calinb

This is an old thread but figured I'd weigh-in on this mod. I just turned and threaded a new drive stud (it's not a bolt) on my lathe, cut the groove for the clip, and used a 4mm-0.75 tap to cut the teeth. I haven't tested PLA yet but the drive force is much improved over the Ultimaking Ltd. V3 part when printing ABS. I have 0.25, 0.35, 0.40, 0.50, and 0.70 nozzles and the 0.25 and 0.35 nozzles previously only worked with PLA. Now the 0.35 nozzle works with ABS too but I've not yet tried ABS with the 0.25 and the new drive stud. I also have the newer Bertho style feed mechanism. Actually, the standard 0.40mm nozzle isn't all that great with ABS either and I'd adopted a 0.50 for ABS printing. Now I can run smaller nozzles! By the way, I used this tool and it turned out great: http://www.thingiverse.com/thing:3780 Update: When I do reach this filament drive's limits and it grinds, it's really a PITA to clean and I must use a magnifying hood and a dental pick It's nothing like the V3 stud which only requires a few quick strokes with a nylon brush.

Agreed! I have a heated 5/16" Mic6 tooling plate for a bed and I often use a bubble wrap tent to create a heated build chamber. The Mic6 is also very stable with temperature changes. I'd like to build the entire Z-stage from it someday (bearing mounts, arms, and table).

I think you are correct, Lars. I follow this thread occasionally, because belts are a PITA--especially MXL belts that don't seem to be terribly available in the U.S.. I plan to throw away my short belts someday and install GT-2 pulleys and long belts for this reason. We are pushing viscous liquids around and, given reasonable expectations for the precision of this process, I think the performance of the X/Y drive system outshines the Z-stage stability, regardless of how we move the print head. I have a heated bed made from 5/16" Mic6 aluminum tooling plate (quite dimensionally stable with temperature changes) and I'd like to build the entire Z-stage, arms, and bearing mounts from Mic6 someday. I've improved the return to zero of the bed height adjusters too. I hate to reduce accessibility, but I agree that another Z axes rod in front is very compelling.

By moving the filament feed stepper motor off of the print head to a stationary location, the mass of the moving head can be reduced (everything else being equal). The idea is the movement of the head is then more precise at high speeds and it can be accelerated more quickly. It's tough to prove the significance of any actual benefit, because comparisons between 3D printer designs are always an "apples and oranges" comparison, due to so many design, setup, and usage differences. The Ultimaker is known to be a very fast and accurate consumer/hobby 3D printer, however. No need to worry about under extrusion. The Bowden tube does nothing to cause that malady. You should work to anchor it well at both ends, however. Any play in the endpoints creates hysteresis when using retraction (filament "pull/push" or "suck/prime"). In my experience, ABS cannot be printed as fast as PLA (ABS is more viscous) so the benefits of accurate high speed movement may not be realized with ABS. Similarly, the highest quality prints are generally a relatively low speed activity. At the speeds I typically print, I'm confident there are many printers with moving extrusion stepper motors on their printer heads that can print just as accurately as my Ultimaker. BTW, I've used 0.25mm, 0.35mm, 0.40mm, and 0.50mm nozzles and the 0.25mm nozzle works with PLA but not ABS, because the very small orifice presents more resistance to ABS than PLA and the filament feeder cannot supply sufficient force to drive ABS through the nozzle. I seriously doubt that any of the current 3D printers on the market have a filament feeder that could do it reliably. If you compare the UM to a Rep2, I suspect the setup and "tune" of the printer will have more of an impact on the results than the model of the printer! As an example, all of the printer nozzle sizes I mentioned above have a marked effect on the nature and performance to the printer. Nozzle size is always a compromise and the best nozzle depends on what you are trying to accomplish, as do a plethora of process parameters and settings!

