abstract

OK, so today I started testing based on the theory that the 'clear' nozzle size is less than 0.4mm. I also wanted to see what temperature does. So, instead of the default setting (210 ? at 100% flow), I tried the following: 230 degrees, with 90-95% flow (manually fiddling) 240 degrees with 95% flow interestingly it ejected that little thread of black (from 3 prints ago) halfway through the print. although it might have come from the *outside*of the nozzle (which still has lots of black on it). Given that I measured the threads at 0.38mm and also this filament is about 2.9mm, one might expect the flow rate to be required as (0.38/0.4) * (2.85/2.9) * 100 which would be about 93.5%. Clearly either the rate or the temperature has improved the printing (marginally...), but I feel like we are getting a bit more scientific with matching the available flow rate with the supply rate. Also, I have noticed that we get some minor under-extrusion (*without* a motor kick) as the print rate changes in this model - looks like the speed of movement increases before the flow rate has increased - this leads to one 'thin' section of one of the early layers in each section. I think we can ignore that for the purposes of testing. Gonna take a closer look inside the head next. Mark

So, on the theory of the nozzle size. I measured the little tips hanging out the top of these: They all come up as 0.37mm and 0.38mm diameter. These are now cooled and they came back up through the nozzle when I pulled, so I am not sure whether we expect them to be 0.4mm in this state. Any comments ? Anyone else tried this test and measured them ? Thanks Mark

OK, Tomorrow the head is coming apart. The more I think about it, the more I buy the 'nozzle opening' theory. This is (by its very nature) the 'bottleneck' of the entire system, and if its smaller than 0.4mm, presumably the filament *has* to back-up if the flow rate is based on a 0.4mm hole. I have heard people talking about needles of the appropriate diameter to clear the nozzle. Otherwise I guess the 'burn-it-out' method sounds plausible. I guess I would quite like to get a 0.4mm needle through the hole just to be sure. Will let you know how it goes tomorrow. I have requested new teflon and nozzle from UM. Thanks for all the encouragement. Cheers Mark

Yes, all 3 fans work as expected. here is a photo of the head so you can see the general condition. Please could someone describe in detail how I should investigate the state of the teflon piece ? Thanks Mark

OK, so here are the results so far: 1) Tried the Hot Push - Cold Pull method: Removed bowden tube from head, heated up to 180 degrees, pushed in filament manually until it extrudes, set temp to 90 degress, wait.... at 90 Degrees, sharp pull filament out of head. Repeat until happy. Below you will see the various pulls, in order You can see the first one was really nasty, and it got much better with the other pulls. Doing this test you really start to realise the implication of the curve on the filament - in order to get it into the head from the top you *really* have to straighten it, or it wont go in - so I can start to appreciate the work the bowden tube is doing to resist the curve on the filament. So, having done that I am feeling really confident, so I go for another print. However on first attempt to load the filament, it doesn't come out of the head. Looks like its jammed. So, I retract the filament, make a new clean break, really straighten the first few inches, and load it up again. The resulting print is below.. Hmm... even worse than before.... So, next, I tried a 260 degree flush - again, removed bowden tube, heated to 260 degrees and pushed through as hard as I could. Now, I reload the filament. Again, it doesn't load properly. Worse this time - I can't retract - its stuck in the tube. So, I take the bowden tube out again, and here is what I see: This is not the first time I have seen this in the last 6 months, so this problem seems to recur Anyway - managed to tidy that up, and I am starting to get a bad feeling about this filament. It is the last 20% of a reel of Ultimaker Black PLA, and it has been on the back of the UM2 in my garage for the last month or so. When I try to snap it, its alot softer than some other PLA (like the stuff inside the house) - so maybe its damp ? Anyway, I am giving up on this reel of filament, in case that is the cause. I have a short length of Faberdashery RED PLA which has never been tightly wound on a reel, and it has been nice and dry inside.... Fingers crossed - I load it up and go for another print: No better. It felt like the flow was improved when it was good - but the extruder still skipped, even part way through the 3mm section. If I pushed the material REALLY hard through the extruder motor, it DID work better, and didn't skip so much, but it was very tough. So, I am no further on with results, but I feel like I am getting closer to the problem. It seems that the fundamental problem is that I can't get enough material off the reel and out of the head. There would seem to be 3 possibilities: 1) The Head is still blocked in some way. 2) The Bowden tube system is constricting the movement of the filament 3) The feeder motor cant push hard enough to overcome the resistance to the head. When I have had the bowden tube off the head, I have used 'move material' and it doesn't seem to have much trouble pushing the filament as far as the end of the bowden tube - so I would like to think I can exclude the actual tube itself. So, I think we have constriction at the head, causing a bottleneck and the feed motor which is moving at the requested rate has to skip. The constriction seems to 'build up' after the print starts - the first 30 seconds seems OK. OK, there are my observations. Hopefully the good folks here can apply some analysis and come up with a hypothesis and we can try something else. Thanks for your input. Mark.

