gr5

After you get the power supply, if you don't use it after a year, sell it. I would guess you can get $400 if you include the power supply. It's a really tough printer. It can print 10,000 hours. You can check the hours once after you get the power supply, lol. Yes it may need some work to get going though. There are indeed many chinese printers for under $150 but they tend to be pretty crappy and won't last as long as the um2go. I have 3 um2go's. I use them the most. I don't need any more though, lol. You may have to replace the teflon part as well ($20) if it has more than 1000 hours. Anyway you won't know until you try it. But *any* 3d printer you get will be a learning curve. I think it takes about 100 prints before you can consistently print something and know it will turn out really nice and not fail or something. It's more like learning to use a CNC router than learning to use a 2D printer. There's stuff to learn.

oh. well the bad blocks are rerouted pretty quickly. Once you write a new gcode file you are writing only into blocks that have never failed before. It may be something completely different. Ultimaker support in Netherlands know much more. I know some of those people and they are pretty good. They have lots of inside knowledge. They will want to see your log files.

That has more power than you need. The um2go doesn't come with a heated bed so you can use a lower value power supply. Actually I sell heated beds for um2go and you still don't need to upgrade the power supply (there's a trick if you want to use the smaller power supply and also a heated bed). Anyway, here is the official power supply part information: company: Mean Well part number; GS120A24-R7B If you are in USA I recommend digikey.com. $50. Her school likely has the power supply. They just lost it. Maybe she can find it? Google it to see what it looks like.

Only once, I had to take apart the core. You can use a hot knife to slice off the filament above the "trumpet" (top) of the core, then slide out the core, power of the printer and disassemble the core. On youtube search for "um3 core disassembly". But I recommend instead just buying a new one. They are considered "consumables". Like filament but cheaper since they last much longer.

In the middle left menu, click on the left or right filament (depending which you need to get out) and then select top right "..." and then select "MOVE". This is one of 2 ways to heat the core from this menu. Wait until the temp is above 120C and start pulling. Make sure the print head is in the corner so you don't bend the rods. It's okay to lift hard enough to lift the printer off the table. If that doesn't work, leave it at full temp for a full 2 minutes and try again. Sometimes the filament gets into cracks above the heat break. This is rarely a problem but sometimes. If this happens a lot then replace the core as it probably has a defective piece of teflon in there (sometimes it's cut too short and there's a gap for filament to get into).

With a working printer it doesn't matter if you have 3X the gcodes for supports. It really doesn't matter. You can do millions of moves. These printers are work horses and can normally print pretty much nonstop for years. Someimtes a slip like you have is caused by friction so push the head around by hand and then put (only) 1 drop of oil on each of the 6 rods sliding the head around to spread the oil. Only one drop. Make sure you get all 6. Then see if friction feels different. Usually not. Usually it doesn't need it. Sometimes it does. Although usually you get more gradual layer shifting on many layers with friction. With a sudden change like that it's usually a loose set screw on a pulley but that is pretty much unheard of on an S5 as the factory has really good quality control with torque wrenches and so on. Overheating of a stepper driver can also cause a sudden shift. Is this a hot working environment where the printer is? Of course flakey firmware can cause an issue like this as well. Due to a dying SSD perhaps. Whatever is causing the error 84 is probably somehow related to the sudden shift. Not cura. Not gcodes.

Dustin knows more about "error 84" than I do (I know almost nothing). It sounds like you want to get through to UM support in Netherlands so fill out that form above that Dustin linked to and choose "hardware issue" and mention "Error 84" in your description of the issue.

Personally I'd wait until it happened twice in the same month before changing out the linux board. They sell it and the other pcb as one assembly so it's a bit expensive I think to replace it all - something like 500 euros maybe? But if it's the company paying for it then maybe it's worth it. My S5 linux computer SSD got a bad block about 3 years into my ownership and it was much worse - I had to redo the partitions. It has been good every since. I want to say it's been about 3 years since then but I'm not certain the S5 has been out for 6 years?

