Jump to content

geert_2

Ambassador
  • Posts

    2,084
  • Joined

  • Last visited

  • Days Won

    34

Posts posted by geert_2

  1. I haven't printed with flexible material yet. But have you tried removing the bowden tube, and feed some filament manually into the nozzle? And then see if it flows through easily? This would make sure there are no blockings, no old material stuck somewhere, no nozzle-clogs.

  2. You need to adjust the temperature to the material you are trying to remove and to the material that you use for pushing/pulling.

    I usually do the cold-pull / atomic-method a little bit different:

    - place nozzle in front corner,

    - remove bowden tube and filament,

    - heat the nozzle, insert a piece of filament and push through some material by hand,

    - dial down temp to zero,

    - gently push a few more seconds,

    - do a small manual retract of a few mm (this prevents a big blob from forming in worn-out couplers on the UM2, which would make pulling it out impossible or very difficult),

    - hold the filament there until cool,

    - use compressed air to cool the nozzle faster to room temp (from a compressor only! Never use spray cans: they may contain explosive gasses, and you don't want that on hot nozzles!!!),

    - let it sit at room temp a bit, so the inner part of the filament also gets time to cool well,

    - gently wiggle and turn the filament: this tends to loose debris that is stuck to the nozzle's inner walls,

    - heat the nozzle to 70°C (for PLA) or 120°C (for nylon), or whatever works for your material,

    - gently keep wiggling and turning the filament, and gently pull until it releases.

    So this method is much more gentle than the usual method. The big advantage is that this gives no risk of damaging the rods, or of displacing the coupler and nozzle, contrary to the brutal pulling of the classic atomic method. Also, cooling down deeper and longer makes debris come loose easier, and it solidifies the inner core of the filament: so when heating up again, only the outer layer will be soft, while the inner is still hard: this reduces the risk of the filament breaking half way.

    Sometimes I also poke through the nozzle from below with a needle (very gently, because steel needles are harder than brass nozzles, and it could do damage); or I wipe debris on the inside of the nozzle with a brass M3 screw thread (must be brass, no steel: this would be too hard).

    This method does not work with my modified ABS+ (too weak: breaks), and does not work well with PET (sticks too much). Then I use PLA or nylon instead to clear the remains of these materials.

    This works excellent on the UM2, better than the traditional atomic pull, but I can not say if it would be suitable for the UM3? Especially, can the UM3 nozzles withstand the wiggling and turning action to loose debris?

    • Like 1
  3. I don't have an UM3 and haven't printed with CPE, so the following is from experience with PLA and PET. Usually this sort of defects comes from overextrusion. Or, if it is on the first layer only, it comes from the glass being a bit too close to the nozzle: the advantage is that the layer is squeezed well into the glass, usually with better bonding as result.

    I also see brown spots on the print. So I would guess that the blobs come from molten material that is accumulated on the outside of the nozzle while printing (thus due to overextrusion or the nozzle being too close to the glass), and that sags and falls off on the print. The brown is when it starts to get burned.

    Usually big light brown chunks and strings come from burned material on the outside of the nozzle, while thin black chips come from the inside, in my experience.

    Try printing a small test piece, for example a cube of 15mm x 15mm x 15mm, 100% filled, and see if it still happens after a few layers?

    Stay with the printer and watch what happens under the nozzle.

    Bad bed adhesion also causes a sort of bobbling, but as far as I experienced, it looks a bit different. If it would be bad bed adhesion, that should clearly be visible from the bottom: there should be indents, like the shape of water drops, but then hollow.

    All this is a guess, but it is an educated guess. :)

  4. Each method seems to work in a lot of cases, but not work in some others. Maybe a combination of all good ideas would work?

    The area above the metal bed clips can not be used for printing anyway. So maybe it could be used for priming? What about making a small trashbin in silicone (=heat resistant) that can be clipped onto this metal clip? A size of 10 x 15mm and 10mm high might do. Then prime in this bin, and wipe the nozzle against the edge of the bin, so the nozzle is clean and does not drag anything into the print? After completion, the primed stuff can easily be taken out of the little silicone bin, as it does not stick.

    This would be a simple add-on (a "clip-on")that does not require hardware changes.

    Or something in this direction?

    Priming in the skirt would often work, but sometimes a skirt is not desired. And if the prime contains some dirt, or some oil (e.g. from wiping the nozzle with oil to reduce build-up of molten filament, or from lubricating flexible filament), then that could reduce bonding on that area of the glass plate. Thus if "skirt style priming" is done, it would be best to print a thin plate in a corner, I think. For example: 15mm x 15mm x 1 layer, or a 15mm diameter spiral?

    Anyway, for me the FillCatch thing shown above works fine on both of my UM2, as it also does some wiping.

  5. If you haven't done so yet, I would suggest you try DesignSpark Mechanical for 3D-design and editing. This is freeware from RS-Components (big electronics supplier), and is a limited version of SpaceClaim. It is very good for geometric shapes (not for organic shapes). It only requires registration on their site, which seems an acceptable price to me, regarding the quality of the software.

