Jump to content
Ultimaker Community of 3D Printing Experts

cubic

Member
  • Content Count

    10
  • Joined

  • Last visited

Posts posted by cubic

  1. Both Torgeir and Sander thank you for the input. It really helps. 

     

    It much looks like Sander's suggestion regarding the faulty layer one third up was correct. I added some weight to the printing table and the problem was gone. Thank you.

     

    And Tongeir, thank you for explaining the thermocouple operation. I did know that but good for completeness. For my work I use thermosensors (melting plastics 😉 ) on an industrial scale. When I have to choose, I will never use a thermocouple. I simply cannot trust millivolt signals in high power installations (although I know the majority uses them). I rather use Pt-100s even when I know that Nasa engineered them out of their spacecrafts for reliability reasons.

    Today, finally, I got the possibility to measure the temperature with a proper Pt-100. When set to 250°C and the controller indicating 250°C, the side of the Alu block measured 210.2°C. I made several attempts and have been using thermal compound between the alu block and the sensor but 210.2 was the max. Very likely the actual temperature may be somewhat higher but it certainly will not be 250.

    And I can see that a too low temperature causes the problem I experienced; too little flow.

     

    Reconnecting the thermocouple wires has made no difference. They were firmly screwed down and are again now.

     

    So the conclusion is that the actual temperature is much lower than indicated. In my opinion it is either the thermocouple or the compensator that is in error. How to find out? Or solve this differently and switch to Pt-100? (easier to check)

     

    Thanks for any comments.

    Cubic

  2. Hi Sander,

     

    It took some time; life got in the way.

    This photo, I think, shows that the heating cartridge and the sensor are in okay: front.thumb.JPG.c7413d13b80073749254118731b89d45.JPG

    The side view shows a more general impression.  I noticed that the front (left in the picture) screw is not tight, but there is no way I can tighten it more than this. Looks like the ring of the Peek bush is thicker than the space provided. The "blood stain" on the Alu block allows for the thermal camera to be used.  Recently I disassembled the nozzle from the Alu block and found that is was no contact between the Alu and nozzle in the surfaces around the thread. Now there is as you can see.

    side.thumb.JPG.9cca83bd1143834fb7f9f5df24d950cf.JPG

    And from the rear the fan duct  is in the way but I hope this shows what you wanted to see.rear.thumb.JPG.940ca90e96ea8179e40a21376c127471.JPG

    I've no clue what you hoped to see, but okay. 

    Maybe I should inform that the temperature controller seems to work alright in that the setpoint and process temperature are always close together during printing. But I have the feeling that if the controller measures, say, 240°C, the actual temperature is something like 210°C. Do thermocouples wear out?

     

    regards,  Cubic

     

     

     

     

    front.JPG

    rear.JPG

    side.JPG

  3. Hi Sander, 

    Thanks for the invitation, although my first reaction was "the weight is wrong, how to photograph that?". But anyway it may be a good idea that more experienced people give their view on this. Please find two photo's attached. The left one is what started this search, so the prints before any  of the actions described above. The right one is what comes out at 240°C and at 110% material in Cura. I am not unhappy with that, but it is still the weight is too low, Cura says 19 grams and it is 16 to 17 grams. Close but no cigar.

     

    Photo top view:

    topview.thumb.JPG.a1e96bb037818c49b1991cf1566a0d53.JPG

    What I think I see in the top photo; at the left part you recognise lose outer walls (right of the right leg and at the bottom at the left. And I see the top not being filled, where it looks like things go especially wrong when the printer is in a hurry. To me it looks like the temperature is getting lost and less and less material is provided.

    At the right part all looks a lot better, maybe even over extruded in the left leg but remember this is 100% material. Maybe even I should not complain about this. But wait, lets look at the side:

     

    Photo side view:

    sideview.thumb.JPG.dee285437603d4a6d9529e649bb04967.JPG

    The left part shows signs of under extrusion. It warps. And the top half of the wall looks good, but remember it is connected to nothing inside. It is a 0,3mm free standing wall. If you wanted this, you'd not be able to design and slice it. The "dashed layer" about 1/3 up is the same all around the object, you see this in all walls.

    The right part has been damaged a bit when removing the brim, sorry, just ignore. Still it warps a bit. I'll save up some money and get me a heated building plate. But the "slits" worry me a bit. Never had that before. Maybe it has to do with the temperature. At this point I am not all to worried about that.

     

    What worries me most is that with the 240°C setting the weight is too low. And only 82mm is extruded when I ask for 100mm. Knowing that it does extrude 100mm with the Bowden cable out of the printhead.

     

    Thank you all for your time, regards, Cube.

  4. Hi All,

    Although I am the first owner of my UMO, I am no way an expert. I have been making things, but I am not happy with the results. In hindsight printing always has been troublesome. Let's say mixed results. But more and more I am using the printer in projects and I have seen much better print quality than what I get. Often I have vertical outer walls not connected, or the top layer not showing a closed surface. So a simple print quality topic? No!

    Recently I printed 3 clamps and Cura told me they should weigh 59 grams. I doubted that and so put the clamps on the scales. It turned out to be only 29 grams. And I  should tell you that I am printing at 110% material and 210°C. So the 29 grams would have been only 26 grams at 100%. So close to 50%. Oh, the feeder is wrong? Maybe the stepper is doing half steps?

