neotko

This is for the users that want to get their Zebra Stripes removed (almost completely) What does it do It fixes the Zebra stripes from the stepper motors that UMO, UM2 and UM3 use (as it today 6 August 2017) Left is WITH the TL-Smooth, Right is WITHOUT. Printed on the same day, same spool, same gcode Before and After installing them. The print test used is the Curve Test by @Valcrow https://www.youmagine.com/designs/curve-test-print The issue is documented on this post https://ultimaker.com/en/community/16357-why-does-my-print-have-tiny-zebra-stripes On that post you will find that one way to fix it is by doing soldering on the SMD board, a delicate endeavor not for everyone. So a month ago googling I found this TL-Smooth chips that are being sold on a aliexpress shop, bough them and test them. So far it's been a week since I installed them and I don't see any issue after using the printer that has them for the whole week. Does that mean that is safe? No. Any mod ofc might have side effects, and since isn't a UM product, who knows. But IMHO it works nice and doesn't have side effects (that I can see so far). If I find any I'll update this post ofc. The guy that found that this kind if circuits fixes the Decay-something that makes the problem, was this guy http://www.engineerination.com/2015/02/drv8825-missing-steps.html The china guys sell it without mention the guy that made all the hard work, but also he never did sell them or offer them, so is a middle ground grey area, where the users benefit from this knowhow. Ofc the little boards are around 8€ for 4 steppers, so it's quite cheap, and they send it quite nicely (except the cables that they sent me wrong like they show on their aliexpress page, so at least they warn about it). So... What you get when it arrives You get 4 plastic thingys, to put the chip in, they seem to compress with heat, but I didn't tested that since doesn't seem important (or is it?) First thing to do, is to check the wire cables. The seller warns about this, and it's important to do it right. That's why I made this simple video showing the process of rearranging the cables and, ofc, twisting the pairs to avoid electrical noise. How to install them Turn off the machine (doh!) Remove the Power Plug Now, without the power cable in. Turn ON and OFF the machine. This will discharge the capacitators, so there's no residual power on the Board. Turn the printer so you can access the board and remove the cover Now, one by one, unplug the motor from the board, and plug them on the TL-Smooth. To remove the cable, use your nails from the corners and push, don't pull the cables to avoid damaging them: Do it on all your steppers, close the cover, and arrange them so they don't disturb you. Warning I take zero responsibility if anyone gets his board damaged, or if their version doesn't work. I don't sell nor promote this stuff. I just found it, bought it and share my results. Important Warning Triple check the cables are right 'in' and they don't go out when pulling them (gently). If that happens you might have them wrong when you reinserted them. Update Last week found an important issue with torque with this stuff on. I need to debug it when I get time but basically if your printer is at the torque limit (for x/y/e) this will affect the prints. Specially the extruder since it's more sensitive to this. I'll test other TL-Smooth version but probably will just move to a Duet3D board and forget about um2 boards.

I don't use mm/min but I mean the values on the tap. Umm. Did you try the same on 3.X and prints better? Cos if that's the case it could be easy to run a gcode check to see where the issue is.

Add. Can't edit the main post due max words So I add it here. for the "In theory" main part. Forgot to add that, the theory in my mind when doing this, was to keep the same mm3/s of extrusion, that's why I bump up the speed, but also because the faster you do the moves the easier is to have a 'sand' effect. So that's why it could be interesting to have a proper math formula behind to keep the extrusion ratio at the same level of the normal layers when doing this, to avoid changes on the extrusion flow that could lead into unwanted blops, extrusion errors, etc. Also that's why when using 0.1 (extreme and just for visual testing) it gives wrong perimeters and bad infill, but ofc, it gives the most smooth top layers you could want (mind the print errors due used on a process on s3d 4.0 without being able to split the perimeters from the toplayer, because 4.0 has a bug that breaks processes without perimeters).

