When 3D Printer LINEAR RAILS upgrade MAKE SENSE? (vs POM wheels)
ฝัง
- เผยแพร่เมื่อ 26 เม.ย. 2024
- Get professionally LASER CUT parts for your projects at PCBWay (5$ OFF for your first order) - pcbway.com/g/584K9t
STEP file of this project - bit.ly/HeatBedBracketLK5
STL files - bit.ly/BracketSTL
👉🏻 In this video I am testing how much print quality improvement you can expect from 3D printers' Y-axis linear rail upgrade. I designed it for the LK5 PRO 3D printer.
🧡 Support my work at PATREON - / diyperspective
📋 RELATED ITEMS TO THE VIDEO (Amazon Affiliate)
- MGN12H linear rails s.click.aliexpress.com/e/_DeN...
3D printer:
- Longer LK5 Pro amzn.to/3chMvud
Filaments:
- ANYCUBIC PLA amzn.to/3W3MmwE
- filalab PLA www.filalab.lt/home/pla-1kg.html
📷 VIDEO SHOT ON:
- Fuji X-S10 amzn.to/3Zar9BW
📢 OTHER MENTIONED THINGS:
- LK5 Linear Rail MOD www.printables.com/model/3709...
- Gcode analyser www.gcodeanalyser.com/
- TT ACC test prints (modified) www.printables.com/model/3709...
- GJB Test Cube www.printables.com/model/2931...
- Low Poly Bunny www.printables.com/model/2888...
🕗 TIMESTAMPS:
00:00 - Why?
00:43 - Reducing the weight
01:53 - Important notes before tests
02:54 - Test prints
03:28 - Upgrade TLDW
04:01 - Smooth & Rigid!
04:12 - Analysis
06:34 - Summary
07:23 - Your thoughts?
🔗 YOU CAN FOLLOW ME:
Twitter: / diyperspective
Instagram: / diyperspective
#3dprinting #3dprinter #upgrade - วิทยาศาสตร์และเทคโนโลยี
![[TH]FFWS SEA Spring 2024 Knockout Stage - Day 10](http://i.ytimg.com/vi/4QaEryUGWQw/mqdefault.jpg)
Excellent video and testing!
Would be interesting to see if the same tests could be done on a laser cut G10 or Fr4 bracket. Its essentially fibreglass and used in PCBs and has a fraction of the weight. Or try carbon fibre.
That is actually a great point, with enough thickness it could be a possibility. :)
@@DIYPERSPECTIVE I would I also recommend the same thing, FR4 is great, the reduced weigh is enormous, I would recommend 1,5mm
There's is no way 1.5mm bracket from FR4 would be enough. You can easily bend even smaller pcbs at that thickness. The aluminum one weights ~300g which is quite low for the insane rigidity it offers. But the biggest amount of mass is at the heating plate + PEI and magnetic base. That's the real problem of the weight.
@@DIYPERSPECTIVE yeah, it's not pretty rigid, but I still get a fair amount of flatness with the clips on the side
@@DIYPERSPECTIVE I took off my heavy Alum plate with older Garolite and used a new G10 Garolite Bed with NO support Plate. Seems to work fine. Clean with Alcohol between prints, sticks well and releases within 5 min of cooling off.
Thank you for the nice and clear explanation!
That was very informative, lots of work went in to this tnx!
Very informative video 🎉
Thank you for enabling CC on the video. My ear wasn't tuning to a specific word, I was so confused 😅 It was "Values" I was hearing "Valleys".. I just woke up though.. Coffee has not kicked in yet.
I am excited to see you shed more light on Klipper for people who have yet to experience it.
I recently upgraded my e3v2 to klipper and it’s great(once I know what I’m doing)
Good day, Sir-- What length are the MGN12H linear rails for the LK5 Pro? Thank you in advance and thank you for your vidoes!
I have a Duplicator i3, for me a massive improvement was designing a center mount that goes on the screws for the springs to mount firmly on. This prevents any kind of shifting with will both offset the bed and also help stiffen the spring. Then the next major improvement was adding wool around the edges to act as thermal insulation but also to dampen the springs.
*i also added a teflon sheet on top of the bottom frame the bed goes on, both to reflect the heat back up to the bed and as a side effect also help even distribution of heat on the bed
I have not had to level the machine after over 2,000+ hours of printing, or about a year
Side thought; I don't think i've ever check a slicer for X/Y specific jerk/acceleration settings, that way you could run more better/different X speeds and Y speeds by tuning each independently.
