3poro

Right... I haven't yet had time to play a lot with PLA. It looks like I need to use higher temperatures than before in order to use same retraction settings as before. I don't know whether this is logical or not - maybe I should start trying different retraction settings. What comes to ABS, I'm very happy. As ABS is more important for me, I have used much more time to try it out. Here are a few photos of my most challenging print so far: There's quite a bit of retraction involved - also some overhangs are quite tricky. Prior to my closed chamber hack, I also suffered from warping. Now the geometry aspects seem to be in order. The photos above have been taken using sunlight and flash in a way any problems would get accented. In real life the piece feels fairly smooth and very solid. BTW - with raft, the size of the basis is 203x203mm. In order for me to get it printed, I needed to bend the clips holding the glass bed in its place - otherwise they hit the print head.

I'm actually rather surprised the 2nd hot-end makes such a big problem, as the original UM2 printing head leaves so much room for quick fixes. One quick fix I made in my hack (which I forgot to mention in my original post) was "burning" the stainless steel isolator: ...as you can see, the hollow bolt is not shiny but it's rather dark - which again is good for heat dissipation. AFAIU, there would be sufficient space underneath the lowermost aluminium block to add thermal insulation. The absorption of radiated heat from the nozzle unit(s) should be possible to bring down quite a bit without major changes. Also, the contact area between the Teflon part and the stainless steel isolator would be possible to make significantly smaller. I believe you have tried many tricks already, but at the same time I'm always cautious about "more invasive changes" - as they normally come with side effects.

I actually ran into a similar problem with Robert's feeder. It took me a while to figure out what the problem was; in my case the yoke was barely long enough to press the bearing against the knurled wheel. I don't know if the feeder has bent slightly due to heat from the stepper motor - in my case it's closer to it than normally: http://umforum.ultimaker.com/index.php?/gallery/image/6106-feeder-close-up/ I first recognized retraction problems with softer filaments - then random under-extrusion problems appeared. I fixed the problem by removing some material from the yoke and the feeder arm in order to allow the yoke to travel an additional 1mm or so. In my case the feeder arm was also a bit loose fit to the feeder body, so I added a washer to make it tighter. Otherwise the arm turned slightly inwards our outwards causing the bearing to touch the filament only by its edge. Now the feeder works like charm again I've been very happy with the freely revolving feeder stepper motor hack. I haven't had any problems with the filament path anymore and I have better consistency of print quality throughout the print area.

Which part of the Teflon gets too hot? Is it the lowermost part of it inside the hollow stainless steel isolator bolt - or the visible part right above the bolt? When I faced a meltdown due to Pt100 failure, my PTFE part was slightly malformed in the way it didn't fit anymore to the stainless steel isolator properly. It look very much like what Kris describes in this photo: http://picload.org/image/liacpoo/picture_t3.jpg ...of course, in my case the temperature of the nozzle unit got probably a bit out of scale. I have experimented a bit with additional cooling: http://umforum.ultimaker.com/index.php?/topic/6307-um2-twin-tornadoes/ If I had a need to cool the printing head further, I would probably add a thermally resistant plate under the lowermost aluminium block (right above the nozzle unit). Now the nozzle unit radiates heat more than necessary to the aluminium block right above it.

Actually, this comparison I should have realized to do already earlier - here are thermal images from the front and from the rear, taken from the fans: I took several images from front and from rear, having the printing head in different areas of the build plate. The temperature difference was consistently 10C. Given that I have 2 fans (each 5.13 CFM - I know they don't equal to having a 10.26 CFM fan) replacing the original 3 CFM fan, I think the temperature difference is significant. At least the PTFE part gets much cooler air than it would get with the original setup. Of course, my closed chamber makes my setup difficult to compare with the original setup without the dome. My firsts tests with challenging (from warping point of view) ABS prints are very encouraging. I'll try to get some photos uploaded tomorrow...

