Dadkitess

Do you happen to have something specific at this 4 hours threshold ? Maybe some fast contours for instance ? You might be close to the max flow rate allowed for that given material, and UMS5 is really really bad at flow rate extrusion. If it always happen at a given spot of your print, it's maybe because of a geometry that is eager to go faster / wider / thicker layer. You say "I get an under-extrusion error of about 0.3 units under the minimum threshold" : I don't get it, what is this "0.3 units" ? Do you have some kind of feedback I don't have, when it comes to UnderExtrusion sensor error ? Maybe it's in the LOG file ? I really really often face this under extrusion sensor trigger... Be it with PVA on BBCore, or PETG / PLAtough on AACore, when I print "fast".

Ok, Thanks for your feedback, I definitely need to understand how to fix that issue on our side... By any chance, did you use the Ultimaker PETG as well ? I'm trying to switch to this material to replace the T-PLA, which is too limited when it comes to heat resistance.

Ha, I thought he answered "It used to work the way I described. I've done it many times." regarding your screenshot of the 5.7 new version, like "Yeah, I already have that menu for months" haha.

Sooo... You actually manage to get high successful rate when using the PVA as "interface support" only, the rest of the support being made with T-PLA ? Because I really won't exceed 60-70% of success when doing so, it tends to not stick the T-PLA for the floor unfortunately. It would help a LOT since it would reduce the PVA printing by 90% probably, reducing also the time, the scars left because of PrintCore changes, etc. Never found proper advices / tips about it unfortunately.

How the hell did you get that option before the 5.7 ? 😮

I don't have a Bumbulab at work / home myself, but do have a UMS5. I would tend to approve every single line except the speed single material that I would set to 1 : it's horribly slow and there is nothing you can do to accelerate it, its max flow rate is really poor. You might want to add the "perfect first layer" that would get an honest 5 with UMS5 : I've never ever ever ever failed a single first layer, be it with Polycarbonate, PA12-CF, PETG, PVA, etc. But on the opposite it lacks something about overall design flaws, the chance to get a failed print, which is definitely WAY too high if you use PVA as a daily support. Damn, that thing is horrendous, especially the Ultimaker official one. And the dedicated printcore, the BB, don't seem to help at all. It will jam so often, it will trigger the flow sensor, it won't stick to itself, to PLA, it... It's endless. You really have to have proper "basic" simple shape, with very classic parameters, and a re-dried spool to ensure a long print to success. Unfortunately, this 2 printcores configuration is the reason to be of the UMS5, there is nothing else that set it apart, nothing, except maybe that incredible first layer. So yeah, I would mitigate the overall review of UMS5 with a note on the PVA / 2 printcore configuration that is an absolute must BUT it will work only time to time and will lead to a lot of frustration. And since, apart from that and the first layer, the print speed is atrociously slow and will have ghosting even at 25 mm/s + very expensive spool + very expensive printer at all + quite noisy + really not ergonomic + multiple errors + CURA that tends to be the best AND the worst slicer sometimes + build quality issues, etc, it's hard to recommend except when, well, Soluble Support. I'm really looking forward a Bambulab IDEX. And I need to see how a Prusa XL would not do exactly the same job as the UMS5 while also correcting most of it flaws by a good margin.

Yeah, I see, I can understand that, it is capped by its PrintCore design. Well, it's quite "bad" news but it is information nonetheless and this is what I came for, feedback about what could be wrong, so that I do not miss a parameter or something that I can actually improve. It just appears that, definitely, UMS5 is made to print low speed, low flow rate, and that's it. The question remain about the 0.8mm haha. Being able to print twice as fast in 0.3mm height and 0.4 width, is equivalent to normal speed in 0.3mm height and 0.8 width, flow rate wise, and then you won't get any advantage. Lower speed would mean a bit less jittery, ghosting, but 0.8mm width means way less X/Y precision for "small" features. There is not net print speed advantage, which is the main objective of large print core. I don't get it.