Just found this video and will go see the others. I own a 5 year old CR 10S Pro with lots of upgrades. I put in Z axis Synch Belt, Z Axis Struts, Garolite Bed, Micro Swiss Hot end, Micro Swiss DD, BL Touch, OctoPrint in Pi 4B, last Week Klipper via OctoKlipper & OctoDash. I design with Fusion 360 and slice with Simply 3D. A friend who has the twin to my setup just installed Micro Swiss NG DD, so I ordered one too. Then I heard back he did not see any improvements. I recalibrated again and reprinted the 20mm Cube, X and Z were near perfect but Y had ghosting and fat ends, I contacted my 3D Parts Supplier and canceled the NG and ordered Y Axis Linear rails. They also agreed that was the weaker link But they did suggest a Pancake motor to replace the original Creality extruder motor in my MS DD. Thinner and lighter would help for much least that the whole NG DD. Then I see this video..... Seems like I am on the right track. Trying to adress issues with a budget it not easy. But Klipper was by far the single best upgrade I made. I had the Pi for OctoPrint already, the 7 inch screen was $50, Klipper is free, and now the linear rails precut and all the hardware was $125. This will be the most expensive upgrade yet. But i see the Y axis ringing and Ghosting so it made sense. I will go on and see how it turns out. need to consider the pressure calibration too. Thanks Dennis
Great video, thanks 👍
Is there a recent upgrade (not the bltouch) to get rid of the springs so you won't have to manual adjust the bed no more? This was a real pain on my ender.
I am looking at buying a new printer to compliment my ender 3 s1 pro that I have klipper for. I realize I need a bigger print volume and would like it to have linear rails for all axis. I think for me its more about adjustment and no more flat spots over the wheels. so benefits of performance would for me be also about stability of the bed and rails over time.
Thoughts?
Hi ... i have done a Y-Railupgrade to my CR-10 and have y-layer shifts. Could you tell me your max movement/print speed?
I ordered the parts for my linear rail conversions just before I found your video. I'm not so sure Longer LK5 Pro can even be called a Longer anymore. I look forward to seeing what else you do to this printer. I've done the dual Z axis upgrade, the SKR3 board replacement, BTT SFS, Micro Swiss Direct Drive NG plus tensioners for both x and y.
That is true, when you make so many mods to a printer it does not really deserve the original name.
I am going full linear rail on my Neptune 4 Max. It’ll be interesting to see if it can hold the print quality at higher speeds.
Have you ever thought about using a counter weight when the bed moves in one direction and weight moves int eh opsite cancling out the swing
Once again, you gave us an amazing information and content. You're not going to make too many friends with other 3D printer YTers :) Talk about myth busting!! There are other advantages to linear rails even if they don't necessarily translate in print quality. Or more precisely within a hobby setting. The benefits to a less maintenance system could be of value in a print farm or similar setting.
Looking at the dynamics of a bed, it is somewhat easy to understand why those changes doesn't make a large difference. The rotational forces are front-back, not side-to-side (like you showed with your dial). In this fashion, the wheels, rails or simple rod bearings all do a good enough job to keep the deflection small enough to almost make no effect on the print.
And I'm surprised why everyone is fighting to make the bed lighter with spending $$$ as opposed to just switching to a larger motor and belt system - again, all assuming you need to make a change there in the first place.
Anyway - excellent content and great presentation. Job well done!
Thanks! I am not trying to make enemies, just trying to provide better-controlled testing results. ;D Yeah, even cheap linear rails are more than enough for upgrades like these. I even specifically ordered multiple just to check the quality, kinda YOLO it with two orders and all of them had no binding and wasn't too loose. Yeah, less maintenance and especially no wear in a long run. Heat beds with POM wheels just give me anxiety... It is such a flawed design from an engineering standpoint. X axis and Z are fine, but Y... They ride only on one side and wear way more quicker the heavier the bed is. I keep constantly checking them... LOL
So yes, the main wobbles come from another direction, I perfectly understand that. It was just an easier way to show that it is not a rigid system overall with no deeper thought. However because it is so wobbly to the sides, even the light X-axis could affect it with the fast movements. But turns out that when the heat bed weighs so much it is too much to handle for POM wheels with a design like that. Another thing that could affect the results is the rolling resistance. With linear rails, the bed moves like on ice skates, with POM wheels not so much. But it is hard to say how much that could affect the results.
