Jump to content
Ultimaker Community of 3D Printing Experts


  • Content Count

  • Joined

  • Last visited

Posts posted by das_enginer

  1. Just ordered - I went with 3D Universe (US) as they were showing the fastest expected "ship by" date (3/18). Hope to actually have it before the end of March!

    I'm also really hoping this is the end of my under-extrusion saga - I've never really been able to print reliably faster than 40mm/s (0.2mm layers), and even then I'm running at 245C and with exclusively Faberdashery filament (tighter filament curvature on the rear spools kills it). I love Faberdashery, but it's quite pricey in comparison to others, and I'd love to be able to use a rear-spooled vs. my floor-fed turntable. Feedback so far sounds promising, and the updates seem to address all these issues, so my fingers are crossed!

    In terms of the pricing, I've had UM2s since late 2013, so I'm not going to complain that I need to spend 13% of the original purchase price to replace 1/2 of the machine's critical components. Yes, it's a bit steep to fix something that never worked perfectly in the first place, but it's not like Ultimaker did any of this on purpose - and they could have never released this upgrade at all.

    • Like 1

    I would hate to spend that amount to upgrade only to have a UM2+ Dual come out afterwards which would allow me soluble supports...


    As we said before; Don't expect an official release for a dual kit for the 2. There are too much problems to fix (fixing them would require massive redesign of the machine, which is simply not feasible for a kit).


    Absolutely! I was intending to say that I wouldn't want to spend $X00 on an upgrade kit if there was a $X000 new machine that I would want to purchase instead. The line of logic that follows 'why would I upgrade my old machine if I was to just replace it entirely'.

  3. The plot thickens! I'm now VERY curious about the price of the kit to upgrade my UM2 to a UM2+. Before this, I wouldn't have blinked an eye at something around $300 USD for the upgrade kit, but I would hate to spend that amount to upgrade only to have a UM2+ Dual come out afterwards which would allow me soluble supports...

  4. No additional holes needed. It comes with an adapter plate that is mounted with the four screws that hold the feeder motor.


    Interesting! So the conversion kit functionally makes a UM2+ out of a UM2, but it's really a 3rd variant (in that a converted UM2 and a new UM2+ aren't identical). Are there any other deltas you know of aside from the extruder housing adapter plate?

  5. When I originally read about the 2+ on Engadget, I was initially rather upset - but then I remembered Ultimaker isn't Makerbot and found this post including the announcement of the upgrade kit.

    Thank you for being so inclusive of everyone in the community and not excluding those with a UM2! I will definitely be getting the upgrade as soon as it's available!

    • Like 1
  6. Thank you Simon, you've once again saved the day!

    I've updated my pre/post fixes (below), and aside from the initial table raise, the only thing left to figure out is why when using this code, the extruder motor skips once right before it begins its skirt path. It's not a big deal as long as your skirt is large enough to allow it to re-prime fully. I've tried many different E lengths and F speeds, but it seems to constantly do it no matter how much/little I prime or at what speed, whereas Cura does not, even at lower temps. The full-stop retraction at cycle end should be 21mm, and I'm priming 18mm quickly, then slowly priming another 10mm, so theoretically only 7mm should be getting pushed through.

    Disabling de-string stops this skipping, so I'm guessing that because of the travel length from the prime spot to the print area it's doing some form of suck and prime, which is running too fast and causes a skip. That said, I plan on dealing with the skip for now since I'd rather not disable de-string.

