gr5

Okay printing something different is actually helpful in diagnosing the issue. I see a LOT of heavy stringing. I don't trust your filament dryer. How hot did it get? Did you unspool enough filament for a cube? How long was the filament in the dryer? For PETG, 8 hours at 50C should be enough to dry any unspooled filament. Within a week the filament may be "wet" again - I'm not sure as I keep mine pretty dry. PETG seems to take longer to get "wet" than nylon or pva. To dry the entire spool it could take many days as the humidity may be far towards the center of the spool. I do like Torgeir's idea of maybe printing some PLA to verify your equipment (such as feeder steps/mm as mentioned). Also are your issues only when there is support material involved? Is everything printing beautifully when there is no support? For parts that have potential to string (islands of printing on each layer)?

Yes! There is a solution, but first: I haven't had any trouble getting all manner of filaments and colors. I have heard the most complaints about colors. For example in carbon and glass filled filaments there are probably over 100 to choose from. I do have one spool of very old 1.75 and I was able to print some recently using a new trick so I'll share that. It was quite simple. This works on S3/S5 but not on UM3 because the UM3 feeder just won't grip the 1.75mm filament. It was a shade of brown that I didn't have in stock and I didn't want to buy more filament when I had a spool already. I got some PFA tubing that was "4X2" meaning 4mm O.D. and 2mm I.D. I stuck it down into the printcore and marked where it stuck out the top. Then I cut the tubing there. I reduced retraction by about 2mm but the 4X2 bowden seems to stay pretty well inside the printcore (I was afraid it would slide up). It worked quite well. I also changed all the "flow" values in cura to 265% (which is 2.85^2/1.75^2). I also had to hand-insert the filament. Which I always do anyway so no change there. You might have to remove the normal bowden at the head to get the filament started into the smaller tube. I forget if I had to do that or not - I definitely removed the bowden at least one time. Other than that it worked pretty well. I printed 2 boxes that were each about 2 hours.

You will get much nicer results (typically but probably not in your case) if the letters are inset instead of sticking out. That may be good or not for cookies - not sure. You can definitely get those filled in. It looks like you have quite a bit of underextrusion - the flat layer below has lots of gaps. There should be no gaps there either. It may be the same problem (underextrusion). Try printing at half the speed you were printing at and/or clean out the nozzle. I think underextrusion is mostly your only issue however it may also be slicing this way on purpose. There are so many settings that can affect this. It's good to look at the print in PREVIEW mode in Cura to know if it will have gaps. Instead of me listing 30 to 50 settings for you to check, please post your project file so someone can look it over to see if anything needs to be changed in your cura settings. To do this, in Cura, do "file" "save project file" and post the resulting file here please. We will be able to see your STL file (your model), where you placed it, if you scaled it, what printer (machine) profile you have, material profile, and all of the 500 or so settings you are using.

Go through the menus. Pretty quickly you can find an option to enable/disable the filament sensor. I have an S5 but don't use the filament sensor and I have never run out of filament. It's not that hard to make sure there is filament on the printer but this a nice feature sometimes.

I recommend you learn a bit about getting things to stick to glass first but yes, there are 3rd party companies that will make various alternative sources. All of them should be able to handle a hot nozzle just fine. Some of the companies don't specifically mention the S5 but if you contact them they usually have that option for quite inexpensive prices. What material are you trying to print? Did you put anything on the glass (e.g. glue)? Oils from your fingers can get on the glass. Have you tried washing that off with dish soap or glass cleaner? Higher temp materials (e.g. ABS) are more difficult but I've printed pretty much every material out there on an S5 glass bed. So tell us more about what material you are printing. Personally I have the opposite problem - things stick too well and sometimes I pull up (spall) some of the glass off the bed. yes, really.

