destroyer2012

Nice! That's exactly what mine looked like until I started using a heated bed and all my PLA upgrades melted :(. But for ABS I think it doesn't matter as much that the aluminum plate is cooled because the ABS is more rubbery/solid when it's heated up so it doesn't form plugs as much. Also screwing a bolt onto the bowden tube helps keep it from riding up as well. Almost as good as threading the PEEK

Ok here's another Pyruvate Dehydrogenase! This one is closer to what humans have. 10 hour print! Support removal took way too long on this.

lol! It's actually a crystal structure of the catalytic core of pyruvate dehydrogenase from Azotobacter vineladii, see here http://www.rcsb.org/pdb/101/motm.do?momID=153

Sure I'll add my current config: Feeder: -custom tap-hobbed bolt -ball bearing feeder upgrade http://www.thingiverse.com/thing:25490 -Bolt-based bowden bracket http://www.thingiverse.com/thing:26686 -Custom filament reel modified to hold larger diameter http://www.thingiverse.com/thing:8317 Bowden: -PTFE 4mm ID tube -Bowden clamp modified to use bolt screwed onto the bowden tube http://www.thingiverse.com/thing:11864 -Short section of 3mm ID tube before the V1.0 hot end; another bolt screwed onto the short tubing and clamped between wood and aluminum plates leading up to the nozzle. Etc: -chicken wire used to tighten belts; also http://www.thingiverse.com/thing:19892 -Fan shroud http://www.thingiverse.com/thing:21182 -Panelmax viewtopic.php?f=6&t=906 -build platform homes to Z max -Heated platform made from 1/8" thick aluminum plate, heated by 120 ohm of alu-clad resistors JB-welded to underside, powered by 120VAC via solid state relay -Kapton tape with hairspray on top for build surface -oven bags taped over the holes in the sides to hold in heat -motherboard fan replaced with radial fan from old dell power supply Oh also, I print with ABS now

I printed this yesterday: http://www.thingiverse.com/thing:30323 For the first time, absolutely no adjustments! I just hit the print button and scraped the finished product off my cooled-down bed four hours later! Sorry for the poor quality of the photo, my camera is having trouble focusing on this object, no doubt as a result of its color.

I'm still not clear why you don't use AC current in your HBP; seems more efficient since the bed is a resistor and they don't care about the frequency component of the current... Yes there is a "dangerous voltage" issue but surely it's not hard to make an enclosure for the relay/wiring? Just get one of those project boxes from radio shack, or print something. I use AC and a couple of solid state relays for my heated platform (using 120 ohms of aluminum clad resistors JB-welded on an aluminum plate) and it works great, plus it was super easy to make.

I feel like the better solution is to just use 4mm ID tubing. That way it doesn't bind entirely when the filament is slightly the wrong shape. My printer is running on 4mm ID PTFE tubing and it works great! Super slippery on the inside, held by threaded-on bolts at both ends (heater/feeder)

You can buy radial fans and replace that fan. Here is a thread I made about this topic: viewtopic.php?f=6&t=733 On a $2k machine one can afford to spend $50 on a fan right? I think so. Well I got my fan from a junked computer but that's how much new ones cost. That said my motors are still really loud.

does the relay still click like crazy? And nothing bad has happened yet eg getting worn out? I am using a solid state relay for mine...

EDIT: silly me Non-roundness is probably caused by loose belts. Tighten the belts or print out belt tensioners. One way to tension belts is to wrap chicken wire around the moving axis blocks and stretch the belts that way. Make sure to check the short belts that go to the motors. Next time you should resize your photos or post links to imgur because I can only see about half of each photo.