    G21 ;Metric Values

    G90 ;Absolute Positioning

    M82 ;Extruder to Absolute

    M106 S0 ;Fans Off

    G28 ;Home All Axes

    G1 X5 Y10 F9000 ;Move Head to ~Origin

    G1 Z20 F9000 ;Raise Bed to Head

    M140 S<BED> ;Set Bed Temp

    M190 S<BED> ;Wait For Bed to Reach Temp

    M109 T0 S<TEMP> ;Set & Wait Head Temp

    T0 ;Select Extruder 1

    G92 E0 ;Zero Extrude L

    G1 F100 E18 ;Forward 18mm @ 100mm/min

    G1 F50 E28 ;Extrude 10mm @ 50mm/min

    G92 E0 ;Zero Extrude L


    T0 ;Select Extruder 1

    M104 S0 ;Head Temp Off

    M140 S0 ;Bed Temp Off

    G91 ;Relative Position

    G1 E-1 F300 ;Retract 1mm

    G1 Z+0.5 X-20 Y-20 F9000 ;Drop, Move Head

    G90 ;Absolute Position


  7. Hi All!

    After working with the amazing Simon & fbrc8 quite extensively, I finally have a printer that I think I can run without worries of under-extrusion so long as I stick to certain filaments (with more testing to come).

    I was running KISSlicer back on 14.03, but it's been so long since I've had a truly functional printer that I haven't dealt with the new UM firmware until now. I'm trying to mimic the 14.07 Cura startup sequence, but either I'm missing something or I'm no longer able to override all actions on the printer with 14.07.

    Basically I have the prefix almost down to decently match the Cura startup. I need to adjust the extrusion prime lengths, but the main issue is that the z-axis raises to ~z = 15mm before my home command even gets reached. It's almost as if there is a "G1 Z15 F9000" in the firmware that is superseding my GCode prefix. The machine will raise to the z height, THEN home all of my axes, which means the startup sequence is: bed up - home bed (bed down) - home head - bed up. Something very similar happens if I simply change Cura to output RepRap/Marlin instead of UltiGCode.

    My postfix is in even worse shape; it seems to run my postfix just fine after the print (hard to tell exactly due to speed, but what I would expect), but then retracts the filament even further and re-homes the axes again. Here it's as if there is a "G1 E-5 F300" and another "G28" concatenated to my postfix.

    Is there any way to get the UM2 to stick to just what is in my prefix / postfix, or is the result going back to an older firmware?

    My prefix code is:

    G21 ;Metric Values

    G90 ;Absolute Positioning

    M82 ;Extruder to Absolute

    M107 ;Fans Off

    G28 ;Home All Axes

    G1 X5 Y8 F9000 ;Move Head to ~Origin

    G1 Z15 F9000 ;Raise Bed to Head

    M140 S<BED> ;Set Bed Temp

    M190 S<BED> ;Wait For Bed to Reach Temp

    M109 T0 S<TEMP> ;Set & Wait Head Temp

    T0 ;Select Extruder 1

    G92 E0 ;Zero Extrude L

    G1 F200 E5 ;Extrude 5mm @ 200mm/s

    G1 F50 E7 ;Extrude 2mm @ 50mm/s

    G92 E0 ;Zero Extrude L

    G1 F9000 ;Set Axes Rate


    My postfix code is:

    T0 ;Select Extruder 1

    M104 S0 ;Head Temp Off

    M140 S0 ;Bed Temp Off

    G91 ;Relative Position

    G1 E-1 F300 ;Retract 1mm

    G1 Z+0.5 E-5 X-20 Y-20 F9000 ;Drop, Retract 5mm, Move Head

    G28 ;Home All Axes

    M84 ;Steppers Off

    G90 ;Absolute Position

    Any help is appreciated!



  8. Have you ever removed the bolts on a car wheel, only to find you have to hammer the wheel off because it's stuck to the hub? Two dissimilar metals, as illuminarti points out, can corrode and bond to each other (which is what the aluminum or magnesium wheel is doing to the steel hub).

    The copper grease here is an anti-seize compound that reduces/prevents galling, lubricates the joint, and prevents the two metals from seizing to each other. I usually only spec the copper version in more expensive delicate machinery, but it's nice to see it here!


  9. I have a habit of finding weird ways to break things in cars; it's apparently carried over! :cool:

    When the extruder stepper is at full speed with 1625mA during the few millimeters of priming, the motor doesn't skip and lightly chews the filament. When the extruder stepper slows down and is printing the cylinder, it's definitely skipping backwards just as it does in the higher volumetric speeds with lower currents. I have a small red mark on the end of the extruder stepper shaft so I can watch this instead of guessing (it should be almost visible in the video).

