gr5

The print in the upper photo looks quite good. That's the typical quality I see on most 3d printers. The print in the lower photo looks bad. It is severely underextruded. The gaps between the filament are more than the amount of filament so you have less than 50% extrusion. maybe 30% extrusion. So 70% less than nominal. This is severe. If you print 3X slower it will help a bit. I don't know what is causing your underextrusion as there are literally hundreds of causes: feeder issues, filament jammed somewhere, semi clogged nozzle, printing too cold, printing too fast, layer height too thick, bad cooling fan above heat break. That's an extremely short list of common causes. People who know your patricular printer will be able to help you much more than I as I've never touched an ender 3 of any type.

Excellent photo - you don't have the silicone thing installed quite right. Here's a video explaining how it should be installed and how to install it. It's harder to do than shown in the video. And easy to tear. Fortunately they cost almost nothing. https://www.youtube.com/watch?v=9IX-X1yPIBg

This is normal. This machine can print two different filaments. One on the left and one on the right. The problem is that the inactive print core always leaks a tiny bit. About 1mm of filament. So the printer keeps the inactive print core about 2mm above the active print core. When the right print core is in use, the right nozzle is lower. Only the right print core moves up and down. You can do this with power off by sliding the switch above the right print core towards or away from you. It's okay to leave the print head in either state as when it powers up it will move the right core to a known state.

You can go horizontal like this for one nozzle width as well. That's typically .4mm. And you can "bridge" across long gaps. google "3d printing bridging". The bigger overhang definitely needs supports. As you have discovered. The support only has to support the outer edge of the rim as it can bridge in between (refer to "bridge" earlier in this post). Usually I can design the part to have 45 degree angles under the overhangs. Not always. Sometimes I flip the part over to a new orientation. Sometimes I divide the part into 2 parts and screw or glue them together.

Find a metal needle .25-.35mm in diameter. Maybe from a metal brush where you clip off one piece. While at 200C, poke that in and out of the tip of the nozzle dozens of times scraping the inside and pushing gunk up into the nozzle. Then cool to around 90C and pull that out as your next cold pull to get any gunk out. After you can see through the entire core and out through the nozzle hole, take a small piece of filament and stick it down into the print core while cold and not in the printer. Feel for extra high resistance in the path.

No need to print an entire print if it fails the "12 second test". If it were me I would take apart the whole print core but I have lots of spares. So maybe that's not good to do. Can you borrow a core from someone with a working printer? The cores in the S3/S5/S7 and UM3 are all the same. Can you order a print core? They are about 120 euros/dollars I think. Actually I would do some cold pulls manually (not through menu system). I would remove the bowden from the print head and insert the filament directly in the head. Google around about the procedure (same procedure for any printer type). I would put the head in the corner so as less likely to bend the rods that go through the print head. I would do at least 5 cold pulls each at a colder and colder temp until you can't pull it out. At least 3 clean cold pulls at the coldest temp where you can just barely pull it out. I would check how the filament gets to the feeder to make sure it doesn't have extra friction. If you do take apart the core here is a video: https://www.youtube.com/watch?v=Ln_tMz8Dwd0 Right now this printer is not functioning properly. I would consider it "broken" even though it somewhat works. There is a teflon part inside the print core. It may need to be replaced. Are you the person who said there was only like 300 meters and 300 hours on the print core? That's nothing if that was you - it should be like new if that's true. I'm a little stumped right now but it's got to be the "hot end" or the "cold end" (feeder) and it sounds like the feeder is okay but it's failing the 12 second test so something is seriously wrong.

This thread is very confusing because Cura has so many damn features. Yes, some settings are in the comments at the end of the gcode. This is not the best source of settings but it may help out sometimes. Then there are project files and there are profiles. Profiles are a set of settings. To make things worse there are machine profiles, material profiles, and profiles that inherit from other profiles. And then there is "the" profile. You can create your own profiles, save them, load them, change settings on top of them without saving the changes and so on. Some people love profiles. I hate profiles and love project files mostly because historically profiles had more bugs. It's like two different religions. You can use either profiles or project files for the same purposes. We really don't need both features but they each have their minor advantages. SO TO ANSWER YOUR QUESTION Use project files for now. Everytime I save a gcode file I immediately also save the project file. The project file contains: the STL the profile machine profile material profile overridden settings And those are all separated out so I can see what settings you overrode AND I can see what printer you are slicing for (the machine profile) AND I can see if you messed around with the machine profile or material profile. ALSO if you rotated your STL, resized it, if you printed multiple objects... it's all visible in the project file. So if you have a cura bug, I can load your project file and should get the exact same situation as what you see in your cura (if same version of cura). So what you should do in above example 1) load STL, slice, save as gcode, save as project file P1 2) change profile, change settings, delete STL 3) load P1, STL will be loaded, back to original profile, back to original settings overrides 4) delete model by clicking it and hitting delete key, load second STL. All settings preserved from P1.

