olivierc

Ok I think I found a bug, I don't know if it has been reported yet, but here it is, I attached a video for you to see. I used cura 13.06 to slice this one. I'm printing the legs of a character, turned upside down. Because I'm kinda in a hurry (this is for a gift) I used cura's supports instead of building my own like I usually do. During the printing, I noticed an unusual amount a stringing while I have retraction turned on, and I'm not using a very high temperature (210). I have printed models with much more retraction involved and very little to almost no stringing. For a sec I thought maybe I forgot to enable retraction, but I can hear the retraction noise. So I observed the pattern of the head and paid close attention to when the retraction happened, and I notice something: It does not retract when moving from the support located on the left side of the object to the support on the right side. Problem is, the object is right between the two. However it retracts correctly when moving from the support on the right to the leg on the right and from the leg on the right to the leg on the left. So I'm wondering, does cura check if retraction is necessary when moving between supports? Look at the video and listen to the retraction. Sorry, I filmed with my phone Sorry if this has been reported before EDIT: Another problem that I noticed only since I have been using the beta: when the printer automatically slows down when the area get smaller at the end of a print (like the top of a sphere), it does not adjust the flow rate too, and it creates big blobs of PLA

Ok, first of all, before saying "Why would you want to go under 0.1mm? it's overkill blahblahblah", yes I'm aware of that, for most part I print, 0.1mm per layer is perfect. I just ran into a couple of long, thin objects (the branches of the glasses for a character I am printing) that if I print vertically, it looks ugly, just a stack of blob, and if I print horizontally at 0.1mm, it looks better, but not so good, you can see the layers because the objects are long and slightly curved. So I want to try to print horizontally but at a higher resolution. The objects are small they only took 2 minutes to print at 0.1, so I can afford doubling the printing time So I often read about the the ultimaker "you can print at 0.1mm and even lower than that if you machine is well tuned". So my question is, what does "well tuned" means? What really needs to be done to print at 0.05mm? My belts are tight, my axis are straight, bed is level... I'm going to try anyway, I mean if a 5 minutes print fails, it's not the end of the world. But I was wondering, what is it , that prevents higher resolutions to be officially supported by the ultimaker?

Hello Albert, oui c'est un modele que j'ai fait moi meme, que voici: Comme tu peux le voir, une fois assembler, il sera tres gros. Mais les autres pieces ne sont pas encore pretes à imprimer et j'ai un autre perso à imprimer d'abord pour l'anniversaire d'un copain, donc il est pas pret d'etre fini Il a ete realisé il y a plusieurs années, pas du tout avec l'intension de l'imprimer, donc il a fallu modifier pas mal de chose pour y arriver. Oui je compte le peindre. D'abord je cleane avec une lame, l'enleve les cacas. Apres que ponce d'abord avec un papier moyen puis un fin pour liser. Il faut y aller doucement avec le moyen pour ne pas effacer de detail, mais le fin à lui seul ne suffit pas pour affacer les lignes des couches. Ensuite je passe une couche de primer en bombe, j'ai deja essayé, ca marche très bien. Blanc ou gris, ca depend ds couleurs que je vais passer par dessus. Pour des couleur vive, il vaut mieux du blanc. Ca permet aussi de voir des defauts qui etaient pas visibles avant, et de poncer d'avantage, pour repeindre ensuite. A partir de là, tu peux peindre par dessus avec ce que tu veux. Pour ma part j'utilise de la peinture acrylique toute bete, meme pas de la peinture pour maquette, juste de la peinture en gros tube achetée dans une boutique d'art. Je pense que des peintures enamel pour maquette comme Humbrol marcheront bien aussi, c'est du plastique apres tout. Je prefere l'acrylique juste parce c'est pas cher et que je les ai deja sous la main. Et aussi parce que l'enamel a tendance a ruiner mes pinceaux. Pour ce qui est d'imprimer les deux peices separement et de les coller ensuite, je ne pense pas que ca marcehra aussi bien, car la surface en contact avec la plaque a souvent tendance a se contracter en refroidissant, si tu n'utilises pas de raft. Tu te retrouve avec une surface non plane et legerement plus petite que prevue. donc bonne chance pour coller ca apres.

