alaris2

I don't understand this, so someone please explain it to me - I calibrate steps per E, using the Cura approach. I extrude 100mm of filament with E=821.29. it's a V2 bolt extruder with Bertho's extruder mod applied. 100mm comes out. all is good. I attach the bowden tube again, and this time extrude through the hotend. procedure now is to run the following to bring me up to working temperature - M107 M92 E821.29 G21 G90 M109 S230 G92 X0 Y0 Z0 E0 G0 E20 G92 E0 and ensure the nozzle is primed. then I mark the filament as it goes into the extruder and execute G0 E100 mark again. withdraw the filament and power down. what I expect to happen is 100mm or thereabouts is extruded. but what actually happens is - 1) when the nozzle was empty (after heating) the act of priming it with 20mm of filament actually only moves 10mm of filament through the system. 2) when the nozzle is primed, pushing 100mm of filament actually moves 25mm of filament through the system! Daid says my filament is slipping. He's a clever guy so he's probably right, where else could the filament be going? but WHY is it slipping then? is my extruder really so puny? did I totally misunderstand the g-code? my observations are that in order to get good quality prints I have to set the material amount to 130% of normal to ensure I don't get under-extrusion. maybe this is why I've never had a bowden tube pop on me...

from UM. it appears that they suffered two unfortunate events at the same time for machines shipped early this year - the first was their filament suddenly started varying dramatically in diameter across a roll (I measured 2.6 to 3.1mm) and the second was a fault in ID of bowden tubes (which were too small - mine was 2.92mm). I expect it accounts for many of the problems people have been having but can't have been good for UM reputation. anyway, if you ask them nicely they are good people and will replace either if defective. the new tube they supplied is much better at 3.20mm ID and I can now use all the thick filament I had put away. but what's your opinion on the grease vs oil vs other? which do you use and what have you found?

so i broke the first rule of engineering and took something apart that wasn't broken. I decided to upgrade the bowden tube from the supplied 2.92mm ID (yes you read that right, it was limiting the filament I could use) to a nice new 3.20mm ID one. I had been waiting for the inevitable bowden popping. however it's never happened to me and didn't look like it was about to, so decided to risk taking the whole gantry, axes and head apart in the process, just for fun. I have to say that I recommend this. the first time I knew nothing about UM and had to follow the instructions. this time I followed my gut (and some experience). the result is I can now push the head around with one finger, just like in the video. new tube fits perfectly, still no popping. I also fitted some newly designed tensioners (more in another thread) and just generally spent more care and attention with alignment etc. so after printing a 2hr print, it occurs to me, I forgot to re-oil the rods. I had wiped off all the old grease and oil (and all the dust it collects!) but not re-oiled. so UM recommend that thick green grease. our current preference is for light machine oil. but someone recently suggested no oil or grease should be used on account of it collecting dust and dirt (and I can attest to that). and I remember reading somewhere that machinists use some sort of silicone lubricant that doesn't attract dust. unfortunately I forget the name and can't find the article now. if anyone can tell me what it is I'd be grateful.. so what's the opinion of the experts on lubrication?

it's a great idea and the wider tape is useful to know too - fed up with lines on the bottom of my prints! thanks Benoit

it feels like an old vacuum cleaner head would work - I'm just surprised that the idea works so I'm going to have to try that once I've reconstructed my head

no slipping. it's because my bowden tube to hotend interface adds a lot of friction. as the filament increases in diameter, the friction also increases, but that's another problem. everything works well at 1400, as long as the filament doesn't exceed 2.95mm or I get a blockage.

do you have a hole in the wooden underside of the z stage then? there's only a small gap between the acrylic and wooden base on mine?

agreed about the quality. this is the replacement silver from UM - I've already sent one back :( but faberdashery has similar problems - I have rolls from them that vary enough to give problems. the Bertho extruder mod has made a big difference in this regard tho. maybe I get through a lot more filament than others? I've gone through 4.5kg since April. I get very different values for E depending on which method I use, which is why I was trying to understand the motivation - about 890 for the 'Daid' method, vs. 1400 for the 'destroyer' method. since the latter matches more closely the machine in a working state, it seems more sensible to adopt surely? or have I totally missed a point somewhere?

can someone explain to me the logic behind steps per E calibration? my understanding is we are supposed to pull the bowden cable out of the quick release at the extruder and are thus measuring the steps required to make the extruder extrude filament a fixed distance. this seems to assume that changes due to filament diameter variations, friction in bowden tube, print speed, effects of temperature and friction in the hotend are all nil. are we saying that marlin or Cura is going to automagically fix this then? otherwise, shouldn't the procedure be to NOT pull out the bowden cable, but 1) heat the head to intended working temperature 2) extrude some material (to make up for all that which just leaked out during heating) 3) make a mark with a pen on the filament 4) extrude 100mm or whatever 5) measure how far the mark moved 6) maybe do it again to be sure having just lost yet another print due to Ultimaker Silver thickness variations I can't help thinking changing the filament diameter value in Cura (or any other slicer) is pretty futile since I have no idea what to measure if the filament keeps changing during a print. I've been compensating with M221 S110 and similar as I observe the variations (this print failed because I left it unattended)

I lost track of how many 'my windows are like this' posts you made Ian if you send me a link to the file I'll have a look for you.

it's good, but as I said earlier, do you have the model available? it would be interesting to see others print and compare the same model. also do you know that's the best a dimension 1200 can do? or was that just 'coarse' mode? because I see a whole lot of imperfections on the UM print and wonder if we can do better still. that wouldn't be enough to convince my boss to buy an UM I know - surface consistency is more important than thickness of layers.

for those that haven't seen it - http://www.creativemachines.co.uk/shop/ ... ucts_id=51 I think the reprap boys are using these j-heads and they come highly recommended but for those of us (like me) that don't have access to lathes cnc machines and all the other tools necessary to make mastory's very nice design, here's something you can just buy. of course, now the question is.. how to adapt it for use with an UM?

