Jump to content

dojan

Dormant
  • Posts

    33
  • Joined

  • Last visited

Posts posted by dojan

  1. If you want to try different hardware configurations, I can certainly help. Iv'e got access to some 50 different machines (~10 laptops) of all makes and models, mostly disassembled. Ages for complete systems ranges from Pentium III and Athlon XP to first gen Core I3 and Phenom II, so not the newest models, but I could easily put something together specifically for testing, if you know some specific hardware you want to try out. I prefer to use Linux (most distro's, I'm not into Arch just yet however), but I could pull some windows from some very fully legit indeed source around here somewhere. While I'm not a programmer, and can't help you out directly there, I certainly not afraid of looking at the code, or at least mucking about config files and the like :)

     

  2. FWIW, I made a V2 brass tube replacement out of SS and used it with a V2 brass nozzle (the V2 nozzle threads into the aluminum heater block and meets the SS tube inside the aluminum block) and it didn't really work at all with ABS (never tried it with PLA). I fear you will not have a long enough SS section. The all SS hot end designs that I've seen are pretty long (Prusa).

     

     

    Interesting. Why didn't it work? What happened? Did the molten ABS stick to the SS somehow? Please tell me more!

     

    I have seen that the all metal hot ends usually are very long, and they of course have to be to dissipate all the heat, but since the original UM hot end kinda works with just the insulating plastic and no heat sink, my idea with this design is that by adding a small heat sink, but keeping the plastic insulation it should make a radically sharper thermal gradient, and be kinder to the plastic and wooden parts at the same time. This is a good observation though!

     

     

    Also copper is difficult to machine whereas brass machines easily. I would think brass is good enough!

     

     

    Ahh, this may well be true! I have never tried lathing copper, and may well fail :)

    Brass obviously works, but I do think that copper would heat up and cool down more uniformly, and thus make the electronics respond quicker. You might well be right though, the difference may well be very slight :)

     

     

    Joergen and I both print a lot of ABS and we are using longer 36 mm MakerGear "barrels" (brass tubes) and BigHead nozzles. We both use an aluminum heatsink under the PEEK (though of different design). We have the full MG set of nozzles but I think we both like the 0.50 for general use.

     

     

    I would love to see pictures on how that fits together!

    I have so far only used the original 0.4 mm Ultimaker nozzle, but trying other diameters is something I have been longing to do. Does changing the nozzle size change what layer height you can use?

     

     

    I have a long SS tube design in mind that will run the enitre lenght of the wooden print head box (yes, it's tough to fit it in) but my current setup works so well that I'm not motivated to try anything else, ATM. I commonly print Ultimachine black at 275 C and I have no discoloration of my PEEK whatsoever, as I did before the change.

     

     

    That's a brilliant idea! Though I think you would still want a heat sink at the bottom to create a sharp temperature gradient for the filament... That configuration would save some bowden as well :)

     

     

    I can't really tell you what works best for PLA. I always thought the stock UM setup worked quite well for PLA, but admit that I have limited experience with PLA.

    -Cal

     

    Well your input is most welcome in any event :) I'm afraid that the stock UM setup haven't worked to well for me, though I haven't tried the V2.

     

  3. So, I've been thinking about making a new hotend for a while, and the recent discussions on the matter inspired me and made me actually come around to it. I haven't actually made the part yet, but I've put a lot of thought into the design. This is where I'm at for the moment. Any and all feedback appreciated!

    Dojans Hotend - Blueprint - Photo 1

    So the sketch is pretty, um, sketchy. Its designed to be made on a lathe, and is here seen as a cross-section through the middle.

    The general direction of my thoughts here are that you want the hot part to be as short as possible, and as close (in the thermal sense) to the actual nozzle as possible.