It could be the power brick or it could be the cpu. I suspect it's the power brick as those die more often. You can prove it's the bed by not using the heated bed. You can use painters tape on the glass, clean the painters tape with Isopropyl alcohol to remove the waxy surface of the tape (otherwise parts won't stick). Re-level as the tape has a thickness to compensate for. Try to do a print. If the computer doesn't reset then the problem was with the power supply. Make sure the heated bed temp is set to zero when doing the test! To me it really looks like the power supply. They cost about $90 USD. If you want a cheaper solution, note also that the printer is using the most power when it is heating up. You can force the printer to use less energy with some special firmware for the UM2: https://github.com/TinkerGnome/Ultimaker2Marlin/releases/tag/V19.03.1 Download the hex file about 5th blue link from the bottom. Make sure it says "plus". NOT "dual". Tinker-MarlinUltimaker2plus-19.03.1.hex You need a USB connected from the printer to your computer to install firmware. You can use Cura (or arduino ide) to install custom software. It will ask for this hex file. Once the new firmware is installed, find the "power budget" feature. I believe the default budget is: nozzle: 25W bed: 150W total budget: 175W Only mess with the total budget. Lower that by 50W (50 watts). That will reduce the power only to the heated bed.

According to the link below, it's time in seconds. I think it's the time since you turned the printer on? If you read the post I link to below, it mentions that there are about 10 temperature samples per second.

the "20" command seems to get the more recent temperatures. I suggested you hit the red button to stop those as they just keep going every few seconds. Yeah if you want more than 10 minutes of data it gets more complicated. It occurred to me just now that you could look through the code for that web page (no idea where it is - probably in one of those ".js" files) and the temperature data is almost certainly appended to some array at some point in javascript. You could dump that array after many minutes/hours of collecting data. But I don't know the name of the array or anything - if you aren't a programmer and don't have a good friend who is great at javascript then you are out of luck.

Unlike PLA, PETG absorbs water from the air and if you don't keep it dry it will "string" a lot. What happens is the water boils and foams the filament slightly which causes it to expand. If you have a heated bed, the easiest way to dry PETG is to unspool the amount you will need, place that on the bed with the spool on top (no need to cut the filament as it's still attached to the spool). Cover with a blanket or towel and leave at 50C for at least 2 hours. 2 hours is enough to get the unspooled portion. 8 hours will get some of the spooled as well. After printing with the dried filament, put in a sealed zip lock (I use 2 gallon) and include a cup (1/4 liter) of color changing desiccant. Recharge the desiccant when the color changes (typically once every few months). Your "flaky surface layer" might be related to overextruding which might be related to "wet" filament. I've never had this "flaky surface layer" with petg - just thick stringing. For me "stringing" is the only symptom I get when I have wet PETG. Other than that the quality is okay. When I dry the filament, the stringing goes away.

How much did they cost? I seem to remember someone bought the "more expensive" version and it was $3?

Good analysis, Slashee. That 110% flow is definitely a potential cause of the problem. speed Again, I haven't looked at the profile, but what about speed? cura profiles tend to do different speeds for each type of "line". e.g. infill, support, outerwall often all have different speeds but I don't like that - it lends itself to brief over and underextrusion. The engineering profiles all have constant speeds. temperature CPE and PETG are pretty similar (same thing?) so if there is not PETG profile I'd start with CPE and modify temp. CPE and CPE+ are Ultimaker-only brand filaments I think. I print a lot of PETG. I use 230C. I just checked. fan I've never touched fan on CPE/PETG which is at 20% which is STUPID as my UM3 and my S5 both are at 100% when you set them to 20%. So I've been fine with "100%" fan but the next time I reslice one of my PETG prints I will lower that to 2% (if printed on UM3 or 4% if printed on S5). I would try 20% since it sounds like you don't have an Ultimaker printer and that's what is in the CPE profile. Fan probably has nothing to do with your underextrusion (it's possible - sometimes air bounces off build plate and cools nozzle especially on lower layers). If you have trouble with adhesion, lowering fan can help especially with higher temp materials. If you have ugly overhangs you want more fan. PLA usually needs more fan than 90% of printers out there are capable of delivering so with PLA always go 100%.