The steppers on that 3d printer were already ran at 2A RMS at stock. So with TMC2209, it is the limit. (EDIT: I checked today and I have no idea why I said 2A RMS stock because it was set to 0.78A RMS, so yeah there is quite a lot of wiggle room). I have 48mm 2A motors for further testing and will order a 9mm belt that with some modifications should fit. But the lighter the bed the less the whole frame shakes. And when the bed weighs 2.2kg the shaking is quite insane for those not that high speeds/acc/jerk values. It is pretty annoying if you have it like on a table without any dampeners. So yeah. ;D
@@DIYPERSPECTIVE Ah good points about the max motor size and bed weight- very true. Guess that's where corexy have the advantage. I just can't see building many corexy for my farm to gain that much in speed and maintain extreme surface finish. Although the "Rook" homemade printer has some promise.
What I like about your channel is debunking mainstream ideas. Even I have a hard time accepting the results!
Maybe after P1P release we will see a bigger shift into corexy 3d printers from mainstream manufacturers. It is superior design after all. I am tired of seeing Ender / Prusa clones...
I love the design idea of the Rook so much. Hopefully it will reach a more complete state. Would love to make one and make some print quality comparisons.
Yeah.. there are a lot anecdotal evidence or "from experience" without any actual testing and detailed comparison basically "just trust me bro" ;D When there are so much variables unless you properly evaluate and limit them, you are only speculating. Good luck that I actually love testing and triple checking and even retesting if the results are not what most of us are used to hear. :)
It would be interesting to see how print quality holds up as the height of the printed object increases. Once you start printing taller objects with a large format bedslinger, it seems like you would reach a point where the object itself will start to flex during the bed's rapid acceleration / deceleration (depending on its shape, of course).
It depends on the shape of the object but yes you can't really print slim and tall objects on bed slingers with high acc/jerk values. I added this point to test in the future video.
You’d have to be really cranking, and by the time travel stops the model should be back to where it was.
The biggest issue I have is when printing TPU (about all I print) and getting higher up on the models allows a lot of flex to occur in thin walled parts
Hi, exactly the video I was looking for thanks for the great content.
The Kp3s Pro is shown in this video and I wondered what was your linear rail results for that printer ? As I tested mine with an adxl345 and the max recommended acceleration was higher with the rolling wheel bed than with the linear rail, the linear rail had half the recommended accel (1800mm/s^2). It would be really helpfull if you could share some of your results.
The problem with cantilever 3D printers in general is that you cant push acceleration/jerk too much as the X gantry will start to wobble from the fast Y-axis movement. Kingroon made mistake by using only MGN12 linear rail on Z-axis instead of the MGN15 or two MGN9 ones. I didn't make any proper testing yet so I cant tell.
6:34 Why is the Y on the X axis? lol Is that to get the depth of the letter done with the Y axis? That would kinda make some sense. But, I've always seen it the other way around.... Now I know though.
I have thought about what the springs that support the bed can do with ringing. Since springs that are part of accelerated mass that stops must go somware. A spring that compresses will rebound with about half the force or a bit less. Have you tried the silicone buds that is stiffer under the bed to see a difference?
If you tighten them a lot (at least those yellow ones) it is very hard to compress them more, even with hands. Another question is do they keep their tension after hundreds of heating and cooling cycles.
I have one set of silicone ones, but I never even tried it. For smaller heat beds I really doubt they are needed but for the big ones probably would make the most difference. Maybe I will test it after I finish with all the upgrades for this 3d printer.
@@DIYPERSPECTIVE Silicone is a pretty fair insulator. Better than steel I'm sure. Might be worth it for better heat distribution in the bed.
Kind of a related question I've been wondering about but haven't tested completely: Toothed idlers or not? "They" say that smooth idlers have larger bearings (good), but that high belt tension causes vertical artifacts.
This video explains what actually causes VFA - th-cam.com/video/-et5eMyLlUs/w-d-xo.html
I did klipper on my voxelab aquila and it's so much better. It's like night and day difference.
i have an lk4 pro and an lk5 pro. awesome printers. I'm definitely subbing to this channel. you seem like you know your shit plus we have the same printer so i can do the same upgrades with exact parts. awesome vid.
Awesome! Thanks for showing before and after! Too many gimmick upgrade videos out there. I’m just curious why you had two carriages on the left and 1 carriage on the right 3:48 ?
I am interested in doing this for my printer. Thanks!!!
So because it is a large heatbed minimum of 3 carriages are suggested for it to be more stable when moving back and forth. Maybe you could get away with only two, but I wouldn't recommend it unless it is a small heatbed.
