gr5

If you set shell thickness to less than nozzle width then it purposefully underextrudes. This is acceptable down to about 75% (shell 0.3mm) anything lower and you get ugly underextrusion and weak layers. And holes.

I think UM2 XY is quieter because of the mounting and because of the material used. Wood makes for a great sounds board - think violin, guitar.

That sounds wrong. That sounds like pillowing. Pillowing can usually be fixed with more layers. Please post a photo. Here is information about pillowing (post #10): http://umforum.ultimaker.com/index.php?/topic/1872-some-calibration-photographs/?p=17300

Are you printing in a dusty environment? Most dust will not melt or burn until much hotter than 260C. If it gets on the filament it can get carried through the feeder, through the bowden and get stuck in the nozzle. Alternatively the ABS that the feeder is made out of can get ground down by the PLA filament and tiny flecks of black ABS can also make it's way to the nozzle. Usually 260C is needed to get that ABS out. I don't know if it is "high temp" ABS or what but I have a theory that it causes many clogs that people have been seeing with UM2.

I listened to both videos and hear nothing unusual. I think it's fine.

I don't know about command line. But there are 2 ways to make solid layers - one is set the skin to say 200mm. That works fine and gives you concentric fill. Or you can set the infill to 100% which gives you diagonal fill which alternates each layer. 100% infill is not much (any?) stronger than 50% infill but is heavier, uses more PLA and takes longer to print.

If the walls are .4mm thick meaning it only makes *one* pass and never goes back to the same spot on the same wall a second time and if there are no holes in the walls of the piece and no non-printing moves and no retraction then you can probably use the "spiralize" feature which adjusts the Z position very gradually along each layer so that there is no Z seam - every move includes a tiny increase in Z. However if your part isn't incredibly thin then you can't use spiralize and I'm not sure what you can do. I guess I recommend tuning the Z axis - try doubling the Z acceleration. Each doubling of acceleration means the movement is 1.4X faster (square root of 2) which means the blobs are 1.4X smaller. Increasing Z jerk also speeds things up. But if you increase acceleration too much you will lose steps and your part will be the wrong height. That's why you need to experiment - make long 10 to 50mm moves up and down and verify that the Z comes right back to the same starting point. I recommend pronterface to play with acceleration and to do the movements: printrun/pronterface/prontrface download: http://koti.kapsi.fi/~kliment/printrun/ Another way to shrink the blobs is to print slower so that there is less pressure in the print head. Try printing at half speed and see if that makes a difference. I don't know if your reprap has marlin but if so you can do this on the fly with feed rate control. I have found that printing at 30mm/sec with .1mm layers makes the Z seem almost disappear completely due to the pressure in the nozzle being so much lower. Another way to shrink blobs is to lower temperature so that the PLA flows more like toothpaste and less like honey. For example 190C might give you smaller blobs than at 220C (although this also raises the pressure which could paradoxically make the blobs bigger so lowering XY speed might be better).

It's the same procedure except it is very easy to destroy your bowden tube when you try to remove it. So my instructions are more detailed.

I think the default is retraction off. At least it was at one time. So that's definitely not good. Grab the bowden and lift it at the print head and push it down again. Mine won't move at all. But when my UM2 first arrived it moved up and down by about 1mm. Measure how much it moves up and down and then add 4.5mm and set that distance to your retraction. Originally I used 5.5mm retraction but now that the bowden is immobile I use 4.5mm retraction. Other than that use the settings Illuminarti recommends. Exactly those settings. One more thing - 60C can cause ugliness if the fan is off. Consider having the fan come on at 100% relatively soon - maybe by 1mm. By default I think the fan doesn't come onto 100% until 5mm. This is a very difficult print and I'm not confident you can get it much better but after trying everything Illuminarti said and after doing 100% fan on a lower level, consider trying .2mm layers to see if that helps. Sometimes you can get better results with .2mm when you have overhangs (and sometimes it's worse). 210C should be cold enough for this color blue and 30mm/sec should look noticably better than 50mm/sec. So do that also.

