gr5

Ah. Okay. Well if those 2 short spots were about 5 inches long then I'd say you definitely have the "heat" issue. The left side of the filament looks perfectly normal after going through the feeder but those flat spots will get stuck in the bowden tube when they move along a bit further. Retractions tend to make the issue worse because then the same bit of filament goes through the feeder more than once. 75F (24C) is a bit warm. But it does get that hot in my house occasionally. Still I'm thinking more and more it's the "heat" problem. One woman with this problem - her house was 30C (86F)! The fan thing worked great for her. Also this particular blue filament I believe softens at a bit lower temperature. For some reason this is much more of a problem for some people than others. I don't know if UM is shipping a new brand of stepper? Or the electronics or firmware changed slightly?

1) The corner panels are very easy to remove. Where the edges touch the side walls of the printer each edge has one threaded hole to receive a screw. On the outside of the printer there is one screw lined up perfectly with each edge - those come out in seconds and the panel is now loose. The entire printer is designed to come apart easily. It's kind of the opposite of most modern equipment. Ultimaker's history is about printers that are easy to take apart and fix. The only part that may violate warranty is clearly marked with a sticker (the print bed connector). 2) Has it been warm in texas at all lately? Do you run it in an air conditioned environment? I'm still thinking this is a heat problem but there are other possibilities. Many people have had heat issues - they tend to happen 30 to 300 minutes after starting the print. So taking off the back-left cover and having a fan blow at the stepper motors might fix the whole issue - then we can think about what to do next so you don't have to always run the printer like that. 3) Your print seemed to have a ton of infill. What infill % did you use? Usually zero infill is enough for prints of people's heads if you do shell thickness of 1.2mm although 20% is an okay compromise.

Marlin itself is very well tested and I doubt there are any changes in the movement code in Marlin. It should be the same code on every reprap machine out there. Same as on UMO and UM2. It doesn't know what layer it is on so if there was a Marlin bug regarding flow it would happen on every layer. Much more likely to be a bug in Cura. And even more likely a operator/hardware error (something causing underextrusion). On top of this I print about half my prints .1mm layers and half .2mm layers and they all have good top surfaces. I have looked at all kinds of complicated, fascinating bugs, and there have bugs in Marlin but none like this.

Corect - "marlin/sprinter".

You really make it sound like the problem is with the gcode. The gcode is very easy to test. If you post the gcode somewhere I will be happy to check the top layer using excel. It's really not hard to do. Actually even better would be if you posted both gcode files so I can make sure the .1 version isn't *over* extruding. Which I doubt but we should test everything. Please don't reslice with cura. I want the *actual* gcode file that you printed that has a problem (it's possible that the tiniest change in the profile could change things - for example "cut off object bottom" or a plugin.

Could you also please post pictures of your prints so far? Even though you say they all fail the same way I'd like to see the quality up to the failure point and see other things. Like do they fail more at the same time or at the same height? Are there other issues visible... stuff like that. I've been reading this forum for 2 years now so I've probably seen it all. And finally did you try the fan thing I posted about? Or at least look at the filament inside the extruder? Was it flatened at all? Stripped? ground down? anything? how was the "tread"?

I have found that when people describe a noise and then later record it in a video there is absolutely no correlation. So please, record the sound. It might be completely normal. It might be the most important clue so far.

I think those fan ducts do more harm than good. The fan ends up sending out almost no air at all. I tried a few of those and went back to the original design. Some of them are okay but if you constrict the flow too much you end up with almost no flow at all. The fan just isn't designed for ducts. Try measuring the flow with and without the duct. I did and was a little surprised.

It could be that certain colors are near the end of the spool (blue) and so they have too much curvature for the white teflon isolator - it is amazing how much force it can take to get through the thing if the PLA is bent but how little force if the PLA is straight. So question: of the colors that do worse at .2mm, are any of them < one quarter remain on the spool?

Sounds like basic underextrusion which commonly gets worse over time on the UM2. Fortunately the various fixes aren't too hard. You can test by printing the test cylinder (see my 3rd link in post #205 above) at 230C.

If you are lucky the test print is still on the SD card. In cura choose "load settings from gcode..." and select that gcode file. This will load all the settings used for that print into Cura. Test prints change from month to month. I don't know what the latest test print looks like. I got my UM2 almost a year ago.

Ah, okay. Sorry if I don't always get the context from older posts. I guess I read too many different topics and loose the thread in my mind and am too lazy to reread each time. It's simple enough to analyze the gcode - part of the final layer to make sure that the Z setting is correct (.2mm movement) and that the extrusion is correct. I have a little excel formula that calculates the amount of extrusion expected for any given move and compares that to the actual extrusion and cura produces excellent results normally. But it seems possible the top layer could be somehow messed up due to the height of the object not being integral with the layer height. Or some other bug. But I've tested very similar things before and Cura was usually correct. Someone should just check the "bad" gcode. But keep in mind that at .2mm you are DOUBLING the printing volume so of course you will get less extrusion than ideal (more underextrusion). Some printers (or filaments) skip backwards, but other filaments/um2 printers instead slip a bit and kind of quietly underextrude.