    I took the Bowden cable fron the head , cut the filament flush and had it extrude 100mm. And it was exactly 100mm.  The gear moves about 4.25 revolution. All good.

    However, when I re-mount the Bowden cable, preheat the thing, and mark the filament 100mm before the feeder and then extrude 100mm only 60mm went in.

    I have listened closely but I do not hear steps being lost. So it is slip then? Maybe a bit more force in the feeder? No! it does not make any difference.

    Knurled bolt blunt? No!

    Maybe dirt in the nozzle? Well it is obviously not new but I took it out, got the material out and could look through. Used dentist  tools and had a good look inside the nozzle; it looks like a new one. When taking the nozzle out I saw that the nozzle was not really against the aluminium block, so no good thermal contact. I remounted it so that the nozzle now is firmly in and against the Alu block. After all this about 75mm go into the feeder when I ask for 100mm.

    Next I elevated the temperature because I saw with a thermal camera that the temperature of the nozzle is only 185°C. I ended up setting the temperature  at 240°C. And that for PLA. It should come out brown, what ever colour you put in, isn't it?

     

    Is my thermocouple gone?

     

    But still at 240°C.... it extrudes only 82mm when I ask for 100m.

    I am at the end of ideas. Please help.

     

    Regards,  Cube.

     

     

  5. Hi All,

    Today I made a fresh install of Linux Mint 18.1 on a PC. Next I added some software like FreeCAD and Cura. Clicking Cura seems to do nothing. When I start Cura by typing cura in a terminal window, I get "/usr/bin/cura: regel 3: python3.4: opdracht niet gevonden". Good Dutch for "command not found".

    This is obvious because python3.5 is installed. When I modify the /usr/bin/cura script to python3.5 I get a traceback ending in:

    from . import multiarrayImportError: cannot import name 'multiarray'

     

    Anyone a good idea?

    To me it looks like Cura should be using "its own"python3.4 but is finding the system installed pyton3.5 but I have no idea how to set this right.

    .

    Thanks, Cubic

  6. @ gr5: True, no doubt about that. However the printout with-no-fan is even better to the point than the first printout. Something was very wrong in that first printout but I am not able to reproduce it.

    @suminvent: Please go ahead. My first question would be "does it print the same (wrong) thing every time?" As my printout was wrong just once.

    Cub.

     

  7. Hi again all,

    If you look at trouble as an opportunity to learn something, making mistakes is becoming a must. But so is learning from them. So I put in some time to try some of the suggestions. And again the outcome is unexpected.

    It seemed a good idea to just reprint the object exactly like I printed it yesterday. Same Gcode file; the SD card has not been out of the printer. The printer operates stand alone. The left one is what I got:

    triplet

    It is my impression that the left hand side is even better than the right hand side the way the ball extends from the stem. This may be explained because the fan was on the left.

    Now I have to admit that I do not trust the fan. It sometimes makes odd noises and then is running much slower. I suspect the fan bearing is worn due to the accelerations it is subjected to. I could not remember hearing that yesterday but you never know.

    And since it was suggested that the high print speed and faulty fan operation may produce this result I sabotaged the fan. The middle one is printed without fan. Sure, this is not good but still it does not have the wide bottom under the ball. The way the ball starts extending from the stem is stll better than the first print.

    Just to be certain I printed a third one. Also all right.

    It looks like we could conclude that the print result was a one-off and close the case. Or should we learn more? Just keep it in mind for in case more turn up.

    Cub.

     

  8. Hi Again,

    Thanks all of you for taking your time to look at this item. Here are the requested views. In my perception there is no problem with what is shown on screen. It seems logical but the output....

    basic settings

     

    advanced settings

     

    xray view

    slice 30:

    layer30

    slice 51:

    layer51

    slice 65:

    layer65

    slice 67:

    layer67

    slice 71:

    layer71

    slice 80:

    layer80

    slice 90:

    layer90

    slice 100:

    layer100

    slice 110:

    layer110

    What tools can I use to investigate the gcode file?

    Cub

     

     

  9. Good, it is a project in its own to get a picture in the forum, but here we are:

    after importing the .STL into cura I saw:

    incura

    And this was the result:

    printout

    Not only is it different, but I would find it understandable if this could not be printed at all due to the large overhang.

    Cub.

     

  10. Hi All,

    I am using an Ultimaker 1 and Cura 14.07. I have some experience printing either self designed objects (using FreeCAD on Linux) or Thingiverse things like gears. All well, never had unexpected results. But today I designed a “mushroom”, so a 5mm stem (stick) 5mm tall and a 12mm football on top.

    I was a bit anxious for where the ball should start widening from the stem since it is an overhang but since the subsequent layers extend only slightly I hoped it would go.

    I was more than surprised to find in the printer a shape where the stem is OK (5mm diameter) but then the next layer, where the minimal overhang should begin, is 9.2mm diameter. The next layers remain the same size until the ball becomes bigger than 9.2mm and from then on the rest of the ball is OK.

    The imported STL file looks OK in Cura. Inspecting it on a layer level show nothing special. But the output is remarkable because this is quite an overhang and it looks good but is far from what was intended.

    Thanks for your thoughts

    Cub

    as shown in Cura:

    incura

    and this is the result:

    printout

     

×
×
  • Create New...