Nano particles are dangerous also because of their small size, but they also have their chemical properties. They also can stick to each other and to surrounding surface materials. The more aggressive they are (chemically), the more likely they will do so. Therefore, in real world environments, time does help - but estimating the impact without thorough analysis is very difficult. Hmm... That's a very interesting idea... What do you think about devices like these: http://www.amazon.co.uk/High-Efficiency-Electrostatic-Precipitator-Filter/dp/B00J0O39T6 http://www.amazon.com/Kmall24-Amazing-Portable-Purifier-Travel/dp/B00L525TVW http://www.amazon.com/Hewe-HC-727-Home-Ionizer-Purifier/dp/B0014X6R22 I'm now thinking about placing something inside the dome.

I agree - smell and nano particles are not always correlated. However, if the burning/heating process produces both sensible smell and nano particles and if you isolate it in a way you sense 90% less of smell, I would assume also the number of nano particles has decreased. Whether the decrease is 90% is a different question. If candles flicker, the burning process changes, but there is no such change in case of printing with ABS - unless nozzle temperature would be changed significantly. At the moment the (by far) biggest gap in my dome is the one needed for the Bowden tube and the cables to come out from the printing head. The size of it is 3x5cm (15cm2) and it is well below the level of the printing head. As the hot stepper motor is right under the gap, the fumes and other particles need to do some acrobatics to come out from the chamber. I'm not saying the problem is now solved, but I do believe there's a clear improvement over the starting point. One reason why I think the fumes are less of a problem is the delay after which the ABS smell appears. When the printer was open, I could smell the fumes in a few seconds after starting a print - now it takes at least a minute before I start smelling anything. I believe during the delay also the fumes and the nano particles lose part of their aggressiveness. Again, I'm not claiming the time is enough to make them harmless, but some impact it should have.

You are right about the dangers of the ABS fumes. The room is quite well ventilated, but when my Ultimaker was open, even the maximum level of ventilation didn't help. Now, even if I turn the ventilation off, I don't feel discomfort in the room. I understand this doesn't yet make the room safe to stay in, at least for longer periods of time. However, I assume there is at least some correlation between level of smell and level of dangerous fumes and nano particles. As my next effort I'll make my dome even tighter by closing any unnecessary holes/gaps - this serves the temperature management as well. I'll also place there some soft material (an old cotton cloth or something) which absorbs small particles better than the hard plastic walls do.

Now I have (finally) gotten a chance to test the hack a bit more extensively with ABS (I guess I should thank Ultimaker at this point or something). It looks like there's quite little need for vapor extraction, at least regarding odors. I still can smell I'm printing ABS rather than PLA, but now the room feels fresh - not like before, when smell of ABS was fairly dominant. At some point - maybe - I might consider a fan and a filter combo inside the dome. Now it doesn't feel like a priority, though.

After initial tests, I feel quite good about the hack... My starting point is a bit peculiar due to the closed chamber: As the outside surface temperature of my dome is around 38C (in air-conditioned room), I believe the air inside the dome must be over 45C - and above to the heated bed naturally more. The image above was taken when printing ABS at 260C, having the build plate temperature at 90C. Unfortunately I don't have before and after images of my printing head, but here are some thermal images after installing the Tornado fans: I need to install an additional lens to my thermal camera to get more precise images... However, if the front fan temperature is around 65C and the PTFE temperature varies around 63C-74C, it looks like the ventilation is quite efficient. Unfortunately I have no way to observe the temperature on the back side of the PTFE part. Also, there are some tricks with the measurement, as I need to open the front window of my printer in order to capture thermal images. This will naturally start cooling the chamber immediately, unless I placed the printer to my sauna. At that point my wife would probably start calling around for some professional help. Further updates after further tests - now it looks like retraction works without issues and ABS warping is well under control.

I got this setup ready for testing last night: http://umforum.ultimaker.com/index.php?/topic/6307-um2-twin-tornadoes/ Now I need to find time to perform some tests...

