gr5

Your power brick is probably weak. The power brick is sophisticated and has it's own computer. It monitors the voltage and such and will shutdown if it is unhappy with anything. I believe the UM3 comes with this power supply part number: GST220A24-R7B There is another version of this supply without the "T" in "GST": GS220A24-R7B Both versions of these power supplies can age or have "infancy failure" (meaning it dies within the first week or so) and it's quite common for these supplies to no longer put out the current/power they were capable of initially. I prefer the GS version as those consistently put out more power. Ultimaker prefers GST (if I remember right - it's because it meets more safety requirements and is probably mandatory for the CE certification - but my memory sucks on this point). Anyway you can buy either version as a standalone part. There are probably good deals on ebay. The last I checked (probably 5 years ago) it was about 95 euros for either version. For a new one (not used). I have to tell you that your description of the issue is very critical. It might be better if you showed a video. But if it powers down before the print head reaches either end and the printer reboots then it's probably the power brick. Also try pushing the print head around yourself to make sure there isn't double the friction on one axis versus the other. X and Y should be similar resistance.

I'ts important to lift the bed up when you do this test. You can just pull it up. Firmly gripping towards the rear of the bed. The amount of force to get it started is almost the amount of force needed to lift the printer so don't hurt your back. Also look under the bed and you will see this long screw sticking down and look at the hole it sticks down into. Often a piece of plastic (from earlier printing) is in that hole messing things up.

Also going forward, try to stick only with millimeters both in CAD and during the "export to STL" stage. This will save you time and pain.

I have to put cura view into x-ray mode to see your holes at all. They are completely hidden inside a wall. I'm sorry but this is a blender issue and I don't know blender at all. You could try an stl repair service like the free one from netfabb but there is a good chance it will just remove all those "holes". The basic problem is that the face of your part is covering the holes. Blender isn't great at making 3d models that can be printed. Blender lets you create infinitely thin walls with no need to make them have depth/thickness. CAD software that is meant to create real things won't let you do that. At least not easily. If you were really good at blender I'd say stick with it and maybe google "making blender models manifold". But since you are a beginner, if you are using a PC, then I strongly recommend DSM (design spark mechanical). It's free and easy to use and it won't let you create models like this where a wall is covering a hole. There's just no way to do it as your parts are always manifold (real, complete, solid models).

So there are many possible causes of horizontal banding. temp One is temperature swings if the nozzle is 5C hotter for one layer then it will extrude a bit more and a layer sticks out. Then if a layer is 5C cooler it will stick in. z movement But the most common cause by far is Z movement issues. If the Z axis moves just 10% more than usual when going to the next layer then you get a layer that is 10% underextruded because the distance from the nozzle to the print is farther than normal. Then if the next layer the Z axis moves 10% less than usual you get an overextruded layer which sticks out. Layers that stick out and stick in by a little bit causes the banding you see. The quick solution is to clean your z screw. Spend 5 minutes with a toothbrush and some wd 40 sprayed on the toothbrush (not the the screw as some will get on the print bed!). Or just use a paper towel and a toothpick. Clean up the few cm that are used for the bottom of your print and these horizontal lines should go away. I recommend printing a test print 1cm by 1cm by 10cm tall before and after cleaning. Sometimes cleaning is not enough and you need better quality z screws and nuts. There was a time when Ultimaker had 50 cent Z nuts and if you moved up to the 80 cent z nuts the quality was so much better. Z rods and z bearings can also cause binding and issues with stick/slip. Sometimes putting a brick or lead weight attached to the portion of your printer that moves up and down can help as well but this is just a debugging tool and not a permanent solution. This can reduce stick/slip issues.

@tomecko - your cura settings have the fan off at layer 0 and gradually turning on on following layers. Is that what you wanted?

Are you putting your filament into something to keep it dry while at the same time printing? I find that after a few hours on the back of my printer the filament needs drying again. Once I put the filament in a zip lock bag on the table behind the printer with dessicant (think 1 cup/quarter liter dessicant) and with the ziplock bag unzipped just enough to let the filament out. I haven't done that for years and I don't remember if that worked or not. I tend to do 30 to 60 minute prints so don't have to worry about this too often. There are companies that make dry boxes that go behind the printer so you the filament is only exposed for a few minutes on it's way from the drybox to the bowden.

