Jump to content

gr5

Moderator
  • Posts

    17,520
  • Joined

  • Last visited

  • Days Won

    373

Everything posted by gr5

  1. 10 months late @Mangee but thanks for helping out!
  2. @acp - I have learned important information from Ultimaker employees and will be updating my notes on my unbricking page soon!
  3. The point of the project file is so that we would have the exact same settings (adaptive layers versus not, bad model versus good, .01 layer versus whatever). And also same placement of the model in the same location (in case you rotated it or moved it) and so on. Also there is "fixed" and there is truly fixed. Did you create the model yourself? If so what CAD software? Anyway this is a model problem and not something Cura team will fix. You are correct about that. It would be good though for you to know how to avoid this in the future. This particular issue is impossible in most CAD software but is common in a few such as blender and sketchup. If you use one of those I have suggestions on how to fix next time.
  4. Oh. I hope your printer is still "square". A common problem in bad shipping is that the sides of the printer are no longer perpendicular and the gantry is now "banana" and impossible to level as two diagonal corners of the gantry are high and the other two are low. Check to make sure everything visually looks square. And push the glass up until it just touches the nozzle but with a gap and with the other hand push the head around to see if the glass and the nozzle are parallel.
  5. I still think you should watch the video (again?). The most likely issue is "squish" but it could be any of the 4 or so issues I list. Lifted corners is... very very common when you are still learning how to do amazing bed adhesion. I explain what causes it and how to fix it. When you get good bed adhesion you will be asking a different question: how do i get these parts off the print bed?
  6. You call it "support blocker" but I think the official name is "modifier mesh". When you first created it, perhaps it was called a support blocker but as soon as you changed it to do something else it became a modifier mesh. I think.
  7. Stopping for 30 seconds is bad as the nozzle is sitting on the print somewhere during this time and melting it and also the filament expands (because it's just sitting inside the nozzle heating up) and squirts out a bit and you get this... button/divot thing. Also there's thousands of other products that might upgrade and reboot windows at any random moment.
  8. Not sure but I think you either have "spiralize" turned on or infill turned off. Most likely spiralize. If you click on the star next to the profile name and click all the tabs (usually just extruder 1) you can see what you have overridden versus the default. Typically they are italic if you changed something.
  9. Okay so one in 256 errors (if errors are on average at least two bad bits in the data) can make it through. I vaguely remember watching printrun print through USB and seeing that there were retransmits extremely often - like every second. My memory could be wrong - this is probably 10 years ago. I don't even know if retransmits are supported in Marlin and printrun. But the idea somehow got into my brain and I vaguely think I saw it happening - a lot: scary amounts of retransmits and I thought "1 in 256 of those errors are probably getting through to the printer". If 90% of the errors are only 1 bit then none of those 90% but 1 in 256 of the remaining errors should survive to be "printed". And looking at the hardware on the UM2 and UMO, the USB connector is mounted right on the circuit board and data only travels from that connector through short traces to some "usb chip". So I was wrong about the ribbon cable as well (that was SD card printing) So I don't have a good explanation for if/why "USB is unreliable". I do remember windows. F**ing windows. Screen saver bit me. Powering down hard drive bit me. windows update bit me. Being afraid to use the computer while it was printing was the last straw. Even now in 2021 windows 10 sometimes the computer is so so so slow that everything seems locked up for 30 seconds - this only happens once per week but still - what will that do for printing over usb? I mostly use Ubuntu but I have a lot of computers and most of them have windows. Ultimaker customers commonly do multi day prints. A 24 hour print is not at all unusual. Ultimaker doesn't want to explain to all customers that they need to disable windows updates, disable screen saver, disable power saver, and there's more - sometimes other auto-update software reboots the computer. Maybe it will be adobe? Or edge? Who knows what random software on the computer will cause it to reboot. So I think Ultimaker has this attitude that printing over USB is "unreliable" and I'm not 100% sure where that comes from. It's not that the Cura programmers don't care - it's that they have other priorities and are probably always months behind on the features being asked for.