    It is very easy to learn: it works similar to SketchUp but produces much better quality 3D-models. On Youtube you also find a lot of good manuals and modeling tips. Have a look there to see if the DSM-workflow appeals to you.

  6. If moisture is the problem, then consider making a box with disseccant where you can put the spool *while printing*.

    Find a sealed food box of suitable size, a bit bigger than a spool, make a spool holder in the center, drill a small (3.5mm?) hole for the filament to exit, and put a big bag of silica gel in it. Then place that sealed box with PVA-spool behind your printer while printing. I haven't tried it myself (no need to), but I have seen people using this for printing nylon, which is also very sensitive for moisture.

    Silica gel with color indicator can be found in car shops: these bags are used to dry car interiors to prevent condensation. As soon as the indicator changes color from blue to pink, it needs to be dried in a microwave or other oven at ca. 100°C.

  7. ...

    I am kinda on the side of not turning Cura into a full 3D Package. It reminds me of the days, way back when, at a time when every program was trying to shoehorn functionality into everything.

    ...

     

     

    I am kinda on the side of not turning Cura into a full 3D Package. It reminds me of the days, way back when, at a time when every program was trying to shoehorn functionality into everything.

    ...

     

    Yes, I also prefer compact, separate, dedicated programs for each function. It is better that a program can do one thing excellent, than 100 things half-way. Because in the latter case, you have a bunch of bloatware, but you can't use it for anything.

    For example, Microsoft Word got such a nightmare, compared to the old versions '97 and 2000, that it takes me five times more clicking and searching through the mess of ribbons than before. It really decreases my productivity. And to prevent it from destroying my layout, I have to disable every automatic function (which I also did in Office '97 and 2000 of course). Microsoft calls this "autocorrect", but their idea of correct seems to be different from mine. :) So, for writing bodies of text, I switched to Notepad++, and I add the layout as the last step only in Word. And I do all corrections manually, in the way that I want it. Thus MS Word now takes gigabytes of disk space and RAM, but is less functional for me than a 20 years old Word 97, and even less than Notepad++.

    I hope Cura will not fall into this bloatware trap. At least that's my viewpoint, but of course, anyone has the right to see things differenctly. :)

  8. It was me that posted about my FilCatch addon for my UM2.

    ...

    I originally thought of making this using the wire from a large paperclip, but thought better of it when I realised that it could possibly get snagged with the nozzle/print head if it was accidentally displaced - hence the safe option of making it in plastic.

     

    Yes indeed, that was where my version was based on.

    Being of spring steel (I am not sure if that is a correct English word, meaning: "steel for making springs"), the wire easily moves down when the nozzle passes it. So I deliberately placed it a few mm too high, to get a sort of wiping action. I only need to make sure that I do not position my models in that corner. But even then, it would probably do no damage, or not much (although I am not gonna try it): most likely the build plate will be pushed down when the nozzle tries to put down the first layers, similar to what you have when corners of big overhangs are curling up. The wire is 1 mm diameter.

  9. If you try to avoid standard glass from warping, don't forget the effect of the print cooling fans.

    Some time ago I bought an infrared thermometer, and to my surprise the glass plate temp was *very* uneven: on a small print, in some areas it was 15°C cooler than in others.

    So I tried heating the glass without printing, and then the temp was very nice and even, except at the last centimeter at the edges, but that is okay.

    Only then it occured to me that the cooling fans blowing at 100% made all the difference. Their job is to cool whatever is in their path, so that is what they do. .. :-)

  10. For PLA, try putting a desktop fan (ca. 30cm diameter) at a distance of 3m in front of the printer, and let it blow at lowest speed towards the printer. It should not create a storm wind, but just a nice breeze. This extra air circulation prevents/reduces local heat build-up: it helps a lot for me.

  11. At the very beginning, I also noticed that in dry weather (freezing cold, blue sky) bonding to the untreated glass was reasonable. But on warm, rainy weather, it was almost non-existent. Putting down a straight line of filament on the glass is one thing, but I have models with lots of tiny circles, which are much more difficult to make stick.

    With Ultimaker and colorFabb PLA now I get a very good bonding with my "salt method": first clean the glass with isopropyl alcohol and then with warm water. No soap, no detergents, no window cleaners, as they all reduce bonding (nothing sticks to soap). Then using *salt* water, wet a paper tissue, and wipe the glass plate with this. Keep wiping very gently while the water dries. So you get a thin mist of salt stuck to the glass plate, almost invisible; it just looks a little bit dusty.

    Heat the plate to 60°C for PLA. When hot, this gives a very strong bonding: I can lift the whole printer by a pulling on a model of 10mm x 10mm x 10mm. But after cooling down, the models pop off by themself, no force required.

    This salt method does *not* work for ABS, and it works a bit but not really good for PET, NGEN, etc.