It skips, not grinds.

I am in London, England. I am going to try the hot push cold pull thing, and also the 260 degree flush thing you mentioned. I am partly tempted to dismantle the hot end, but since its not 'broken' at present I am a little reluctant. Maybe I can buy a replacement teflon and print head ? if I am going to dismantle I might as well replace. The manual feeding of filament and pushing hard certainly helps, but not beyond 4mm - so I can see that tidying up the filament feeding will be a longer term project. For now I really think I must have an obstruction preventing me going past 3-4mm - so I will focus on clearing the head first. Will let you know how this first step works out (hot in cold out, and flushing at 260). I will also put on a new spool with full radius. I really think this 10mm test thread has done a long way to rationalising and understanding under-extrusion. Often if you can measure something like this up to a guaranteed point of failure, its much more helpful than trying to diagnose random failures from time to time. Thanks. Mark

Thanks for the tip - certainly I think the feeder has something to do with the problem, but even if I really babysit the feeding and push hard, I still have problems at 4mm Incidentally - having read may other posts, it seems that UM2 is supposed to come with some sort of filament guide which you go around between the spool and the feeder. ** I did not get one with my UM2 - it was quite an early one, so maybe this was added later ** So, currently my filament comes off the spool and direct into the feeder. Should I have this filament guide ??? Where can I get one ? Thanks Mark.

If anyone wants to guide me through solving this - here is my current problem: I have had the UM2 for nearly 6 months, and I have got reasonably experienced. However, recently I have been sensing a few problems with under extrusion - most recently I noticed that the first layer was sometimes skipping, presumably because it was a higher rate (0.3mm x 20mm/sec) compared to all my prints (0.1mm x 50mm/sec). I print PLA at 210 degrees. Overall as a printer I am fairly happy with the results, because I print slow - but clearly my UM2 has issues, revealed with this test print today - first time I have tried it. Thoughts: 1) I am getting near the end of a reel of Ultimaker Black 2) The one time I tried printing ABS I gave up after a hour, maybe something is still in there 3) My UM2 is in my garage. Its pretty dry, but not like inside a house. If anyone wants to take a shot at the best first thing to try - fire away. I would dream of 10mm clean print…... Thanks Mark

So, I found the UM2 extrusion volume test print, and gave that a try, after levelling the bed and making sure the screws were tightened down far enough to give the springs something to do. I think this more or less identifies the problem. It's struggling to even do 4 cubic mm per sec. I found that if I really straightened out the bowden tube at the motor and the head, and pushed hard on the filament entering the extruder, it helped, but that really only got me up just beyond 4mm3/sec Anyone got any tips. I feel like my UM2 is deteriorating since its original performance - what are the maintenance steps so restore the performance ? Thanks Mark

So, I just tried changing the first layer to 0.2mm instead of 0.3mm and the first print looks good, very flat and full traces. Since the bed offset is the same, and only the head offset vs bead size has changed, it would suggest that the problem before was the effective flow volume - ie. it was not giving me enough PLA for a 0.3mm bead. 0.2mm at 20mm/sec is fine, and equally 0.1mm at 50mm/sec is fine for the rest of the print. 0.3mm at 20mm/sec at 210 degrees should be fine I presume, so if we think this was the problem, might it suggest my head is slightly blocked ??? it can't do 6 cubic mm/sec reliably ? (but can do 5). The kick in the extruder is presumably an effect of resistance in the feed through the head ??? Thanks. Mark