I have been using the beta for a bit now. The slicing speed improvement is amazing, no question about that. But on the other hand the software is very laggy in general, it keeps freezing for a second or two for every single action I make whether it's changing a value or just navigating in the viewport. This was not the case in previous releases. At first I thought it was because the model I was using was too heavy, but it also happens with small models or even when there is no model at all. I'll keep using that one over the others, but I was wondering if I was the only one experiencing that. I run 13.06 on windows vista 64bit

That's what those screws were for !! I was wondering. I ended up with the little bag of unused screw, it did not say anywhere in the doc what they were for. I'll definitely put them on, because those pulley ruined several good print of me.

Hello, je vais partager avec vous un petit hack que j'ai reussi a mettre au point avec l'aide d'Owen. Tout d'abord, la petite histoire. Le weekend dernier j'ai voulu imprimer une tres grosse piece, le torse d'un perso que j'ai modelise, pour faire une tres grosse figurine de 40 cm environ. C'est un projet sur lequel je bosse depuis un mois, ca m'a pris tres lontemps pour preparer les pieces, les decoupers, les fusionner, les cleaner.. Cura m'a predi 31 heures de print rien que pour celle-ci. Et oui C'est deja tres long, car il s'agit d'une piece non seulement grande mais aussi large, avec plein de supports. Et ce que je ne savais pas encore c'est que Cura s'etait completement plante dans ses estimations, puisqu'au final, l'impression aura dure plus de 40 heures. Je lance donc l'impression, le dimanche soir, un peu nerveux, car la plus longue que j'avais fait jusqu'a present ne depassait pas les 7 heures. Ce soir la, je n'ai pas dormi chez moi et le lendemain je suis alle directement travailler, donc je n'ai pas pu surveiller l'impression pendant 24h. Je vous dis pas le stress en rentrant chez moi le lendemain soir. A ma grande surprise, je trouve ca: Tout se passe bien, je suis content, d'autant que le plus difficile est passe, il n'y a plus de support et de jump dans le reste de l'impression. Bon le petit probleme c'est que d'apres l'estimation de cura il ne devrait rester que deux ou trois heures d'impression, or il reste un bon tiers a faire alors que l'imprimante a tourne plus de 24h. C'est inquietant car j'habite dans un petit appart' et meme si je mets l'imprimante dans la salle de bain pour qu'il y ait moins de bruit, je me vois mal dormir avec ca dans la piece a cote. Donc j'attends, mais rien a faire, a 3h du matin il n'a toujours pas fini et impossible de fermer l'oeil. Je dois me lever a 7h30 pour aller travailler, super. Finalement je me m'assoupi vers 4h, mais je me reveille vers 5h, a cause du bruit. Bon, comme je suis reveille, je vais jeter un coup d'oeil a l'impression. Et la c'est le drame: la tete d'impression est environ 2 mm au dessus model, plus rien ne sort de la buse. JE suis encore endormi, mais je reagi vite, je sais que c'est foutu, je coupe tout. Je regarde de plus pres, le bowden tube a saute des deux cotes, la tete et l'extrudeur, et j'ai un gros bouchon dans la tete. Je pense que c'est le tube qui a saute au niveau de la tete, car a l'extrudeur j'avais installe un clamp. Ca a provoque le bouchon, qui a ensuite entraine l'ejection du tube de l'extrudeur malgre le clamp, car l'extudeur continuait de pouser le filament qui etait bloque a la sortie du tube. Bref, c'est la que j'ai regrete de pas avoir monte le clamp pour la tete d'impression, que j'avais pourtant imprime il y a un mois. C'est maintenant chose faite, mais un peu tard. Bon je vous cache pas que j'etais un peu au bout du rouleau, 34 heures d'impression et une nuit blanche pour rien. j'avais plus trop la motive pour recommencer. En regardant mon print fichu de plus pres, je constate que la buse a vraiment du se bouchee subitement et pas progressivement. En grattant juste un peu quelques cacas, j'ai une coupe tres net et propre: Aussitot, je me dis que je pourrais presque coller la partie manquante dessus si je l'imprimais separement. Puis je me dis, pourquoi ne pas imprimer directement par dessus puisque je n'ai pas decole le model de plateforme. Pas de probleme de colle et de surface non plane. Il est deja pile au bon endroit, il suffirait de reprendre l'impression