Oh and your other question about the hex diameter. It's 2mm. That will cover 99% of all screws on the printer, the core, the print head, under the printer, in the feeder, pretty much every screw is a 3mm screw (that's the thickness of the screw including the threads) also known as an M3 screw. Mostly "button head". And all of them have a hex 2mm socket. I like this tool: https://www.mcmaster.com/catalog/128/3088/ It may look like that link takes you to 100 different hex drivers but one of them should be highlighted.

So you can manually heat both cores. In the menu system on the front of your printer. Look around in the menu system. Set them both to 200C and use pliers, tweezers, toothpicks, etc. Be very gentle around the two wires that come out of each heater block in the back. Nomenclature: The things on the back of the machine are "feeders". Your issue is called a "head flood". You can get those silicone covers very cheap at fbrc8.com in the spare parts of their website. Something like 3 for $5. This is a very minor version of a head flood - it can be much worse. Consider buying a heat gun or possibly using a hair dryer (not sure if the hair dryer would work). This might be a bad idea as you don't want to melt the other plastics in the area. I've never tried either of these tools against a head flood because I've never had a head flood. Actually I think a soldering iron where you can set the temperature to 150C would probably be helpful in cleanup. Once you get the cores out you can go at it more easily. Remember you have to retract the filament a bit before you can get those cores out. The bowden tubes come of very easily - slide off that horseshoe clip and push down around the outer collet while pulling up on the bowden. Don't pull up on the bowden without also pushing down on the collet! I just use my fingers but needle nose pliers are better for pushing down on the collet on both sides at once. AVOIDING HEAD FLOODS I've never had one even with over 10 years of printing. They have 2 causes: 1) print head door flops open while printing. Test that it takes more than say 3 grams of force to open the door. If it takes less then examine the two magnets that hold the door closed. Pry them out by maybe 1/2 mm and the door should stick better. 2) part comes loose (90% of the time this is the problem). If your part is partly printed and it comes loose from the print bed *and also* the part is wider than it is tall, then it will start sliding around the bed like a hockey puck on ice. So the nozzle is stuck in one spot of your part and filament is still coming out. For hours. The filament only has one place to go at some point and that is back into the print head. So learn how to have your part NEVER EVER COME LOOSE FROM THE GLASS. I have a video on how to get your parts to stick very very well. It's a long video but packed with info: https://www.youtube.com/watch?v=t58-WTxDy-k In the video I even pick up the entire printer by lifting on a tiny part the size of a vertical thumb. If you are doing active leveling then concentrate on the sections that discuss "brim" and "glue" and cleaning off any oils (from your fingers) and dust on the print bed.

White is sometimes caused by bubbles (super tiny bubbles) which are usually caused by steam which is usually caused by moisture. It's a lot harder to get PVA dry than you might think. And in my experience after about 2 hours of printing the PVA is already getting bad again (nylon also). Here is a succesful Nylon/PVA print from a few days ago. I suppose it does look a little white but also kind of transparent. Final picture is after dying the nylon black.

So there is something almost identical to profiles but appears to have fewer bugs. They are called "project files". So in your case you would load the project in older cura, load zero models, and then save it as a project file under "file" "save project...". Then tell your users to open the project, add their STL and slice. I stopped using profiles many years ago and people who use them seem to go through a world of hurt compared to people who use project files instead. Project files are usually (always?) forwards compatible as well. So a project file for an older cura should work for the latest cura (although it seems much safer to stick to the latest version of cura).

time versus money. If you'd rather just spend the money then just get another supply. 4 years ago they were about 90 euros/dollars 🙂

As you extend the build plate size it also moves the "center" to the right and back a bit so your parts are no longer actually centered in the printable area. So you can't extend it very much. A few mm, yes. so his part is like an upside down pyramid. He is only putting brim around the pointy end but the code that checks if your part is printable is looking up at that base size and adding the brim to that. So... bug. I'm familiar with this bug.