This picture is so much more helpful!! So it seems to be printing reasonably well once the support has stopped. So for now you should only be worrying about the portion where there is support involved. I can see that this side is severely underextruded - probably <50% extrusion. Is it the same on the back side? It seems like your settings probably change between the support and the print. I didn't look at your project file but I would make all the settings the same although it's almost certainly either line width or print speed. When you are printing at a certain volume per second, the pressure in the nozzle takes a little while to equalize. This is particularly a problem for bowden printers but all printers. If you, for example, speed up, the pressure is too low in the nozzle and it takes a few seconds to equalize - during that time you get underextrusion. Then overextrusion when you slow down for a few seconds until the pressure in the nozzle lowers enough to equalize again. Similar with changing layer height or line width. Also retraction amount, if too big, can get air in the nozzle.

The problem with the above is that it appears that perhaps the printer only stores about 10 minutes of data (that's about how much data came back from my printer). The graph will continue to grow however as your web browser will store much more data. I don't know the best way to get more data. Maybe you could let the browser continue to grab the data and then dump out the array that is storing the accumulated data from the javascript command line in the browser debugger. Or maybe you could look at the code for the "temperature" web page and change the code to suit your own needs.

Step 1: find out the IP address. Through the menus put the printer into "developer mode". Once you do that you will see the IP address of the printer at the top as shown here: Developer mode might not be needed if you know the IP address. But maybe you need developer mode for the next step? Not sure. Next go to that address in a browser: And click on the link "temperature graph". Now to get the data in a CSV format - it's actually in JSON format. The above browser must be on a computer (not a phone). Press the F12 key, then make sure the screen on the right is in network mode as shown (see blue line under "network"). Then hit F5 and once the graph loads click the red button to stop network requests (see the button is gray to the left of "network"). Click on the line that says "5000" as shown and click on "response" (see blue line under "response"). Now you can see the data in json format. You can cut that and paste it into a website that converts json to CSV. Or write a few lines of code to convert it from json to csv.

I've never removed the nut on the S5 but I know it comes off pretty easily on older models and in general, Ultimaker printers are pretty easy to take apart and put back together. The threads appear to be about 2mm so with triple helix, one rotation would be 6mm. If you remove the Z screw then you can slide the bed up and down to see if there is some strange friction related to the two vertical rods and their bearings. Maybe those bearings need to be cleaned? They are linear bearings with balls inside. You shouldn't need any lubrication but some people do lubricate which can attract dust and the dust can get in the bearings. Another interesting test is to put a brick or similar weight at the back of the printer (after it starts printing). If this improves things then there is some kind of dirt/friction causing problems in the z screw or z rods.

Whenever I see that pattern it's almost always related to the zscrew. Did it do this from the beginning? Or did this gradually happen over many months/years? I know you said you cleaned the z screw but sometimes you need to take it out. I believe it's attached to the stepper so it all comes out (not as hard to do as you might think but you have to get to the 4 screws that hold the stepper), put it on newspaper, spray with wd-40, clean with toothbrush, dry it off, put a pea sized drop of grease on it and then put it back together. When you cleaned the z screw did you use a toothpick or something that gets into the cracks? Did you realize it's a triple helix and you have to get all 3 threads? If you measure the distance of 3 threads on the z screw (one rotation) is that similar to the height that the pattern sort-of repeats? If so then on thread is dirtier than the other two. Maybe. The issue can be with the nut as well but in my experience, cleaning the z-screw very well is usually enough. As a test you can just clean the 4cm of z-screw near the top that is used for prints shorter than 4cm in situ (without removing). You can use a paper towel and a very small screwdriver and a friend to spin the z screw. Or put rags around/under the area and use a toothbrush and wd-40 (and paper towels or rags to dry).

A more common error mode is when the X and Y axis are not perpendicular. This happens usually in shipping where the shipping company apparently has an elephant tilt the printer on one corner and step on it. Is there any chance the error is more on a diagonal? Often this error in orthogonality can be seen visually or with a square against the outer rods in the gantry.