    The first thing I wanted to do towards that end was to not have a separate nozzle and heater block, since that creates both a thermal block at the interface between the to, and a much longer distance for the heat to travel to keep the very tip hot. This part should ideally be made from a material that has a very high thermal conductivity, but also a high (volumetric) specific heat capacity, which will help its temperature to remain stable even if extrusion is not. So I looked up Specific Heat Capacity on Wikipedia. Intriguing read. It turns out that the material that best fits this description is probably diamond. Other than that though, copper really stands out with around 380 W/(mK) (Watts per meter kelvin difference, unit of thermal conductivity) and 3.45 J/(cm³·K) (Joules per kelvin difference per cubic centimeter, unit of volumetric specific heat capacity). If I can't get hold of any copper that fits my lathe though, I'll try some brass.

    The heaterblock/nozzle connects to a M6 threaded rod of Stainless Steel (one of the least thermally conductive metals, at around 18 W/(mK)), drilled out with a 4 mm drill to become a threaded tube. This part will act as a backbone and connect the other parts together, and also act as a thermal bottleneck, hopefully not allowing to much heat to creep upwards and create plugs or melt the PTFE parts. The threaded pipe screws about 5 mm into the copper nozzle, where it should be possible to make a molten-plastic-proof connection, given that copper is sometimes used as a gasket material. On the other end it screws first into a aluminium heatsink, and then into the PEEK replacement, separated by a few layers of fiberglass insulation. The bowden tube screws into the PEEK replacement from the other direction, contacting the threaded tube hopefully well above the point where the plastic being printed starts to melt. The exact parameters here will have to be experimented with.

    The above mentioned heatsink is shaped like a couple of wide washers stacked on top of each other but with a distance between, much like the heatsinks on the all-metal hotends, but made of aluminium or copper for improved heat conductivity. I'm leaning toward aluminium since it is lighter, and I already have the material I need. Exactly how large the heatsink will have to be is also something I will have to experiment with.

    The top part, which connects to the Ultimaker, and replaces the PEEK will be made from PTFE (Teflon, same as the bowden). It is largely the same as my current bowden, but has threads all the way through, and screws onto the bowden. The choice between PEEK and PTFE is pretty arbitrary, but I think I have some PTFE laying about.

    Between the PTFE part and the heatsink I plan to put some fiberglass exhaust wrap as insulation. I also plan to wrap the heaterblock/nozzle in fiberglass wrapping (This stuff is fantastic, I have some on my current, ordinary hotend to test, and with only one layer of 1mm I can hold my finger right to the hotend while it's printing at 220 degrees C. It is rated for 800 C, with peaks up to 1'300 C :)

    I have gone through a lot of design changes before getting here, this picture shows the rough stages of evolution:

    Dojans Hotend - Evolution And messy desk

     

    So what do you guys think? Am I reinventing the wheel? Am I onto something?

    For more detailed photos of the sketch, see the gallery under my profile. Also, beware the nerdy mess on my desk.

     

  4. Awesome! I figured you had something like that planned :)

    I thought of something else :)

    Under the "Speed" tab in the printing window there is "Outer wall", "Inner wall", "Fill" and "Support". (Obviously it would be nice to be able to define that before printing, although I like having real-time control; If I recall correctly that was a limitation of SF though.) But what about travel speed? And fan speed for that matter. Possible?

    Although I have to say; Don't take time from working on your (probably) awesome slicer to do this stuff :) As an end user all I see is the gui after all, so don't listen to closely to what I have to say :)

     

  5. Man, I'm out of the loop for a few months, and find everything changed. (for the better, Daid; Awesome.)

    Anyway, I had a few thoughts on how to improve on the awesomeness:

    I use the "Fan speed min - max" setting. My fan (not original) runs happily and very quietly at around 25 - 28 %, however it won't start at anything below 45%, and not always then. This is a problem in combination with "Fan on layer (1)" and the project planner. [edit] A simple solution would be to give the fan 1 second of full power before settling on the given percentage. [/edit, added proposed solution :)]

    In the project planner, I for one would like back the ability to manually place parts. I would also like the ability to override if parts are placed to close together; If printing really low parts one cm in each direction is plenty, and changing the "Machine head size" every time is tedious. I Imagine a popup saying something like

    "According to the current settings, the machine head will crash into printed objects x, y, z [just pull the names of the red ones], are you sure you want to continue? Do so only if you know what you are doing."