The current ultimaker is made using a laser cutter, with off the shelf parts. Thus there are almost no tooling costs, besides machining the nozzle. What additional cost would be incurred by redesigning? We are doing the designing (here in the forum right now) so that's no more cost to YOU for us to keep having this conversation. We are trying to design a system that is better, without any stipulations on complexity. I posit that the current system is non good enough thus it doesn't matter how simple or complex it is. Having the bot made of laser cut wood is a great reason to always redesign everything. You have no added cost on making something different, and all the advantage of staying ahead of your competitors. How does the bowden tube attach to the top of that white piece? I still think mastory's design is simpler than what you have in that picture. Three pieces instead of six (counting the white clip not shown as one piece..) and an attachment point that doesn't depend on some wooden parts (assuming the bowden tube just sits on top of the 'white piece' the same way it does on the current PEEK part).

Why not redesign everything? What do you mean by a connector piece of teflon? Just a short piece of tubing? If plugs still happen then it doesn't fix anything! If the teflon is attached well to the nozzle (as in mastory's mod) then plugs won't happen. But to me that is a more significant mod than just "a connector piece". I think a really simple mod would be just to thread the PEEK all the way through (instead of having that smooth bore section at the top) and then thread the end of the bowden tube, as was talked about in an older thread about how to eliminate plug formation. Ideally you want the junction far from the hot end, so plastic cannot leak out of there. Here's how mastory's mod can be used to make a quick release head: Make the "bowden tube" that sticks out like two inches long. Then, screw a bolt on the end of the bowden tube coming from the filament feeder, and make a bracket to fit the orifice of that tube onto the short one coming out of the nozzle. Because the junction between the two tubes is cold it doesn't have to be water tight so you have some pretty wide tolerances on tube length. Then when you want to switch out the head just unclip the nozzle-and-short-tube portion, and replace it with something else. That would be pretty sweet. I almost feel like the current gantry setup does not facilitate that. Because of the crossing axles there's not a whole lot of room to mount something without losing a LOT of build area. If only the ultimaker gantry looked like this

Awesome nozzle mod! Now if only I had a machine shop in my garage... How hot does it get at the top of the PEEK section? It seems like the aluminum block is closer/more thermally connected to the PEEK insulator. What's the longest print you have done with this setup? How long is your hot zone and how does the oozing compare to stock? How long is the nozzle orifice? With this design all you need is a quick release clip at the top of the PEEK and there we go, modular low profile nozzles!

Very cool. I am wondering whether it is worth it to make an ultimaker style gantry with such a small footprint. All those rods make everything very crowded! You should consider making it like the mojo, which has a gantry like a darwin but one continuous belt path which goes in an H-shape. The motors are mounted on the back (stationary) and driving both motors in opposite directions moves the head up and down along the H (y axis), moving the motors in the same direction moves the head side to side (x axis). With that design you need only four rods instead of six, and it seems to me you get all the advantages of ultimaker (light weight head) without the excessive complexity. The two transverse rods could be mounted one on top of the other to further reduce the build area taken up by rods. Another thing you may consider is using steel cables instead of belts. Steel cables are way smaller and fit in smaller spaces. I am not convinced that they stretch more than belts, and besides with a thin cable the tension mechanism would be much less bulky (guitar string tuner??). Besides, with the H-belt(cable?) setup there are naturally two idler positions at the bottom of the H (motors at the top) which lend themselves to spring-loaded tensioners. Keep up the good work and consider making this printer have a heated chamber. Smaller volume = easier to heat up!

yes I am experiencing this too. I think it's because the tube is too long but I haven't bothered to trim it yet as I don't really see it as a huge problem. Yes I ended up doing this too. So now I have two nuts, one holding the 4mm ID PTFE tube on the top, and another nut holding my 1/8" ID PTFE tube on the bottom. The bottom tube is for making sure the plastic doesn't creep up (as i've explained previously in this thread) and the top tube is for allowing any diameter filament. It's sort of like my idea for attaching the tube directly to the nozzle so I think I'm going to stay with it for now. The only problem is my method of screwing the nuts on the tubes by hand causes the tubes to get constricted. It's not a big problem for the 4mm ID tube but I think it makes the 1/8"mm ID tube considerably harder to get the filament through. I haven't passed a 1/8" drill bit through that tube but I'll probably try it soon. I got some dies but I couldn't get the dies to carve threads in the tubes because the tubes were deforming and digging in to the dies. How do you people do it? This would all be way way easier if my plastic was just the right size, but oh well. I do have tons of it though.

