eriksw

That looks weird. Something is going wrong with the slicing here, so you can rule out printer problems for now. Never seen this happen in Cura (nor in Repetier, but I don't use that a lot). You could download Cura from the Ultimaker site (it's free), and see if the same thing happens when you load the model in Cura.

If these bits are caused by the slicer then you would definitely see it in Cura's layer view. If you don't see anything like that there, it is not caused by Cura or a "faulty" model. In that case the problem is in the printer, mechanical or firmware. Or perhaps it is caused by cheapo Chinese PLA, but then I would expect to see a lot of other shit in other parts of the model as well. Still, don't use that stuff. Always get good quality PLA if you want good quality prints. Another tip for delta printers: accuracy of those is worse in the centre of the platform, and better towards the edges...

Never seen anything like that on Ultimaker with Cura, so I doubt this is bad slicing. Good to know, and PERHAPS this might be causing it: a model of a thin walled cylinder is never 100 % round. The cylinder has sides and although it might seem round when there are many sides (120 or more for instance), it actually isn't. Curves don't really exist in a digital system, they are always "steps".... As a result, the thickness of the wall is never exactly the same all over. This is always the case in all curved walls. So if you have a straight wall that is exactly 0.4 mm thick, when you put a curve in it there will be areas where the wall thickness will be less than 0.4 mm, which usually results in Cura "skipping" these areas. When this happens, an easy way to solve it is to scale up the model just a tiny bit. 0.5-1% is usually enough. Or you can change the setting for the nozzle size a bit, that also works. In your case it might be the other way around (so that some bits are just OVER a certain wall thickness) in which case it might help to scale the model DOWN just a bit. And you only need to scale in x and y direction, no need to scale the height as well. Again, I doubt whether this is the cause, never seen defects like that, but it can't hurt to try scaling to see if it helps.

PS I just noticed that gr5 already mentioned that this could be the cause, back in june. See page 1 of this thread, I quote: "Another solution (maybe) is to use a copy of Marlin that uses "slow fan PWM". You can do that by unchecking the 4th checkbox from the bottom here. This won't eliminate the problem but it will cause less frequent spikes in the voltage so that hopefully you get less frequent "bad temperature" readings. http://marlinbuilder.robotfuzz.com/ " (It's the 5th box from below actually, if you want to try it out.) Somehow I missed this first time around - too busy replacing all the parts I suspected, I guess....or perhaps I did read it but did not feel like messing with firmware (yet), which is tricky... and you can damage the printer if you don't know what you are doing, just so you know.

Quite a while since last I posted on this thread, but here's an update. I FINALLY found out that the temperature swing is caused by a firmware setting called "fast PWM". This setting causes much more PWM noise when you print with any fan speed other than 0 or 100. This noise interferes with the thermocouple signal. I made a version of the firmware with the "fast PWM" switched off, and now I only get very moderate swings in the temperature reading (2 or 3 degrees max) and (more important) the actual nozzle temperature actually stays in the neighbourhood of the temperature reading, instead of dropping some 10-20 degrees below it. (You can check this by first printing with reduced fan speed, then set the fan speed to 100 on the controller, and then quickly go back to the info page to check the temperature reading.) But HEY - I installed the new default firmware with Cura 15.2 and it looks like that also solves the problem ! I'll do some more checks later, but it seems like it is back to normal now. So perhaps the "fast PWM" has been switched off in this version again, or perhaps something else was altered, but I'm not a firmware expert so I'll leave that to those who are.

In case you're wondering: the smoke accident happened when I was tinkering with the print head, and I did a test print without clamping down the plug of the thermocouple (and probably didn't plug it in properly). Plug came loose, the electronics were getting no temperature reading, and that caused the heater to stay on. By the time I stopped it and stuck that plug back in, the temperature was 320 degrees...and that took a while, so it got hotter than that. Big PLA smoke cloud in my shop. So this is not bound to happen very often, however I understand that the connecting wires can break sometimes, and that could happen to anyone, even without tinkering. Has not happened here yet though - which is quite amazing, been printing almost 24/7 for 1,5 years on UM1. That's a WHOLE lot of shaking going on...