    The extruder motor is the same size as the others, around 42mm a side.

    I've also uploaded the test files I used here (https://dl.dropboxusercontent.com/u/10655012/Test%20Files.zip) just in case I mucked them up again.


  10. I’ve run some more tests with increasing current, and found a rather odd behavior. The print improves at 1.5A, but taking it any higher causes the print to decay rather quickly.

    1250mA – Print starts to skip at 6mm^3/s, best was 7mm^3/s

    1500mA – Print starts to skip at 8mm^3/s, best was 9mm^3/s

    1625mA – Print starts to skip at 5mm^3/s, very quietly

    IMG 3325

    1750mA – Printer skips immediately after priming (video of it a few seconds into the print)

    IMG 3321

    1875mA – Printer skips immediately after priming

    2000mA – Printer skips immediately, never primes

    I ran another 1500mA cylinder after going through these to make sure something hadn't fried in the process and created the error, and it failed exactly as the first 1500mA cylinder (eg nothing was fried). The only point at which I had filament grinding was when the printer was priming the nozzle at 1625mA, but this was very minor and the printer carried on of its own accord.

    Thoughts? :???:


  11. Hi gr5 - that's good to know, I didn't know if the stepper driver circuitry was different between the 4 steppers, or if it was 4 identical drivers merely being controlled differently.

    I do agree that there are concerns with the feeder housing and internal friction; I'm actually trying to design my own with a completely "frictionless" path up until the bowden tube entry by making use of grooved roller bearings as entry and exit guides. I'll definitely share for review as soon as I get time to wrap it up and can source some 'reasonably' priced grooved bearings in the spec I'd need.


  12. About the filament, I have always had this in the back of my mind.

    Different colors require different temperatures, different suppliers have different stiffness.

    There are so many different gradients in just 'PLA', that it is hard to tell what you get from which supplier.

    Especially I think what we are suffering from now is the stiffness of the filament.

    Has anybody ever checked the filament during/after under extrusion if it was still round or oval?

    - How does this compare with a piece of filament that has not went through the feeder yet?

    I think if you store the filament near a heat source the moist can evaporate, and it will make it more stiff or perhaps brittle? Does that make sense?


    Hi Sander,

    This is something I've been looking at closely with the filaments I've had, albeit only 3 suppliers. My Faberdashery filament is the most pliable of all of my filament, but still fairly stiff. The Ultimaker filament that came with my printer is stiffer than the Faberdashery, but the black I got from MakerGeeks is the stiffest of all of them (though not MUCH stiffer than the Ultimaker). This matches the ease of printing with them; the Faberdashery is the easiest, then the Ultimaker, then the Maker Geeks, but this is probably skewed by the Faberdashery having a much larger bend radius than the others.

    I did a quick measurement on my filament, and a strand extruded in air is pretty round (0.39mm to 0.41mm), where the input Ultimaker filament is also round, but less so (2.85mm to 2.92mm). I didn't check my Faberdashery yet, and have stopped using the Maker Geeks for the time being.

    In terms of removing moisture, I've heard this is beneficial for print quality as it will produce less defects since any water in the filament will steam when it's extruded (and make a small popping noise). I don't know how heat will affect this, since it can change the structure of the plastic of its own accord. Personally I store my filament in a bag with a large desiccant pouch I bought from Amazon, and this keeps most of the moisture out, but I've not noticed a change in stiffness.


  13. Okee dokee, I ran the test cylinder at 1.5A, 1.6A, 1.7A, 1.8A, 1.9A, and 2.0A.

    IMG 3319

    Then I realized I had flubbed the G-Code hack and had just run 6 cylinders at nominal current because I’d put the M907 in the wrong place, which explains why they all looked basically identical. So I fixed the G-Code, and redid the some of the tests. This time I ran the 2.0A before anything else to make sure I noticed a difference in the printer’s behavior and had gotten the current increase correct.