All of those prints are severly underextruded. This is most likely a printer issue. Underextrusion is a common problem with many printers. I only know how to fix them for Ultimaker printers.

maybe it's slipping in your feeder? Or grinding in your feeder? Maybe the diameter is larger and it gets stuck somewhere (check with a caliper). I don't know. Ask on a forum just for your printer type. Someone probably knows exactly what the issue is.

I'm pretty sure you can just change the feeder and motor and use UM2+ firmware and you will be fine but to be safe I would use the "tinkerGnome" firmware for the um2+ because it lets you change many more settings in case I forgot some detail and you need to change part back. When changing to the UM2+ you definitely have to reverse the E axis and change the steps/mm to a different value. Other changes in the plus are: fans on head different orientation easily change nozzles on plus higher wattage heater in print head better teflon part (but it's the exact same shape and size - just a better type of teflon) replaced spring with aluminum cylinder (I still like the spring better) smaller spool holder that holds only one spool better glass clips that are less likely to cut your fingers So the heater is different which means the PID values for the heater are a little different in the um2+ firmware but I don't think it matters much. Worst case you will see the nozzle temperature oscillating and if so you can alter the PID values. Tinker Marlin is here: https://github.com/tinkerGnome/Ultimaker2Marlin/releases

It looks fantastic. No underextrusion. Great surface finish.

You don't need to dry ABS but PETG prints better (less stringing) when dry. Nylon absolutely must be kept insanely dry. Desiccant works great as long as you recharge it. Get the kind that changes color. I'm hearing good things about alumina dessicant but I just have the silica kind. Get a $10 humidity sensor as well. Try to keep it well under 20% for the Nylon. I aim for 16% humidity at 21C. I use about 1/2 cup (1/8 liter) per spool (much less needed for the alumina). Even with all this I sometimes have to dry the Nylon on the print bed.

If it's PLA I doubt moisture is the issue as I have 10 year old PLA, never stored in a bag, works fine. What kind of printer is this? This kind of issue and the solution tends to be printer specific. It's tricky to build hot ends that don't get stuck.

So it sound like sometimes the active leveling is off. I don't think the 90 degrees thing had anything to do with it. But it sounds intermittent which makes it harder to debug. Start by going to maintenance and testing the leveling sensor. It's a very quick test. It gives you some numbers I think (different versions of the firmware show you the numbers differently and I lose track). Lower numbers are better. 8 I think is the passing threshold. larger numbers means there is more noise confusing the sensor. The most common thing messing up the leveling sensor is the front fan. It can both be the cause of electronic noise or the fan coil can act as an antenna to pull in radio noise from around you. Make sure electronics like computers, phones, toasters, etc are at least a meter from the printer. Especially motors like refrigerators and vacuum cleaners. Keep those 2 meters away, lol. Let us know the returned value when you do the sensor test.

yes, definitely. engineering profile turns off jerk and acceleration control. I'm pretty sure the primary purpose of jerk and acceleration control is to reduce ringing. I can't think of another reason for those. Some people want accurate parts and don't care about how they look visually. For that you want engineering mode. Engineering mode assumes you don't care about ringing as it's too small to measure with a micrometer, yet it is visible to the eye because of the way tiny tiny tiny changes in surface slope reflect different areas of the ceiling (including lights in the ceiling) and show up extremely visible (especially on dark filaments). I love engineering mode for 99% of the stuff I do but sometimes you want to do letters or something just like you are doing. For that you don't want engineering profiles. The accuracy of your "ruler" won't be affected by not using engineering mode. Engineering profiles are more to do with the accuracy of the diameter of the part. The accuracy of the thickness of cuboid parts. Things like that.

"we transfer" link gives me an error. Can you check it? 30 seconds sounds extremely slow. almost 4X too slow. If the bar is in the middle, as in the photo, I can't imagine how it wouldn't have the right tension. however, the feeder comes apart pertty easily and quickly and it's important that the tensioning screw is *inside* the housing. Most people (including me) screwed (pun!) up the reassembly the first time. fbrc8.com has great instructions on this kind of thing. I don't know how strong you are - can you pick up the printer with one hand? If not then you probably can't do the test one handed either. I don't know. It's too hard to quantify. Ideally you do the test with a "good" printer and also with a "bad" printer and then you just "know" for the rest of your life. So do the test with other printers when you get the opportunity some day and you will eventually be an expert on feeder strength. At this point I recommend you take the feeder apart and make sure it's assembled properly.

It looks like you have an S5. Make sure that the belts are all equally tight (top and bottom half). Sometimes a belt slips a tooth and half is tight and the other half is loose. Make sure you aren't using engineering profiles. Make sure accel and jerk control are on. A quick fix is to drop the speed to around 20mm/sec when you get to that part of the print using a feature called "mesh modifiers". Google "mesh modifiers cura" and read about that feature or watch a youtube video. I'd probalby drop the acceleration to around 1000 and the jerk to around 5 all at the same time. If you can characterize the ringing and determine the frequency, I can help you with the ideal amount of acceleration and jerk to counter the ringing 90%. You do this by printing a cube, then measure the distance between waves with a ruler/micrometer and knowing the print speed you can calculate the frequency. Then you want the exact same speed increase in the first half of the period of the waves as the second half. This will counteract the ringing.