la ou elle s'est arrete. Evidemment plus facile a dire qu'a faire. Je me dis que ca doit etre possible en editant le gcode existant. La logique est la suivante: retirer du fichier tout les layers deja imprimes, apres la mise a zero de tous les axes, ajouter une commande pour mette le lit au bon niveau, et mettre d'extrudeur a la bonne valeur egalement, car le gcode pour l'ultimaker utilise des valeurs absolues. Donc pour chaque extrusion on ne specifie pas une petite quantite a extruder, mais on incremente la longueur total extrudee depuis le debut de l'impression. Il faut donc dire au programme que l'extrudeur a deja extrude toute la quantite des layers precedant, car si on le laisse a zero, a la premiere extrusion, il va passer de 0 a cette nouvelle valeur et extruder un spagetti de 200m. A part cela, il faut pas oublier de remettre le ventillo, car par defaut il ne s'active qu'apres le premier layer, cette commande saute avec tous les layers qu'on supprime Le seul inconnu dans tout ca, c'est le layer a partir duquel il faut recommencer. Dans un premier temps j'essaye de trouver a vu de nez dans Cura, en faisant defiler les layers un par un, celui qui correspond le plus. C'est pas evident, mais je tombe sur le layer 958. En cherchant ce layer dans le gcode, je trouve que l'axe Z est a 96.10mm. C'est la que je realise que les valeurs sont toutes absolues et pas relatives et que donc il y a plus simple et plus precis que de chercher le layer dans Cura: l'ULTICONTROLLER. Et oui, grace au fonction du menu Prepare, je peux remettre tous les axes a 0 avec la fonction Home, puis ensuite, on abaisse le lit d'impression jusqu'a ce que la buse d'impression effleure le sommet de l'objet imprime. Il suffit de lire la valeur de Z sur l'ulticontoleur et de chercher cette valeur, precedee de la lettre Z, dans le gcode, et voila, vous trouvez le layer exact ou reprendre. Ou plutot, le dernier layer imprime, il faut reprendre au suivant. Donc, je suprime tout ce qui se trouve entre l'initialisation de l'impression et ce layer et je me retrouve avec un gcode qui ressemble a ca: ;Generated with Cura_SteamEngine 1.0 M109 T0 S210.000000 T0 ;Sliced at: Sat 08 Jun 2013 09:12:05 ;Basic settings: Layer height: 0.1 Walls: 0.8 Fill: 20 ;Print time: #P_TIME# ;Filament used: #F_AMNT#m #F_WGHT#g ;Filament cost: #F_COST# G21 ;metric values G90 ;absolute positioning M107 ;start with the fan off G28 X0 Y0 ;move X/Y to min endstops G28 Z0 ;move Z to min endstops ; ---- MODIFICATION A PARTIR D'ICI ;G1 Z15.0 F9000 ;move the platform down 15mm ON SUPPRIME OU COMMENTE CETTE LIGNE QUI ABAISSE LA PLATEFORME DE 15MM G1 Z96.10 F2400 ;ON AJOUTE CETTE LIGNE POUR ABAISSER LA PLATEFORME AVEC LA VALEUR Z DU NOUVEAU PREMIER LAYER G92 E0 ;zero the extruded length G1 F200 E3 ;extrude 3mm of feed stock ;G92 E0 ;zero the extruded length again ;ON COMMENTE OU SUPPRIME CETTE LIGNE G92 E30189.53316 ;ON REMPLACE PAR CETTE LIGNE. LA VALEUR DE E EST COPIEE DANS LA DERNIERE LIGNE DU LAYER PRECEDANT NOTRE NOUVEAU LAYER DE DEPART. ON PRENDRA SOINT DE LA COPIER AVANT DE SUPPRIMER LES LAYERS G1 F9000 M117 Printing... M106 S255 ;MET LE VENTILO EN MARCHE ;ON COMMENCE AU LAYER 958, ON SUPPRIME TOUS LES LAYER PRECEDANTS ;LAYER:958 G0 F9000 X130.51 Y69.41 Z96.10 M0 ;ON AJOUTE UNE PAUSE ICI, POUR VERIFIER QU'ON S'EST PAS PLANTE, IL SUFFIT D'APPUYER SUR L'ULTICONTROLLER POUR REPRENDRE ;TYPE:WALL-INNER G1 F3000 X130.67 Y68.83 E30189.53692 G1 X130.51 Y69.41 E30189.54068 G1 X134.75 Y70.65 E30189.56841 G1 X136.84 Y71.34 E30189.58220 G1 X137.39 Y71.56 E30189.58595 G1 X137.67 Y71.76 E30189.58808 G1 X137.90 Y71.93 E30189.58984 ... Voila. Si le nouveau layer de depart commence par G1 F2400 E... C'est une retraction, il faut supprimer cette ligne 4 jours plus tard, je retiens mon souffle, je lance l'impression et.... TADA! Alors oui, il y a un joint visible, probablement car j'aurais du commencer un layer plus haut, je n'etais pas sur. Mais bon, rien qu'une lame et du papier de verre ne puissent effacer. Ce n'est pas parfait mais entre 10 minutes de poncage et 40 heures d'impression, le choix est vite vu. Pour une impression de 1h ou 2, je ne pense pas que ca vaille le coup, il vaut mieux recommencer proprement. Mais s'il s'agit de tres long print, que vous avez un plantage "propre", et que vous comptez poncer votre model (ou que vous vous en foutez), c'est une solution interessante.