I have no idea what error 84 is. It's not on the list of possible errors anywhere I can find so I have no idea what to do if you get that. Do you know it Dustin? "Detected move outside of normal printer volume" however I am familiar with. It means the X,Y, or Z position commanded in a gcode is outside the printer volume. That would most likely indicate bad bits in the gcode file which can happen over the network, on a hard drive, but most likely on the SSD as it is known to get these occasionally. I have no idea what messing with the print head cable would have to do this as the movement gcodes are not sent to the print head. I'm ignoring the 84 as I have no idea what that is. If you are good with linux you can put the printer in "development" mode, then ssh into the printer and run linux utilities to read every block on the SSD but I wouldn't worry about it too much. Bad blocks are rerouted automatically in the SSD chip. So anytime one is found it is immediately fixed such that you just have to print something new and it will be written to the good blocks on the SSD. I would also contact support - it's free to ask questions. If you are within the warranty period and you get these errors a lot they might send you a replacement computer.

Did the "detected move outside..." come with an error code that you can look up on the ultimaker website? I'm 90% certain that is a gcode error. Probably not cura's fault - probably, somehow, the gcode got corrupted somewhere. Where was the gcode file stored (<option A> on USB stick? If so I'd read it and check it against the orgiinal or another identical sliced version). Or was it through the network? (then it may have been corrupted on the <option B>"hard drive" in the printer). Or maybe through the <optionc C>communication between the 2 computers: the linux computer sends the gcodes to the computer with Marlin firmware on it that controls the servos. I suspect option B and part of your "hard drive" (it's SSD) got corrupted. Fortunately these SSD's detect bad blocks and reroute the bad blocks and so you probably won't see any problems going forwards.

No way can you solder that and expect it not to break again on the first try. You can buy that PCB pretty cheap. Contact 3dshop.ca Probably < 100 CAD

Oh!! The "connect". I missed that. Reversing the direction is easy. Fixing the steps/mm is harder - change "flow" in cura every time you slice. Here is how to change direction. Video is cued. Watch only for 20 seconds! swap only two wires as in the video. Swap only on twisted pair or only the other. This only fixes direction. There is a strong chance you also need to fix steps/mm. You can't fix it in UM2+Connect firmware. This is the *only* printer from Ultimaker where you can't change the firmware. You may experience 2X or 1/2X extrusion. You should calibrate the extruder somehow with a ruler. Once you know the error you can compensate with "flow" in the Cura settings. Not a great solution as loading filament may insert 2X or 1/2X desired amounts. Retraction setting in cura may also have to be adjusted by 2X or 1/2X. Or maybe it's fine. Maybe feeder steps/mm will be fine.

80mm/sec with .2 layer and .4 line width is pushing things a bit. Are you really doing .2 layer? Thats .2*.4*80 = 6.4 cubic mm/sec. The section #1 above was written for the UM2 *original* and the UM3 can print much faster. I doubt this is your issue - if you are using the profiles for layer height, temp, line width, and speed, then #1 in my list is not the issue. Did you look at the other items in the list? I never know what people are going to say - sometimes they are like "oh, the printer is in a room that is around 30C". Most of the things in the list are unusual but worth going through the list. Underextrusion has to be either the cold end (the feeder), the hot end, or the bowden. It's possible it is the bowden but not likely. Above is a test to be absolutely sure it's not the feeder. Finally it's almost surely the print core. Since the printer came with 3 you can just try another. The BB core can be used to print other materials. You just have to hit the ignore button when it complains that you have the wrong core. You can indeed buy another printer but it seems like just buying a print core is cheaper. Note that all your filament for the Ultimaker is "3mm" and Bambulabs uses 1.75mm filament so they are incompatible. I've heard both fantastic and horrible things about the bambulabs. Amazing printer. Tends to fall apart with loose screws and other issues that are very frustrating to figure out. But more than half of them are absolutely fine. Prints very fast. Unfortunately it's confusing, but there is the Ultimaker S3 and the Ultimaker 3 (and UM3 extended). You have the UM3ext, right? Can you use S5 profiles on the UM3? I don't know. I never tried it. You can use UM2go profiles on the UM2; or UM3 profiles on the extended. But I've never tried S5 on UM3. Your problem is almost certainly not Cura. Well, unless you are using version 5.4. There are some bugs in 5.4 that look similar to underextrusion but I don't think that's what you have. Worth trying an older Cura I suppose. So I would slice for core #2. That way you are using a different feeder, a different bowden, and also try a different print core.

