1 very briefly. For dual color printing or for PVA support printing on the UM3 no other slicer can compete with Cura I strongly suspect. S3D (simplify 3D) comes very close - probably good enough. But even S3D won't do the preheating-idle-nozzle feature last I checked. And if you aren't doing dual color or PVA support prints - get the UM2. It's better at single extrusion prints than the UM3. In my opinion (of course - everything I post is just my opinion).

Sketchup isn't so great at making manifold models but it also takes a long time to learn a new cad program. So if you are going to continue to use sketchup I recommend you read these tips for sketchup users who do 3d printing: https://i.materialise.com/blog/3d-printing-with-sketchup/

Please show a screen shot of what you are talking about. Also please confirm that this has nothing to do with plugins, right?

200,000 triangles is a lot. It's good practice to learn how to remesh. Read this article. By the 3rd time you do this it will be very fast. Also slicing will be faster if you reduce. Try a 10X reduction but look at the text carefully after to make sure none of the letters were ruined. you might not have to reduce the triangles so much as move them around so you could try only a 2X reduction as well: http://www.shapeways.com/tutorials/polygon_reduction_with_meshlab What about 13C? I was asking if the nozzle temperature changed much. If it's moving 13C above and below the goal temp that - well that's unlikely. But you fix that kind of thing by fixing the PID values. the nozzle temp should be very steady - so for example if your printing temp is 210C you shouldn't see it go above 211C or below 209C but it would have to move to around 205C or 215C to see a difference in the walls (horizontal layers). More likely it's the Z screw that's not perfect such that some layers the bed is moving more and some less. When it moves less you get more extrusion sticking out on that layer. When it moves more then that layer would be somewhat underextruded.

This is a pretty good print. You can do better but there's nothing really easy to fix here. This is a typical quality. Horizontal banding is usually caused by Z issues where the Z axis isn't moving consistently. On an old machine this means it's time to clean the Z screw. It can also be caused by fluctuating temperatures - watch the temperature while you print. If it varies by more than 3C then that will cause banding issues. But this is pretty good - I don't think you can get this better without buying all different Z hardware. maybe increasing stiffness of bed. The zits - I think I know what you mean. Those look pretty damn subtle. I don't know if you can fix that. I would try cura 15.X to see if that fixes the "zits". There are so many changes in cura and it changes so fast -- I think there might be a setting to help with those but not sure anymore. I think cura 15.X might give you a better surface. It's possible (check layer view) that your model has too many triangles coming all to the same point near those zits. if so you could remesh the model. do you have a mesh view of your STL that shows all the triangles in it? I'd like to see that in the area of the "zits". If you have the issue I'm thinking of you will see one point on your model with hundreds of lines emanationg out from it. When cura slices near here it creates dozens or hundreds of line segments all within 1mm. Printing 10 or 100 line segments within 1mm will give you "zits". it should be smoothed out - line segments should be at least .3mm long for the most part. Preferably mostly at least 1mm long. That's 2 of your issues - maybe you could zoom in and circle an issue as we might be talking about different things.

You make good points. We are all guys responding on this thread which means you probably will get "solutions" and not sympathy. Sorry. One solution is to switch to Ultimaker 2s. Another is to not let students print anything other than PLA. Waiting for UM to come out with a video algorithm to detect blobs is not a solution. I suspect that will take more than 2 years. The spaghetti solution is easier but less useful to you. I don't think that will come out within a year either. I've worked for many companies and you get this "move forward" attitude. This is the kind of thing that gets fixed in "the next product". Then *maybe* they'll have time to port the solution back into the older product. Um is not full of computer vision experts. Even though openCV should be able to do this it's tough because filaments come in many colors. But it *is* doable and hopefully they will hire some summer intern to do this. So you'll need to come up with something else. Another solution is to disable autoleveling on all UM3 printers, have an expert do a manual level once every 50 or so prints. Enforce a rule that cura should make the bottom layer 0.1mm (this really helps a lot - it's another solution to help get that first layer to stick extremely well). Make sure the person who does manual leveling does NOT use the calibration card to get the filaments to stick real well. I'm not sure this will work with Nylon. Nylon is a tough one - I've had trouble with Nylon sticking before (years ago) - I don't print nylon much. I've printed about 10 nylon prints in the last 2 months but they were all small and had lots of brim and the bed was very hot I think (I'd have to check my notes). Maybe if they are going to print nylon they need a dedicated printer for that - one with a different build surface? I'm not a Nylon expert so I'm not sure what build surfaces might work well with Nylon. I've only tried blue tape and glass. I prefer glass. In my town we have a local middle school - children aged 11-14. They tried putting a printer in a teacher's room and having a teacher in charge. That didn't work so well. Instead they put the printers in the IT office of the school (it's a large school with about 1000 students). I trained the IT people and they do a print service for the teachers in the school. All over the school (e.g. history students can print things like medieval artifacts such as war machines). Maybe this model/concept can be used where students and teachers don't touch the machines. I used to think students should "learn" to use the printers but have since changed my mind and I think the students should be learning CAD and design. Not how to use a printer that will be obsolete in a few years. The other advantage of having these printers scattered around the IT department is that people are constantly walking by the printers and if a head flood were to occur they would notice within 10 minutes.

This really sucks when it happens. It's called "head flood". There are other discussions elsewhere. Basically heat up the left core to 200C and the right core to 100C and use a heat gun or hair dryer and remove that slowly. It will take a few hours. This happens when your part is wider than it is tall and then it comes loose from the glass bed and gets dragged around like a hockey puck but won't fall over (because it is wider than tall). The solution is to get your parts to stick better. This is a complicated subject so I made a video. The thing people get wrong the most is squishing the filament more onto the glass but glass preparation is also very important.