Agreed, all electronics can cause fire. But this machine also has an electric heater that is melting plastic, quite different from most other equipment I would say... People WILL be leaving their printer unattended. Or am I supposed to sit and watch the machine when it is doing a 24 hour print ? I can watch it for hours (printer meditation...) but I need some sleep as well. And I did not suggest building a smoke alarm into the UM3. The system I described above would hardly add to the cost of the printer (if at all), and would definitely make it MORE safe. 100% safety was not what I asked for. By the way: I doubt that the hot end would catch fire in case the heater stays on continuously (anyone ever tested this ?), but this WILL produce quite impressive clouds of smoke, I can tell from my own experience. Fortunately that time I was there to switch it off within a minute or so, but it would be great if the machine would do that for me, in case this ever happens again. The last thing the industry needs is press reports about houses burning down due to a printer catching fire. And the last thing I need is my shop burning down. So ANYTHING that can help minimize this risk would be a GREAT advantage, I would say, especially if it can be added without extra costs. To me 70 dollars seems reasonable for this Kickstarter Smoke Signal (if only to prevent my shop filling up with smoke), but of course 20 dollars would be better...let's hope it gets picked up by some Chinese manufacturer.

That's the one. So printers CAN burn your house down.... I wonder how often this happens, looks like chances are very low, but still... some people win the lottery as well. All the more reason to add the function as described above to ALL printers from now on. It should be simple. In the meantime I'll be waiting for that Smoke Signal to arrive... thanks for the link.

PS Google only comes up with some patents for "smoke detector power cut-off systems", but apparently there's trouble from insurance companies. Can't find any such thing on the market...

Still very happy with the UM1... what a great machine, and many ways to make it work even better, which seems a lot harder with the UM2. I really like the ability to manually move the feeder wheel (while starting prints, changing filament...), which you can't do on the UM2. Would be nice if that was possible on the UM3. Or else some dedicated buttons to do this. Could do with another 2 buttons to quickly lower/raise the platform. For safety it would be great if there was something built into the electronics to check the behaviour of the power going to the heater. It is possible to get a malfunction of the temperature measurement (in any part in that chain, but mainly broken wires due to metal fatigue from vibrations are my concern) causing the heater to stay on continuously, leading to excessively high nozzle temperature, causing clouds of smoke (had that once) or perhaps even fire (although I never heard of that happening, fortunately...). And this of course ruins the insulator if you're not there to switch it off immediately. Normally the heater switches on and off several times per minute, except during heat-up (or when the temperature gets changed by a tweak). So it should be possible to have some monitor system that would "notice" that the heater is on for (let's say) more than 30 seconds, even though the printer is not heating up for printing or due to a tweak. In that case it could either switch off the printer directly, or give out some alarm sound first for a while (to prevent any "false alarms" ruining a print) before switching off the printer. As the UM1 does not have such a system, I have been thinking about building a smoke alarm that can switch off the power to the printer when it is triggered. Anybody know if this exists already ?

@UltiArjan: nice thread about that exchangable nozzle. Very interesting, but not very accesible to most users yet, I guess...A good one for the "what do you want on the UM3" request ? And yes, 120% flow can be good for watertightness. However I have noticed that when the flow gets too high this actually makes the bottom leaky again, as the upward ridges between the print lines become too high, creating more pores again. Also, very high flow can make the outer edge of the bottom layer look ugly, so I often set the TweakAtZ plugin to increase the flow gradually during the first few layers. Another thing that has not been mentioned yet: print temperature should be quite high to get stuff waterproof. You need nice "runny" PLA to "seal" the bottom, for good layer adhesion in the wall, and it also helps to make increased flow happen more easily.