  10. All words, all statements are a simplification of reality. It's impossible to be 100% thorough on any statement. So I was simplifying above. A checksum doesn't catch every combination of errors. I think Marlin uses just a one character checksum and it's not even 256 states - it's something like a number 0 through F maybe? So it's only 4 bits. So 1 out of 16 bad corrupted gcodes (I believe) make it through USB and get "printed" by Marlin. [correction - it's a 256 bit checksum as stated in the next post] But I was seeing so many errors that this meant a few gcodes per print job were bad. Usually it didn't matter as it would just skip that gcode but sometimes it would retract the extruder by 100mm (and then recover on the next gcode) or move outside the printable area (and fail) and so on. So *some* printers *occasionally* had very unreliable printing through USB and Ultimaker chose to stop supporting it. I think it had to do with crosstalk maybe? On UMO, UM2 the USB signals go through 2 flat ribbon cables that touch. Or an overloaded processor maybe? The trouble with intermittent errors (one error per hour on only some printers) is that they are hard to debug and I think Ultimaker just lost interest and decided to stop supporting it in all their printers. Even though it was very reliable on most printers. Just that the number of printers where it was unreliable was probably too much to support. There are lots of volunteers but mostly Cura is created and improved and maintained by Ultimaker and there are tons of other priorities to add features for their latest and future products. It's true that not all printers use Marlin. But most do. Something like 99% and that number is dropping but all those cheap chinese printers use Marlin and Ultimakers use Marlin (sort of) so that's what Cura is going to support the best. Having said all that, Cura does support USB printing - it just isn't as good as Octoprint, printrun, etc. and it could be 1000X better (maybe video-over-usb and everything that Ultimaker is doing over Ethernet in their latest printers - hell you can do tcp/ip over usb).
  11. I'm going to give you a partial answer. Hopefully someone else can give you a better answer. So every cura setting has a variable that can be inserted into the start g-code script. Including initial layer height. So you can do something like this: G1 X0.1 Y20 Z{initial_layer_height} (I think that works exactly as is - all lower case with underscores but I'm not 100% certain) So that solves your issue of hitting the glass. however you need to adjust the extruder amount - otherwise it's only going to extrude enough for a 0.3 high, 0.4 wide bead of filament. To do that you need to multiply: {initial_layer_height}*{initial_layer_line_width}*1.2345 where 1.2345 is a constant depending on filament diameter and length of line printed. I just pulled that number out of thin air as I'm lazy HOWEVER, you can't do math in these (or can you? 90% sure you can't. @GregValiant what you think of all this?). So I don't know how you would do that. A good compromise would be to leave the layer height alone, and double the extrusion amount. A 0.4mm nozzle can print a 0.8mm line usually just fine. Especially at these extra low printing speeds. And then even if you had a 1.2mm nozzle you should get enough filament to come out to hopefully stick to the glass. Another solution is to have multiple "printers" each with a fixed nozzle size and you switch between "machines" in Cura when you switch nozzles. This is not an issue for Ultimaker printers because they don't do a purge line (I really like the idea of purge lines by the way). I wish purge line was a feature in Cura instead of hidden in start g-code. So that's another solution: get rid of purge lines and stick with purging in the corner in mid-air followed by a retraction. I mean the skirt/brim option takes care of the purge line concept for he most part.
  12. If only it were always so easy that you just have to restart Cura! I mean usually it is that easy but sometimes you have to install the USB-serial driver. Sometimes (one windows) the COM port is a very large number and Cura can't seem to find it.
  13. I agree with everything. "best printing software" is a bold claim. Maybe best slicer? Cura really is the best or one of the best slicers I think. But it doesn't print over USB very well at all. I think the history is that it worked well on the Ultimaker Original but was unreliable enough on the UM2 that Ultimaker stopped supporting it. That was a long time ago - maybe 10 years now? They never looked back. They felt USB was unreliable. I think it is unreliable when used with an Arduino style processor (AVR processor). But USB can be made to be quite reliable. But not with such a wimpy computer that is already overloaded. You might want to look into Klipper which adds a raspberry pi and has custom software on the raspberry pi and also replaces the Marlin firmware on the AVR computer and uses the USB to do really impressive live communications (both ways) with protocols that are very resistant to errors. Or just octoprint on raspberry pi. That's a really nice interface and works great.
  14. @fbrc8-erin - have you ever seen this where you have to push on the cores extra hard for them to be seen? (this guy is not in USA)