When the PLA passes through the feeder it gets compressed into an ellipse such that the longer axis of the ellipse is longer than the average diameter before compression. So a 3.00mm filament might be 3.1mm wide after passing through the feeder. This is dangerously close to the 3.125 ID considering the 3.125 error is typically high (can be as low as 3.0 ID). But typically if your filament stays below 3.01mm diameter you should be fine. In real life situations, if the diameter is nominally 3.00mm it *will* get stuck eventually and you will end up throwing it away. Someone on the forums took a piece of sheet metal, drilled out a 3mm hole, then passed their entire reel of filament through the hole. Then they threw the whole thing away. You might want to just skip to the "throw it away" step as getting the filament out of the bowden means you might have to remove it from both ends and put it in hot water to pull out the stuck filament. The bowden is *not* symmetrical. The end at the feeder was drilled out to a wider ID. So if you remove it completely, mark which end is which. It sound slike the filament you ordered in India is *not* 3mm but actually 2.85mm? Don't forget to tell Cura the nominal/average filament diameter before you slice - or for UM2 you can do that on the UM2 instead under "materials" "customize" menu.

75C Bed Temperature is too hot! 60C to 70C works great. 75C bed temperature caused the problem in this photo:

100mm/sec is a reasonable print speed for parts that don't need to be perfect but you can definitely go faster. I recommend higher nozzle temps for higher print speeds. Setting feed rate to 200% won't print twice as fast because the acceleration and "jerk" settings aren't affected by feed rate. You mentioned you use "the default of 75C" bed temp for PLA. This is a horrible default and results in uglier parts for the bottom 10mm due to the filament having trouble cooling down between layers. Typically the parts bow inward in the bottom 5mm. I recommend no higher than 70C. I've printed many excellent parts at 70C bed temp and at 60C bed temp and see no difference so I usually print at 65C. Most people on the forums seem to use 60C. For very large parts that are likely to warp/lift at the corners I like 70C which keeps the part in the glass state where it is soft enough that it can't lift the corners. If you try to remove a part from the bed when the temp is 70C it will be like clay and get all bent and destroyed so you have to wait longer for it to cool before removing.

I think this is caused by air in the nozzle or maybe an air bubble sealed in solid PLA retracted up into the head away from the heat a bit. When a nozzle is hot but not extruding, filament drips out and eventually presumably air gets in. Or maybe when you retract at the end of the previous - maybe that is when the air gets in. When you heat the air, it expands and pushes some liquid PLA out. That's my solution also. I'll even do "move material" cold and move it until I see the PLA touching the top of the bowden.

The glass isn't compressible. And the bed will bend a bit if you push down on the glas. So if you put a dial indicator on the side of the head, an indicator that touched the glass and then raised the glass until it touched the nozzle you can identify this because the needle stops moving. You can then note the dial position and then lower the plate until the indicator has moved 100um. This could be very useful if your first layer is .1mm or thinner. For example someone printed a tiny frog the size of a pea and they had a .25mm nozzle and did something like .02 layers and got fantastic results but they had to level very carefully. I still think that with practice, doing it with paper, or by eye, or by just looking at the first layer you can get excellent levelling.

No design problems. But if you want higher quality than .05mm you should get a smaller nozzle first. The nozzle hole diameter is .4mm. The radius is therefore .2mm. You can't get good details in X/Y any better than .2mm without getting a smaller nozzle. So there is not much point in going higher than around .06mm for Z axis. I never go thinner than .1mm. Also I have a print that took about 40 hours at .1mm. If I went to .02mm layers it would take 200 hours or 8 days to print. More than a week! That's too long to tie up my printer.

I went a whole month without releveling my UM2. It doesn't change. My UM Original I have to relevel every week. Minimum. When it was new I had to relevel it every day.

I don't know how this happened but my copy of 14.03 has some strange stuff in the start gcode that slows down the print quite a bit. The main weird thing is that the XY acceleration was set to 1000. This makes my prints go much slower. I took that out. The other thing that is strange is it moves the bed by .08mm after homing such that homing techniques used previously now don't work anymore. I took that out also. I don't know where this code came from. It might not be part of the default release of 14.03. It might have come from me somehow by accident when I did a "load profile from gcode" on some strange model.

If you guys are too lazy to send daid ALL the settings and too lazy to send him an STL it's amazing he listens to you at all. The settings can be saved with "file" "save profile" or you can just include the last line in the gcode (that's safer as it guarantees it is the same settings as that you made the gcode).