Ghosting is indeed a common name for what you see. "ringing" is also. Last night I actually read some papers about the many ways to reduce ghosting (by like 99% reduction). What kind of printer are you using? If you are using an Ultimaker S3/S5/S7, make sure you are *not* using the engineering profiles and also make sure you have accel control and jerk control enabled. Beyond that initial advice, I need to know what kind of printer you have. Ghosting is a complicated topic. It is caused by vibrations in the print head. When you print those indentations, the head moves a tiny amount, but also sharp 90 degrees. The head rings/vibrates. Typically around 50Hz (50 times per second) but it could be anywhere from 20Hz to 500Hz depending on the printer. For many printers that have belts, it helps to tighten the belts a lot (not applicable to UM3/UM5/UM7 which has belt tighteners hidden in the blocks). When you pluck a belt it should be above 400Hz in pitch. You can test this with a guitar tuner app on your phone. I have a video on youtube showing how to do this. It also can help to lower acceleration, jerk, and max speed. Or lower the weight of your print head. some people put an accelerometer on the print head, determine the harmonic frequencies and then you can adjust speed, jerk, acceleration to perfectly counter the vibration. Here is a video I happend to watch last night that shows a massive exaggeration of the problem by slowing it down by 50X and increasing the distance of vibration by about 10,000X: https://www.youtube.com/watch?v=gzBhTrHv0-c In the video, the problem is so exaggerated that you might think it has nothing to do with your ghosting. But it's the same exact problem from an engineer's perspective. Klipper firmware, found on many 3d printers has these anti-ghosting (aka input shaping) features built in.

good stuff. Probably less viscous than regular PLA by a small amount. So it flows nicely. Less underextrusion. But your underextrusion seems a bit extreme. Assuming you did the 12 second test correctly.

I'm not surprised. It's a specialty tool. Some people manage to find a metal wire brush and cut off a bristle. Your core is so new, I'm not convinced it's the core yet. But it's a common problem. It could be a defective core. Not likely but often we only hear from the unlucky people on this forum. The core has some nylon inside the heat sink and if it isn't cut to the exact correct length it can cause different issues that result in underextrusion. It seems unlikely both cores would have the same issue. Someone, somewhere knows where the 3rd printcore is. There's other items. A few tools, some grease, a gluestick, some oil maybe. I forget what else comes with the printer. Someone knows where that stuff is. The person who wrote the purchase requisition for the printer probably is the person who unboxed it and knows where that 3rd core is.

Can you make a 10 second video of the 12 second test? I'm not sure I understand what you mean. Let me explain again. Get it extruding with MOVE. Experiment with the fastest you can spin without it slowing down extrusion (if you spin too fast it slips and you get reduced extrusion). Then put a kink in the filament so you can see it's movement. For example pull the solid (just extruded) filament sideways at 90 degrees. Maybe hold it like that at 90 degrees. Or pinch off/break the filament. Then maybe start timer and then start spinning the MOVE COMMAND at the correct spinning speed for the maximum extrusion speed possible (if you go too fast then extrusion slows down). Stop timer when the kink gets to the build plate (which should be in the lowered position of course). Should be under 12 seconds. Typically more like 8 seconds. 60 seconds is a problem. 246 meters is brand new. I've seen cores with 2000 meters (2km) still working just fine. I've seen cores at 6km finally needing refurbishing.

I'd go back to the AA 0.4. The BB cores tend to (partially) clog easily as PVA caramelizes in just a few minutes (20?) of sitting at printing temps and it raises the BB core temp everytime you do active-leveling. And there could be a tiny drop of PVA in there. AA cores also tend to get a caramelized coating on the inside of the nozzle. I scrape that off (while nozzle is hot) when the 12 second test fails. I use a hypodermic. It's hard to get a nice sharp piece of metal that fits in a 0.4mm hole. If you go into the maintenance menu, you can get it to tell you how many kilometers of filament you have printed so far. I'm curious what is reported for each of the cores you have access to. That UM3 came with two AA 0.4 cores and one BB 0.4. Do you still have all 3 or did one get lost?

And like Slashee-the-cow says - try lifting a weight of some sort to gauge the force. I think the UM3 is around 10 pounds maybe?

33 seconds for half way is horrible. Really horrible. That's the problem: the printer. Try the "fight the extruder" test again. It takes pretty much all my strength to fight the filament and have it slip. If you can't get it to slip with all your strength it's probably not the feeder. It could be the filament I suppose. I'm a little wary of a new filament I've never tried. But EVERY different type of PLA I've ever tried including HT PLA filled PLA and many others have all had about the same printing viscosity.

To do the 12 second test, do not stop spinning the dial. You have to spin it quite fast, but not so fast that it starts slowing down again.