So here it is, finished So you definitely see the seam, but I think it's moslty due to me getting the wrong layer: I printed the same layer twice and the same height. But I already manage to erase most of it with a blade. Once sanded and painted, it should look good. Overall, I would not recommend using this hack if a 2 hours print fail, you're better off starting over. However, if like me, your prints fails after 35 hours and you don't mind having a seam in your print, it's a really cool trick to know

So there we go it's printing. Unfortunately, looks like I was a couple of layers too low, so there is a big seem visible. Even though I had prepared 3 files starting from layer 955 to 957, I could not really see the result until it had printed a few layer. Oh well, nevermind, a bit of sanding and carving with a blade should fix it. I'm still happy. I'll post a photo when it's done. Thanks for the help guys!

Yes! that's what I got! I just need the rods, I tried to order from conrad like you suggested but my visa payment could not go through, because the shipping address is in belgium while the address attached to the card is in france. Some sites do these kind of security check. I must ask some co-worker to order it for me but keep forgetting

Ok, so I measured with the ulticontroler, and I'm able lower the nozzle down to 95.8mm. At this height I can move the head with no problem, if I go 0.1mm lower (really hard btw, you have to old the wheel between 2 clicks), it's scratched the surface of the print. So, do you think the first layer should be the one that starts at 95.8mm, or the one on top at 95.9 (yes, I'm using 0.1mm layers)? I would say the one at 95.9,but I'm not sure I'm glad that I did this check, because with my previous estimation just by looking at layer in cura was at 96.1, I was close, but I could have got some air-print. Unfortunately, I might not be able to test tonight because my girlfriend is coming over, but definitely tomorrow If I'm successful, I will post a recap of the prcedure in the Art of Printing

Yes they are called Makerbot, but too bad they lack in the QA department

awesome, will try that tonight and keep you updated. This is what I'm trying to printing. The torso was the hardest part to prepare and print. If that works, then move onto the legs, boots, head, hands... this is going to be a very long print, but if I'm successful, it's going to be huge

As for having this implemented directly in cura, I think it's a little too advanced. The reason cura is great is because it's simple and accesible to new user. This is more of a hack

Hi guy, thanks for all the help. Thank you GR, I like the pause idea, that's precisely what I was thinking of last night: I'm not 100% sure I got the right layer to start from, all layers between 955 and 960 seem they could fit, but only one has the correct height. it it's too high, it's ok, it will print above the model, I'd just stop it. However, if it's too low, the burning hot nozzle would do some damage to the the model. At first I thought maybe I can just make a few gcode files, that only contain the init code with a low temperature, where the printer level itself to the target height and then stop. I could make one for each of those layers and see which one matches the best. But then I realize, why not just use the ulticontroler to to home the printer and then position it to the perfect height, read the Z value and find the matching layer? What do you think, is the ulticontroler accurate enough? I could not try last night because was not home, I need a break from the printer.