There's 2 computers - the unix computer and the "arduino" computer that has Marlin. They talk a lot. I'm not sure but I think the python code on the unix computer sends this kind of thing to Marlin on boot up. You could grep through the python code for the steps/mm command: "M92". The nice thing about all that python code is that it isn't compiled so all the source code is visible. Note that any changes to python will get replaced if/when you update the firmware. Even just 1mm seems like a LOT for an S5. I'd characterize it a bit with a ruler measuring actual nozzle movement and not measure the plastic which can vary by quite a bit due to cura settings, over/under extrusion, warping. For example if one of the gantry pulleys isn't drilled out in the center and is off center by 1mm that could cause that kind of error but UM knows this dearly and has higher quality pulleys than the old days. Maybe binder clip some paper to the bed, move the head around to specific positions and mark the positions with a pencil then remove the paper and measure with a good quality ruler.

This part of the forum is in English only please. There is a german and a dutch langauge section. The clip that holds the bowden in place has 4 tiny razor-blade-like knives. One or more may be broken. By now you have scraped the outer part of the bowden a few times. Do you still have the horse shoe shaped clip? Remove that, then push down on the collet (outer ring) and with other hand remove bowden. Then cut off 1-2mm off the end of the bowden so that when you insert it, the knives will be able to touch fresh bowden. Also consider taking apart the feeder (2 screws - watch instructions on assembly as almost everyone messes up the tension screw). Remove the bowden collet and look inside at the blades. If one or two are not sticking into the center as much you can push on those firmly with a tool or buy some new collets.

I think you want tensions somewhere between these two - more like the one on the right. By the way, that's from an ultimaker feeder on 2.85mm pla. I don't remember which printer. Probably UMO or UM2 classic. This diamond pattern is typical of UM printers. UM claims that the diamond shape knurl is less likely to clog the feeder gear with plastic and that the diamond knurl allows powder and small chunks to fall out (self clean). I'm skeptical. Non ultimaker printers are more likely to have a horizontal lined pattern.

.5% across 200mm is 1mm and is huge. I doubt you have that much error. 0.5% across 20mm I believe. How did you measure the error? Did you print something and then measure it with a caliper? Better to measure such that you avoid the corners as those tend to be off due to slowing down and other reasons. There are so many things that cause errors, e.g. if the Z axis doesn't move enough, then that layer will be overextruding and it will stick out in both X and Y, or one of the pulleys could be out-of-round and you get +.3mm error and then as the pulley rotates you get -.3mm error but over the course of the print area it should average out. These are only 2 of about 50 things. Before trying to compensate e.g. Y scale, I'd first check if the movement of the print head is also off by 0.5% and is it off by that much from one end to the other or does the error vary (e.g. the pulley thing).

I'm not sure what I'm seeing but I don't think I'd call it "warping". It's really hard to say what I'm seeing - yes the blue and red areas look "different". Not sure what I'm seeing. Maybe the blue area is underextruded a bit? Some things that can cause a small amount of underextrusion: - If you messed with the feeder tension (keep it in the middle - don't mess with it) - If the front fan failed (not side fans). It can get clogged up with strings. Heat the nozzle, by the time it gets above 60C the front fan should start spinning - it's very easy to hear. Open the door and look at it. - Print cores get partially clogged. Try some "cold pulls" from the menu. You might not have any underextrusion. It's hard for me to diagnose from this photo.

If you ever choose to post this on thingiverse - this one screenshot shows you what comments to put around variables you want in the customizer and in the faint background on the left you can see what those comments generate: https://cdn.thingiverse.com/assets/da/03/ff/97/ff/Screenshot-02.jpg

I'm so glad you were successful with openSCAD. Now you can change one variable value to make it a different inner diameter or qty of holes or whatever.

yes. No idea what is going on there. I have seen .06 layer height with .4mm nozzle and it looked pretty damn good. But not every printer can do that - the Z axis in particular will likely move .07mm one one layer and then .05 on the next and this creates a lot of ugliness. In my opinion .2mm layer looks the best but on a really well made printer with super clean and well greased Z screws and with a highly accurate extruder, 0.06mm can look better than 0.1. I would try thicker layers (start with 0.2) and every other setting the same to see if the issue is indeed layer height. Instead the issue could be print speed - no idea. Or crappy printer - no idea what causes that weird mess on your "calibration cube".