  15. Do your cores have gold contacts? Or are they covered by silvery solder? If solder - get new cores. If gold, try cleaning them a little with isopropyl alcohol. Clean the contacts in the head also. Make sure not to leave any fuzzy cotton behind. Did you try changing the PCB in the print head? Maybe something is loose in there. Bad solder joint or bad ground or something like that. I've never heard anyone with your issue so it's rare and I suspect you may just have to slowly replace everything until it goes away.
  16. If the first layer is lifting then this has nothing to do with cura. Here's a video I made about this - it's long - sorry - but I had to pack in a ton of valuable information:
  17. To clarify about "slicing anyway". After you slice you click onto PREVIEW mode and scroll through the layers to see if it is doing what you want. Always look at your part in PREVIEW before printing it. Always. Always.
  18. Did you try slicing it anyway? Sometimes there will be no issues. The things to look for are holes filled in - areas filled in that should be open or things missing. It's usually pretty obvious. By the way, I have never had openSCAD create a non-manifold object - openSCAD is great. I use it all the time. By the way you should rotate those parts onto their backs I think. Large flat area rotated down (you can rotate the part in cura easily - I suspect you know this). Other tools to make a part manifold (there are dozens): netfabb free repair service is here (you have to create a free account first): https://service.netfabb.com/login.php Here's another service - drag and drop mesh repair service: https://3d-print.jomatik.de/en/index.php Some people recommend tinkercad - it's free - you import your stl into tinkercad and then export it to a new stl and tinkercad will fix many issues.
  19. Ooh! If you do please post what you learn and also direct message me as I get so many forum notifications I'm bound to miss your postings. So I would really appreciate some DM message simply saying "hi gr5, I posted something you might want to read". I can then easily and quickly look up your most recent posts by simply clicking on your name (just as you can look up mine - don't bother - nothing worth reading, lol).
  20. The top of the core has a clear plastic part. When it's all the way in you can hear a click. Does it click? Something seems to be blocking your cores from going all the way in. I'd do some more visual inspecting. Also maybe it's the connector on the top of the print head. You can remove the rear 2 screws on the top - they are very very long screws. Nothing will fall apart as long as you keep the other 2 front screws untouched. Then the top opens up - a little pannel on the back half of the top and inside push down on the cable connector. It's common for this to get pulled upwards and eventually lose connection.
  21. Yes. It's both complicated to learn how to do this and buys you potentially nothing and may actually make quality worse. It also creates MORE work for your printer firmware which for 99% of printers is bad but I have to admit there are very special cases (and this will happen more in the future) where that's a good thing. I guess the only case I can think of is: Klipper firmware That's partly why I asked about your printer. Klipper is really amazing. Hopefully it's the future. But I suspect less than 0.01% of printers are using it. I haven't tried it. It typically runs on a powerful multi core with floating point computer (e.g. a raspberry pi or beagle bone black) combined with a slower computer such as an arduino. It handles G02 (arcs) natively. It may not handle them well at this time - I don't truly know - but it probably already performs better than Marlin on arcs. It's the only firmware I can find that smoothly clocks all axes independently (no bresnehan crap like in redeem firmware. No fixed rate interrupt crap like in Marlin). I also asked about your printer because if you got a $200 Chinese printer but are using $4000/year cad software there might be a mismatch in quality expectations between the 2 products. From an engineering point of view this is almost a meaningless point but I tried to make a point anyway: yes STEP is a better format but most 3D printers are running Marlin firmware which, although it's pretty amazing, it has issues and is kind of limited on how well it can handle G02 gcodes.
  22. https://github.com/TinkerGnome/Ultimaker2Marlin/releases
  23. What kind of printer did you get? How much did you pay for it? The most common technology right now is you convert those beautiful solidworks curves into triangles. Sorry. But it works pretty well. And you want a pretty low resolution because if you have too many points on the curve your printer (with it's wimpy arduino style processor) will probably freak out and stop and stutter as it tries to process too many points in too small a distance (too short of time). So paradoxically, if you want good looking curves on the final print you should have fewer points. There are some tricks to put curves in the gcode but I recommend you DO NOT do that.
×
×
  • Create New...