No because nozzle height from the print surface affects the pressure inside the head. THus for example if you start a bit too close to the bed then the pressure builds up during that layer and subsequent layers that all happen on top of each other so when you travel to a smaller area or you start doing an overhang you will see some blobbing. A similar effect occurs when making solid infills if you happen to be even slightly over-extruding. The counter argument being, just set your nozzle height perfectly or input nozzle height into the equation, and I do think this might be the better solution. However, with a heated bed it is much harder to set the nozzle height (at least for me it is) because the bed expands upwards as it heats. Does this mean we should rig up a complex pressure sensor just because I suck at setting z=0? Maybe an auto bed leveler would be a better solution. What are some things you can do with a pressure sensor that you cannot do currently? Detect when the filament is about to be stripped and stop extrusion. Sense when extrusion is not happening and pause the print. Ensure consistant extrusion. Do these things justify the complexity? Idunno

I have been thinking this way for a while as well. The problem I see is, how are we going to fit a pressure sensor in the head without making it leak? The best place to measure pressure is in the zone where plastic is liquid. Pressure is generally measured by the strain of some elastic material like a spring. Where will we fit said tiny spring in the head and still have a relatively short (read: correctly sized) hot zone? The pressure is likely to be quite high so maybe the best way to go would be a strain gauge on the outside of the head. The most workable thing I can think of is measuring the strain along the length of the bowden tube. While not a direct measurement of the pressure that's important, at least it gives us some handle on the pressure applied to the system, and if the bowden tube is screwed directly into the nozzle (not allowing for other flexible components) this might be the best sense we can get of the pressure in the hot zone. But this measurement might be liable to noise from different head positions causing the bowden tube to be differentially stressed. If there's any friction at all along the path of the filament then the pressure in the bowden tube becomes more isolated (mathematically) from the pressure in the hot zone. It'd be like a spring with a damper and than another spring attached, and we are measuring the second spring hoping to get insight on the first. This all sounds nice but I think you don't really need to measure the pressure as long as you build the system so it's pretty rigid. Our biggest enemy is low pressure, since that results in oozing and slow flow. So if we just drive the system at max pressure all the time we shouldn't have to worry since volume in equals volume out, assuming a calibrated extruder. The only times when I have had problems with building pressure is when I was overextruding. Once I calibrated my E steps/mm and measured my filament, assuming I started at the right layer height everything was fine. Thus we should be wary of overengineering the problem.

I think the main reason why people want to get things as soon as they are announced is they still associate ultimaker with an experimental open source project, and "ultimaker quality" means nothing to them. So maybe ultimaker needs to do something about that image. I suggest more blog posts as the answer, as well as selling completed fully functional kits. Until you guys release 'no fuss' printers, ultmaker will keep being understood as experimental and open source and will never be as successful as makerbot. Ideally, the system would be designed so that in the case of a clog the filament drive gear is what would fail first i.e. scrape a big gouge in the filament and stop pushing. I don't think the nozzle design should be intended to be taken apart and re-adjusted after every print as you suggest. Cutting off the tube is a very poor solution and the most it ever does is band-aids the problem. I'm not convinced that having a wider tube is better because I now have a 4mm ID bowden tube and while the friction is gone, the filament shifts around inside the tube and causes retraction to need to go farther, which is hard since our filament feeder is limited to ~40mm/s max speed. I ended up having to attach a short piece of 1/8" ID tube to the nozzle itself so the plastic doesn't melt upwards and cause a plug inside the bowden tube, but this still increases the friction on my slightly oversized plastic (ABS) a whole lot just as it enters the nozzle. The real solution is get better plastic that is exactly 2.8mm or so, then use 3mm ID tubing. Well I mean, the hot zone needs to be optimized is all I'm saying. Perhaps we need some data from a bunch of different nozzles to see which length of hot zone performs best. So what we need to figure out is how much molten plastic you actually need to do fast prints. How fast are we talking? I'm not convinced anyone has measured the max cm^3/sec of our current nozzle (which I imagine is huge based on me manually turning the feed gear really fast), and whether the machine can actually move fast enough to take advantage of that. Another important parameter is cm^3/s^2, aka the acceleration of extrusion. How fast can our nozzle lose pressure? Probably a function of hot zone length and plastic material properties. You know, UM, if you guys published this information that would be really awesome. You are starting to phase out the current nozzle right? So if you've done this research tell us about it to at least contribute to reprap development. Really I can't see why sharing it would compromise anything as ultimaker's nozzle is NOTHING NEW.