@UltiArjan: Nothing wrong with a big nozzle, but I assume the majority of UM users don't want to be bothered with changing nozzles (and releveling the bed) for each print... @Kirash4: please read my reply again: you can have the bottom as thick as you like (up to 20 cm ) but when you set the shell to 1 mm (using "spiralized") while using a 0.4 nozzle this means that your infill lines will be 1 mm apart, and there is no way you can extrude enough PLA through that 0.4 mm nozzle to get that bottom watertight, no matter how thick it is. Apparently you can file down the nozzle to create a larger "shoulder" so the above MIGHT become possible, but I haven't tried that yet as it seems a bit crude. Cups/vases with only 0.4 - 0.5 mm shell will be pretty strong already, by the way. But if you don't mind changing nozzles then UltiArjan is right, go for a bigger nozzle. And of course you could print without "spiralized". Then you can print your wall any thickness you like, because it does not have to be a "single line wall" anymore - but you will get a small vertical seam at the spot where the platform moves down.... however if there is some 3D pattern in the wall (as opposed to a smooth cone or cylinder) this seam is often hardly noticeable, so perhaps that's a better way to go for your print.

That's funny... I tried that trick some time ago, but it did not work then. I have checked, and indeed it does not work in 14.01 (which I have used for most part of last year) as that one also produces "inside skirts" (pretty annoying...) but it DOES work in later versions. Never noticed those inside skirts had disappeared, as I had set all my skirts at quite some distance around all objects with cutouts since then, because that removed the inside ones. So now I can put those skirts back to 1-2 mm again, which is nice. Thanks. Still, would be nice to have a checkbox for "inside brim", so we don't have to use this workaround, which is not obvious to many users I guess...

Using the "spiralize" function, the wall of your cup will always be a single layer, because the spiral starts at the bottom and then spirals up on top of itself, like a coil/spring. So you should usually set the shell to the same size as your nozzle (standard is 0.4 mm) when printing spiralized. You CAN increase the wall thickness by using a higher setting for the shell when using "spiralized". That way there will be more extrusion, so although it is still a single layer, the extra extrusion will be forced out of the nozzle and will be spread out sideways by the "shoulder" of the nozzle (= the flat part of the tip around the hole). This works fine with shell set at 0.5 or 0.6, but 1 mm is way too much to get a good quality print. HOWEVER: with the shell setting a bit higher than the nozzle size as described here, Cura will increase the distance between the lines of the infill. This is what causes the gaps between the extrusion lines as you mentioned (mainly because you are bound to get some under-extrusion when printing this way, definitely with shell at 1 mm). With shell set to 0.5 or 0.6 the gaps will be smaller (or non-existent), and you can close them up by increasing the flow a bit (try around 105-110%). That way the wall of your cup will be a little bit thicker as well. OR you can set the shell to the same size as your nozzle, and then use the TweakAtZ plugin to increase the flow at the height where the wall starts, so 3 mm in your case. Keep in mind that this increased flow will remain in your printer's settings after it has been changed this way, so either switch your printer off-on before the next print or set the flow back to 100 with the controller (but you can only do this AFTER you start the next print, unfortunately, as you need the TUNE menu to do this, which only appears while printing...) So if you really want your cup to have a wall thickness of 1 mm, there are 2 options: either switch off the spiralize function (watch your print time double or triple...) or use a much bigger nozzle, which is not a very sweet option. One more thing: the fact that you did not get the gaps between the lines in the first and second layer probably indicates that your nozzle is a bit too close to the platform.

Check whether you have set the right setting for "combing" in Cura 15, as it doesn't properly copy the settings from ini files made with earlier versions. Not sure whether it is causing this problem (probably not) but it's worth checking anyway... My intuition tells me it makes more sense to do the shell before the infill, as there are bound to be small imperfections at the starting points of the infill lines (blobs). These will be in the way when the shell is built, and the shell will be pushed outward at these points. When you put down the shell first, these blobs will be stuck against the inside of the shell and will therefore not affect the shape of the shell. Or a lot less, at least, I would assume. I'll check to see if I have the same trouble, haven't used v15 a lot yet - it's great to get all the improved versions, but often there are new troubles appearing, so for production I usually stick to a version that has proven to work well. It's becoming quite a skill to know what version of Cura works best for each print, it seems...