You're not alone. I've always had this problem too. It certainly causes the TC signal to "command" erroneous PID controller adjustments, but I just live with it. However, it is a shame to have a nice, linear, and accurate TC (instead of a mere thermistor), and then toss the advantage out the window by tolerating this unsolved engineering problem but, in my opinion, the electronics and cabling are not the Ultimaker's and Ulticontroller's strong suit! Thanks for your posts, Ducky. Your findings indicate that the problem is likely not caused by stepper motor load induced "ground bounce" but rather it's caused by EMI/RFI.

You could always try coating the TC with some non-conductive thermal grease (heatsink compound) before you insert it into the aluminum block. Get some good stuff, however (premium computer CPU / GPU heatsink compound). The cheap stuff hardens up and might make it tough to remove the TC. I've not looked up the specs and even the good stuff might cause problems with removal for that matter, but look for maximum operating specs up to 260C, if you can find some.

I assume you are using volumetric mode in netfabb. The problem with every slicer I've ever used is no single calibration object covers everything you are going to print. Without going into the possible reasons behind my observation, I can only recommend that you scale each model when you find unacceptable dimensional errors. Unfortunately, even this customization often results in imperfect dimensions because not all of the dimensional errors found in a print are equal--percentage wise. If the errors are not in proportion, you can't apply a global fudge factor for each of the axis to correct all errors. You must compromise. I have an Ulticontroller so I perfer to temporarily change the X-steps and Y-steps (I've not found Z-steps to ever require changing). Be aware that I've also found larger dimensional errors with nf when a part is printed on the diagonal of the bed instead of square--more so than when using Cura or Kisslicer, for example. nf appears to be somewhat "Pythagorean challenged!" :(

I don't think I'd spend that much money on a Kickstarter project. It's too easy to get burned. Also, my experiences of supporting a Kickstarter project at a lower level is I'm a second class citizen when it comes to delivery schedules. Just my $0.02 and your mileage may vary. Also, Kickstarter supporters are more into drinking the Kool-aid dispensed by the project owners and less into being critical consumers and customers. Not my cup of tea. I like the extra size of the Zortrax, particularly in X/Y, but it doesn't seem to offer the features I'd like to see in a new Ultimaker. The Zortrax head is massive and appears to include the filament drive motor. If this is correct, it's a significant departure from the Ultimaker concept. The Zoltrax probably needs its double printer head rails, because of sag implied by a single supported beam (these rails must also be accelerated during head motion). If I were designing a larger Ultimaker, I'd investigate the use of a stiffer rail shape (like an "I" beam) to keep the mass down. Of course the bearings would need to change too. I don't see any signs that the stage is any stiffer (or features better return to zero after a disturbance) than the Ultimaker statge and the X/Y steppers are still within the cube frame structure (not good for enclosing and heating the build chamber). It's easy to move the steppers to the outside of the box on an Ultimaker and maybe it will be equally trivial with the Zortrax. In this application, the Zortrax ball screws are a big "yawn."

If you have an Ultimaker Engine License and the Pro version, the Engine will work with Pro. I've used nf extensively. It is difficult to learn and complex but provides more control than any other slicer over a variety of things (layer-by-layer control of countless things, for example). It is also very buggy and you have to be prepared to work-around the bugs. No one slicer works well for everything I make. I have and use them all! Even a freshly installed demo version will not create gcode or connect to a printer, as I recall. You must activate it to actually print with it (or generate gcode).

Yep--for big ones, here's what I do: http://umforum.ultimaker.com/index.php?/topic/1925-modification-to-the-hot-end-for-abs-printing/&do=findComment&comment=14558 I've often thought about installing an "X/Y bellows" to keep my AD595 TC amplifier in cold air, but it's handled the heat so far and it's probably easier to just solder a new SOIC amp on the board when it burns out!