Well, I was printing at 100-140 mm/s 0.25mm layer height with a Prusa i3 back in 2016-18, actually, which is really not in the same price range, nor the same construction, etc. Sooo yeah, while I "can" admit that the two printcore head can induce low print speed to avoid quality issues because of inertia (even if, huh, ghosting remains even at 20mm/s xD), this has nothing to deal with thermal performance, which should at least be on par with the contemporary printers, indeed. Again, it's a 5000+€ machine, not known to print fast for some valid and some less valid reasons, but I would have guess that the heater would not be the limit. Maybe it's not, though, I'll have to try some tests with low temp material to see if it's still problematic or not. I don't see what would be the benefit from 0.8mm if the max flow rate remains the same so I guess it's improved.

Thanks for your answer. You speak about the "limits of the thermal capacity in the Print Core" : I don't get it. It is to say that the printcore don't have the power (hence, the time) to fuse the filament when the flow rate is too fast ? If so, how could it achieve some 285°C PC print ? I guess it's 300°C capable to give some margin for that high-temps material that I've print successfully by the past. Of course, it was not a speedy profile neither an high flow rate but I guess that 235°C at high flow rate equal a 285°C sustained low flow rate with, again, some margin so that the printer won't be too inclined to stop at any occasion ? Does a 0.8mm Print Core has a higher max flow rate ? Like, if's thermal related, does it have a more powerful heater ? Because at 0.8mm, I would print, well, 0.8mm width and possibly 0.35 or 0.4mm layer height and even at 30mm/s it would reach just short of 10mm^3/s which seems to be the problematic threshold 😕 There is no point to print even slower to compensate : anyone printing with a large 0.8mm print core, does it do speed up a very long print which won't matter much of the coarse quality. This is all about flow rate. Please, tell me that UMS5 are not bound to below 8mm^3/s flow rate ?... I would prefer to read that I'm missing something and / or need some new printcore / cleaning.

Je suis preneur aussi de la réponse si elle existe, et d'une manière générale de toutes les astuces qui permettent d'avoir autant que possible une continuité dans l'impression avec le moins d'interruption possible. Mais je ne crois pas que ça existe vraiment, en dehors du Vase Mode qui n'est pas applicable ici...

Hello there, It's quite a basic one but I'll admit I will take any advice or feedback about it. I'm having recently too much of "No more material in XX spool" while there is still, because of the triggering of the flow sensor. Our UMS5 is quite picky when it comes to high-flow prints. I know, I know, Ultimaker are almost advertised as slow printers "to guarantee best results" and every topic about their printers will at a moment or another mention something like "it likes to print at 20-30 mm/s". It's really way too slow by the today standards and it was way too slow by the 5 years-ago standards as well, especially for a fully enclosed expensive printer, and I'm totally okay to trade off some quality for speed when I consider that the contours, the continuous lines, are well doable at higher speed (think of a 100mm diameter wide tube, it can print past 200mm/s on any printer or so, there is not difficulty, not inertia problem, etc.) like about 50-60mm/s for internal contours for instance. So that I can divide the whole print time by half, that's it. Since I want to reduce Print Time, I also set 0.3mm layer thickness, obviously !... Which lead to "high speed" (cough) and thick layers, or even, I dare, 0.5mm layer width 😮 ! Which, individually, is perfectly fine of course (perfectly might be a superlative) but altogether represent some kind of High Flow. I'm using a 0.4mm AA print core, don't have the 0.8mm by now, I know it would suit this case but I want to be able to deal with 0.4mm that are 98% of what we use and should be perfectly doable. So, 0.3mm thickness, 0.5mm width, 60mm/s = 9mm^3/s, I'll round at 10mm^3/s. I can't find it back but some day ago, I read that the max would be 25mm^3/s for the UMS5. My research was with AA 0.4 printcore but can't say for sure that this max was not for 0.8. But I would say that this is a "normal" value considering the ability of other brands to print way way way faster than this. I've raised the print temperature to help a little bit, but it did not solve the issue. I've slow down the print using the Live parameters when printing, and It worked. So yeah, I really have the feeling that Flow Rate is the issue here... Is it simply OK, like, this printer can't print at high flow rate, maxing at about 6-7mm^3/s maybe and I will deal with it ? Would a 0.8mm PrintCore definitely help at reaching better max flow rate by reducing the constraints and pressure ? I've especially had this issue with the official Ultimaker PETG (grey) that I'm currently experimenting, but also vaaaaaaaaaastly with PVA (but PVA, well, uh ? It just jams everywhen it can ^^) and also had it very rarely with T-PLA (Black) with the same situation of "high" speed, 0.3mm layer thickness. I'll admit that for the last print with the PETG, I did not check and clean the nozzle with the atomic procedure, maybe it's required, it's been some time I would say. But the question remains, do we know about a concrete maximum flow rate for our beloved AA 0.4mm print core ? Your ideas about it ?