If you read my original post above, in bold I said: So for many parts you can print with no gluestick as long as the bed is above 45C. (again 60C is recommended for other reasons). One more piece of advice. For getting parts off I have changed to a new technique (months ago) and now my favorite is a putty knife sharpened to a razor blade sharpness. I used a metal file to get it very sharp: I usually have to hit the rear of the knife with my other hand palm to get the knife started, then I can be very gentle and slow and the part pops off a section at a time.

I prefer a thicker first layer - usually .3mm so that my bed levelling doesn't have to be perfect. With many printed piecs it's *good* to have it crazily stuck. You can try printing with no glue at all if you want. All these glues wash off with water just fine - it's easier to do it in a sink as you need a lot of water. But you can do it in place if you let the water soak in for 30 seconds each rinse cycle but you may need to rinse 10 times. Sink is much easier. This is recommended. The software assumes your paper is .1mm thick which is typical. It should be pretty gently "stuck" or not really stuck at all but more friction than when the nozzle is farther from the bed. There's no difference in stickiness from 30C to 75C. Many people noticed that 75C causes problems because the part warps a bit and is quite ugly (doesn't cool fast enough) for the bottom 10mm. This has nothing to do with sticking to glass. 70C is cool enough that this problem goes away for me. Most people prefer 60C for PLA. I usually print 65C PLA which is plenty cool enough. For very large parts that warp enough to lift corners, 70C is better than 60C as the part is too soft to lift the corners. But again this has nothing to do with stickiness.

The end near the feeder or the end near the print head? If "print head" then definitely make sure that 3rd rear fan is working. It's on the print head - not the side 2 fans - the 3rd one in the back. It should come on as soon as you power on - even before the lights come on. If it doesn't work, slide up the black mesh until you expose all the wires and connectors - the 3 fan connectors are black/red wire pairs. This area of the printer fails for many people during shipping - I don't know why. Two of the connectors are connected together with a short 50mm wire - those are the side fans. Tug and push on the wiring to the 3rd fan. Also if it still doesn't work, remove the larger cover under the machine - only 2 screws hold it on.

The normal design of the Z stage allows the Z nut to move back and forth but I built mine very tight so the nut will not move. Same with the side arms - the left side arm normally can move but I made everything stiff and tight.

Your z movement is not linear. If you graphed commanded-z-position versus actual-z-position it should be a diagonal line but it has this same sine wave pattern. It's not easily measured. What happens is for the thick areas, you command .2mm Z movement and it moves .18 and extrudes as though it is .2mm and over extrudes by 10%. Then 1.5mm later you command .2mm z movement and get .22 and get 10% underextrusion. I guess you could calibrate it and fix it in software by tweaking all the Z movements. The proper fix is to rebuild your Z stage. I put a bunch of shims in my Z stage and tighted it up and locked it very tight. I even added some hot glue. I had to play with it for hours to get it so that it moved perfectly. Of course it's also possible that your Z screw is not straight. Take it out and put it on a table and make sure you can roll it and it is straight. If it is bent, then bend it back to straight.

USB just seems very unreliable. I've had problems with it - maybe 20 or 30 times a print aborted on me due to USB being unreliable. When you get up to 20 hour prints - it's just very dangerous to use USB. Daid, the developer, uses many computers and many printers and he's experienced problems with USB being unreliable also. The UM2 is even more unreliable with USB and officially is no longer supported. If you call tech support saying you can't print USB on a UM2 they probably won't help you as it isn't officially supported. People have told me that simply turning a light on or off in the same room during a print will crash the print if over USB. So UM is getting away from these features. I know Eric Zalm who is the lead developer for Marlin is working on a Arm based controller that can print over wifi. This will replace USB. Eventually. That will be a nice upgrade for the UM Original. There are lots of choices though - octoprint, pronterface, older version of Cura, doodle3d. Doodle3d will even transfer the print over usb to the SD card and print it at the same time. At some point about 1/3 of the way into the print the transfer is over and you can power down the doodle3d I presume. Also I think doodle3d features are built into Cura - I think it locates it automatically. So that's even better. Doodle3d is about 80 Euros.

Possible but unlikely. These printers are tough. More likely they will hurt you than themselves.