I don't think it has a minimum travel distance setting like Cura for this, so it may go mad with retractions in very tight spaces. I did have one of these today, but it doesn't seem to be causing failures for me, perhaps because I have Bertho's extruder mod installed. as with all software, there's no such thing as perfect

thanks Joergen, and good point about the models too. I think de-rating by a factor of 8 is quite fair then!

someone asked me what the theoretical max resolution and real max resolution we could print with UM was. I thought the first question would be easy, but can't find the answer anywhere - it would be calculated from the gear ratio, microsteps and steps per degree for the motors of course. does anyone happen to know the answer tho? the second question I was going to say 0.1mm in z (but there are occasional reports of people doing better in extraordinary circumstances) but I'm not sure what to say for X and Y. I realize repeatability is an issue here - is it fair to say 0.1mm in all axes perhaps? what do others think?

did you switch on 'round infill' on the first dialog tab? if not, it will probably shake - I haven't tried it off I admit.

k'slicer does circle infill by default i think. I'm not sure if you can change it or not, but it's definitely better than the grid infill of Cura which was shaking my UM to pieces and sounds like a sewing machine, but I do like the line infill and it may be missing that. top/bottom is about the same for all slicers i've tried. which is a shame, because it's the one bit that everyone complains about :( those settings are about what I'm using, I made a couple of minor tweaks. I don't think i can post .ini files here can I? suppose I could c&p the contents if you get stuck tho Ian, let me know. k'slicer is not as friendly as Cura. new users should always start with Cura. but when you're man enough to move on to awesome print quality and speed, it's time to switch to k'slicer.

the thing I'm not seeing anyone talk about (but wish they would) is detachable heads. (not nozzles) I'd love to release two clamps that take the whole head off the gantry (but leaving the rods in place) and snap on an ABS head with a larger nozzle, and snap on a gigapixel imaging camera and snap on a laser for laser engraving and snap on a milling bit.. etc etc. UM is an awesomely accurate XYZ table and can do more than just print. I suppose someone will say, just release the gantry at the wooden blocks. but it's not a quick operation to do this as it requires re-squaring each time. I saw a cassette style multi head on thingiverse recently which sort of went some way towards addressing this but isn't quite there yet.

I should add that I tried a hot air gun too. this works very well (for a different purpose) provided you are super careful with the temp. the different purpose was that I was trying to smooth the surface of a finished part without sanding it - you can semi-melt the surface PLA which reduces the layer strata lines and also gives a nice gloss finish. I haven't tried this for melting in an extra piece of PLA - but it might conceivably work.

I have. not very well is the answer - it makes a horrible mess of the tip too. but on a plus note - if you ever need to 'drill' holes in PLA, use a soldering iron. it's awesome at it.

if you find a good answer, I'd like to know too.. I've taken to sanding and painting as the only solution for now.

I have the same problem but tend to design around it wherever possible. I'm not sure it's fixable unless you go to a heated bed. my parts are perfect until I prise them from the bed - then all that latent curling energy kicks in and bends the corners up, especially close to where I prised it.

Ian, that's exactly the sort of print I had at one time too. now fixed I'm pleased to say by: 1) switch to Bertho extruder mod. no more variable filament problems gets rid of the sparse perimeter 2) check belt tension by all means, it causes alignment errors but shouldn't cause the spareness problem 3) switch to kisslicer. despite the silly name the difference in quality and speed will blow you away. (sorry Daid) you should be able to achieve AWESOME prints at 100mm/s. it's speeds above this that will cause problems (I get that same effect now when I reach 150mm/s) nik

Joergen is correct, and we had a related discussion here. I'd like to think my machine is good and oiled and the temperature is certainly high. I performed a number of tests with the standard 0.4mm nozzle up to 200mm/s (before Bertho's extruder) very quick summary was: at 100mm/s prints come out perfect at 120mm/s there's occasional misextrusion on perimeters at 150mm/s misalignments (ie. head positioning problem, backlash?) becomes very apparent, more misextrusion, infill becomes blobby and structurally unstable (switching to grid infill helps a bit) at 170mm/s I couldn't print anything useful. (Joergen's results show what you can do with a 0.5mm nozzle) I wouldn't recommend 120mm/s for something you want to show someone and impress them at how good UM is, but if it's for a 'quick and dirty' then it's fine. I intend to re-run the tests with Bertho's extruder mods and see if there's any change. I did already notice if I up the E value I can print very successfully at higher speeds but at the cost of overextruding on fine details. I guess it's possible to change E within the gcode however, so that suggests more speed is possible. there's something to think about Daid? post-process the SF g-code to adjust E on the fly for faster extruding - we'll love you for it

i guess i'm too ambitious then! 90% of what I print is larger than that and I often wish I had just a little more space. that's not to say what you've done isn't cool - au contraire. if you paint it pink and put fur trimmings around it you'll become a chick magnet too who couldn't resist such a cute lil UM be interested to see it finished and what quality prints you can get! keep up the good work Daid