He does say though : "the support slices with gaps between the layers and when I print the PVA doesn't stick to the lower layer" ==> I don't know if this representation has not any impact on the actual Gcode and travel. I find myself having hard time to print PVA that won't stick to itself and I would definitely say that it's especially the case when Support Layer Thickness is at 0.3mm when everything else print at 0.1mm, that would be ignoring the causality fallacy : the actual model i'm using this setting with, are the most difficult and very prone to fail because of tiny patch of PVA, discontinuous lines, etc. So it might just be that instead of various layer thickness being at root cause.

Yep, still exist. Try to just move / rotate a place slightly and it might do the job, as weird as it sounds. It fix 50% of my issues, even if it requires to try multiple time with loooong slicing operation before showing a fail, so not very convenient. It seems to me that playing with the horizontal expansion makes things worst. But it might be a bias, since I use this parameters on quite complex geometry to "weld" and reinforce tiny fragile connection.

Manquerait plus qu'UM en vienne à ne plus respecter ses diamètres x)

No... I don't want more support for the sake or more support, obviously, as stated in my message. Thanks for the advices, i'll definitely try the Cross connected support lines. Gyroid has helped me a bit but did not do much for tiny patch of support (around 0.5 cm²). Do you know if PVA and other brand alternative, are supposed to stick well onto the PLA Tough ? It actually does when the PLA surface is a big flat one, but I can't manage to get something reliable when it's a PVA interface onto a PLA Tough Support. It would need some kind of pre-interface in PLA Tough providing 2 whole full surface so that PVA can stick to it.

Attention, ça ne changera rien au fait que malheureusement, le PVA est une vraie galère à imprimer correctement, et c'est quasi impossible pour des toutes petites géométries : tu vas délaminer ou avoir des points de cumulations, de chauffe, etc. Les "nuages" sont caractéristiques de ce problème : le filament n'adhère pas suffisamment à lui même et au reste pour prendre des angles prononcés sur des petites distances, même à vitesse réduite ça colle un peu à la buse et du coup tu fais paté sur paté. Ajoute à cela les problèmes d'humidité qui font te ruiner localement une couche avec quelques bulles / irrégularités de débit, et c'est le reste au dessus qui risque de partir en sucette. Le PVA d'UltiMaler est globalement le pire des filaments à dissoudre. Tu peux réduire ces problèmes via un PolyDissolve S1 ou un BVOH (pas encore testé mais apparemment mieux en tout point). Mais ça ne sera jamais miraculeux. CURA est en bonne partie responsable du problème : l'impression des supports est très saccadé, rempli d'arrêt, de reprises, de mini-segments, limite des points, des rétractations à tout va, etc : il n'y a aucune continuité pour les supports de toutes petites géométrie, CURA ne lie par les surfaces entre elles pour avoir un beau débit continu qui limiterait grandement les galères. Ca bénéficierait d'ailleurs à tous les types de support, y compris ceux qui ne sont pas soluble. Courage, c'est pas évident, seche bien ton filament à fond avant une impression, imprime lentement, élargi les surfaces de support via l'Horizontal Expansion (de mémoire), trouve la forme de support qui aura le plus d'impression continue, type triangle ou gyroid. Et malheureusement, non, pas d'impression Support Soluble en Interface Seulement, car ça n'accroche vraiment pas bien, du moins chez moi... Jamais vraiment réussi. Et le PVA aime pas rester dans une buse semi-chaude pendant des heures le temps d'arriver à son heure de gloire. J'ai écrit à ce sujet sur la partie internationale, j'attends de voir si y'a des astuces et je reviendrai ici en parler, parce que ça aiderait beaucoup, de n'avoir en PVA que les interfaces ^^

