Anders Olsson

May I ask if the quality issues are mainly related to alignment of the nozzles or to oozing?

Don't worry, I don't mind other people working at this. In fact I could just have given away the step-file by now if it wasn't for the scientific paper we are writing where my heater block is featured. And just as both of us concluded, a heater block for the Ultimaker 2 more or less has to look this way if you are not going to modify other components. The pictures in this thread shows version 1.2 of my heater block by the way. Since then there has been version 1.3, 1.4, a few more poorly documented iterations and then version 2.0 which is being manufactured right now. It was not necessary for the discussion to publish drawings or photos of the later designs though, since the basic concept hasn't changed, so I haven't written about all the refinements of the design that I have made. If I am giving you more details now, the guy who wants to manufacture and distribute these might not like me anymore though, so I am going to keep my mouth shut :smile:

I haven´t really thought about that, but it is worth investigating. If you go for stainless steel you more or less end up with something that looks like the E3D-V6 hotend: http://e3d-online.com/E3D-v6-Documentation In my design the bits would have been too tiny and the threads too short, so I abandoned the idea of a two piece heater block at some early stage of the design. In theory this would help, but I have not thought about doing something like that yet. Mainly because it requires custom tools and custom nozzles. My design was based on what I could get made in my local workshop with tools they already had combined with the design challenge of not having to modify any other component of the printer. That is why it ended up like it did. :smile: The same compromises were made regarding the material. I could have chosen something else than brass, we actually have a lot of rare high performance materials. But I did not immediately see any much better option and I know brass is easy to machine and safe to use. The temperature differences inside the heater block is not more than a few degrees, so I don't think you have to go for lower thermal conductivity really. But it is fun to think about the options of course :smile: I will reply to you about nozzle geometry in the other thread.

I think another property of brass that is important is that it does not corrode easily. Few other materials has that combination of price, thermal conductivity, hardness, easy machining and corrosion resistance (in air). Pure copper for example would corrode heavily if you got finger prints on it and kept it heated to 260 C in atmosphere for an extended period of time. For alternative heater block designs, I have some comments on gixxers design. Don't read this as criticism. I am just using your drawing to explain the design challenges with this part of the printer a bit more in detail: - You can not make a hole threaded and get a nice seat for the nozzle the way it is shown in your drawing (think about how a thread tap looks, it does not cut complete threads all the way to the bottom of a hole). This is why I had to complicate my design with a milled section inside the hole. Or where is your sealing surface? (both the nozzle and the hole are 7.5mm) If you select to use the hexagon surface for sealing you will fill the threads with plastic. - There needs to be a screw at the same place as the original set screw. The set screw is not only supposed to fix the heater and the sensor, it also prevents the heater block from rotating. This is particularly important if you are going for dual extrusion, as you would adjust the individual height of the nozzles by rotating the heat break tube which is screwed onto the M6x0.75 thread. - A much larger mass of the heater block is likely to trigger the sensor monitoring function. Daid recently decreased the sensitivity of this function: http://umforum.ultimaker.com/index.php?/topic/8634-curas-1412-firmware-is-screwing-my-um2/?p=82766 but if the mass is twice or more, I think the monitoring will still be a problem. Modifying the firmware is not that difficult, but you would have to keep modifying it for each updated version of the firmware. - If I made a thick heater block like that, I would separate it in two pieces: A heater block with M6x1 through thread and a connection pipe with M6x1 in one end and M6x0.75 in the other end. That would make machining much easier. So, I am not saying that my design is the only one possible. But if you want to keep it completely plug and play, no other hardware or software modifications, there are a whole range of very limiting constraints for the Ultimaker 2 which severely restricts your design options.

I did some further tests on the properties of the heating. My custom heater block, which has about 40% more mass than the original block, gets from 50 to 150 C in 41 seconds. To investigate if thermal conductivity between sensor/heater and heater block affects the heating speed I greased both sensor and heater with copper grease. After heating the nozzle to dry the grease, I repeated the experiment. I get exactly the same heating speed with copper grease applied. However, without copper grease, 10 degrees of temperature increase (18-28 C) takes 10 seconds. With copper grease the first ten seconds takes 7 seconds. So, it might be that insufficient thermal contact between sensor/heater and heater block is partly to blame for causing a delay that can trigger the monitoring function. (Even if the sensor/heater are still in place)

Strange, it should not differ that much, I will do some more tests on my machine this evening. I am pretty sure I have never seen 100 C increase in 20 seconds on my machine, not even when it was brand new. My machine probably does like 40-50 degrees in 20 seconds at low temperatures. However, the issue is mostly the delay when the temperature is dropping and you start heating again. If doing that at 200 C, the triggering function starts counting from 200 C. The temperature then drops to 195 C before it starts increasing again. On my machine that means it now has to do 25 degrees in about 15 seconds, which is about what it can do at that temperature. I am a bit suspicious if thermal contact between sensor/heater/heaterblock can influence this somehow.