@tinkergnome

Some people lower feeder tension for very flexible materials like ninjaflex which is similar softness as the material in a rubber band. But the answer is "No, tension doesn't change for different nozzle sizes". If tension is too high you can grind up the filament into powder. If too low it slips. So the 0.8 nozzle is 2X the diameter and 4X the area and has 4X less "friction" and so you can pass through 4X the volume per second (typically half of that is taken up by double line widths and the other half through thicker layers). Try a different AA 0.4 core (every printer comes with two AA 0.4 cores - maybe it's still in the original box?). Your core may need some cold pulls (read about it in the maintenance menu), or maybe you are printing too cold or too fast. Or maybe 10 other things. It's common for the inside of an AA 0.4 nozzle to get a layer of caramelized goo inside it such that it seems "fine" but it's really now a 0.3 nozzle and should probably be replaced or burnt out or it needs some very serious cold pulls. Ultimaker priced AA 0.4 cores "lower than filament costs" and it's meant to be an expendable part. Like filament - sort of. When I say cheaper than filament that assumes an absolute minimum of 4 spools of filament bought for every print core bought. Having said all that, cores tend to last a very long time. I'm guessing 30-100 spools.

By the way, Nylon is just as bad and I love Nylon. But both PVA and Nylon need extra care keeping dry. PLA you can store in a tub of water if you want, lol. Well maybe not but I have 7 year old ABS and PLA that print fine. PETG I need to keep dry but it's not as big a deal as PVA/Nylon. I've never printed BVOH but I assume that since it is water soluble it probably absorbs lots of water from the air just like PVA. And people say BVOH dissolves much faster in water than PVA. I think that's one of the reasons people love it over PVA.

PVA is a pain because it absorbs water in just hours. I'm sure BVOH absorbs just as much but I've heard that BVOH is much easier to use than PVA. Store your BVOH in a sealed zip lock with a cup (1/4 litre) of desiccant. Not one of those stupid tiny desiccant packets. And use color changing desiccant and recharge the desiccant every few months (when it starts to change color or when the humidity gets above around 20%). If you have trouble printing when things were fine earlier, then redry the PVA. Desiccant alone will keep the filament dry but I haven't been able to use it to recover partly-ruined PVA even after a week at 16% humidity (I keep a humidity sensor in the bag with the desiccant and filament).

It's possible to print over USB but not recommended. To do it that way you have to change some of the "machine settings". Printing over USB tends to be unreliable and not suitable for prints longer than an hour (in my opinion) and even though it's not officially supported you can do it. But you have to change some settings and I forget exactly what you have to change. I think you have to switch to Marlin mode maybe in machine settings. But please don't try it unless you have a very good reason. And unless you have disabled all screen savers, OS updating, power savers, disk savers, and don't use the computer for anything else other than printing and then disable all other updates of software such as pdf viewer, chrome, office, well everything installed - you need to disable the update features. Heck you should probably just take the computer off the net as well to be safer. Also use a good quality USB cable and make sure it's short. And you still may get bad bits thrown in there. Basically if you are going to do USB printing you should probably buy a Raspberry Pi and use that as the host and install octoprint software on the raspberry pi. That tends to work pretty well. And again, you have to change machine settings.

Please post your project file from within cura. Do "file" "save project" and post that file here. Do you use an SD card to print? Or did you try printing through a USB cable (not supported).