OK, Let's see if I understood... The only thing I miss is turning the fan on. Should I turn it on immediately or should I wait for the first layer to complete, this way it might merge better with the existing part? I'm not going to try it tonight anyway, so take your time. My printer has been running for almost two days non-stop, I want to spend a quiet evening. ;Generated with Cura_SteamEngine 1.0 M109 T0 S210.000000 T0 ;Sliced at: Sat 08 Jun 2013 09:12:05 ;Basic settings: Layer height: 0.1 Walls: 0.8 Fill: 20 ;Print time: #P_TIME# ;Filament used: #F_AMNT#m #F_WGHT#g ;Filament cost: #F_COST# G21 ;metric values G90 ;absolute positioning M107 ;start with the fan off G28 X0 Y0 ;move X/Y to min endstops G28 Z0 ;move Z to min endstops ;G1 Z15.0 F9000 ;move the platform down 15mm COMMENTED THIS LINE FROM ORIGINAL HEADER G1 Z96.10 F2400 ;REPLACED WITH THIS LINE, Z VALUE FROM LAYER 958 G92 E0 ;zero the extruded length G1 F200 E3 ;extrude 3mm of feed stock ; I KEPT THIS BECAUSE I LIKE IT ;G92 E0 ;zero the extruded length again COMMENTED THIS LINE G92 E30189.53316 ;REPLACED WITH THIS LINE, LAST E VALUE FROM LAYER 957 G1 F9000 ; DO I STILL NEED THIS LINE? M117 Printing... ;G1 F3000 X131.77 Y70.16 E30189.53316 LAST LINE FROM LAYER 957, KEPT FOR THE CORRECT E VALUE. NORETRACTION USED AT THE END OF THIS LAYER ;NEED TO TURN FAN ON SOMEWHERE AROUND HERE ;LAYER:958 G0 F9000 X130.51 Y69.41 Z96.10 ;TYPE:WALL-INNER G1 F3000 X130.67 Y68.83 E30189.53692 G1 X130.51 Y69.41 E30189.54068 G1 X134.75 Y70.65 E30189.56841 G1 X136.84 Y71.34 E30189.58220 G1 X137.39 Y71.56 E30189.58595 G1 X137.67 Y71.76 E30189.58808 G1 X137.90 Y71.93 E30189.58984 ... Edit: thank you for the explanation GR5, it makes sense, I'll be more cautious before starting a long print in the future

I'm not as experience as Snowy and I'm not asking for as many things because I'm not particularly interested in dual extrusion or printing ABS, but must agree that since I have been using my ultimaker, the pulleys getting loose and bowden tube popping out have been an issue. Don't get me wrong, I'm very happy with my ultimaker in general and the other point Snowy stated, I know that 3d pritning is not "plug n' play" yet, but it's frustrating when a print fails after 10 hours because a pulley got loose

Awesome! thank you so much! So I don't even have to calculate the Z value, it's already in the layer. I thought that since the layer height was constant, it would automatically increase by the layer height for every new layer, I did not think it would require an absolute value every time. Anyway, that's cool, that was the one thing I was worried to get wrong. I'll give it a shot tonight and I'll keep you updated, can't wait to try. If it works, I will owe you much EDIT: man, threads like that should be pinned somewhere, I'ms ure it can be very useful to many people.

Xeno, yes I had found your post on the same subject, and will do that ultimately if the other way does not work (I will imediatly shut down the printer if it goes wrong, it should not damage it too much) Owen, thank you, that's exactly what I wanted to do. Yes I used Cura, the beta version, I had to, because version 14.04 could not slice such a big model, it ran out of memory. I understand everything you say except zeroing the extruder, which number correspond to what? I will definitely post here tonight. I'm at work now and I did not think to bring my gcode with me. I think I do have the model here and I could slice it again, but I'd prefer to use the same gcode as the original to be sure it's a perfect match (I might have moved or turned the model before slicing). I'll keep you updated. Also, on a side note, since we are in cura's subforum, I'm sure Daid already knows that, but I notice the longer the print, the less accurate the estimation of duration was. For a small test object, it matches to the minute, but when I print something that takes half an hours, I begin hiving a couple extra minutes. For bigger stuff that take approximatly 7 hours, cura is generally about half an hour off (it predicts something around 6.5 hours). And for this one, cura predicted 31 hours, which should have ended last night around 2am, but when it failed, it had been running for over 34 hours and as you can see, not quiet there yet, probably at least 2 more hours. I'm printing at 100% speed. Maybe it's because I have a lot of supports, combined with retraction