I am using a slightly modified version of the 15.01-RC1 firmware now. This time I reduced the triggering criteria from 20 degrees to 8 degrees in 20 seconds. That seems to be fool-proof. I can not trigger the monitoring even if I try my best now. I don´t know if Daid would be open to a similar change in coming firmware, or if there is a very good reason to keep it at 20 degrees? It is clever to use this kind of monitoring function, even if the main reason to have it seems to be to reduce the risk of disasters caused by the non optimal design of the sensor locking screw. If the monitoring is too sensitive though, it might cause more support work for Ultimaker that if the monitoring would not have existed at all.

We have two Stratasys printers at work, one uPrint SE and one Dimension Elite. I have compared prints, and even printed the same object, with my Ultimaker 2. The Stratasys machines has the advantage of a heated chamber, which makes large components easier to print. For the Ultimaker, large ABS things are always a bit tricky to print (they want to bend and come off the platform). With some experience you can mostly get it to work, but some designs are almost "unprintable" in ABS without a heated chamber. For the print quality, the Stratasys machines seems as accurate as the Ultimaker in the X-Y-plane, BUT they offer considerably lower surface quality in the Z-direction I would say. This is partly because the rather thick layers (0.254 mm for uPrint and 0,178 mm for Dimension Elite) but also the surface quality of the top layer always seems much worse on the Stratasys machines for some reason. For what I do, the rather thick layers of the Stratasys-machines prevents me form printing certain things, like fine pitch threads, which I find very convenient being able to print on the Ultimaker. I would say that for reliably printing rather large prototypes in ABS, Stratasys is a good choice. When it comes to strength, one could assume that the heated chamber makes layer adhesion better on the Stratasys out of the box. This sometimes seems valid for small components, but I have not done destructive testing on Stratasys prints to compare yet. Playing with the settings on the Ultimaker can probably get you quite far too though. My printed sway bar link is still working fine for exampel: http://umforum.ultimaker.com/index.php?/topic/7240-examples-of-truly-practicaluseful-prints/?p=70778 For what I am doing though, the Ultimaker is a way more flexible machine which, can archive better quality at a much lower cost than the Stratasys-machines. Things printed on the Stratasys machines typically looks like good prototypes (accurate and robust, but not very fancy). The Ultimaker prints can go beyond that, to a point where your fist reaction is that you could probably sell them for a reasonable amount of money just because they look beautiful. When I bought the Ultimaker 2, one of my demands was that I would not immediately be able to tell that the object was 3D-printed if held on an arm length distance. I would say that the Ultimaker 2 can do that but the Stratasys machines we have can not. But that is probably just because Ultimaker and Stratasys are focusing on different customers which will use the printers for different purposes.

Check if the heater block is in contact with the fan cap. That can potentially drain a lot of heat and trigger the monitoring. I have this issue too by the way. Eventually I modified the firmware to only require 10 degrees in 20 seconds and this more or less fixed it. I am of course using my custom heater block, which has slightly larger mass and therefore heats slower, but I still think the monitoring function is a bit too sensitive. As I understood it, it is changes in PID parameters made after 14.09 rather than changes in the monitoring parameters that has increased the sensitivity? I would kind of prefer if the monitoring was a bit less sensitive. I think I can trigger it with the original heater block too if I try. Where it normally goes wrong is when the heater is cooling after a finished print and you select the "move material" function. The monitoring function then starts counting degrees when the temperature is already falling and with current parameters the heater does not not catch up before 20 seconds has passed. EDIT: I made a calculation which suggests my custom heater block has about 40% more mass (sensor, heater and nozzle included) than the original one. So the larger mass of my heater block should not be enough to trigger the monitoring function if the safety margin is three times.