I'll second everything Joergen has said here. I mostly print black and yellow ABS and Joergen's estimate of roughly a 20C nozzle temperature difference is what I've found between the two colors too. The problem with black is the 20 C sometimes pushes you into problematic territory with a stock Ultimaker! Fortunately, both Joergen and I are running similar simple hot end mods and I think we are both happy with the results. I was planning a major stainless steel rework of my hot-end but may never do it, because my simple changes are working very well. I print Ultimachine black at 275C with no visible effects on the PEEK or Teflon hot end components. PEEK and Teflon are both limited to 260C for long term service, BTW, but a small additional heat sink and heat isolation (PEEK air gap) can keep them cool. Joergen is also right that "it's complicated" and there's no general rule that works for every print, with every plastic (or every color, per the above). For small parts, I'm printing on bare glass, like Joergen does most of the time too, I think. It's just too convenient (also for PLA at lower temps than ABS). For big, thick, blocky parts that must be as strong as possible, I pull out all the stops and use every trick I know. I set aside my glass build plate and I tape-up my Mic-6 aluminum heat spreader plate with Kapton tape and wipe the Kapton tape with "ABS juice" (dampened paper towel rubbed on ABS). I start the print with a hot bed (120C or better) and, as soon as I know the first layer is going well, I erect a bubble wrap and blue tape enclosure to heat the build chamber. A 120 C bed produces 75C+ ambient temperature at the same height as the bottom of the print head, which I've actually found to be too hot for the entire print (results in areas with shrinkage and dimensional inconsistencies) so..... 1. I reduce the bed temp to 90 C after four layers. I've found this to be better for Kapton to bed adhesion too and it prevents Kapton tape lifting. 2. Use a slow fan for about layer four to 5 mm or so. This is after the initial layers are well bonded but while the bed and ambient air temps are still dropping from the bed temp reduction. Without a light fan, the part is usually slightly shrunken in X/Y in this region. 3. Fan off at 5mm and it stays off, except when needed for very fast print layers (small area layers). 4. Turn off the heat and remove the bubble wrap tent after print completes, When fully cold, I power-up the bed for about 20 sec (my bed heater is 1100W and AC wall socket-powered) and wait for the part to warm a bit on the bottom. 5. Work like hell to pop the part off the bed (sometimes requires resorting to using a razor blade and putty knife wedge, which isn't good for Kapton tape life usually). The above may not work for you at all (no generalizations, like Joergen said) but I find that it results in no curling, no lifting, and minimal internal signs of stress (partial layer delamination or dimensional excursions) in the print. In general, (and there are no generalizations ), I've found that a build chamber that is too cold results in delamination or partial delamination between layers. This assumes the part remains fully affixed to the build surface. If the part warps and separates from the build surface at all, this relieves the internal cooling stress in the part and you may not see any delamination effects or they may be minimal. I consider any separation of the part from the bed (lifting) to be a print failure and it's a failure I never see when following the steps above. Sometimes I'll accept a small amount of lifting when printing ABS on glass but, if it's excessive, I pull out the Kapton tape! By the way, Kapton tape on glass works too, but it doesn't stick as well to glass as aluminum. Kapton tape on glass offers a compromise between the convenience of printing on glass (hot swapable glass plates) vs. the delay of waiting for a Kapton-taped aluminum bed to cool after a print completes. Glass plates can easily be broken when removing large parts from an ABS-juiced and Kapton-taped glass plate though. I speak from experience; I've broken two! If the build chamber is too hot, you will see "waisting" or shrinkage in some areas and the part will have excessive dimensional errors. Partial delaminations are easy to spot in dark ABS colors (like black, which is most prone, due to its higher temp requirements). Any lighter striations along the layers is a partial delamination and it means the part is not as strong as it could be! When it comes to the issue of stress and strain associated with cooling, I always say "the plastic has to cool sometime!" The trick is to make sure it cools in a manner that distributes stresses (spatially and over time) such that the strain is minimal and evenly distributed too. I just received five 12" x 12" sheets of PET tape from Lulzbot but have not tried PET (polyester film tape) yet. It's thicker and looks tougher than Kapton but will have to prove to be tougher to be worth the price!