Ok, I'm having the same problem, my print failed after 34 freaking hours!! I got up a 5am this morning, and the printer was "air-printing" or more like not printing at al, il was clogged, and the bowden tube had popped on BOTH side, eventhough I used the clamp upgrade and the extruder side. I don't know if it's the bowden tube that popped-out first and the caused the clog or the other way around. The funny thing is that is failed maybe 5 minutes before I got up, the head not even 1mm above the print. I had a clamp for the hot end too but never mounted it. now I did, maybe a bit late. Anyway. if it had been a 4 hours print, I would have said "screw it and start over" but here, it has been printing for a day and a half and it looks really good, it only need a couple more hours to finish, I don't want to start over. Sure I could print the missing part separately and glue them together (I will do that in the worst case scenario), but I manage to scratch off the couple messed up layers, it looks very clean now, I could print over that. I'm not sure I understand what Ian did. I get the boolean part but then I'm confused. I'd be more interested by a strictly gcode approach. Here is what I want to do: load the sliced model in cura to find the exact layer where it failed (I'm printing a detailed character, so I could easily find it, it's not like finding a layer in the middle of a cube). Then I would edit the existing gcode and delete all the layers before that one. Finally, in the initialization, after the homing command, but before the start of the print, insert a command to lower the bed to the right level (number of layer*layer thickness) I don't know shit about editing gcode, it's my first time, I don't even know if that feasible. It's probably not as simple as it sound. I would only have one shot, so please help me! I don't want to mess it up! EDIT: By the way, this is for a HUGE project, I'm printing one of my own characters, fully standing. I cut the body in several part and will assemble them after, the whole thing should stand about 40cm tall. The part I was trying to print is the torso, which will be about 15 cm EDIT2: Here is a photo of the print, as you see, it's pretty clean for a failed print, I generally have spaghetti of PLA stuck everywhere. And yes, there is enought room to home the print head

Pour de la modelisation "artistique" (comprendre par la, lorsqu'on cherche a faire un truc joli avant tout et pas un truc qui reponde a des criteres techniques precis), il y a blender comme cite plus haut, mais beaucoup moins connu et pourtant excellent, il y a Silo, pas gratuit, mais pas trop cher (environ 150 euros) qui est vraiment un excellent modelleur polygonal. Il ne fait que de la modelisation et rien d'autre, mais il le fait bien avec de tres bons outils. Moi qui travaille dans maya et 3dsmax tous les jours, je prefere celui la pour ce qui est de la modelisation pure et dure. Apres le top comme package complet, c'est modo, mais il est de plus en plus cher a chaque nouvelle version :( ( mais il reste moins cher que des logiciels comme 3dsmax) Pour la modelisation de type sculpture, il y a biensur Zbrush, la reference en la matiere, il coute environ 700 euros. Ca peut sembler cher, mais c'est avec ce soft que sont realisees la pluspart des creatures dans des films tels que le seigneurs des anneaux. Il y a aussi mudbox, a peu pres au meme prix, moins puissant et moins complet que Zbrush, mais beaucoup plus facile d'utilisation. C'est celui que j'utilise. Apres, dans les "budget", il y a sculptris qui est gratuit et assez sympa, mais assez limite, on peut pas tout avoir. Et puis il y a 3d coat, un modeleur a base de voxel, qui coute environ 300 euros. J'ai jamais essaye mais il a l'air sympa.

I think it's time for the direct drive upgrade, the fewer pulleys and belts, the better

Alright, I feel stupid I was just coming just to say that I found my problem, when I saw your message. it was inteed a pulley, but the one on the stepper, the only one I forgot to check. I always check the ones on the axis because they get loose more often, but I forgot to check those two. I'm very sorry for wasting your time with something that stupid. It's not the first time I get a loosen pulley, it was the first thing I checked, but missed that one. The funny thing is that I still manage to print a couple of test cubes that looked ok., it's only when I tried something bigger that I saw it happening again.

Thank you for the detailed answered illuminarti. the infill looks rough because of those last three or four layers that are all over the place, it looked much nicer before, all perfectly aligned. I had cleaned and oiled the axes just before starting this print, since it was a long one. I don't have any problem sliding the X and Y manually. There is a bit of resistance overall, but I can't tell if it's right, maybe the belts are a bit too tight. I can't feel any difference between X and Y I will try to build a custom version of mariln with the setting you recommend, I have always been using the default one, I have not upgraded since april.

That's a possibility because I had many supports around it for the arms of the character. But I would not expect it to offset the head by a whole half centimeter

Hi, I need you guys help because I have no idea what's going on. I have read about objects printing completely sideway because of badly tightened parts, but here its different. I was printing a big model, cura predicted a 30 hours print. It printed well for over 10 hours. But then out of nowhere, the layers started to shift to the left, badly, like several millimeter offset for each new layer. The layers themselves seems they would have been printed correctly, but at the wrong place, like the head did not position correctly at the beginning of each new layer. I immediately stopped the print of course. I then checked the printer, pulleys, belts, axis... everything looks normal, I don't see what could have failed. I did a quick test cube after that and it printed fine. Could it be the gcode? I used cura 13.05.2 beta The reason I'm really clueless is that it happened suddenly and not progressively, and that I can't find anything wrong with the hardware I attached some photos so you can see for yourself