I was worried about that too, but it has worked fine so far. The thread inside the heater block is very short which means the nozzle can tilt quite a lot before it is tightened. I think this helps it align with the seat inside the heater block. Both sealing surfaces are machined in a proper way too and the sealing surface on the nozzle is not very wide, so I would say the design is not that bad as a metal to metal seal. Here is a picture of how the nozzles look after some use. The right one has been printing about 200 hours I think, using all sorts of strange plastic. Interesting information! So if I would claim that I am now running comparisons between a normal nozzle and a prototype with an "anti ooze valve", would that suggest that I might still be ahead of Ultimaker on this part of the printer? :grin:

The nozzles are standard nozzles from E3D: http://e3d-online.com/E3D-v6/Extra-Nozzles So I can not take any credit for that design :smile: As you can see they are not that expensive, which was one of the reasons why I selected E3D. That also means one heater block and three nozzles would be more like 80 Euro. (depending on how expensive nozzles you select) The internal shape of the E3D nozzle is, even though it looks complex, a bit primitive in the way that it seems optimized for manufacturing with standard tools. That is why it has that two step shape as I understand it. The flat areas are about the same size for the E3D nozzles as for the Ultimaker nozzles I believe. Larger hole means the flat area is increased accordingly for the E3D-nozzles. The original UM2 nozzle internal cone shape is a more "professional" approach to plastic extrusion, but requires a less common tool for manufacturing. In theory it should have benefits over the E3D-design, but in practice all I can say is that my heater block with the E3D-nozzle can reach higher flows than the original nozzle. That might be because the hot zone is a bit longer though. What I particularly don't like with the original nozzle is that small step where the cone becomes a 0.4mm hole. I had a lot of issues with clogging where it seems like the clogging particle was hiding on that edge and was sometimes very difficult to get out of the nozzle. I have no idea about how well the E3D nozzle works for dual extrusion compared to the original nozzle. The Ultimaker design might have advantages when it comes to for example oozing, but I have not tested this at all. Now over to some important information: - I am rather busy already, and selling heater blocks really is not my main goal in life. I have a job and a private business as a consultant which really would need more attention and is much more beneficial than selling heater blocks. So as much as I like to help all of you, I really would not mind if someone else distributes the heater blocks. Manufacturing them is much less than 60 Euro each. - I would prefer to make a small initial series of let's say 20 pieces to validate the design and I think I just might found an opportunity to do this without increasing the price. - There are rumors that Ultimaker are performing experiments with their own design of some new nozzle related component. I have tried as hard as I could to get information about this, but so far the result is that I can not get anything confirmed really, not even if there will be some Ultimaker approach with exchangeable nozzles at all, ever. This situation is a bit inconvenient for me. I don't like the idea of selling my heater block to people and then two weeks later Ultimaker presents their own version I also don't like the thought that I might end up with 50 unsold heater blocks when Ultimaker suddenly releases their own design. But on the other hand we have been waiting for another famous upgrade from Ultimaker for quite some time now :smile: - I have considered letting the design completely free, but before that happens I would prefer to get two things done: 1. Finish writing the scientific article where this design is featured and get it published. 2. Get a small series made and tested by people around the world to confirm that the design really is 100% working.

Your enemy is basically temperature gradients in the printed object. Professional printers solve that problem with a heated chamber. I have like 95% success rate printing ABS on the UM2, but I generally print small things and/or things which are suitably designed for printing in ABS. I just wrote a rather long reply about that in another thread: http://umforum.ultimaker.com/index.php?/topic/8157-percentage-of-prints-completed/?p=81116 If you insist on printing difficult objects, like large thin walled cubes, in ABS, you should try to mimic the heated chamber concept. That means: - Hot buildplate (To stay above Tg and to heat the "chamber" - Fans off - Front of the printer covered - No open windows or fans next to the printer. A top cover should improve it further, but I have not tried that myself yet.