I am using Cura to produce REPRAP format, which is printed through Octoprint, but basically all defaults. Although the main speed default is 50mm, the first layer speed is 20mm/sec on defaults, so the real volume on the first layer, is actually 0.3x0.4x20 = 2.4 cubic mm/sec - I guess this should be OK ? I am printing PLA onto 75c glass at 220c I did re-level the bed, but from illuminarti's mail, I guess maybe the gap is too *small* ??? Looking at the head, it's not 'squeezed' against the glass on layer 1 - if anything its too high - the gap seems to be about the height of the unsqueezed bead - maybe 0.3 to 0.4mm ?? When the extrusion barfs, the bead gets very thin, and then its not extruded touching the glass and it waves in mid-air until the extruder kicks in again with full flow. The 'piece of paper friction' method is a bit vague - can anyone tell me the actual desired bed offset in mm ??? Could this be a friction setting issue on the extruder ? I am guessing not. looking at the prints, this does seem to be primarily a 'first layer' problem - since the rest of the prints seem pretty perfect. Thanks Mark

Folks Ultimaker 2, generally working great. I have noticed a problem recently where the extruder motor skips every now and then, on the first layer, and this leads to under extrusion for a few seconds. This ends up with gaps or very fine sections. I notice it mostly in the first layer, which is set to 0.3mm. I print with the standard settings, 0.1mm layers, 50mm per sec. First layer is 0.3mm Using Ultimaker black PLA Photo example enclosed: Any ideas on what to try ? Cheers Mark

After a couple of months experience using UM2, and Octoprint, I thought I would share some of the info. I have my UM2 outside in a garage, connected to a Raspberry Pi running Octoprint (octopi). I try to reduce the number of trips in and out using Octoprint to do some of the work. Primarily Octoprint is used for the Camera, which is a logitech WebCam which hangs from the top front of the UM2 from a little bracket I printed. The camera works pretty well, although I can't see the while bed from that position and it can't focus on the front part of the bed. However, it gives me a pretty good picture of roughly what is happening, including failure and completion. I have previously posted about my challenges with cold startup - sometimes the UM2 refuses to start because the temperature of the print head is 5 degrees or less. In this case I have to warm the head a bit with a hair dryer - only takes a few seconds, then UM2 starts OK. The major thing I try to optimise is how much time I have to spend stood in front of the UM2 waiting for the bed to warm up. This can take quite some time to go from 5 degrees to 75 degrees. One technique is based on starting the UM2 (optional hairdryer sequence), and using the Advanced menu to manually warm the bed to 75 degrees. (this is trip 1 outside). Then I go in and wait for a while, perhaps prepare an SD card. Then on trip 2, I can often insert the card, and hit print - because the bed is already hot, I can generally wait a little while and babysit the startup of the print. I try to do this due to fairly common issues with startup such as nothing coming out (too much PLA dripped out during head warmup) or avoiding mess contaminating the print when the pre-extrusion PLA gets involved in the print. Another option is to simply start the print, and let the UM2 sort out heating up the bed from cold, and the new firmware now delays the head heat up. This can often work, although a number of times I have found that the print simply doesn't start (not sure why) and I go out to find the head front left, bed up, but no printing started. Hence, my babysitting. Now, to the Octoprint : This is a great tool, and running on a Raspberry PI it is very convenient. I use the Octoprint for 3 jobs. 1) Camera 2) Manual Warm up, before a UM2 front panel print 3) Complete Print jobs. 1) I have already talked about above 2) If the UM2 is already turned on, from a previous print, I will often connect with the Octoprint and set the bed to warm up to 75 degrees. This allows me to monitor the warm up from inside, then only go out and start a front panel print (babysitting startup) when the UM2 is almost ready to go 3) Sometimes I will do the entire print from Octoprint. Note that this requires you to make a machine setting in Cura which outputs RepRap instead of UM2 code, since it requires some extra steps which the UM2 front panel normally inserts. In this case I manually warm up the bed, then manually warm the head, export RepRap GCode from Cura, drop into the Octoprint web page, then start the print. Quite often this will work nicely. I have observed that the retraction in RepRep mode is not as good as the native UM2 language retraction, so I usually only use method 3) when I don't need any retraction. Also, by default, the RepRap code doesn't lower the bed, although no big deal. Overall, the Octoprint adds a lot of extra flexibility and limits the number of trips I make out to the garage. Once you get used to all the variables its pretty easy to juggle these different methods. One thing which I can't really get around is the need to wait for the bed to cool before removing a print, then warming it up again. This adds quite a bit of time between prints. Trying to remove a print with a hot bed is tricky and usually results in you damaging the print. Pops off nicely when cool tho (especially in a cold environment). I clean the glass in warm water once in a while, then apply a full coating of gluestick, then carry on with that for a few days - I don't bother re-applying it, since it seems to leave a residue even after a print which serves the next one. I have tried PLA from a few companies so far. The UM2 blue was nice, good prints, I then got 100M of Black and White PLA from Faberdashery and was astonished to find it didn't come on a spool. So I ended up unspooling all the blue and manually winding on the Black (that was a chore I can tell you). So, a tip - make sure your PLA comes on a spool !!! I also got 100m of white PLA/PHA from ColorFabb (which did come on a spool) but that led to some blockage, and I have abandoned that reel. I have also got some 10m samples of PLA and ABS from 3D Filaprint, but not tested those yet. The Faberdashery 3mm PLA material has printed *great*, but I still need to find a good UK supplier with quality filament *on a reel*. Outside of the UM2 ecosystem, (I am mostly making engineering parts) I am using Sketchup Pro on Mac, with a couple of plug ins : susolid and a STL import / exporter. SuSolid is a great tool and let's you easily visualise model glitches, which can be essential if you want to make 'solids' which can be combined by Sketchup. I had been using Blender to convert .stl to .dae to import stuff into Sketchup, but I have now found a good STL Import plug in which is working great. The very best thing I have bought (apart from the UM2) is a 3DConnexion Space Navigator 3D mouse. This works really well with Sketchup and *transforms* the experience of 3D Modelling. Hope all this info helps people new to Ultimaker.