Thank you Kris for inspiration - I have worked on the topic a bit and I'll update after some tests: http://umforum.ultimaker.com/index.php?/topic/6307-um2-twin-tornadoes/

As I have built a closed chamber for my UM2 (http://umforum.ultimaker.com/index.php?/topic/5921-um2-greenhouse-edition/), I have been particularly interested in improving cooling of the printing head. Kris posted some really inspiring analyses here: http://umforum.ultimaker.com/index.php?/topic/5953-idea-on-underextrusion/ ...and I felt like trying my luck with more powerful fans. These were the most powerful ones I could find, given the size constraints: http://www.yeahracing.com/catalog/yeah-racing-tornado-high-speed-25x25mm-ya0201-p-2520.html http://www.yeahracing.com/catalog/yeah-racing-hacktronic-tornado-high-speed-ball-bearing-25mm-htn304-p-3471.html In an attempt to maximize the heat dissipation, I got the aluminium blocks black anodized. The anodized layer is removed from the joint surfaces between the two lowermost blocks in order to improve thermal conduction: To avoid excessive absorbing of heat from the nozzle unit, I removed the anonized layer from the bottom of the lowermost block: ...also, the holes for the nozzles are drilled slightly bigger from underneath (except for the last 1mm or so) in order to reduce the absorption of heat from the nozzle unit. I installed 2 of the Yeah Racing Tornado fans to the printing head: Both fans are blowing from the front to the rear - in other words, I inverted the airflow. Arctic Silver was used between the two lowermost aluminium blocks. I had to make small dents to the lowermost plastic block in the printing head for wiring purposes. If i wouldn't have the dome around my printer, the noise from the new fans would drive me nuts. Fortunately the high-frequency noises are absorbed quite nicely by the PC and acrylic windows, so the increase in noise level is subtle. My first impression is - I certainly need to find a new balance for retraction settings. I'll post an update after some testing.

Yesterday I received the rest of the parts I was in need of - specifically, I needed a new PTFE part as the original one had gotten slightly malformed. I believe the malformation was because of overheating of the nozzle unit during the meltdown. I also ordered some spares for the spare parts in an attempt to avoid such a lengthy downtime with my printer. It turned out to be possible to pay for the parts, although not online. IMHO the PTFE part should have been covered by the warranty, but I don't mind too much - the prices of the spares were mostly within some reason.

Today, 9 weeks after opening the ticket, the spare parts were delivered. OK, some parts were delivered already earlier, but only today I got the full set I needed to get my printer running again. Now printing tests...

I agree the current heat sink is not helping to cool the PTFE part. It does, however, help in keeping the lower aluminium block relatively cool. I think the cooling of the isolator (the hollow bolt) needs to be improved - for that, another heat sink is needed. My guess is - it's not optimal to combine the two into one. The aluminium block absorbs a lot of heat from the nozzle unit and I'm afraid it will remain quite hot.

I guess the nozzle unit heats the aluminium block significantly, as the nozzle unit is between the steel plate and aluminium block. I would guess the new heatsink could be cooler than the aluminium block. For estimating the role of the slotted aluminium block at the back, I really think we would need to understand the impact of the rear fan. I think the lower aluminium block gets quite a lot of heat from the nozzle unit, also to the rear parts of it. Therefore, I don't think the slotted block is that far.

