alaris2

Sander, first it's good to see real UM people on this forum. it's been far too long. I hope it's not just a flying visit. regarding print quality - I have been trying to showcase how awesome the UM is to potential buyers here in the UK, but the comments are always the same - the reliability and quality are just not good enough. they would prefer to buy Makerbot or stick with Shapeways. the reasons for this are: the problems seem to be threefold. first is caused by variation in filament diameter which can cause grinding, blocked bowden tubes, variable layer quality, poor adhesion etc. perhaps improving the hotend and extruder can help this, we are waiting for details of what might be happening officially regarding this. second is caused by spikes on the power supply which cause the arduino to hang. leave a long (>6hr) print to run ,especially overnight, and expect to wake up to a finished model? not happened yet. I can't start a print and walk away because something will always go wrong. I can give you a much longer list if you really want. PS. the filament I now use is from faberdashery, which is far better than (I'm sorry to say) that supplied by yourselves. The first roll supplied with the machine had terrible tolerance (2.65-3.05mm) and the second roll you supplied as a replacement after the 'bad batch from suppliers' had been sorted, was almost as bad (2.61-2.98mm). faberdashery filament is usually 2.84+-0.1 and it still has some problems but not so many. (on the plus side, your customer support is awesome however). third is the software. it makes a massive difference to the quality of the print I have found. repG is unusuable, Cura is better and a good starting point for beginners, but slow and only average quality (not Daid's fault but SF). there are other slicers which produce superior results but have no official support and are therefore too complex for new users (eg. kisslicer). this might not seem like a UM problem - but the first impression for a new user and for those being shown the UM in action are very much dependent on that print quality. so the problem here is getting the right settings to people who are using x,y,z slicer so they can use an UM. these points are made as constructive criticism only. I bought an UM because i believed it was the best of what was available. I still believe that (although the advantage is quickly being eroded by other printers) but I have difficulty convincing others that UM is best and they should buy it. best regards, nik

grumble grumble, split community across multiple forums again grumble.

depends whether you live in Florida or not. I've tested several models in a hot garage which regularly reaches 35-40C and those models are still OK. I accidentally left one in a car on a warm (not hot) day, no thermometer but probably 50-60C at a guess. the model softened and did a Dali on me with lots of droopy arms. would have been a nice effect had I wanted to do that..

same problem here Ian. I've now given up using UM for prints at work and taken a more reliable route. it's a shame, but the UM isn't robust or reliable enough to be used this way - you can get a great print one day, and an awful one the next.

that's the one. silly name, ghastly UI, but does a good (9/10) job on pretty much anything. Yes i've had the exact same problem with netfabb, my finger is hovering over the 'uninstall' button..

I'm afraid my 6 months of printing experience sides with Joergen on this. the V2 bolt is actually pretty good, rarely needs cleaning and I've found no reason to use the V3 bolt. the 'faults' lie mostly with the hotend.

completely agreed - I think you meant to say PFA rather than PEEK tho. note this infers the size of hole in the PFA tube to be <= that in the brass tube or you can't push it down. unfortunately as the filament grows in diameter (which by virtue of the molten plastic working its way up into the tube it effectively does) friction increases and we can't push it down. this may be hampered by the fact the PFA tube has some elasticity and can expand slightly whereas the brass tube won't. why isn't the hotend all metal? eg. like http://forums.reprap.org/read.php?1,145069,page=1

k'slicer has a particularly nice g-code viewer. very helpful for 'proof reading' and it just found me a bunch of inside-out geometry that allowed me to correct it before wasting 7hrs print. netfabb, on the other hand. took my inside-out geometry, part fixed it, ignored the rest and somehow found a way to increase the print time by a factor of 1.2. not cool.