It's looking like my experiences may not be unique! http://umforum.ultimaker.com/index.php?/topic/2047-resetting-firmware/ I did not have to rebuild the firmware in each of my three recoveries. I simply re-flashed with the same firmware. I'm sorry that I did not record which versions have been loaded when my three failures occurred. I'm currently running Illuminarti's "Marlin-Menu_fixes" with only a few personal code tweaks to suit my preferences, and it has not exhibited this problem, but I've only been running it for 2 or 3 weeks so it's early days yet! Even if one finds a version of firmware that does not fail in this manner, I suspect the firmware is not the root cause of the problem, but rather there is a signal integrity problem or some type of data corruption occurring in the Arduino and, by chance, may not manifest itself or be as likely to manifest itself with certain firmware code. In short, I'll wager that, at its root, this is not a software/firmware issue. It's a hardware issue.

My printer can print 210 x 210 x 205 mm and that's what I've put into netfabb setup. Nonetheless, nf displays a grid with 22 x 22 squares. In the netfabb graphic, the top row (the one across the back of the printer by the lead screw) and farthest right column of squares are "out of bounds." I just place the part manually and don't use that real estate. I don't know how far off your parts are placed by nf, but the center of the actual build area many not be where you think it is!

Yeah--I noticed that too, but didn't mention it and figured I'd leave it to someone else. Yes--very cool and I think it's very complementary of your work, Daid!

Hot off the servers in 2.67: http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.67/Print_Tools

Yes, this is only why I mentioned it here, but it's a tough problem to debug. And now that I think about it, it may have been on the second layer for me too, because the netfabb gcode files have a nozzle prime/test pattern on the first layer and it printed okay but then it went to the wrong place to print the part. Very strange indeed! I tried some Cura gcode on it too, but the print had a very small footprint and, when I saw the head go to the wrong place, I knew it wasn't netfabb related (plus I'd printed these files previously with no trouble). I wasn't very hopeful when I resorted to re-flashing the firmware the first time but it cured it. When I saw the problem the second and third times, I knew what to do! After twice previously fixing solder problems (a poorly soldered through hole that resulted in an intermittent open, and a solder splash outside of the solder mask, resulting in an intermittent short) that I decided to just de-solder the display and reflow/resolder all the solder on the boards! Sorry that I was still stinging when I started writing, from most of a weekend's worth of Ulticontroller rebuilding, but my machine is running very well now. I've been printing non-stop since the weekend. By the way, I'm still very happy with your menu fixes! It really makes the UC work as I originally expected it to function.

My nut was extremely tight in the wood holder, with no discernible movement. However, my rods are very straight (no discernible wobble when rolled on glass) and I'd rate my lead screw as fair (some wobble). The lead screw wobble is, no doubt, why it is not anchored (bushed) at the top. My rods are straight and firmly mounted in the frame. My lead screw nut has no play. Thus, my wobbly lead screw is forced to track to center by the other, dimensionally correct and well-fitted components. No biggie, really. Crappy lead screws are a problem with cheap Chinese hobby mills and lathes too, but a good machine will have very straight lead screws. I had to upgrade one of my Chinese machines for this reason too.