Hey there, We all know that PVA is a real mess and is hard to print for various reasons. It is also something that will reduce the overall quality of your print because of the PVA getting slightly stuck when dripping, because of pauses, flow being inconsistent, etc. I've tried by the past to only print the 80% dense interface rather than the whole support, so that layer affected by the PVA are way less important, and it would avoid the famous BB Core jaming. But I quit, as it never really allow for good adhesion, leading to failures. I tried once again yesterday, with the PolyDissolve S1 which is way better than the UM PVA, and... Well, I can't rely on it for small surface of interfaces. It won't stick enough on the PLA Tough to allow a confident use of it. It kinda work for 70% of the interfaces but it delaminates for the 30% other places. I'm not sure if a bigger surface would change something actually. Do you guys have hints about it ? It really should be something possible and clearly interesting for many reasons ! Maybe a 3D interlocking ? Also, one of the biggest issue with Soluble Support is the very short travel and retractation : CURA won't ever allow for large continuous optimized support path, unfortunately, even playing with the various parameters... I am sure that we don't care a 20% more support material being printed if it allows for successful path which are not jerky little points, dots and tiny segments that only want to delaminate and accumulate, even when it's PVA on PVA... There is a lot of way to reduce the whole Support Quantity, be it inverted conical shaped, tree, stages, etc, so it's not a big deal to increase the base support a little bit. And it would actually print faster, I always have a LOT of useless tiny sections and retractions all the time.

I second that, there is a LOT of useless / not continuous / not homogeneous / not relevant support, just like the overhang rule is actually not really followed. The OP has good examples : some conic continuous slope having some support at some point, but not everywhere else while it's the exact same overhang, the exact same geometry, etc. And it's missing a lot of required overhang area too, unfortunately. And... well, the support strategy tends to be very lacking, not built to be used properly with PVA for instance : lot of very short path, lot of quick retractation, it's totally not PVA friendly and it would be "good" either for normal material. Supports needs to be as continuous as possible, and most of the time the geometry totally allow for it but computation won't achieve to something practical, even using the whole palete of experts / custom settings.

Well I've been told that this is an issue they are very aware of, so I assume it's generalized, even if in some way, I hope it's not, since CURA is the best Slicer 🙂 Unfortunately, most of the parts I can't slice are not shareable, they are confidential and when I try to simplify / adjust them so that I can share them, well, they got slicable haha. But they are definitely not weird or exotic, basic automotive industry shapes. My PC can be part of the issue, though I changed it some month ago and I already had the issue before with the other one.

I am using an alternative version in 4.x, 4.23 IIRC (don't have my PC right now) and it globally never fail at slicing when 5.1 ; 5.2 and 5.3 does fail a LOT unfortunately. With the same STL. I don't remember if I actually tried 5.4 but I read the comments about it, and about the 5.5, hoping from improvement on that topic, but it did not seem OK. And my IT department is quite sensible and won't allow installing all the versions haha, better stick with one and then update every 3-4 versions. I really 5.X arachnee when it works and I tend to avoid using the 4.x version for that reason.