@@DIYPERSPECTIVE Makes perfect sense. Last quick question. Do you see any problems with going on 4 Carriages? Two each side?
Thanks again! I’m a stickler for symmetry but only if it doesn’t hurt haha
4 would be even better, but it can be more difficult to align them so that they slide without any binding, plus extra cost.
@@DIYPERSPECTIVE Awesome you the man! Thanks!!!
I just found your channel, it's very good and I hope you keep going. I have two suggestions for a follow-up: 1) I wonder if you might be able to improve your Y-axis acceleration/jerk with a heavier-duty stepper motor. One crazy guy equipped his Ender 3 y-axis with not just a heavier stepper, but a dual-stepper system with a stepper on both sides of the axis. I think he got his acceleration up to 10,000 mm/s/s, maybe higher. 2) Would you see any improvement by removing the bed springs and mounting it rigidly? Teaching Tech has a video showing that springs aren't really needed anymore if you have BL-Touch bed mesh and some hand-tuning. It's possible that the springs are still helping a bit by vibration dampening though, so I'm interested to see if it makes things better or worse.
Glad you like it. I love doing it and have no plans to stop it.
1. Most fast printers usually run lower jerk and high accels values. I have a 48mm length 2A motor (run_current: 1.000 A) on the Y axis and I can quite easily run 8K accels with 8 SCV on this printer. I feel these are more than enough for such a big and heavy heatbed, the more you push the more you will impact the longevity and increase noise levels.
2. Yup, they are not needed if everything in your printer is assembled squarely. Currently, I am running with silicone spacers, but I haven't seen any difference compared to compressed yellow springs almost to the max. Basically, you want the most rigid possible setup, because that eliminates any wobble and provides the least ghosting. Then you only have to worry about the belt and the motor.
Do you a print profile for different materials for the LK5 Pro you would be willing to share? I have an LK5 Pro and want to get good profiles
I never had an acceptable print quality to my standards with this printer, so I only printed PLA.
Would be interesting to see a comparison of linear rails affects on small printers vs larger printers.
The smaller the printer the less difference you will see at the same acc/jerk values. But linear rails are not all about the quality of the print, it is also about longevity. The heavier the bed and higher the speeds the more wear will be on those POM nuts. And if these results were when they were properly tightened before test prints. I don't even want to see how they would look with the loose nuts. You would have to constantly check if they are tight enough, it would be a nightmare.
@@DIYPERSPECTIVE Im a CR10S5 owner. 500mm bed slinging around.
That's a big boy!
where can i find the n4plus cad model?
I know this is not a recent video, but since i'm planning to upgrade my LK5 PRO in the same way i'm asking you what is the length of the rails i need to upgrade both X and Y axis?
Thank you very much, this is such a great video ☺
Sorry, forgot the link www.printables.com/model/370968-brackets-for-lk5-pro-y-axis-linear-rail-upgrade
2x 450mm for the Y and 1x 350mm for the X. But you will have to drill and tap holes yourself for the X axis, plus cut a little bit of the metal tensioner to fit the X rail.
@@DIYPERSPECTIVE you’re the best, thank you 🙏
How much is a plate like that from PCBWay?
Btw, you can use the Arc Welder plugin to help the 8 bit board keep up with curved surfaces. Just play with the values and it'll be just fine at almost any speed. (Of course Klipper can do much better without breaking a sweat, but still.)
Yeah, but at this point, I am fully committed to seeing what brings this complete 3D printer overhaul.
@@DIYPERSPECTIVE Sure. Then Klipper. (Well, I put Klipper on my printers regardless of how they're functioning out of the box.)
After looking into the Marlin source code, arcs still fill up the command buffer because the firmware converts the arcs into linear segments internally. When I experimented with heavy stutters/pauses while printing cylinders at 200 mm/s, arc gcode didn't help at all. Admittedly I have a 32 bit board so maybe it's a different issue.
@@MiG82au The command buffer is between the gcode reader (e.g., the display module which has the memory card reader) and the mainboard (e.g., running marlin). Once it's on the MCU on the mainboard it has much higher bandwidth to the stepper drivers, so that shouldn't be a problem. If you have stutters even then I suspect the problem lies somewhere else.
(When I had stutters, before I switched to Klipper, it did help to switch to Arc welder.)
@@marcus3d I don't think that's correct, firstly because I've read numerous discussions about calculating steps which get put into the buffer for the motor drivers to execute and secondly because my SD card is on the main board, not the display.
Was the linear rail here only on the Y axis and not the X ? Curious if replacing the other axis would erase the ghosting ?