    Being able to designate a 3 dimensional machine head size comes to mind, but I don't know if that would be practical...

    Also, why is the project planner in its own separate window? Couldn't we have the ordinary Cura window, and the side panel from the project planner available (collapsible?) on the right? This would be more intuitive I think, and enable a more harmonious workflow... As it is now, I find myself flipping between sometimes up to 5 Cura windows (Project planner for setting up; main window with all the settings; The progress window for slicing the next print; The currently printing window, and sometimes one of the "Expert settings" "Preferences" and "Project planner preferences" windows), which is a bit... messy. I have dedicated a computer with dual monitors, and I still have to flip back and forth a lot.

    Further, is there any way of looking at the layers in the project planner? I can't seem to find any way of running gcode generated at some other time either, or getting at the printer controls (jogging, temp etc) without first slicing some model, and then printing that.

    If I'm giving you a hard time it's only because I really like the stuff you're doing :)

     

     

  6. Are you sure it's not skipping steps when you rise the speed, because of either lack of lubricant/oil on the rods; lack of belt tension; or bad alignment of the rods?

    I have had strange stopping/starting behavior, in a few cases I believe because of to small baudrate on Linux, and in a few cases it fixed itself with a re-slicing.

  7. I can spot a few things that should make things better. First of all, for printing ABS your temperature is way to low: Try 250 C instead, that will probably help with both the extruder missing steps, and delamination. You should also know that the first model is a very difficult print, with those overhangs. You might want to try Curas built-in auto generated supports, but be prepared to waste alot of plastic and time over it. Know that supports are generally not used for supporting the weight of the model during printing, but rather to be able to print parts 'mid air' so to speak, and very steep overhangs. Adding a few millimeters if retraction (try setting the 'distance' setting to 4 under 'retraction' in the advanced settings) will help with the stringing, but might increase the risk of grinding.

    Is there some particular reason for printing it with 100% infill? I usually print with 20% or less, perhaps 30% and 1.2 or 1.6mm wall thickness for a structural part.

    Curling is a well-known problem with ABS, but I'm no expert on that, having almost only used PLA. But to make successful prints in ABS that are that large, I suspect you will need a heated print-bed. Know that the Ultimaker is primarily built for printing PLA.

    Anyway, good luck to you!

  8. So I just tried installing Cura 12.7 on my secondary computer running Debian 6. After extracting and running "sudo apt-get install python python-serial python-wxgtk2.8 python-opengl" it starts and everything seems fine, but when I hit "Slice to GCode" I get "Something went wrong during slicing!", and "Show Log" reveals "/home/dojan/3D/linux-Cura-12.07/pypy/bin/pypy: error while loading shared libraries: libssl.so.0.9.8: cannot open shared object file: No such file or directory"

    What should I do?

  9. I have been thinking about posting a list of feature-requests for some time now, but I see you added most of them anyway :) The most important for me being the retract-on-support, smarter combing, and the min-max fan option. I also like the smarter skirts. Thanks!

    One thing still on that list though, is I would like to be able to print the infill and supports at, say, 0.3 mm thickness, inner wall at 0.1mm, and outer wall at 0.05mm. I know about the "skin" option, but it really doesn't cut it for me.

    Also, the auto-baudrate option seems to mess things up for me: Only about 1 in 20 times it manages to connect properly. Changing to a fixed baudrate fixes it however. (Linux)

    I have started seeing a new bug, but I'm not sure it's related to the new version; When I place an object right in the Home corner in the Project Planner, It will start with that object no matter the order specified, and worse, will not lower to the printbed after the heatup-home-lift z-extrude routine, but will instead start printing the object mid-air.

    Also, the project-planer is smart enough to not re-slice a copied object, which is great, but sometimes it hangs on the "it took 20 seconds to export the file" part.

    I would love the opportunity to help you locate any of these bugs, if I can...

×
×
  • Create New...