Well Ultimaker better hurry up because they are being left in the dust... Lack of updates on ultimaker blog is very offputting. I severely dislike the google group so I will continue to post here in hopes that eventually people will migrate. Maybe starting a "hardware development" section of the forum would be a good start. Ultimaker claims to be open source? Then let's start developing it like that instead of this secrecy. Together we are stronger. Consider how much faster the development could be made if we were all working together on improving the nozzle design. It's really a shame that buying stuff from the UM shop is so expensive for me (shipping to USA! Also everything seems overpriced). Thus I'm probably not going to wait for UM, and in any case the whole reason I spent so much money on this kit is so I could mess around with it. You guys should all read this thread http://forums.reprap.org/read.php?1,139866

This forum is sorely lacking a "wishlist" section where we can post what we would like improved in the next ultimaker iteration. This is why I have decided to make this post. After endless struggles with the nozzle/ bowden tube interaction and other features of the UM nozzle I would like to voice my opinion on what I think the next UM nozzle should be. 1) Screw the bowden tube directly into the nozzle assembly. Pretty much everyone is having the "plug" problem, and there are way too many opportunities for that tube to slip upwards and allow the plug to form. Screwing it in solves that problem entirely, since the tube is no longer held by friction but by leverage (threads), which is much stronger. 2) Smaller hot zone. One of the primary reasons why ooze is so hard to get rid of is the fact that the UM hot zone is too long. Molten plastic stays in the nozzle even after retraction and results in dripping. A shorter hot zone allows the retraction of filament to have more effect on the pressure at the nozzle tip. 3) One-piece nozzle. The major reason for leaks is the fact that the nozzle itself is screwed into a brass tube which is screwed into a PEEK insulator, not to mention the bowden tube held on by sheer luck at the top. Too many joints result in too much leakage potential, so go with the J-head styled design and make the metal parts of the nozzle all one part. 4) Thin-walled heat isolator. The arcol hot end has this. Basically, a portion of the heater barrel is turned down so the metal is very thin. This forms a bottleneck for the heat and prevents heat from spreading upwards towards the PEEK insulator. Yes it does make the nozzle a bit more fragile but I think controlling heat is very important. This also helps to make the hot zone smaller and thus makes retraction a bit more effective. 5)Easily removable. I think it would be cool to be able to switch nozzles in and out. With a screw-on bowden tube this would not be hard. Just screw a short length of tube into the nozzle itself. A part at the top of the head assembly accepts this short tube and acts as a junction between the short tube and the long one coming from the filament feed. Since the junction is far away from the hot end no plugs can form there, and since the short tube is screwed into the nozzle, no plugs can form at that interface too even though no compression is applied to the short tube going into the nozzle. There's a gap between the short and long tubes (within the 'junction' part up top) which allows a tolerance in the length of short tube attached to the nozzle. The aluminum plate has U-shaped holes and fixing screws to allow the nozzle to be slid in sideways, etc. The nozzle and filament feed are two of the major areas where ultimaker needs to improve, so let's make sure this happens. What does everyone else think?