Yes, only Y. But the same all principles apply to X axis too. I made the acceleration test print yesterday with X axis rail with my custom design direct drive extruder head (which weights more) and results were without any ghosting. Linear rails are superior in every way for the print quality but if weight and speeds are on the lower end, POM wheels are just as good.
You are not comparing the sliding mechanism actually. The major change is that you actually made is the connection points. With the linear rails is wider vs the stock configuration, but you can achieve the same thing of you just get 2 more extrusions and mount them the same way as the rails, but with rollers .
The width has very little to do because the bed movement forces are not in those directions. It mostly helps with consistency on big prints. I don't understand when people will realize that POM wheels on aluminum extrusions are not a precision motion system. It is a non-rigid constantly fast-wearing system that needs constant adjustment. And not only all that but the Y axis literally rides and wears only on one side of the wheels which from an engineering standpoint it is a mindboggling flawed design. What amazes me the most is that companies by cheaping out convinced the masses that plastic wheels on v-slot profiles are okay. They are okay on slow dirt cheap machines, but not okay if you actually care about speed, quality and consistency. Don't take this personally or anything, that is just my take on this, cheers!
What is that white hot end shroud?
I wish I could remember the specific model but I only know that it is one of the Hero Me series on thingiverse.
Is the Z axis worth upgrading to linear rails at all or should we stick solely to X and Y?
I doubt it is worth it, unless you are pushing for high acc/jerk where the whole frame shakes like crazy. The only real reason that I see, is because the uneven wear of the wheels as they rotate so little and could develop flatter spots in a long run.
I think your bed springs may affect the results here. (as in rails might be even better?)
3D production machines with many linear rails and guides can be expected to be used for many purposes. surprise!
I'd love to see the differences with a well tuned setup to begin with (Klipper, input shaper) and accelerations of 5k-25k.
it's a bed slinger, with a glass bed you won't ever get over 2-3k out of the bed especially not on heavier prints, i have a Hemera directdrive with a copper volcano block (heavier than the alu, but also better) and i can't go over 6.8K accel on X, no matter wether it is Marlin or Klipper, also FYI Marlin also has input shaping but it's not with a wizard like Klipper has.
it still is better than needing a Pi tho
@@BH4x0r Glass indeed adds a lot of painful mass, but there's no way 2-3k is the limit. Ender 3s without glass can do 15k+ (some have done 25k+) on Y. If you're hitting a limit that low you probably have miscalibrated shaper or bad stepper driver settings like stealthchop.
Stock Ender 3 X can do roughly 100k. A heavier direct drive won't do that, but 30k or more should be achievable.
@@daliasprints9798 hmm interesting, i run mine on Marlin in Hybrid mode (switches to spreadCycle above a certain speed only)
so ye stealthChop under 80mm/s
never thought about that
edit: even with stealthchop disabled i lose Y steps: but keep in mind i don't up the current from the ORIGINAL settings
my motors are completely stock Creality motors
tried 6.8k on Y with and without shaper, and i lost steps either way, on spreadCycle at 850mA which already gets them hot, i should probably swap the 42-34's out with the 42-40 extruder motors (as i have 2 Creality style extruder motors that don't have a press fit gear and they can take more current and have more torque)
@@daliasprints9798 small update, since i still had 2 old 42-40 extruders laying around i swapped them 42-34's out with 42-40's and upped the current from 750mA to 1A, gonna test accel etc tomorrow
There's practically no pressure sideways on the bed, and that's not what causing y-ringing/-ghosting, so that's why the V-wheels aren't so bad, as long as they're spaced far enough apart (front to back) and are tight enough. I think most of the flex in the Y-direction comes from the belt being a bit flexible. If you really want to reduce ringing try swapping the belt for one with less flex.
I will trust my results as the only reasonable variable was the POM wheels and I usually triple-check everything just to make sure. They are more like rubber and far from rigid in any direction on a single 2040 profile. Belt tension did almost nothing, you should have watched the whole video as I said in the beginning. :D
@@DIYPERSPECTIVE Of course I watched the whole video. I saw your results. They don't disprove my claims in any way.
It's no surprise that it moved slightly sideways as you pushed it sideways, but since the front and rear wheels were spaced far enough apart is should've been pretty darn stiff in the front/back direction. Since it still had a problem until switching to linear rails I can only conclude that the V-wheels must've been too loose.
Or the springs.
Try tightening those springs even more and see if it improves further.