    2.0A – Nothing ever extruded; the stepper shuddered and made an awful noise, but never actually completed more than a couple of steps. I’m guessing the motor is oversaturated at this point and simply can’t break from one magnetic field to the next.

    1.5A – I’m kicking myself over this one. The test I had tried earlier in the day that made me think current couldn’t fix things, has definitely made a difference when I can get the code right. The best I’ve been able to get before is failing at the start of 7mm^3/s, but this took it over halfway into 9mm^3/s before failing (usually failing at 5mm^3/s).

    IMG 3315

    I’m not convinced I heard a “skip” when the underextrusion happened though, but I didn’t see any grinding flats on the filament, and I haven’t tested the 1.25A-1.50A range, so this is at least progress. I will run more tests tomorrow or on Tuesday (as time allows) to test the range from nominal to 1.5A, then higher until I get grinding.

    I am curious as to what this could mean in terms of the hardware. The mechanical engineer in me wants to say that this means I have a weaker feeder motor, which requires a higher current to get an equivalent torque as a nominal motor at nominal current. However, the mechanical engineer in me also reminds me I know bupkiss about electronic components and should consult those who do on the matter.

    In the long term I think I’m uncomfortable running at 1.5A all the time, namely due to worry about excess current heat buildup in the PCB driving components and running components out of spec (this was mentioned somewhere in one of the many threads). Would this potentially be corrected (should further tests confirm this is the issue) by replacing the stepper motor, or could this be an issue with the PCB and the current that it supplies to the motor (eg a 1.5A command only gets 1.25A to the motor, etc)?

    Thanks again for being an awesome community!


  14. Okay, the head has been completely disassembled, cleaned, inspected, and reassembled.

    IMG 3288


    • PTFE coupler had a 3.30mm diameter entry and a 3.23mm diameter exit; filament passed through it with no problem.
    • The nozzle had a dark finish on the inside, but was still shiny. I took gr5’s advice and tried to burn out anything that could be on the inside, but nada. It did clean up the outside of the nozzle quite nicely though! There was some very minor thread burrs at the top of the entry; I lightly filed everything smooth. Nozzle tip diameter measured using needle technique; came to 0.39mm.

      • IMG 3295


      [*]Bowden tube didn’t have any significant damage; standard “grab” ring, but nothing that obstructed the flow of the filament.

      [*]Reassembled the nozzle/collar/PTFE coupler such that there was exactly a 2mm gap between the top of the collar and the bottom of the PTFE coupler shoulder.


      • IMG 3302


      [*]Checked concentricity of Bowden – PTFE coupler – nozzle when reassembled, and I couldn’t see any out of alignment.

      [*]Did discover something that I didn’t know; the threaded rod in the nozzle isn’t just a clocking feature. There is a 1.5mm hex key in the top, and actually locks/unlocks the heater and thermocouple into the block. This explains why the top machined plate (that the spring presses against) has clearance holes to access the top of this threaded rod. I only found this because I couldn’t properly reinsert the heater and thermocouple until I had loosed this screw, and once inserted and tightened, I couldn’t remove them either. Knowing this, apparently the heater/thermocouple were never tightened in place since they’ve always just come out & gone in when disassembling the head.



    Reran the test print (230degC / 65degC tray / UM Blue), and it still failed in the range I was seeing before (7mm^3/s). I’m off to bump my extruder current up until it starts grinding the filament away instead of skipping steps. Thanks for putting up with my griping / diatribe troubleshooting, and throwing out ideas!

    IMG 3306


  15. I changed my plan a little; to reduce the number of variables that are changing at once, I decided to fiddle with currents before I disassembled the head. Long story short; no dice. The UM blue that came with the printer is still failing at or between 5mm^3/s and 7mm^3/s. No counts of filament grinding since I didn’t exceed 1.5A. Current was increased via M907 E1500 (http://umforum.ultimaker.com/index.php?/topic/4586-can-your-um2-printer-achieve-10mm3s-test-it-here/page-3&do=findComment&comment=39581).