correct. i got a new one from UM. it measured 3.2mm ID. I put it in the PEEK, printed for a couple of hours, then dismantled to inspect it. the ID was now 3.1mm. it shrinks, probably due to the OD being slightly larger than the ID of the PEEK so the material has to go somewhere and heat softens and reforms it to the new size? dunno, but it happens. whatever you do, don't try to enlarge it. you'll create plugs. i already tried that.. :(

for PLA, it's weak when pulled on the z-axis, strong when pushed on the z-axis de-lamination in x and y occur due to insufficient surface contact across the size of the print. strongest way to print the part in the photo there would be to rely mostly on the infill for strength - use a couple of perimeters and a circular infill. nik

you shouldn't hope for better than 0.5mm accuracy really with UM. Like Joergen says, it is theoretically capable but I have yet to print anything that was better than holes = 0.5mm too small and pegs = 0.1mm too large. some slicers like netfabb (costs money) or k'slicer (free) can be tuned to give quite accurate parts by automatically shrinking and enlarging holes and edges - but if your filament varies in diameter at any point you can expect errors to creep in. most people adjust the design to take this into account..

is this some sort of silicone tubing? my fuel tube here is black, with thick walls and not at all suitable i think. you don't mean fuel tube from a car do you?

you saying something about russian tractors? join the club. loads of people have fans like a tractor. the fans on the bottom are terrible. there have been several threads about this but no real solution (other than replace the fan!)

some good thoughts there tachyio - the heat sink and blower arrangement is what I have been experimenting with too - altho mine is at 90 degrees to yours. I have come to the conclusion that the PEEK serves no real purpose - perhaps someone would care to correct me - why do we want to retain heat in the brass tube near the interface with the bowden tube thus risking the PLA (especially during heavy retraction) reaching the glass transition temp at 55C and forming plugs of doom? we surely need a cold zone above the hot zone, especially now that retraction is in common use.

I've been doing that jcosmo. I find that reducing the layer height on the final layers gives a much better surface. k'slicer also avoids many of those travel line problems. unfortunately there are 2 ways to get a 'perfect' surface, and no slicer properly implements them at present. 1) tilt your model slightly (print on a triangular raft) so that the top surface is angled. for 50 micron printing, tilt at about 15 degrees for perfect surfaces. currently you have to adjust the model for this - or use repG to adjust the rotation and then add support, but the support will mar the bottom surface now.. 2) print first half of model at one layer height (say 0.2mm), then the last part, including top surface or any detailed part, at 50 microns. currently you have to c&p two files to get this.

segmentation faults? that's unfortunate because you're missing a treat. OK, I'll print you some proof as soon as I'm done with my current run. retraction settings are not easy to understand in k'slicer which is ironic given it's name. Cura wins hands-down for user-friendlyness still. you'll want the following settings for good retraction Mr Gecko Sir - style -> de-string = on material -> PLA -> destring suck & prime = 2, destring speed = 70, destring min =5 this is approximately the same as Cura then. note that I do not recommend the higher value of 4.5mm retraction that some people use. it leads to so called 'plugs'. k'slicer does have one bad habit that Cura again beats it with - it tends to retract liberally when it could merely back track over or inside an object and avoid the retraction. this can lead to filament grinding in very small complex areas if not careful. and I'll add that it's not Cura's fault that k'slicer is better, Daid is limited by what SF will let him do ultimately. he's still done a sterling job for new users, so keep it up. but we're well off topic now. let's get back to heated beds. I'd still love to hear of one I can actually build. the best I saw was by Jensa, but ended with the unfortunate caveat of 'it doesn't get hot enough for ABS'. can anyone do better?

I wouldn't waste time or money on netfabb it's installed on my machine for only one reason - it can fix broken stl files from 3rd party sources. when Daid has that feature in Cura, I'll be so happy to hit the uninstall button on netfabb. meanwhile, it's k'slicer for me I'm afraid. the difference in quality just blows you away.