I tightened the wheels before the prints so that they would be tight but I will still be able to turn the wheel with my fingers with some force. Any more tightening would produce excessive wear and is not realistic to test. Springs were tensioned to exactly the same distance, not that far from maximum compression. The top of the heat bed weighs an insane 1.7kg (PEI + magnetic base + heating sheet) so I don't think those POM wheels are able to handle those forces, they are not completely rigid at the end of the day, and spacing of 110mm is not that wide when the whole bed is 310x310mm. Plus if the wheels were too loose we would see way worse results and way sooner. In all my test I triple-check everything, because, at the end of the day, I make these tests primary for myself to gather factual data, so BELIEVE ME, I take my time to make sure they are as properly done.
@@DIYPERSPECTIVE Alright. Maybe the V-wheels on my Artillery bedslinger are different somehow, because I get way different results. But then again, my bed is less heavy, even with its glass sheet covered with magnetic sticker and PEI-covered steel plate. In my case I had to tighten the screws until the springs were practically discs of washers, and then the bed was stable and ghosting disappeared, and I haven't had to re-level it in months. Anyway, looking forward to more of your tests.
BUT - using a better plate and wider setting on the POM wheels should also give you a better stiffness. MAYBE it might not be as good as the rails, it would be very close.
Nice investigation. But why did you choose one of the most expensive ways rather than try others? For example to replace 6mm default belt with 9-10mm. Even move to Klipper looks like more affordable upgrade - besides getting ghosting reduction you will get many other features.
Thanks, but I don't know sometimes it feels like you guys don't watch the whole video before posting comments. From the results that we saw with the 6mm belt at different tension levels, I doubt wider belts would provide significant improvement. A tiny one? Sure. The biggest improvement probably would be to get the right tension more easily. And Klipper will be the orientation of the "part 2" video of this 3d printer's complete overhaul. And at least from an engineering standpoint first, you make sure that your 3D printer is mechanically sound and only then transition to more advanced firmware. Otherwise, it is like a bandage to a problem, IMHO.
@@DIYPERSPECTIVE It depends on what features. IS surely is just upgrade for well-assembled printer. But PA (wich is not "advanced" feature in 2023) with smoothing works much better. And wider belt is not about tension. They correlated, but not direct comparison.
I have an upgrade design you might be interested in. It is a brace for the gantry uprights that attaches to the side, under the lowest level the X axis travels. It features a pivot point reinforced with an M5 screw, and an M6/1/4"-20 bolt to adjust the upright's squareness and lock it in place. I've only had it installed on my Ender 3 V2 Neo for about a week, but the testing is very promising. I'm going to upload it to Thingiverse as soon as I'm satisfied that the design will not fail under normal operating conditions. My printer is new, and I'm new to FDM, so I have a few bug with my printer to work out. I originally made these as temporary bracing to fix the uprights from being out of square by about 2 degrees. It's perfectly square now, but still having Y axis stability problems. Let me know if you'd be interested in this brace system. I've grown to like it more than bracing rods since they are out of the way, but the bracing rods are probably slightly more stiff.
It is so hard to imagine what you are talking about. When you upload it to post the title of it here, I am always interested in ideas and concepts that others have.
Why not 3d printed base instead of aluminum sheet?
Because its weight is insignificant compared to the top part of the heatbed.
I print faster with my printer 300x300 plus you're using marlin and not Klipper you will need another y motor to prevent layer shift or motor losing its positioning. I can do 20K accel at 500mm/s it just shake a lot and has too much resonances at those speed and accelerations. just some things to consider input shaper will help with print at higher speed as well. great video.
Nice :D I honestly felt kinda sad when you didn't show those speeds in your latest video after this comment, haha. Definitely something I need to see in my life.
@@DIYPERSPECTIVE lol you will I'm working on a video will be uploading soon!!
hello!! What type, model or brand of motor would you recommend changing in the Y axis to improve on a 300x300 printer (Ender 3 Max)?
Biggest drawback of wheels is the reliability. Tune them and in few days/weeks issues will appear again. Dirt and loosening screws due to forces. For speed dont use bed slingers, those are cheap and best suited for large and slow prints with thick layers (0.4+) and high flow. For anything faster than 100mm and smaller print, you need coreXY based printers. upgrading bed slingers will only allow to go from 80 to 100-120mm/sec
I agree with the first part but not with the second as you don't need CoreXY to print at those speeds the smaller prints. Properly designed small and medium size bed slingers can do that perfectly fine. Plus it is all about acceleration and jerk with small prints and way less with the speeds.