Got this idea a few days ago, since now Cura has a Neosanding 1.0 (aka Ironing) feature, I though it could be fun to do something along those lines, but without the speed lost of re-sanding the printed layer. So The Theory When using a 0.4 Nozzle (like I do most of the time) If you set the nozzle at a fake extrusion like 0.24 (for example) then, the nozzle path will 'hit' the printed part with half it's tip, scratching (sanding) the just printed part. The practice It does work, but not as 'sanded' as the 1.0 neosanding, ofc in exchange you do all in one pass, saving time. So where's the advantage? The top layers do look better, because you make much thinner lines, hiding the print lines better (on top layers ofc). But to do so, 0.24, means that the nozzle will have to do more travel moves to print the same area as a 0.4, that's why I set the print speed to almost double of the real speed. This should decrease the quality, but because you are 'sanding' the printed part, and doing thinner extrusions, the result it's quite nice. Mind that my top layers are done at 50mm/s. So when I use 0.24 for the sanding2.0 I set that at 100mm/s (for toplayers, not outlines ofc). As you can see on the image the result it's less soft than sanding 1.0, but also, mind that this is the worst angle to see the effect, the texture to the hand is smoother and the normal view is quite matte. The advantage ofc is that I don't lost any print time. The Extreme test 0.24 with a 0.4 nozzle is as low I can go without causing extrusion errors, but I though, well if you use more passes, it will sand even more, so let's test 0.1 with a 0.4 nozzle. (remember I'm talking not about layer height, but extrusion width). So when you do 0.1, the result it's quite amazing, mind that this was a test I just done 2h ago, there's stuff that can be improved, like for example separating the perimeters so they are printed at 0.4 instead of 0.1 So that's it. How to do it First you need to select the last top layers of your print. In this case I used the 4.0 split helper, it helps but isn't all that it could be. Mind that the numbers of the screeshot ain't the ones on the FFF files I uploaded (download at the end of the post). So, you split the process on as many parts as top layers you want to make them 'beauty'. Then, you group the 'neosanding2.0' layers that need the effect and adjust this parameters This, ofc depends ENTIRELY on your print speed. I adjusted the outlines to match the ones I use on the normal process, to avoid loosing quality on the outside, but keeping a very high speed for the last layer solid infill. And that's all. This is how much passes does with 0.1 (mind that one square doesn't use the sanding to have a fast check vs) So is this better than Neosanding 1.0 ? Not really, it's just faster. Weakpoints Since now you relay on the nozzle to sand the previus printed part, the gloss to matte effect changes if the foot print is small or big. If the area is big, it will look more matte, just because there's enough time for the filament to cooldown (cold enough to get sanded). So if the foot print of the toplayer is very small, the filament would not be cool enough to get sanded and will keep a half-gloss-matte brightness. To get a more matte effect, slowdown the 100mm/s speed to something like 75mm/s, that should give you a more matte finish, but will take more time to print than just a non-neosanding2.0 layers. Files for test and check This includes the FFF files to check how it's done https://www.youmagine.com/designs/neosanding-2-0-idea-and-concept WARNING - The starting/ending Gcode it's the one I use on my machines, IT WILL 100% make the hotend crush with the bedclips, YOU must change it to your machine. If you didn't read this, sorry it's your problem. Also filament size, settings of the normal process, all, it's customized to my UMO+ with 1.75mm, Gudo ZGE Direct Drive, UM2 hotends 1.75 and custom fans. So EVERY setting will need to be readjusted to your system. Disclaimer: I share this so users with basic know-how can understand the concept, I take ZERO responsibility if you didn't read this before using the FFF on your printer.