Cura sets the first layer to have different properties - set it to the same height as all the other layers to give better quality as per k'slicer. However external quality of surfaces is far better with k'slicer, so I'd stick with that if I were you. You are correct, heated bed is the way to go for better quality - I also came to this conclusion, but like you realized there are far too many 'solutions' of varying quality out there and no official or even 'most recommended' solution. heated bed would be one of my top 3 wish list upgrades for UM right now. thedude> you're right it is possible to adjust the bed this way - I had mine set up that way too for a long time. but since I've been upgrading the hotend recently I decided to keep the bed level but have the wrong z and just manually adjust it until I'm done with the mods. I also discovered printing at 50 microns is pretty challenging anyway - some form of automated levelling is in order here. random thought - why not a 3 screw position bed (instead of 4) arranged as a triangle with a hole under the bed (instead of the wooden tray) so you can use a hairdryer to 'heat bed' the part off afterwards. someone already reported some success with hairdryers. now I come to think of it. why is the bed acrylic anyway? what's the benefit? why not copper or aluminum and just heat it from the underside when you want parts to stick or unstick? for ABS it only needs to be a hot bed for the first few layers surely, after that it just needs to be a warm bed doesn't it? perhaps destroyer or someone who prints ABS can comment?

yes, retraction is the main reason. especially since people are now using this feature from the word go.

I've stopped using 230 and gone down to 200. I found 230 causes what people are calling 'plugs' I wouldn't recommend such a high temperature to anyone now. we started with 'how low can you go?' and got to 180 and lower, but found structural integrity was a problem, then we seemed to go to 'how high can you go?' and have reached a different set of problems. 200-210 is the safe place I think.

I had that same dilemma with the ABS fan shroud, but your diffuser idea sounds about right. I just tried an experimental 8hr print with no fan at all (PLA) and used external cooling (desk fan blowing down over the bed) which worked pretty awesome actually. (and it was in UM silver for a bonus point) the tinfoil idea is excellent for levelling - I used the sticky kind of aluminum tape that is used for leak proofing garden sheds and have it permanently on the bed. but I find, as you do, that the bed goes out of whack when forcing parts off it.

ah, large cylinders. I printed a 9hr one of those and vowed never again! for a moment I thought you'd downloaded a car.. but you wouldn't do that I am also looking for paying jobs and find the quality and reliability keeps letting me down. I've just finished re-designing the hot end for vastly improved performance and hope to have some results shortly. I already upgraded the extruder to Bertho's mod, which is good, but still not good enough. keeping the table level and to within 50 microns and heated beds are also on my list - if you'd like to share/swap ideas then I'm all ears. PVA support is on my list too actually. I tried the fish tank air idea already - I had a spare bowden tube. it works very well, but you need a flared attachment at the end to spread the air or it just goes everywhere. you're right, dual extrusion is what it's best used for, I had intended to give Daid the design since he was complaining he had no fans now he had dual extrusion but I never got around to finishing it off. (sorry Daid, but you can have where I got to if you like) it's great to see people thinking about upgrades nik

I think I've come across the same problem since upgrading my bowden tube to the new 3.2mm ID one. twice now I've had a failure caused by a plug of PLA forming in the very end of the bowden tube. it happens slowly over a period of some hours and I think it caused by retraction. the plug is about 5-6mm long of PLA exactly the ID of the tube and located right at the end of the tube. this is strange because the filament in both cases was quite small - 2.64mm and 2.70mm respectively. I think what is happening is that retraction causes molten or soft PLA from the hot zone in the brass tube to be brought up into the bowden tube. we can assume the PLA inside the brass tube is close to the diameter of the tube (anyone know what that is? 3mm?) as it comes into the bowden tube it cools slightly. when the filament presses down on the soft PLA it pushes it to the side (hence the large amount of friction in my system requiring an E of 1040 and not 850?) gradually this build up solidifies - leading to slipping at the extruder side and no more extrusion. note this is NOT the typical bowden popping diagram where the PLA sneaks between the brass and bowden tube and pushes it up. I've checked in both cases and the surfaces were flush mating with no problems - the plug is inside the bowden tube. this seems to be a problem where the filament diameter is less than the ID of the bowden. is there any way to avoid it in a good design? the interface between bowden tube and brass tube is the weak point as several have rightly commented. how do we stop plugs?

30hrs!? what on earth are you printing?

what would make it miss steps? or how could I find out whether it's missed steps or slipping? I've tried observing the movement of the filament, but it looks pretty smooth - wouldn't I expect it to be jerky if it were slipping?