    For the full manual, see: https://www.uantwerpen.be/nl/personeel/geert-keteleer/manuals/

    Others have good results with hairspray (spray on a paper tissue, then wipe the glass; never spray in the printer); or with white wood glue dissolved in water (1 part glue, 10 parts water).

  12. Is this caused by leaks in the nozzle? Or by the model coming off the build plate? If the latter, to prevent this from happening again, you might want to have a look at your bonding-method too.

    For example, try a difficult test print? A rectangle of 50mm x 10mm x 10mm, with 3mm chamfers at the bottom, 100% filled. This will be a very hard test for any bonding method.

    Stay with your printer (!!!) until finished, and watch if corners do lift, and if the model comes off. If it does, find a better bonding solution first, before starting a real print.

  13. Put your cell phone and wireless devices far away: they may generate very high levels of noise.

    Try making a well-grounded faraday cage around your printer, in shiny copper gauze, or so... :-)

    But then, if the interference would come from within the printer (e.g. stepper motor noise that is coupled to and radiates from the aluminum build plate), it wouldn't help of course.

  14. What I would recommend: have everyone who will use the printer study the main aspects first:

    - How the technology works.

    - User manual.

    - Troubleshooting and maintenance manuals on Ultimaker site and 3dverkstan site.

    - How to get good bonding of the model to the glass plate (very important, so the models don't come off and damage the print head with molten filament spaghetti).

    - How to load/unload filament in a safe way.

    - How to do atomic pulls in case of blockings.

    - To not let filament sit in the nozzle when heated, so it doesn't burn and clog the nozzle.

    - etc...

    And implement a good polici on how to use, maintain, and report about any issues.

    The correct use of the printer will be more important for good quality and for printer-life than accessories, I think.

    • Like 2
  15. A few outlines that are separated about 10mm (adjustable) from the print are called skirt. This can be enabled and adjusted via the expert settings (at least on Cura 14.09, so probably also on your version).

    A very thin layer connected to the print to improve bonding of the corners and prevent lifting, is a brim. And a thick plate under the print, like a rough floor mat, is a raft.

    • Like 1
  16. Recently I have been testing various things to reduce this effect. Often I have tiny models with lots of small holes (a few mm diameter) at the bottom, so they need to be squished well onto the glass bed, to prevent the hole-outlines from being pulled off.

    Things that help:

    - Round the bottom edges with 0.3 or 0.4mm, maximum 0.5mm. I tried chamfers too, but they worked less good than rounds in my tests.

    - Or if you have access to the design files: if you print the first layer with 0.2mm thickness, then manually reduce width of that layer by 0.1mm to 0.2mm. Ideally this reduction in outer size should be done in the slicer, and on the first layer only. But since that is not possible (yet?), you could simulate this effect manually in the design. If you reduce the first layer too much, it creates a worse effect: ugly overhangs, and a reduction in bonding strength. You need to try the optimum for your settings.

    Make a small test object, and try each of these tactics. For example, design 6 blocks of 10mm x 10mm x 10mm, and give them rounds of 0.3, 0.4, and 0.5mm; and a first layer reduction of 0.1, 0.2, and 0.3mm.

  17. According to Wikipedia (but this is not guaranteed to be correct), acetone can be mixed with water, similar to alcohols. However, I don't see why you would add water in the mix, since that doesn't dissolve ABS? So, at best it wouldn't do anything, as far as I can see? Or did you have a specific reason, for example another additive added to the water?

  18. What about disabling automatic posting on Youmagine?

    On some forums prior to posting we see a warning: "Note: all entries will be reviewed and approved manually by our moderators, prior to being posted."

    And after posting: "Thanks for your post. Your message will be posted as soon as it is approved by our moderators. Expect a delay of a few hours to a day before it shows up."

    Or something similar. This prevents any spam from getting posted at all, and it discourages spammers from trying. But it does not stop automatic spambots from trying, of course. Therefore you may need advanced captchas too.

  19. With a standard nozzle the extruded "sausage" is 0.4mm wide. But then it gets squeezed onto the model and spreads a bit. So, it will be hard to print lines and corners smaller than 0.6mm to 0.8mm. Maybe it could be done with a 0.25mm nozzle? But I still don't think the gear will mesh properly and run smoothly.

    Anyway, for best accuracy print slow and cool (e.g. 185°C for PLA).

    Also, keep in mind that if you print the gear in PLA, and it has some friction or is mounted close to a motor, it will deform from 50°C on, which can be reached very soon in an enclosed housing. So you might need a high temp material (and one that can be printed with a 0.25mm nozzle, but I don't have any experience with these).

    For functional items, I think you are better off buying these gears.

  20. I believe gr5 is right: the first layer should *not* look like round sausages that are dropped onto the bed. But it should look like a very thin plate that is really squeezed *into* the bed. Very different.

    If not squeezed into the bed enough, the molten filament will stick to the nozzle, instead of to the bed. And that is what I see in your videos.

    Especially for tiny details, such as holes, short lines or dots, this is very important in my experience: that is the first place where bonding will fail.

×
×
  • Create New...