    I'm really hoping I find something taking the head apart again; I've had this thing since December and have spent far more time tweaking / testing / disassembling than I have printing actual parts. I could just print slower (as I have been for the few parts I've produced), but the raging perfectionist in me isn't happy until I know I can print my purely functional parts quickly (eg around 7-8 mm^3/s). :oops:

    IMG 3287


  16. That's definitely a possibility, but the UM2 blue is the hangup for me. If others are able to print the test cylinder in the blue that came with the machine, I would expect mine (in a perfect world) to be able to do the same. :) In the end, the only "good" filament I can print with reliably is the gold; which is where I was basing my hypothesis that the gold is easier to print with than nominal, versus the others being harder to print with.

    I do agree that the Faberdashery blue will need a higher temp than the gold for future prints, but with it failing at 6-7mm^3/s and already at 230degC, I would be hesitant to say that temperature alone would be a fix.


  17. Good morning Ultimakers!

    Intersesting turn of events yesterday; I was starting to print IRobertI’s cable chain, and it started to skip like crazy a few layers in. I was running it at 220degC, .2mm layers, and 50mm/s speed. This is only 4mm^3/s, so it should have worked fine given that my test cylinder got through 10mm^3/s with only a single skip. However, my test cylinder was in Faberdashery gold, and I was printing in Faberdashery space marine blue. As such, I immediately aborted the failing print, and ran the test cylinder again. Interestingly, it failed at 6mm^3/s, which was expected given the print was failing at 4mm^3/s, but 10degC cooler.


    To see if something was different on the printer when it’s “cold” versus when it’s been running a few hours (I printed some parts in gold before switching to the blue), I immediately changed back to the gold from the blue, and ran the test. Not a single failure in the entire test cylinder.


    I was now curious if the blue had managed to feed in a manner that was causing binding, so I cut any knurled filament from the blue spool, and re-fed to the same curvature as the gold was just running. This helped a tiny bit, but the cylinder was still failing at 7mm^3/s.


    I left the printer to completely cool so I could completely remove the cold vs. hot consideration, and the cylinder came out almost identically to the previous one; failing at 7mm^3/s.


    I was now curious to see if the gold was really easy to print with, or if the blue was very hard to print with (eg which color is more representative of my machine’s ability). I dug out my old reel of UM blue that came with the machine, and the cylinder started to fail in 5mm^3/s. I babied this test print and made sure filament was feeding without friction from the spool, so I’m guessing that the gold is just really easy to print with, and the two blues are more representative of my machine’s ability. I’m also guessing the increased curvature between the UM filament and the Faberdashery filament is what caused the earlier failure (5mm^3/s vs. 7mm^3/s).


    I’ll spend some part of today disassembling the head and double check the following (if I missed something, please let me know):

    · Nozzle buildup

    · PTFE coupler ID

    · PTFE coupler gaps/spacing

    · Nozzle / coupler / Bowden concentricity as assembled

    · Bowden tube damage

    That being said, since the gold does print well, I’m not expecting anything above to turn up issues. I would expect that since these are all mechanical concerns, one filament to the next would fail the same if there were pathway blockage/constriction (unless it was on the hairy edge of failing and the easier to print gold negates that).

    My only other thoughts are to stepper current, and testing the route that IRobertI found to fix his machine. I go directly to this since his printer was failing at around the same point as mine is, but I could be grasping at straws (correlation does not equal causation, etc).

    While I have it apart, is there anything else that I should be looking for / testing / optimizing?


  18. Crap, they don't have them in the european version of the store. Will have to look elsewhere.

    Thanks anyway. At least I know the size and the other specs now.


    I don't get why DigiKey does this type of thing. I have a very hard time believing that a fan company has export control regulations setup to prevent merchandise from being exported to Europe (especially when some of the stock fans on the UM2 are Sunon).

    The Orion plate fans are available to Germany from DigiKey:


    The PTFE coupler fan should be able to be sourced through TME:



  • Create New...