Do your tool change settings match your retract/speed ? I ask this because if you use more than 1 process, S3D goes bananas and make weird stuff if the tool change doesn't match the same speed and distance of the normal setup.... Apart from that, I seen some twitters of people having that issue on new version, but not me. (I found quite a few bugs but not on that areas)

On S3D you will need to manually select the material on the profile. Cura does work best for dual prints on um3, specially because it uses many tricks to fit the printhead to the slice. S3D is imo a far superior slicer for single extrusion, but cura has an automatic heat/cold calculator that allows to avoid wait/heat/cold when switching from core 1 to core 2. So, even when the slice is better, it will take more time to print. Settings wise, if you are a experienced user of S3D it will be really easy to use the same settings that UM uses on Cura. About the long retractions l, that's the values Cura uses. They are long, they can grind the filament when outside Cura parameters (um2/um3 feeder has some limitations for feed/retract amounts, that S3D doesn't take into account). Ofc Cura is made for um3, but that doesn't matter that s3d won't work, you will just need to carefully select your options and get use to. S3D profile, just like um2 one, is quite bad IMO. I made some long ago but since I don't use the um3 I stopped developing it. So. First test the same object in Cura, just to check the machine is ok. The. Move to S3D double checking all parameters.

Again. Acceleration is a firmware/marlin/gcode setting./command What could be happening is that they ain't sending the correct gcode to change print/travel setting and they just adjust print. You could see that by opening the gcode and reading how they do the M205 etcetc to see if it does what you want to do. Doing some google seems the issue is marked for 2.7 https://github.com/Ultimaker/Cura/issues/2032

Why would travel acceleration have any effect on ringing? As soon it stops at jerk speed and resumes printing, the time from recovering the extrusion (since in theory when moving form a to b it should have retract) then the vibrations from the travel-non-print move shouldn't affect the print. Disable retractions completely and then this changes ofc. Print speed Print acceleration and jerk are the ones responsible for the ringing. Ofc the real responsable is the thin shafts, the heavy printhead, the bowden shaking, the motors noise, the rigidity of the frame, the viscosity of the material, how fast/slow colds... But that's not what you asked. Short answer. Travel speed for non print moves, doesn't affect ringing unless you don't use retractions. If you want to learn about ringing and tests you can check this very old, post https://ultimaker.com/en/community/3532-prevent-ringing-wobbly-surface-after-sharp-corners Also if you remove ringing, lowering jerk to a very low speed you will have side effects. Like you can read here https://printedsolid.com/blogs/news/a-solid-foundation-for-high-quality-corners

I did see this banding on a print a few months ago and the guy told me later than the fix for that banding was the belt tension... Maybe you can just ask him Edit forgot to add the link

Basic and most important step before plugging the TL-Smooth things The rest is just very easy, remove the motor cable, plug it on the P1 and plug the extended cable from the P2 to the board. Also a very visual guide for anyone that ever wondered how to reverse the motor cable pairs Enjoy the music It's 02.00 of the night and I was really tired to find a music XD

On Catohagen tests he changed manually the speed of the sanding (can't be changed from Cura atm) and he got better results. Originally when I made the idea I suggested 100mm/s to avoid overheating too much the surface, but Cura atm uses a lower speed. Ofc it's experimental it needs more experiments. Also check this nice idea Catohagen posted. If we where able to cut the sanding to force a design it could make top layers like the velocitypainting but.. NeovelocityIroning XD or something alike. Check this cool result

Check this thread. Many options contemplated https://ultimaker.com/en/community/22393-um2-socks-well-it-turned-out-to-look-more-like-spandex

Since you don't have a fancap to redirect the air, I would suggest you use one (printed on greentec-platec or a hightemp material using some high temp aluminum tape can make a fancap that last for a LOOONG time). Then, to isolate even more the air hitting the heater blocks, I use the E3D silicon socks, and cut with an exacto some of the thing, and along the fancap hole and the system, all stays put and doesn't move. Ofc if you take into account the sock design, you could make a fancap that also holds the sock better, but so far no issues. Ofc, they don't last forever, more like 4-5 months tops, and with sticky materials they erode quite fast. But even so, they give me a much more controlled heat on the blocks, specially since my fans give much more air than others.