Thanks! Using this machine has proven somewhat capricious as I keep having to make upgrades to make it work. That said now it's printing with ABS on a heated platform, producing comparable quality to what you see in that video, so not all's bad

Be careful that the PLA objects in close proximity to the bed don't sag! Extended periods at temperatures nearing Tg result in plastics creeping. Those cable guides near the bed might warp over time, and also anything like a fan duct that you have, especially if it's PLA. When I made my heated bed (which is powered by mains voltage running through four power resistors) everything made of PLA anywhere near the bed melted, so I had to reprint all my mods in ABS. Of course I am running my bed much hotter (120C). Still I think even at 60C you will have noticeable effects over time.

I tried to print an object the other day and my bowden tube slid upwards again, despite the super-tight clip and despite screwing the bowden clamp down REALLY hard. I think those PTFE tubes are just too slippery for the printed bowden clamp, or maybe my PLA clip was printed a bit poorly (In what way? I don't know). Anyway after removing the fat plug that had formed, I decided I'd try threading the bowden tube. What I did was I took a nut and just threaded it on by hand after compressing the tip of the tube a tiny bit. I don't know what size it was, sadly. I have this giant pile of random nuts and I just grabbed one out of there that looked good. It's probably an American size. Here are some pictures to illustrate the concept. I ended up using a normal nut, not a nylock as shown, but I think they would work equally well. I took the pictures with a spare set of parts after I installed it on my printer. The bowden tube rotates in the nut but it does not come out! You can just screw this assembly into the normal bowden clamp as shown and it holds pretty well. I imagine we can design a better clamp now with the nut being held flush to the wooden top of the print head, but I just set it up as shown for the time being. It works! I printed a couple of ~1.5 hr objects with this, and it seems to be holding up. I don't know whether to be concerned about the rotating. On the one hand it does mean the tube can unscrew itself, but on the other hand I can't think of any forces that would twist the tube enough to let it do a full rotation. It is fixed pretty firmly at the feeder side, and I want to say that the screw holds it pretty well despite the fact that I did not use a die. After all, it got through two prints, including one that failed with the printed clamp that I had before. So Idunno, try this too. It was pretty easy to get the nut on there and it seems to work pretty well so far.

You know I have heard many people claim the fact that the ultimaker is faster than the reprap because it has a lighter head. However, in practice, does this actually matter? I posit that the answer is no, or at least, it depends. If you are printing a plate of mendel parts, then the acceleration and speed matters. On any other print that is not an entire plate full, you have to abide by the "minimum layer time" necessary for plastic to cool down, so at that point it starts to not matter how fast you can go, only how effectively you can cool, and with really fast head motions it becomes very hard to cool the plastic if the head (with fan attached) is moving so quickly. Thus that puts a cap on your speed. So it seems that the next breakthroughs in speed are going to come in the form of better cooling. And all this talk of "my printer can go faster than yours" is really meaningless when you come to see that most people don't print plates full of mendel parts on their ultimakers.

So I know that you guys SET the printer to go to 250mm/s (That's the max speed according to daid's build me marlin. So anyone claiming to build at 300mm/s is suspect) but do you know if it actually reaches that speed? What is your minimal layer time and what is the largest build you have tried at this speed? 250mm/s is VERY fast I mean it's even faster than the speed my UM travels at which is quite huge; I can't imagine how you can print at that speed without massive vibrations happening during infill. Surely the look ahead feature of marlin will slow the printer down automatically so you don't actually reach that speed? Pray tell, do you notice your UM making travel moves with a higher pitched motor sound when you are printing? That would be an indication that you aren't actually printing at 250mm/s. Do you have any videos of your ultimaker printing at 250mm/s?