I have been printing almost exclusively ABS and my success rate lately is probably 95% The biggest improvement in reliability was the firmware upgrades last summer which fixed leveling bugs and improved the priming cycle. Lately I have been using Roberts feeder, but it worked just as good with the original feeder and ABS. Apart from that the only modification I have done is to add a small pulley below the feeder to get a more straight path for the filament into the feeder. I can not recall that I have had a single clog with ABS (my home made filament with particles in not counted :smile: ) If I remember correctly, I had a feeding issue once, which was caused by a thin string of ABS jamming the filament in the bowden. Particles seems not to affect the feeding. With PLA and other plastics, the success rate is a bit lower, maybe 85-90% The reason why ABS wins is that it can take any amount of retracts, when (at least some grades of) PLA are eventually destroyed by the feeder after repeated retracts. I have been printing both Velleman and Ultimaker ABS, both works fine for me. However, when it comes to ABS, if you try to print "unsuitable" designs, ABS might prove difficult to work with. Most things I print are rather small (1-4 hours) low objects with nice smooth corners. If you try to print a tall object with sharp corners and thin walls, you are pushing it. I printed this one for a friend for example: http://www.durovis.com/opendive-howto.html And ended up with these cracks in the lower wall: Now, the print finished successfully and the cracks will not affect the functionality, but I don't like when it looks like that. The thicker upper wall did not show any cracks. I consider this to be a non optimal design for printing in ABS. If I designed it myself I would have made nice round walls, like welding goggles. I am sure it would have printed fine in ABS too then. I printed another even larger thing in ABS lately , it was actually the upper part of a human scull :shock: Due to time constraints I had to leave the printer unattended over night but the scull came out perfectly fine after 15 hours. I could more or less promise delivery next day because I knew that kind of shape would not be a problem with ABS. So, for ABS, there are tricks but it is also about how suitable the shape of the object is for printing in ABS. Large square boxes are typically not suitable. (And I would not select 3D-printing to manufacture such objects anyway) My tricks for ABS are: - No fingerprints at the platform, glue with Tesa Easy Stick when platform is cold. - Make sure the platform is properly leveled, preferably a bit close so there is excess plastic in the first layer. - Platform at 90-95C - Temperature 260C for Ultimaker ABS and 255C for Velleman. - Fans OFF (except for tiny details in the end of the print) - Cover the front of the printer if the object is taller than about 20mm.

Are there any specification of what kind of polymer MPflex45 is? If it is the same kind of flexible polyester as "flexible FPE 45D" from 3DVerkstan I have the following advice: - The only way i could make it stick was by cleaning the platform properly and heating it as much as possible. Glue made it less likely to stick (!) - You need Robert's feeder (the one Didier linked to) and you need to use the filament guide designed for it. - Printing relatively slow makes feeding issues less likely. Printing thick layers (0.2mm) makes it less likely that the part it is pushed loose from the platform by the nozzle. - Retraction does not really work that well, so try to avoid retractions (or try to change the retraction settings) - It seems like the flow needs to be increased about 15% over what would be expected. I used these settings: Temperature: 220C Buildplate temp: 105C (Covered front and increased to 110C after printing started) Diameter: 2.90mm Flow 110% (slightly low) Speed, I don't remember, but something like 20-30mm/s at 0.2mm layer thickness.

I had a reply from a workshop now. If I sell them one by one, I probably have to charge 60 Euro each, one 0.4 mm nozzle included, delivery costs not included. (If I can send many at once, I can reduce the price a bit, so don't hesitate to order several and distribute them in your area) Anyone who wants one, can you please send me a PM so I can estimate if there is a market. (I have to get about 50 pieces manufactured get down to a reasonable price) Please note: I can not guarantee that this thing will work as good as the original nozzle with for example dual extrusion, or that it will be compatible with potential future upgrades from Ultimaker.

Interesting development, a video would be nice! I am also happy to read Daid's post, now I just need to get the remaining parts for the second extruder :smile: My quite extensive tests with full metal hotends really confirms what Daid wrote, PLA and a full metal hotend is not a reliable combination. When it comes to polycarbonate (PC), I have been printing it with my modified UM2 and I also did extensive destructive testing on the prints. PC is not completely trivial to work with. I had to dry the filament in an oven at 130C for a few hours to get reasonable transparency and adhesion. I don't know how often this has to be repeated. I never managed to get the prints as strong as I had hoped. When printing at 300C and the buildplate at 110C they were not bad, but I still had hoped they would be a bit stronger (better layer adhesion). Professional machines like Stratasys has a heated chamber which stays at 120C (I think ?) when printing PC. I think that is the key to really good layer adhesion. So, don't hesitate to print PC, but be prepared to put some effort to get reasonable prints and don't expect the prints to be as strong as a solid piece of PC.

Okay, thank you for your reply, I kind of suspected that. I might have to go for another slicer for objects with this walls then.