Seems like POM wheels provides same results if running CNC slow, makes linear rails not worth the upgrade.
Why not skip the bed platform entirely? Go straight from the adjuster screws to the rails. Since the bottom platform is only to adapt the single V wheel rail to the edge of the bed it is no longer needed if you got 2 rails that can be under the screws.
Did you though how you will prevent rails carriages from skewing? What about the length of the rails that you will need?
Plus weigh-wise it is nothing. My total heatbed weight = 2100g - 300g (heat bed platform) = 1800g + 50g (4th carriage) = 1850g. That is only a 250g improvement.
Just don't even think about weight reduction or speed with a bed slinger... just no, don't do it. It's not worth the effort. Quality and reliability upgrades only.
Are rails a good mod or stick to wheels? Got any tips for mods that increase quality?
@@charlesraes8541 Yeah rails are ideal especially for the bed because most beds have horizontal side mounted wheels that take heavy vertical loads / motion and will wear out really quickly. If the toolhead is heavy (like a direct drive) that will benefit from rails as well for long term quality and reliability.
Fastest benchy is now on a bedslinger
@@imortalmichael1 Benchy is not a good indicator of overall printer performance... it is meant to be used for printer calibration. It's very small so a fast bedslinger can print it quickly but it's not a 1:1 comparison since printing big objects on it will completely change the results as opposed to a square frame coreXY where the printed object's size doesnt have an effect on the performance.
A very good video.
Actually I always wondered abot POM wheels om the bed. Horizontal wheels while gravity obviously is vertical... Of course, v-wheels transfer the force on the V-shape only, so in a 45° surface. but still, a horizontal wheel can only use one side of the wheel, while a vertical mounted would at least use both sides.
Anyway, for y-Axis v-rollers seem a pretty bad solution.
In my opinion the reason for using v-wheels is their tolarance to errors in adjustments. With two linear rails in parallel you have to be very precise, else the motion system can block quite easy.
So quite understandable, why cheap chineese printers use V-wheels.
Esp. on bigger printers (>400by400) you also have to take into account the different expansion of differnt materials while temperature changes. Nearly all 3d-printers are staticaly overconstrained. So you actually can get layershifts while printing, as due to different temperatures the motion system is not perfect anymore. This happend actually to me with an old Tronxy X1. An error hard to find, as you normaly check and adjust the motion system while the printer not actually printing...
In my opinion, a staticly constrained system might be better. So maybe on one side a linear rail directly mounted to the bed carriage, but on the other side mounted with freedom of movement at least in x- direction - maybe even just a plate of igus high temp material directly sliding directly on an aluminium profile.
The whole problem of overcontrainment on bigger printers is actually the reason I am not into the hype of coreXY. Theese printers are nice in smal sizes (were overcontrainment is not a big issue), but going big, cartesian system have their advatages (this does not mean bedslingers, as crossing rods like on the Ultimaker II is also cartesian). On cartesian mashines, you can actually quite easy modify to staticly constrained. On CoreXY this is impossible, as the belts go 90° over pullys, so every edge has to be solid.
It would be interesting to see if PC wheels make a difference. Also, POM and PC wheels come in V and V solid wheel variations.
Lost In Tech did a great analysis video on POM wheels th-cam.com/video/G9i6LccK_qo/w-d-xo.html
I personally, try to avoid POM or PC wheels on X and especially on Y axis on bed slingers, because the most annoying part of them is that they wear too quickly, and constantly need checking if they are properly tightened for consistent results. If you have one printer then sure, not a huge problem, but if you have to do that on multiple ones, that is a punishment lol. And to be honest V-slot wheels on the Y axis are kinda flawed design to begin with because only one side of the wheel is wearing down and not both like on X or Z axis.
Just move to core xy
Those "affordable" rails have poor quality control. Keep that in mind
Thanks for the enlightenment sir :D but it depends on the seller A LOT.
Why not ditch undercarriage alltogether? The table can be leveled with shims and for the rest there is bltouch
I am not sure what you're suggesting here. Ditching the bed bracket that attaches to linear rail carrieges?
@@DIYPERSPECTIVE Yes, remove bracket, springs, wheels it all create additional flex. Make sure rails are parallel and attach bed directly to them (well, need heat insulation layer, wood shims or something like that). It can be leveled down to 100 or even 20 micron with shims or pieces of foil added. From that point it is printer autoleveling only.
@@DIYPERSPECTIVE Additional benefit will be belt mounting point closer to the bed's center of gravity so will reduce vertical forces pushing bed up and down during bed acceleration. Not that it is ever a problem at the acceleration levels consumer printers use... but motion should be smoother for sure.