Yeah this is a bug that I've known about. It sucks. Another brim "bug" is that sometimes it ends up "inside" the part where I don't want it and it is extra hard to remove (I put bug in quotes because in the case I'm thinking of, "inside" is topologically still outside in my specific case because there is a purposeful crack leading to the exterior of the part). Anyway... That's a tough one. My understanding is that other technologies such as injection molding need much more modifications to the design than FFF printing. You even have to change the angle where the edges come together in corners (in other words corners of the mold are not at 90 degrees if you want the final part to be 90 degrees. Crazy. So it's wonderful that you have to change so few things in FFF printing (between desired part and model sent to Cura). Also in FFF I have to make all vertical holes about 0.4mm larger than specification. The difference though with injection molding is that there is a second engineer at the mold company who does all the modifications to your 3d model. I think I prefer 3d printing where the modification is done by the primary designer. Adding a brim in CAD is pretty easy. It can be on a secondary layer that is turned off by default so most people don't notice it? I occasionally add brim in CAD because I can only put the brim on the extremities/corners and this saves lots of time removing the brim later. Sorry I'm not helping with your question. But wondering if you might want to think differently about the overall solution to dealing with CAD models not matching final product and process flows.

If it were me I would do #3 to prove that it is the power brick. These things do indeed go. And when you are convinced this is the problem then get a new one. There are two types: GST220A24-R7B GS220A24-R7B The GST one is almost certainly what you have and puts out a little less current so I prefer the GS one. But the GST has some improved safety features. But again - puts out a little less power. Even though they are both rated for the exact same power, the GST turns itself off when you surpass a threshold and this threshold is lower for the GST. More info about how to setup the power budget in the tinker firmware: To me the power budget feature is very simple but it seems to confuse people. The power budget feature does not know how much power each element uses so you just tell it. Tell it how many watts everything is and what the budget is and it will make sure heaters are turned off or turned down a bit when they would exceed the budget. The bed gets lowest priority. In this version of Marlin is a power budget system. Set the bed to 150W (that's what it's supposed to be I think) and set the nozzle to 25W (if you have 3rd party nozzle such as 3dsolex then set this to what is truth - what nozzle actually is). Then if you set the budget to 175W (150+25) the power budget won't do anything and the printer will work normal. If you lower the budget to 150W then the power budget will lower the power to the bed when the nozzle is on. This changes many times per second (adjustments of power to nozzle). All the remaining power goes to the bed if the bed wants it. So for example if you use 150/25/150 as I mention above and the nozzle is on at 50% (12.5 watts) then the bed will only be allowed 12.5 watts below budget (150-12.5 is 137.5 watts) and so the bed will never exceed 92% power at that time. This changes 20 times per second (nozzle asking for more then less power, bed occasionally restricted a little bit).

It's almost certainly the power brick. Some things you can do to lower your power consumption: 1) Raise the bed temperature. It is counterintuitive but the resistance gets higher as the bed gets hotter so the max current it can pull down is reduced when it is at a higher temperature. Or turn the bed off completely and use blue tape. 2) oil the X and Y axes. Just one drop on each of the 6 rods (count them. six). Push the head around before and after to feel the difference. 3) Install tinker Marlin. It has a power budget system. If you do this make sure the power budget is smaller than the sum of the bed and nozzle power. That way when everything is on it will restrict the bed a little bit. TinkerMarlin is what I use on my 3 active UM2 printers and is here: https://github.com/TinkerGnome/Ultimaker2Marlin/releases

Two thoughts: 1) Do you have machine settings set to "ultimaker 2" or to "gcodes". If you have it set to UM2 then I believe the printer does it's own purge before running the gcodes. So perhaps that is the issue. 2) If you are truly speaking about gcodes in the gcode file then it's probably similar to temperature. If cura sees that you are using one of the temperature variables in your start gcodes then it will not bother adding it's own temperature codes. I'm guessing it's the same deal with the purge amount. I'm guessing if you use the purge amount variable (not sure what the variable is called) then I'm guessing cura will skip adding it's own purge code. Please clarify which machine setting you are using and I'll elaborate some more.

You might think "but I use the desiccant". Well you need about a cup of desiccant and you need to recharge the desiccant every month or so. And you need a humidity sensor as I've thought I've recharged the desiccant but it was still at 30% humidity in the bag. You need to make sure you are getting much drier than 30%. You should be able to get down to 15%.