I would love to but very very little time lately. Heat sinks are just very easy and they should have a ready sticker. You just place it over the stepper chip and done. I bought them year+ ago on a shop in spain. The TL-Smooth are easy to install they come with a cable that also fits the um2 board, so you only need to watch that the cables are in the correct order (they warn that on their aliexpress) and plug the P2 to the extender they sent and plug the motor to the P1 side (this is on the diagram they have on the aliexpress that probably was ripoff from the creator of the idea). So. Complex? Not at all. They even include a plastic to hold the thigy so they don't touch an conductive part. I think they are heat sinking plastics for what I seen on their photos. I will try my best to do a very short video of the install this weekend. But it's really straightforward (except the part to unclip the cables if they are wrong, and twist the pairs).

Maybe is the compiler version? I use a old one for umo+ and a newer one for um2 firmwares... Will check what version I use later on

Indeed on Cura it works since they added the per object settings. For s3d they broke it on 4.0. Visually I prefer s3d over cura and for the prints I do Cura can't manage efficiently the multiple settings I need nor the surface finish. But cura is getting better and better very fast.

Curious. I got an email for this tag.

Yea true me too. No emails for @ tagged name replies or PM messages.

You mentioned this in your first post too. Where do I get this? I must be blind. :-( Is it possible to take a greater PSU or will then break/short/burn the Mainboard? Thanks for the tinkergnome hint - i'll contact him. On one of my printers I have a 280W meanwell gst-280 to use it on my umo+. It's been working perfectly for 2 years+ now. Ofc it would requiere some connector solder/cut/plug to connect it to the um 2.1.X board. https://github.com/AmedeeBulle/Marlin/tree/Marlin_UMO_Unified

On @amedee firmware builder you have a link to the whole Marlin build he made. You can complile his version from that. Two 35W are more of what the psu can manage, but @tinkergnome managed to use them for the mark2, so maybe he has some tips about how they do the heat sequence in order to make it work. Or you can just go to his firmware for mark2 and check it to see his recipe.

It does work indeed! Poated the whole before/after on my twitter account. Clearly this needs more tests, some thermal photos on the steppers after a few hours use (because they seem to get hotter). So if you plan to use this, buy also some stepper chips heat sinks (hyper cheap and easy to install). So yeah, a nice 8€ fix. Ofc there's an almost invisible remaining zebra on some angles, but to remove that most probably one would need to move to a better stepper chip like duet3d, replicape or wait a few years until UM makes a new board (because boards don't pop out of thin airs and that needs time to be developed I suppose). Warning I take zero responsibility if anyone gets his board damaged, or if their version doesn't work. I don't sell nor promote this stuff. I just found it, bought it and share my results.

Interesting enough this doesn't work anymore on 4.0 Simplify3d. If the process doesn't have at least 1 perimeter. I just hope is a bug and not automagic crap

For a fancap, even if it resists 100C try to cover the hottest areas with hihgemp aluminium tape. It will give you much more months of use, specially to avoid deformation due fan weight or high speed moves. A 150C alu tape is cheap and works fantastically. For Pla-tec bed temp is 55C. For greentec is important to use hairspray to avoid adhesion issues. pla and greentec are almost the same but platec is slightly more rigid and greentec is slightly more flexible.

For Faberdashery I used this https://www.youmagine.com/designs/faberdashery-spool-er Ofc was a pain to respool every single one, but for me was worth to avoid tangles, or being dangerously close to become a Loom weaver when a big tangle happen. Faber is great, but they are so behind other filament sellers that I stopped buying to them 6minths ago. Not worth the trouble and 90% of the colors can be found on Eumakers. Ofc faber filament is more precise but the cost Kilo is yoo high to justify the troubles.

I had lots of x/y slips on my umo+ until I changed the pulleys to gt2. For the first 5months of use on my first umo+ I had (2+ years ago) it had troubles and the reseller imakr did sent me pulleys and short belts to replace, but the issue never disappeared completely. With the gt2 it never happened again. About the planner, no idea. Could be nice to know Ultimaker opinion about the planner and if that bugs @gr5 says are documented or fixed on any Marlin release or if just noone plan to update old machine firmwares.