In my case, bed mounted too low would hit the Y motor. Everything you say is valid but personally I rather build voron than commit for such upgrade. The one I made was relatively simple with more than enough quality improvements. I don't think there is a point to go any further than that on a bedslinger.
Why not use carbon fibre plate instead of aluminium? They're cheap enough. Edit: that's already suggested
Yeah it it's definitely a possibility. But the main weight of ~1.7kg is from pei + magnetic base + heating plate. It would be interesting to try like a thick G10/FR4 sheet as a built plate and as a heating 2 in 1. But I doubt I will as the ghosting results even with stock firmware looks great considering how much whole heat bed weights. Plus there will always be a acc/jerk limit for some taller prints, so I don't see a reason to push for the much higher speeds.
@@DIYPERSPECTIVE Yes, you're probably right. I never used scales to check the weight on all parts like you did
Pardon?
You desperately need a pop filter on your microphone !!! Get one !!!
You are so confidently wrong lol. Maybe it is your headphones with way too much bass? Because I have a properly set pop filter.
Or just build a non-bed slinging printer.
Or just buy P1P :)
I don´t understand the premise of those videos. You modify an inferior bed slinger design with this and that...just save that money and buy a Core XY design instead form the get go. All those problems are already solved in a Core XY design, that´s the whole point of that mechanical design. Don´t buy those shitty "i3 Prusa since 10 year alsways the same style stuff"
You can also improve the alu. bed design further with more cut outs or make it thinner with beads and flanges (if possible).
If you don't understand why you even bother to comment? :D Maybe it is fun to modify 3d printers? Why do people modify shitty project cars and have so much enjoyment? Huh? If you don't have fun doing that it doesn't mean others feel the same, the world doesn't turn around you buddy. :D
@@DIYPERSPECTIVE Sure you can have fun doing it. But from a cost-benefit point it just does not make sense to buy a VW beetle and swap the engine to a Porsche one. So pay more money for a inferior product AND also invest time. It´s like all those videos "Oh look the Ender 3 is such a great printer if you only install those 53 mods for an other 800$" Yeah sure, but instead of doing that you could have bought a 1000$ machine which already is a great printer and you have less work with it.
How are you so close minded? This is actually funny to read. :D People who want to get into 3d printing don't want to spend $1k just to try it, so they buy cheap popular 3d printers. But with time for some people it turns into a hobby and upgrades are natural part of it. No one is saying it makes sense the cost wise, but guess what? When it is a hobby and you're getting fun out of it you don't care about the ultimate price/performance ratio. Is that really so hard to understand for you?
@@DIYPERSPECTIVE If you don´t want to spend 1k for a 3d printing, buy a cheap printer. But if you find that you need a better machine or you want a better machine simply sell your Chinese crap instead of wasting way too much time and money upgrading (which is physically not really possible from a i3 frame to a Core XY one) and buy a better machine, like the Bambu Lab for 800$.
In this case if Gaming is a hobby would you recommend buying a 1500$ GPU ? God I hope not.
Of course price is important. If those people didn´t care about the money in the first place they wouldn´t buy a cheap machine.
How you still don't get the point, even when I explained like for a child? ;DD This guy with the only his right ways and truths ;D This is too funny :)
I'm sorry, your heavy accent is very difficult to follow. Would help a lot if you slowed down..... Do people really noodle around with their 3D printers to this level.? I've had my Creality CR 10 for 3 years.... Never had a print issue, apart from those self inflicted as I got to grips with using it.... My prints are incredible, fine detail.... I use a heavy mirror as my build plate,..... All this weight reduction and velocity/jerk talk is for design engineers, not end users...... Think your likely to introduce more problems to resolve for anyone that tries modifying..... I wouldn't consider linear rails an upgrade... Just an alternative.... Nicely filmed video though. Thanks.
I always provide a full script as subtitles, which could definitely help. I agree that these settings are not for most people who own 3d printers but even inexpensive linear rails are the upgrade to POM wheels and the definition of the upgrade shouldn't be defined by users' ability, you know. However, I agree that you must have a certain amount of knowledge and experience to make an upgrade like this work better that POM wheels, that is for sure. I just wish companies used more linear rails in the first place.
I would send you some commas in exchange for those dots. Would help a lot if you slowed down :)
@@DIYPERSPECTIVE You don't have to. I did not notice any accent until that post :) Your analysis was very helpful, I'm going to do a better board now, after watching this.