I've seen some UM3 print heads where that connector in the head gets pulled so hard over many months that the part stretches out. The pins inside, that go in the cable connector, get shorter and so you have much less surface area for the heater power to go through. It's a lot of power (25 watts - about 1 amps) for such a little pin and the pin may be getting hot as well which is heat that isn't making it to the heater. The symptom of this would be that you are only getting say 15 watts out of the heater. It would take longer to heat up (versus slot 2 - hey try slot 2 instead?) and would be more likely to get heater errors. And yes, some print parts and profiles would need just 1 watt less and some might need 1 watt more and those profiles that print slightly more filament (e.g. thicker layer height) would be much more likely to fail.

Great information, thanks. Now I have moved from 80% sure to 99% sure it's not cura, not you, not your part. Other than that the part may be slightly larger than other people's parts and it may stretch the cable that is attached to the bowden above the printer a little more. It's just random coincidence it happened to you more than your colleagues. I know this is statistically unlikely which is why I was down to 80% sure before. Or maybe something about your profile uses just a few watts more than your colleagues. You could slow it down by 10% and that would in theory stop your heater errors but that doesn't fix the underlying problem. Like I said before, this error only has to do with temperature sensor and heater. One or the other (usually the heater) failed at least briefly. It's usually due to a bad connection somewhere. For the heater it's usually at that connector in the head. Please spend the 60 seconds it takes to remove the cover, remove the connector (be careful there is a clip holding it in) and putting it back in. The UM3 is notorious for issues with that connector. I recommend additionally adding a shim (some sheets of paper) around the cable where the cover holds the cable firmly in place. The S5 fixed this issue with a better clamping force here somehow. I'm 80% sure that the heater cables for left and right slot go directly all the way from the core to the underside of the printer and don't pass through any transistors, relays, switches, etc. Just "wiring". However the temp sensor I believe is read in the print head and temp is sent back digitally and so any failure with the temp sensor would be unrelated to cabling (as you would get a different error - communication error or something. There is a temperature graph you can look at. If you go to the home page of your printer - you have to know the ip address of the printer - go to that page and there should be a link to the temperature data. If not you can try googling this forum. Keep the temp graph up or look at it within a minute after failure and you should see the power for slot 1 reach 100% for an extended time but the temperature should be failling or drifting very slowly downward just before failure. If the temp jumps around quite a bit (like 3-100 C temp changes instantly) then the problem is with the temp sensor. But if it's the heater it will be a gradual problem because <see above where I mention the X and Y>.

speed/accel/jerk settings are so that cura can estimate how long the print will take. Only. It doesn't affect slicing. I still think it's worth entering that data as knowing print times is very useful. [edit: I was corrected, the speed/accel/jerk settings that determine the slice times are somewhere else - but they definitely are somewhere in the machine profile] I believe nozzle length to lowest part of the print head is only for a feature called "print one at a time" which is off by default. If you are printing multiple parts in "one at a time" mode then the printer will print one, then bring nozzle to bed again and print another. In order to do this you have to tell cura print head size so it avoids putting parts too close together. Also there is the "gantry height" or something similar. Wait - maybe I'm wrong about "nozzle length...". I'm not sure what that is used for. Something obscure maybe?

?? Are you a bot? Answer is above.

This error occurs if the head is at full power (which is rare while printing and only happens when the nozzle is significantly colder than the goal temp) and it still won't heat up by at least X degrees in Y seconds. I believe Y is 30 and X is 2. So the nozzle is at a lower temp than desired (according to the sensor) and even at full power it won't heat up. I've seen this happen on working equipment when printing with AA 0.8 cores at very high volumes and high temp materials. Is it possible that you are doing that? Even a 0.6mm core printing Nylon might be enough if you are printing pretty fast. So yeah it's possible it's your profile settings. What is your: speed, layer height, line width, fan speed, and printing temp? More likely it's a bad connection somewhere. Did you remove and re-insert the cable in the top of the print heat? That's a common issue. It's super easy to do and once you know how takes about only 1 minute. Instructions (skip the part about removing the bowdens and stop once the cable is disconnected, then reverse the step): https://www.ultimakernasupport.com/hc/en-us/articles/360039150672-Printhead

Well congratulations. The serial cable had nothing to do with it but trying several times had a lot to do with it. I don't really understand it and I would have had some advice but now you don't need it. 🙂