What if you lifted the upper heatsink up by just a bit, so that its fins would not touch the aluminium block underneath? Also, would it be difficult to factor in the impact of the rear fan - just to get an understanding of the magnitude of its impact? Your idea of reducing contact between the PTFE and the isolator is very interesting. Do you think it would be enough to make the surface of the PTFE a bit knurled? Or, what if the slots were vertical rather than horizontal? I'm just thinking how to lose as little of the PTFE part's rigidity as possible. Of course, an efficient way of minimizing the loss is to keep it cool

PTFE has fairly high heat resistance, so the impact of a heatsink would be subtle. Also, the temperature difference between the surface of the PTFE fitting and the chamber (right above the heated bed) is not very big. How about something like this: Sorry Kris for ruining your nice sketch... My idea is that everything below the green line would remain hot - while everything above it would be cooled as efficiently as possible. As the heatsink is likely to be warmer than the PTFE surface, it would probably make sense to leave the front and the rear of the heatsink open in order to let the 25mm fan to ventilate the PTFE efficiently. Also, because of the difference in surface temperatures, I wouldn't add a heat sink around the PTFE as the hot end isolator might actually radiate heat to it. I'm planning to get my aluminum plate (the gray one in the picture) anodized, but not from its lower surface - otherwise it will just absorb heat more easily from the nozzle unit.

The sketch seems to add heat conductive material between the heater element and the top of the hot end isolator. I would keep the conductive material to its minimum - as i is more or less in the original design. Instead, the added heat dissipation at the top is a good idea - maybe it could continue higher on the sides of the PTFE part? Not front or rear though, as it would likely prevent air flow around the PTFE.

I agree; convection is probably not doing a lot to cool the PTFE fitting - but there are not too many options for keeping it cooler. At least if we want to avoid making the whole print head bigger and essentially redesign the whole thing. Heat sink on PTFE wouldn't easily help due to its high internal heat resistance. The idea of reversing the air flow of the middle fan sounds actually promising! Another thought... As the walls of the hot end isolator (the hollow bolt) are quite thin, what if there was a heat sink which would snap into the holes around its rim - and the heat dissipator of the heat sink would be on the place of the (currently) missing extruder?

Kris, your posts on this topic are very interesting read, thank you! Do you have any idea whether the error is anything significant or not? Let me refer to a couple of posts in other threads: http://umforum.ultimaker.com/index.php?/topic/3803-ultimaker-2-dual-extrusion/page-2&do=findComment&comment=53318 http://umforum.ultimaker.com/index.php?/topic/3803-ultimaker-2-dual-extrusion/page-2&do=findComment&comment=55376 http://umforum.ultimaker.com/index.php?/topic/6084-um2-bowden-block-cooling-improvement/ It would be very interesting to know how much the temperature charts would change in case of: Dual extruder Black-anodized hot end aluminum parts More powerful fan and/or changing of its flow As I have built for my UM2 a kind of a heated chamber, the ambient temperature could easily get to 50+C, I'm very interested in ways to mitigate the problems arising from deformation of the PTFE part. I see my teflon isolator coupler having changed its form, although only slightly.

I don't think the hot end is wasting nearly half of the fan output even in the worst case, as the fan is anyway cooling down the aluminum structure. As aluminum conducts heat pretty well, the lone extruder benefits from being the only heat-producing element in the print head. However, I see your point and I'm quite interested in the topic especially due to the closed chamber I have built for my UM2. Because of the closed chamber, I'm now a bit less worried about additional fan noise in the print head, as high-frequency noises get fairly well eliminated by the dome I have built. For this reason I have decided to try one of these: http://www.yeahracing.com/catalog/yeah-racing-hacktronic-tornado-high-speed-ball-bearing-25mm-htn304-p-3471.html I haven't received mine yet, but I'll let you know about my experience once I have incorporated one to my UM2 print head.

Get a UPS. Depending on your location, reliability of electricity network varies... In my location, there are very short power outages once a month or so. I would really hate a long print getting truncated because of one of them. My UPS gives me a minimum of 40 minutes of printing in case of a power outage - it gives some peace of mind.

First of all, if you are using Ultimate Blue PLA as your filament, it might explain some of your trouble. It's not the easiest one to get proper results with. Another thing, commenting the photo you sent, you need to use one of the Allen (hex) keys delivered with your UM2 to tighten the bolt fixing the knurled bolt to the axle of the stepper motor.