Thanks for the ongoing work. Do you have some information regarding a focus on getting a proper stable 5.X version ? All the 5.X versions so far are very-very-very slow / picky when it comes to slicing, with very weird behavior or complete refusal to slice a rather normal part. I mean, there barely is no mention about it and yet new updates and added features, which is nice, but it would be even nicer to know if there is some progress about the Cura Core function stability. It's very important. If you're working on it, and I pretty sure you are, maybe you're not able to advertise about any date, but some of us would be glad to ear about it being worked so that we know that you take it as a priority.

This. Plus, why not having it around your future 3d print so that you don't forget to remove it. I don't like the prime band on PrusaSlicer for instance because it is easy to miss and it's so thick that the head can really force onto the previous prime extrusion. And if it's around the future part, then it's also the occasion to see if the first contours will print well or if does not stick well exactly at the place of the print : grease / deformed plate can be heterogeneous. All of that to ensure that the first layer will be as good as possible, let alone Brim and Raft, and so... In the Bed Adhesion category. And it would not fit better elsewhere, would it ? ^^

Then I was not clear either : I totally get that. I don't say the devs are newbies, that they should work better, that they should trade personal life for a CURA working, like, at all. That's obvious but it's even better when stated, apparently. I just want the team to know that it's not working and it's not "normal" : again should I nuance saying that I obviously don't support the idea to fire the CURA team of make them crunch ? Or can I simply state that it's not okay so far to have 5000+ machines advertised as streamlined and reliable workflow ? It includes Hardware as well as Software, one without the other won't do anything. If CURA was a preferred open-source homemade software among the actual plural software solution, being compatible, well, I would not have these words. But it's not, and we need to use it and the past year has been quite clunky, making it difficult to use. Not "not perfect", but clunky / not usable sometimes. It's a well known problem ? I don't doubt about it, but it does not appear in any way in the communication. Let the consumer know that yep, there is some long-living issues so far, that they are focused on it : it will help. It's funny coz I'm having about the exact same "debate" about a game right now, with exactly the same aspects, the same topics : no communication, not usable, etc, while defenders come to say that it's okay, don't insult the devs, don't harass them, while... Well, you guessed it, there is not insult, to harassments, only people willing to "complain" so that it's well known by the team that it's a major trouble. Anyway, since this is only the first and last words that counts on internet, I'll say and insist : it's not again the dev team, it's not against the Community Managers, etc, it's only to complain to Ultimaker society, to say that it's not right to be stucked with a clunky CURA that is required to get the best of our over-expensive machines, specifically when advertised to be pro-compliant. I'll then excuse if my words were a bit rude, because they were and it lead to your legitimate message : it sounds way too personal and familiar, and lacked of nuance. It was a hot-reaction to the thread while I actually connect to the forum to check the 5.5 official thread and it seemed to me it was particularly okay and then I saw that about slicing issues and was very disappointed. Not a valuable excuse for my message, but rather an explanation of its tone.

CURA is "sold" with UM machines. And it's actually the only way to get proper UFP files. Simplify3D as well but I don't know if it is still the nor if it's okay-ish or not. Since UM machines are using a specific file format, yeah, i'm sorry but I do expect a working software along it, be it free / open source. It's totally fine to get new features needing some polish and all but Cura IS needed to get UM machine working, that's it. Especially when UM S5 machines are sold are easy to use, semi-pro very reliable printers. I would not have the discourse for another printer, of course.

Dude, we're still stuck with slicing not being able to complete ?... Every 5.X version got crippled with this "bug", and it's real pain in the a$$... I'm using it for work and most of the slicing I ask are either taking 5 minutes or more, or won't ever be done, with either no message, just a full stall without taxing any calculation, or a "can't slice" error which will require many tries with weird modif like just a little offset of the part on the plate, or just 25% of infill rather than 30%, etc. It's a bit annoying / irritating. It's really 80% of my print, that are rather complex / big but nothing crazy at all. Like the tubes shown in this 5.5 thumbnail for instance ! Can we have a 5.X milestone that would be Pro-compliant, even if not getting all the new fancy stuff ?