PVA absorbs water like a sponge. Then when you print it, the water boils and causes over extrusion and stringing and foaming of the filament. I've only used Ultimaker PVA but I imagine any PVA is going to have this problem (nylon also). Leaving PVA out on the back of your printer for 24 hours is enough that it will not print as well. And some prints are longer than 24 hours (although if the spool is skinny enough then you are always exposing fresh PVA). do an experiment. In cura in the lower right corner hover over the "i" and it says how many meters of PVA are needed. Unspool that much and place it on the print bed with the spool holder on top. Set the bed temp to 60C and put a towel over it. Let it sit for at least 8 hours. Then do the print. See if you get less stringing. Don't go hotter than 60C as you can soften the PVA too much. No need to go cooler than 60C (at least for UM PVA - you can test this by first cutting some PVA from the end and putting on the bed and set the bed to 70C and 60C and put a towel over it - after 60 seconds at that temp reach under the towel and bend the pva - see if it might deform under pressure of a spool on top).

You can do this with mesh modifiers. You can add a cylinder that only includes the outer or only the inner edge and modify the wall thickness for just that region. The video below explains it. Maybe skip to about 3 minutes into the video. Unfortunately the video is 6 years old so the cura interface changed a little but not much. I think you can figure it out.

Okay so they added speed changes? I hate this B.S. There should be no speed changes for a UM2. I checked cura 4.12 and it also has these speed changes in the profiles. I'm pretty sure Cura didn't used to do this. Also I see accel and jerk control defaults on. This is B.S. The UM2 doesn't need speed changes nor accel control. I have 3 very active um2 printers and I have all these set to the same speed (circled). These speed changes can definitely cause what you are seeing however I still think this is in combination with some overall underextrusion. When you have a very high resistance somewhere in the print head (like the nozzle is no longer 0.4mm for example or the teflon is getting too old e.g. over 1000 hours printing) then the speed changes are more serious because the changes in pressure in the nozzle are more extreme and so the resulting extrusion amounts change more than typical. Also check that your front fan is working. If that fails then you get, well, underextrusion issues. Especially after 10 minutes of printing.

low acceleration could indeed explain why you have low accel in one half and high accel in the other half. Default accel is I believe 5000 mm/sec/sec (5 meters/s/s). for it to accel so slow that one half is over extruded - well that would be very slow. More typically you would get underextrusion in the middle of the travel and over extrusion at both ends where it is travelling slower. For 5m/s/s acceleration you are up to full speed within 4mm (even if printing speed is 200mm/sec). So I kind of doubt accel is the issue. Speed changes from infill to wall and vice versa - yes that could be the issue. I still strongly suspect you have overall underextrusion but that it is showing up in places where you have slight changes in extrusion rates. But I could certainly be wrong!

For UM2 the defaults in 4.X were: 1) accel and jerk control OFF 2) all printing speeds the same (almost all) so top/bottom/infill/wall speeds must all be the same! If there are speed changes you get under and over extrusion. Regarding retraction distance - that isn't set in cura but is in the preferences on the UM2. Too much retraction allows air in through the base of the nozzle. You want just barely enough retraction to releive pressure without actually pulling up out of the print head. If you look at the top of the curve of the bowden you should see the filament start out touching the top inner surface of the bowden and then when retraction is complete the filament is resting on the bottom inner surface. This is all at the top of the arc of the bowden. If you look at the filament where it is just entering the print head it should not move at all there. Typicall 6.5mm is typical for UM2. More for extended. 4.5mm for um2go. REGARDING FEEDER It should push with about 5 to 10 pounds of force. use the MOVE command to put the filament about half way down the bowden (or at least well up away from the print head). Then move the filament slowly with one hand on the controller and the other hand fighting the feeder by pulling down HARD. One thing you can do is hold a 5 pound weight in your hand and squeeze it against the filament so that the filament is carrying all the weight and your hand is just holding the 2 things together. slipping at 5 pounds is not fantastic but typical for a UM2. A UM2+ feeder should be able to do 10 to 12 pounds typically. 4 pounds or less is just not good - something probably wrong. This test helps you differentiate issues with feeder versus the hot end. Knowing that it definitely *is* or *isn't* the hot end only takes a few minutes and tells you a lot.

Before you take the head apart again, do a cold pull to clean out the heater block. then remove the nozzle and let things cool down below 50C and take the bowden off the head and slide some filament though. See how much pressure there is squeezing the filament. If it's less than 1/2 pound friction then the teflon part is probably fine (or maybe leave the bowden on - sometimes the bowden can push pretty hard down on the teflon if you loosen the 4 screws, get the bowden on their tight and then tighten the 4 screws to get the bowden even tighter).

Read the last line. Well, read the last 2 paragraphs also. I guess it's disabled for now. I don't know how to turn it on. @nallath?