Three thoughts: 1) nozzle temp variations. 2) Z movement variations. 3) Loose head. My first thought was #3. You can test this by grabbing by the nozzle (hot with gloves in case the problem doesn't happen with a cold nozzle) and shaking it back and forth to see if the nozzle is loose versus the two rods holding it. Maybe play around the bearings or anywhere else in the head. Then the centrifical force whips the head to the outside of the intended circle. This seems easiest to test and lest likely. #1, nozzle temp - there is a serious known problem for at least some UM2's with the latest software which made significant changes to the heating such that the head temperature can bounce around quite a bit. You can check this by plotting the head temp by connecting with USB or just by staring at the temp while printing. #2 - if the Z doesn't move fully you get an overextruded layer. This seems most likely as it explains why the rings are almost perfect - in other words the "problem" seems to occur on an entire layer but not the layer before or after. Unfortunately I see that every10th or so "ring" only goes part way around. So maybe the Z stage finally moved half way through printing that layer. Grease the Z. Move it up and down. Check for stickiness. Of all the theories, #2 is hardest to test so I would check for the other 2 first. Also #2 is most likely I think.

The robot is not an easy print - it has overhangs and the antennas are very difficult. If you want to just print parts then keep going at the settings you have now. If you want higher quality print slower. So 25mm/sec, 200C, .1mm layer will give you much better quality for the robot than the one in this picture. And make sure the fan is cranked to 100 before you get to the hand supports. But I rarely care so much about quality and I'm not that patient to wait a few hours for a silly robot. So decide what you want and go that route.

I'm not sure what you are trying to say in your pictures in the previous post. I see all kinds of different things in each picture and you don't seem to be saying which parts are "good" or "bad" or whatever. They all have different surfaces for different reasons. For example the bottom/last picture is normal quality - that's what i would expect from a properly tuned printer. The red one also. Do you consider this "bad" quality? The blue square above is definitely underextruded on it's top layer. This can be caused by sagging (top layers are bowed down into the cube) or lack of fan, printing too fast, too thick, or to cold. Some of the white surfaces are just not quite enough resolution to tell but they look like there isn't any backlash. But the blue one again - that one has a pattern of two diagonal stripes followed by a space, and then two stripes again. That is a sign of backlash which is rare on the UM2. And it's only a very slight backlash. It can be improved by tightening the short belts (loosen X and Y motors, slide them down the slots and hold them tight while tightening the screws). Here is a technical explanation of why backlash causes that pattern (post #8): http://umforum.ultimaker.com/index.php?/topic/1872-some-calibration-photographs/?p=14474 Anyway, what is your point? That .1mm is better than .2mm? That's because you are printing less volume and so the feeder is strong enough to keep up. When you print too cold and too fast you get underextrusion (filament is slipping in the extruder). Here is a graph that shows you the max print speeds you can do at .2mm layers for a given temp (you can go twice as fast with .1mm layers of course as that's half as much plastic per second). Try to stay around half the speed of the dark blue line in this graph: http://umforum.ultimaker.com/index.php?/topic/4127-um2-extrusion-rates-revisited/ If you are still getting underextrusion at half of the speeds of the dark blue line then you need to do an extrusion test although I suspect your machine is fine. Extrusion test must be done at 230C to be valid: http://umforum.ultimaker.com/index.php?/topic/4586-can-your-um2-printer-achieve-10mm3s-test-it-here/

I found the opposite. The thinner the layers the faster you can print but it's more complicated than that... There's a few things to understand: 1) Printing too fast. If you are printing at the limits of what the extruder can push out you can of course print faster with thinner layers because there's less thickness to the layers - less plastic. 2) 90% of the quality issues (but not all!!!) are related to the bowden springiness. The pressure in the head can be very high - typically 20 to 100 psi. Similar to pressures in a bicycle tire. When the extruder stops, the pressure is still there because the bowden tube stretches when printing and shrinks when you stop. Actually you don't have to stop to create a problem - simply slowing down printing will cause some overextrusion at first before the pressure balances again and speeding up initially will cause underextrusion. That is why corners on cubes tend to bump out a bit - over extrusion followed by underextrusion at the start of the wall. The fix is to never stop printing - always print at a constant speed. The "jerk" setting is 20mm/sec and the acceleration defaults to 5000 mm/sec (awesomely fast). If you always print at 20mm/sec the printer will never stop (except level changes and retraction hops) and you get amazing quality. Even 35mm/sec is slow enough that you get good enough that you probably can't see any over/under extrusion caused errors. So there is no need to go slower than 20mm/sec no matter how thin your layers are. 3) Bridging - bridging works much better with thicker strands of pla (this includes top layers over infill!). Thin strands break. So there is some point where you just can't go thinner and still get higher quality. To me it's silly to go much thinner than .1mm anyway without going for a smaller nozzle hole - go for a .3mm nozzle and holy crap you will see better quality but you will have to print 2X slower to get plastic through that tiny hole.