>I think it's also unlikely that firmware 'corruption' in and of itself would cause the printer to >print out of alignment on the second layer... Yes. It's puzzling, illuminarti, but I've seen it happen. I'm using an Ulticontroller and printing from an SD card. I've had something like this happen at the very beginning of a print and it didn't matter what slicer / gcode I attempted to print. The head homed just fine, but when the print started, the UM printed offset in X/Y (I think both X and Y) in the wrong place! I think it's happened to me three times since I assembled my Ultimaker last October under three different firmware revisions, including at least once with a .hex file that shipped with Cura. Power cycling, restoring safe defaults from the Ulticontroller, etc. did not fix the failure in any case. The only fix I've found is to re-flash the firmware. I'm using the Arduino 1.0 IDE to compile custom firmware and flash, but you could just use Cura to flash (update) with the standard Cura firmware. <set rant on> The electronics board assembly quality that shipped with my Ultimaker sucks! The only board that exhibits good workmanship is the little TC amplifier board on the print head. I've now disassembled my Ulticontroller three times to correct solder defects and I've spent more time trying to keep my Ulticontroller working correctly than I spent assembling my entire Ultimaker! The last time I disassembled the UC, I had to de-solder the LCD display single row header (very time-consuming) and re-solder it twice before I got the UC working right. Many components on the board cannot be accessed without removing the display and the alignment of the LCD connectors on the board fingers is much more critical than with most LCD displays. I finally installed a very low profile socket for the LCD display to aid servicing, but this also required milling 0.200" deep relief into the wood face place for LCD display bezel clearance. Given the problem I've experienced, I'd prefer kits that DO NOT come with assembled electronics and connectors. I would have spent less time soldering and crimping them together myself than I've spent dealing with the poor workmanship of the boards. No--I'm not drinking much UM user Kool-Aid after my last UC rework session, where I reflowed ALL the solder on the boards. Unless a perspective user is capable, equipped, and willing to rework poorly assembled electronics boards, I would not recommend an Ultimaker purchase at this time--kit or fully assembled. My failures have all been related to poor solder, causing intermittent shorts and opens, and the failures are latent. They are just as likely to rear their ugly head down the road with a fully assembled and warrantied UM (good for DHL, though :(). I suspect that Daid will say my experiences are rare (as he's responded to other similar reports). Given that I have multiple boards exhibiting poor workmanship, I'm not so sure. I have also seen several forum reports of similar problems, or problems that may very well have a root cause similar to my failures.

>Interesting. Why didn't it work? What happened? Did the molten ABS stick to the SS somehow? Please tell me more! The problem was extrusion resistance was very high and it partially jammed at times. It wasn't usable. >I would love to see pictures on how that fits together! It's a simple setup. Joergen and I are both running MakerGear 36 mm "barrels" (brass tubes) and Big Head nozzles. Here is my thread: http://umforum.ultimaker.com/index.php?/topic/1798-got-abs-you-need-a-makergear-50-cal-nozzle-and-longer-barrel/ Joergen's hot end consumes a touch more Z than mine but it features more heatsink area than my hotend: http://umforum.ultimaker.com/index.php?/topic/1925-modification-to-the-hot-end-for-abs-printing/&do=findComment&comment=13401 >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? The rule of thumb that I've seen is maximum layer height is 80% of nozzle diameter. The lower limit is typically a function of how accurately you can feed plastic (though not terribly useful, 10 and 20 microns probably require 1.7 mm filament feeders ) and how level you can get your bed. I usually print 80 to 150 micron layers (with perimeters as low as 40 microns using netfabb "half-layers"). At don't think I've even come close to max. layer height for my .40 and .50 nozzles. >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 Yes--I'll use a heatsink and probably design a fan duct to cool it, using the existing fan, which I've replaced with a stronger 160 mA unit. I'll leverage the aluminum plate as a heat sink and might even mill a replacement for it that features fins. I don't have any drawings but here are the quick design elements. Leverage anything you like as you please, because I won't have time to build it for a while: 1. 303 SS tube (relatively easy to machine). 2. 303 SS tube can be as large as 5/16" diameter (I think) inside the wooden box (Bowden is 1/4"). 3. End of Bowden will be threaded. 4. Top of 303 SS tube will be threaded into a nut coupling, used to connect to threaded Bowden tube. One set of threads will be left-hand (probably on the Bowden tube side) to aid assembly (just turn the coupling nut to connect or disconnect the Bowden tube to the 303 SS tube). 5. Bottom of 303 SS tube will thread into an aluminum heatsink to replace the PEEK and interface with the aluminum plate, providing both upward retention force and heat conduction to the aluminum plate/additional heat sink. 6. Plywood is fairly high temperature tolerant (probably won't ignite until at least 300 C ) but insulation between the SS tube and the wood might be prudent--especially if there is contact at the bottom of the box. The top of the box would be the preferred location for retaining the 303 SS tube on the other end of the "sandwich" (due to lower temperature of the 303 SS tube there). 7. 303 SS tube could thread into existing aluminum heater (only the nozzle would be brass). A hybrid using a brass tube in the heater block could also be investigated. 8. I have no interest in adding another fan. I will leverage my single fan, ducting and heatsinks, Perhaps a new duct with have a mechanical "ABS gate" to route most of the airflow to the aluminum heatsink when printing ABS, which requires little, if any fan cooling.