I guess my suggestion for a fix might be that if the low temp is detected (cold or failed temp probe), the firmware should turn on the heater for 10 seconds, then try again, before failing.

Follow up on Cold temperature startup. So - every day this week I have had to wake up UM2 with a hair dryer. Normally it's just a question of getting past the startup temp sensor error - once I warm up a few degrees, it boots up and we can begin. Today, I did the same routine, then went to maintenance and manually set the build plate to warm up. I went away for ten minutes and came back and the build plate was still at 3 degrees C. So it seemed that the warm up process was failing, even though we got past the boot up. I suppose that perhaps the initial 'warm' with the hair dryer got it to boot but then it cooled again below the threshold and perhaps the same sensor now stops the warm up (but no error message). I dont experience *any* problems printing in that environment - once the build plate warms, the whole UM2 box gets nice and warm. One benefit is that after printing, the build plate gets so cold the models just pop-off automatically ! Anyway - this startup hassle is somewhat troublesome, and it would be great to see some work done on this. Thanks !

On a related topic of temperature, I had a different problem today - the UM2 failed its startup self check and reported a TEMPERATURE SENSOR error / failure. My UM2 is stored in a garage, and I turned it on late at night for an overnight print. The bed and head were reading about 5 degrees C when I manually connected with Octoprint. The UM2's internal code failed to startup and told me to contact Ultimaker. Tried restarting several times - machine is dead. I figured it might be the temperature, so I went out with a hairdryer and warmed everything up a little. After that - the UM2 *did* startup OK and all was well. So, there is a vulnerability in the startup code - if the UM2 is quite cold (e.g. 5 degrees C) the unit may incorrectly fail to startup.