The cabling that runs to the head has a reputation for tiny opens and breaks. Hopefully this is the only issue. What about the servos? Do they work? Do you have an ulticontroller? To test the servos it's easiest to do it through USB. After connecting USB to a computer, you can run pronterface/printrun which is a fantastic, free, (easy to use gui) program for controlling and debugging your printer and can be downloaded from here: http://koti.kapsi.fi/~kliment/printrun/ Or you can use Cura print window in "pronterface" mode which doesn't have nearly as many features but you can still move the steppers. Try moving X, Y, Z axes (E axis won't move without heating head). Also consider heating up the print head to see if that works also. Probably not if the cabling to the head is broken. Or it might be reading "temp error" if the wiring is broken which may disable all steppers - not sure.

Just make sure you buy a "blower" and not a "fan". A fan has about 5 blades. A blower has about 50 blades. That's all you really need to know.

I'm not sure what you mean about "logging issues" but it can take 1 to 2 weeks to get replies at support.ultimaker.com and things are worse if you update your ticket as it goes to the back of the line without the support people noticing. But they *will* eventually "make it right". If you call the international number during normal work hours "netherlands time" you can very often speak to someone immediately and get much faster service. It sounds like your vertical rods or the bearings might be bad. Or the lead screw. 1) Is the lead screw on the Z well greased? Mine was smothered in grease. I thought it was too much but it seems to work fine. If not there might be a green packet of grease that you can add to the screw. 2) The bed should go up and down by pushing/pulling hard. Turn off power and grip it near the back of the machine with two hands and lift up or push down repeatedly past that "10mm spot" you refered to. Is there increasing resistance at the 10mm point? If you could get a video of the problem happening it would be nice. It's easy to move Z only without having to slice and STL. Just connect your computer through usb cable to the printer and get the cura "print window" up in "pronterface" mode which has tons of controls to let you move Z axis by any amount. Play with that and try to duplicate the problem (and then try to make a 10 second video of the problem and post).

I read it when you posted it. I didn't acknowledge it. Do you want me to? I have the same issue - I post good advice (occasionally) and it's frustrating when the person with the problem seems to completely ignore it. Just because one person (Jerry) two people (valcrow and barnacules) didn't read your advice (or ignored it simply because he happened to have only a vibratory one) doesn't mean all of us ignored it. Please believe that people *do* read the stuff on this forum both soon after you post it and also sometimes months or years later when doing google searches. It's not a waste of time. But there is usually no acknowledgement. Also please try to post things that you won't be embarrassed about a year or 5 years from now as people will still be reading it then.

That is too much pressure and not enough volume of air. You need something in between. A blower: http://www.electronics-cooling.com/1996/05/all-you-need-to-know-about-fans/

Fans need to be carefully matched to the pressure. You cant take the existing UM fan and expect it to blow air down a tiny tube - it's not that you need to shrink it down with a funnel. It's the blade shape in the fan itself that won't work with high pressure. Some fans are designed for high volume, low pressure. You need low volume high pressure. You don't need a fan that sucks down any more watts it just needs to be designed to blow through tubes. One of those centripital fans. Like you see for a bouncy castle or vacuum cleaner. Not like the kind in a window. So yes it's a great idea, just make sure you get the right kind of fan - designed for sending air down tubes.

I spoke to a woman who works at shapeways and she said they have a few designers who specifically only print/sell copyrighted objects. That's their business model and it explains a lot about how shapeways deals with this. The typical product cycle is something like this: 1) New movie comes out and suddenly Minions are popular. But they cost like $100 from disney. Or maybe there is an internet meme of Mickey Mouse battling ironman - suddenly it's everywhere on reddit, etc. 2) So a designer makes them for $50 on shapeways and starts selling them making maybe $10 each. 3) Orders start off slow, disney doesn't notice. Word spreads, after 2 weeks he has sold 1000. After 5 weeks there are 50,000 orders. 4) At this point Colbert has one on his desk on a show or it's on slashdot or cnn or facebook or whatever. Suddenly Disney notices and issues a dmca takedown to shapeways. 5) Shapeways tends to have a 3 week lead time so the first few thousand are already shipped and the designer gets $10K (or maybe only $50 depending on how popular it is). Shapeways notifies the 50,000 people that the part has been removed due to copyright violation and returns their money. Rinse and repeat. It's a strange, illegal way to make a living. Also it's strange that the designer wants thousands of people to notice the ads for this product but not millions. You want it to grow, but grow slow enough that there are lots of items already shipped "3 weeks ago" before the copyright owner notices something going on but fast enough to actually make some decent money.

Post your questions! You'll probably get lots of answers from the community of owners.