@ksmilejun - don't worry. No one is mad at you.

Three things. It's probably one of these 3 things. Make sure your third fan is working - the one on the back of the print head. Check your retraction distance on your printer. It should be 4.5mm. If it's more than this you can get the issue you showed. When inserting your bowden - it's very common for it not to go all the way into the white teflon part. From the photo I can see you have a solid aluminium spacer above the white teflon so it's hard to see what's going on so you have to do it by feel. The older printers have a spring instead of that metal spacer so you can see what's going on. Also make sure the bottom of the bowden is cut "square". Meaning perpendicular. It's okay to cut 2mm off the end of your bowden twice to get it square.

I bought some eSun black pla for a large project about 6 months ago. It was fine. It was mostly around 2.9mm. I'm not sure if it's any better or worse than more expensive filament. But it worked fine for me. I mostly use more expensive filament. When you go shopping and you put the 3mm in your shopping cart - at some point before you hit buy - get the specifications of the filament from the manufacturer. most post things like modulus (flexibility) and strength and diameter and tolerance and glass temperature or softening temperature and other things. Among those specifications should be the diameter and tolerance. There was a time when some manufacturers of 3mm filament were producing.... 3mm filament. Yikes. But by now I think everyone got the memo and I believe you won't be able to find truly 3mm filament even if you try very hard.

Something is wrong with your model. Was it created in sketchup? If so zoom into the walls of those holes and right click on each one and "reverse faces" if the wall is gray such that all the walls are white. Also look at your part in "xray" mode in cura - any red areas are a problem and need to be repaired. You could also try repairing your model using netfabb free repair service: https://service.netfabb.com/login.php Also like nicolinux says - ALWAYS check your model in layer view before you print it.

@tinkergnome ?? CCA1 - I'm changing your title to mention mark2 so people with that upgrade will notice this post.

Also does your PVA pop and sizzle? If so it got too much water in it. I've been able to "repair" nylon with water in it but have never tried PVA. Probably put it on the heated bed at 70C for a long time - maybe a 5 days? Under a towel or something. I'm not sure if you can actually get the water out of PVA like you can with Nylon.

Well you probably need larger (wider) towers. You can increase "horizontal expansion" for support. It defaults I think to 3mm but you could go to 4 or 5mm. Also it's important that the bottom layer squishes really well. I'd put an EXTREMELY thin layer pva - invisibly thin - so for example use glue stick and then clean with wet tissue to spread it around (and to remove 70% of the pva). Hopefully the PVA will stick to PVA. Then also when it starts printing turn the 3 leveling screws a bit CCW as seen from below (try a half turn) to move the bed up and squish the PVA harder into the glass. This will stick much better and differentiate the problem of not sticking to the glass versus not sticking to itself. Think of this as an experiment.

You probably could have moved it down less - did you try moving it down by only 0.1?

Oh - and if it's the wall issue, you can also try checking the option in your profile "print thin walls".

There are two likely possibilities: 1) Your walls are too thin. In general the wall should be 2X as wide as the line width. Try setting line width to for example 0.1mm just to see what happens. You can print with a 0.4mm nozzle down to about 0.3mm and there are smaller nozzles available down to 0.25 from ultimaker and 0.1mm from 3dsolex. 2) Your part is not manifold. View your model in "xray" mode and if you see any red then there is a problem with the model. Too many walls or too few walls (holes). There are more possibilities but 95% chance you have one or both of these issues.

It's okay - people call me wrong all the time - I try to take it gracefully. Oh my! Wow. I didn't know that was there - never saw that before. Yes. But there's a lot for them to read. The best place is to post bug reports in github but I can get the attention of a dev for you. @ghostkeeper - how long has cura been able to import profile settings from gcode? Does this work? Does it import across versions of cura?

I also feel that when parts do come loose and if they are wider than they are tall and they get dragged around the print bed, then the resulting "head flood" is much worse on the UM3 than the older printers. It sucks on the UM2 as well (and the UMO) but because the head kind of contains things it's worse. I'm lucky that in my first few years' printing when parts came loose they were always taller than wide and they simply fell over. No head-flooding disaster. Now parts don't come loose. Now I have the opposite problem - I'm often getting chips of glass inside the bottom of my prints.

You should really @SandervG on that last question (I just did it for you) as I can't speak for Ultimaker. Most of us here are users but Sander works for UM. He asked you I think to let him know who your reseller is. Hopefully you sent him a private message about that as he wants to look into how your reseller is treating you. For me I would say much less than every 15 prints come loose. It's more like 1 in 1000. Probably less often than that. But I get your point - you are dealing with a group of people using these printers so it's now getting more indirect. You clearly seem to have a lot of experience and are pretty good about getting parts to stick but at least for me personally (I could be wrong) but I don't think I could achieve that 1 in 1000 if I used the auto leveling. If you really want your parts to stick one of many key things you have to do is to squish the bottom layer more than the default auto leveling will have you do. I really suggest you watch the video (second post from top) but even if you don't... There are many things you have to do to get parts to stick well (such as brim as you mentioned and having the bed above the softening temp of the filament - 60C is a good bed temp for PLA - clean the bed every month - clean immediately if you spill oil on it) but I found that squishing that bottom layer makes a huge difference. I did many tests where I printed with varying amounts of "squish" and used a special test that involved a scale and a screw driver that measured somewhat objectively the force needed to get parts off the bed. The more you squish - the harder it is to get parts off the bed. There's a tradeoff of course. There's always a tradeoff.

If you are in the UK then I hear good things about faberdashery. They sell by the meter instead of the Kg so it sucks for people who do 1.75mm filament. But for 3mm filament that's not a problem.