Not according to my experiences or this data sheet, which says max. long-term temp for both PEEK and PTFE is 500F (260 C). http://www.plasticsintl.com/sortable_materials.php?display=thermal&sort=heat_deflection66 I discolored and slightly deformed my first PEEK running ABS prints at 265-270C. With my current setup, similar to Joergen's hot end, I print Ultimachine black at 275 C. I molded a PEEK replacement out of Hi-Temp Lab Metal compound. Though it's somewhat fragile (it's a filler and molded parts is not its intended application), it's good for 1000 F. I don't really need it right now and if I had some Macor ceramic laying around, I'd machine it into a PEEK replacement.

Yes--I wish you luck, Red! I wish I had the time to build my own UM. I noticed this from your blog: Everything came from printers, the 12mm rods are actually 11.90mm wide, not perfect, but those should do for now. Can you find some bearings or bushings to fit these rods? 0.100 mm is actually a lot of play, when you consider that the UM is capable of printing layer thicknesses well below 100 microns (and I usually do!). Given the scope of the project and the potential to have a machine that's even better than machines built from the kits, I think I'd spend more and get rods that fit the bearings/bushings very well. BTW, the chrome plated rods in the kit are very high quality but, if I were building from scratch, I'd match the end quality at lower cost by using drill rod and a good low friction and wear-resistant conversion coating. I'm looking forward to following your progress. -Cal

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

Assuming you are running a recent version of netfabb, retraction should work. However, it is incompatible with the recent Marlin retraction fix, unless you cap the extrusion speed to something reasonable. I found that 25 mm / sec is a touch slower than the netfabb reatraction speed under the old Marlin, which some people complained sounded mechanically aggressive, but 25 mm / sec results in typical retraction speed and performance and it is working well for me. See: http://umforum.ultimaker.com/index.php?/topic/1734-marlin-bug-the-myth-of-retraction-speed/&do=findComment&comment=13291 I've found that although the netfabb "pull-off" and "push-on" settings can be varied, the retraction speed settings have no effect on the gcode. Hence, you must limit the speed outside of netfabb.

It will take me awhile to lose my inherent mistrust of all Ulticontroller menu settings, because they have been so buggy in the past. However, your Marlin_Menu_Fixes code tree appears to be the best ever, Illuminarti. Thanks!

I've not used Kisslicer in a while but, as Daid and Illuminarti suggested, I found that changing the following line in Configuration.h from //#define DEFAULT_MAX_FEEDRATE {500, 500, 5, 45} // (mm/sec) default settings to #define DEFAULT_MAX_FEEDRATE {500, 500, 5, 25} // (mm/sec) post retraction bug E limit limits the default netfabb retraction speeds (which work quite well) to a sane level that approximates nf's performance under the old firmware. While the nf linear retraction user settings ("push-on" and "pull-off") are functional , the nf user settings to change the retraction speed appear to have no effect on the gcode. Therefore it's necessary to limit the speed in the firmware, as it is necessary in order to run old gcode too. I've filed a ticket with netfabb about the apparently non-functional speed settings.