Well, if you suggest modifying the original heater block there are space constraints both towards the sensor/heater and in the opposite direction, towards the front of the printer: The red lines are the outlines of the original UM2 nozzle and the blue/greenish lines are the outlines of the E3D-V6 nozzle.

Did you try to power cycle the printer? I had this error numerous times with my custom heater block and the latest firmware before i modified the firmware. The error message appears when you are heating up the nozzle and the temperature increases less than 20 degrees in 20 seconds. As I said I have managed to trigger this monitoring function several times even though it was nothing wrong with my temperature sensor. If it is consistent after power cycling though, there is most likely something wrong with your temperature sensor. Did you check the temperature reading in the "heat nozzle" menu?

That is exactly how I wanted it to be designed. (Don't ask me why UM did not design it that way from the start) Just drilling and threading the original UM2 heater block was not an option though. There is simply not enough material in the original heater block to fit a standard M6 threaded nozzle in it. If there were M5 threaded nozzles it could just about work, but those seems rare and I wanted a off-shelf drop in replacement nozzle that is available in different diameters. This was the main reason why I had to make a new heater block. While redesigning it, I also tried to address the heater and sensor fixing issue since the original set screw design really isn't that good.

Yes, I made some slight improvements to the design in order to make it easier to remove the sensor and the heater. I have had hundreds of hours of printing with it on my printer, so I am quite confident it will work now. Regarding print quality and oozing properties, the E3D-nozzle might differ slightly from the original nozzle. This is something that I can or will not do anything about though. In my opinion it seems to work as good as the original nozzle, but I have not done extensive comparisons yet. One of my heater blocks has been living at Ultimakers HQ for some time now, but I have had zero feedback on that. My main issue is now to get a workshop to manufacture a test series of these for a reasonable price. I am negotiating with two workshops, but things are awfully slow here. (and I am quite busy with other things too) I was hoping that I could offer a test series of these to people in this thread before Christmas or so. I asked Ultimaker if they have any opinion on me providing these as upgrades for the UM2, but I had zero feedback on that one too so far, so I guess it is okay then :grin:

This inability of Cura to print single line objects is something which has caused me quite a lot of trouble and I am surprised that I don't read more complaints about this. As far as I understand some other slicers like Makerware can slice single line structures. See this one for example: http://www.thingiverse.com/thing:289650 (Using Cura you will have to print such object twice the size to get the physical properties right :shock: ) Another situation where Curas double walls are really limiting is when printing text on surfaces. You have to go for a rather large bold font to make Cura print it. That has been a major limitation for me when it comes to how much information I can provide for the users on the printed objects. Is there any particular reason why there can't be a "slice single pass structures"-button to make Cura slice single line walls? That would help me hugely when designing tiny components for scientific instruments. I have spent lots of time either trying to trick Cura or editing the G-code manually (which is quite time consuming). Since I got my custom heater block I can print 0.6mm (2x0.3mm). The 0.25mm nozzle just seems a bit small to work properly with the UM2 for some reason though. Also, 0.6mm is still far from 0.4mm, and if I could trick Cura to print single lines I would be able to print 0.3mm walls with my current setup, which would be very interesting for scientific purposes.

Wow! That looks very interesting! Was it difficult to machine? (Our workshop did not like what they read about machining it as I mentioned before) Did you try printing anything with lots of retraction yet? I am very curious how Duratron handles hot plastic since one of it's recommended uses is for components for injection molding machines. That means it could be the perfect material for the thermal insulator, if the friction with molten PLA is reasonably low. If Duraton seems to work, I would switch focus back to that since Kalrez seems very difficult to get and since Duratron is no more expensive than Vespel.

If heat-related deformation of the PTFE insulator is the main problem with dual extruders I think a short Vespel spacer might work. I though another major heat problem with dual extruders was that PLA will melt further up the PTFE insulator when both extruders are heated, causing retraction issues? Personally, I would really like to get a sheet or a suitable diameter o-ring made from FFKM (Kalrez). I think that material has a combination of properties which makes it even more promising than Vespel, even though Vespel is not bad at all. It seems even more difficult to get FFKM than Vespel though.. Regarding Duratron, I had a talk to our workshop, but they kind of did not like what they read about machining that stuff. So I have abandoned that material for now.

You can probably get it back in working condition by drilling the inner diameter to 3.2mm. You have to use a drill which is about 3.3-3.5mm, because teflon deforms quite a bit before when you try to machine it.