I have noticed a trend with the new 1113 firmware for Ultimaker 2. Mine is stored in a garage, and with the winter getting a little colder, often the UM2 warms up the bed from a start of about 7 degrees C. Also with the 1113 firmware the default temp seems to go to 75C for PLA (which works well). So, the bed has a long warm up time, and may take ten minutes to get to 75C. Meanwhile the extruder has got to 220C in a couple of minutes and is busy peeing PLA out the bottom whilst it waits for the bed to get hot. However, with the very large difference in warm up time for the bed and the extruder, I am seeing a situation where the extruder drips so much PLA, it seems to have eaten back into the filament without pushing any more. This can lead to the first print failing to start, since there is a big missing gap of filament above the extruder. I have to manually push the filament to fill up the gap, and then having done that, restart the print and all is well….. Until the next big warm up. I think the warm up sequence should hold off heating the extruder until the bed is within 10% of its target temperature. Then it would not suffer this long wait and the consequential problem.

Re: Retraction. I just downloaded Cura 13.11 and sent the new firmware to the UM2. Then I printed the MiniCalTest.stl (fresh slicing in Cura, which DID show the vertical retraction lines in preview). The results were great - in the sample print which came with the UM2 there was LOTS of stringing, between the pillars. In the print I just did there was ZERO stringing. So, Cura 11.13 seems to be an improvement on retraction. Also, I had been getting some of the issues with prints aborting when it failed to read the SD card - I believe this has been fixed with the new firmware with 11.13

I got my UM2 last week, and I have some comments, following on with this thread. - The Glass plate had slipped in shipping, breaking the back right clip. Clearly this problem is common and I should imagine they have already changed the packing method. I am waiting for a replacement clip to arrive from Ultimaker, meanwhile, I am using it without the back clip. - On the whole I am very, very pleased with the UM2. I had no (reality-check) experience of 3D printing before, so my expectations were very high. Here are some things I have learned, which apply to UM2 and probably in general. - The quality / design of the model seems to be crucial. Sympathetic modelling can work around any weaknesses in the FFD 3D printing mechanism. I am now avoiding any total overhangs (so I can work without support), sticking to an overhang angle of around 60 degrees max. - I am modelling in Sketchup Pro v8 with the susolid plug in. I am vigilant in making sure my models are all 'solid' - by checking that each component shows a 'volume' in the entity information in Sketchup. When the solid breaks, I use susolid to identify the problem and auto fix it, or fix it manually. Eliminating holes in the model makes for strong clean prints. - 0.1mm layer height seems to be what to use. I tried 0.2mm and the difference in quality is huge. 0.1mm = perfect, 0.2mm = disappointing. - Cura 13.10 doesn't show the retraction lines, so its not clear if any retraction is going on. I don't have any experience to tell if I should be getting better retraction - certainly I do get stringing in my prints, so who knows. Selecting "Add printer" then "cancel" in Cura 13.10 makes the lines show up, along with some new retraction settings in the side bar. No idea if I am just switching into UM1 mode???? - The glue stick thing works well. No need for tape. You can wash off the glue with warm water, easily. If you let the bed cool right down and your printer is stored somewhere cold (mine is in a garage), the next day you can just rub the glue off the glass, and it comes off without washing. Then add new glue. Sometimes its tough to get a big print off the bed, but brute force has worked so far. - I am getting a bit of ringing, not sure how to work on that. My prints are very clean and high quality, so you can notice the ringing. Its almost perfect, but eliminating the ringing would be nice. I am printing at the default speed of 50mm/sec. So far just using the standard Ultimaker Blue PLA which came with the unit. I have ordered some Faberdashery PLA, I hope it won't bring any complications. Not tried ABS yet. - The UM2 itself is really nicely built, the LEDs are great, its pretty quiet, compact, and (so far) mechanically-reliiable. - I have got a Raspberry Pi connected, running Octoprint. The webcam feature is working, but so far I have not been able to use Octoprint to control the UM2 reliably. I am guessing Octoprint has not been setup for UM2 yet (its often used with UM1 I believe). It would be nice to send a file to the Octoprint and get it printing on the UM2 remotely, but at present I am still shuffling SD2 cards and controlling from the front panel. I am mostly using UM2 for mechanical engineering, and some architectural modelling. See the pic below of some results of my first week as a 3D printer. Per my comment about the quality of the model, the faceting on the large pillar base is due to the default circle-resolution in SketchUp - not the UM2. The UM2 is capable of much smoother curves, so you need to increase the facets when modelling for best results. I hope my comments and results will give Naz72 some optimism. Using 0.1mm and taking it easy.