3D Printer’s BELTED Z-AXIS - INSANE PRECISION!? or NOT WORTH IT!?
ฝัง
- เผยแพร่เมื่อ 8 พ.ค. 2024
- Get PCB, CNC, 3D printing, and other services for your projects at PCBWay (5$ OFF for your first order) - pcbway.com/g/584K9t Also check out their 9th ANNIVERSARY SALE - www.pcbway.com/activity/anniv...
👉🏻 In this video I am testing a custom design 6:1 ratio belted Z-axis mod with linear rails that completely eliminate the gantry drop. I will compare it to a regular setup with a lead screw and a POM nut.
HERE YOU CAN DOWNLOAD FILES OF THE PROJECT - www.printables.com/model/4952...
🧡 Support my work at PATREON - / diyperspective
📋 RELATED ITEMS TO THE VIDEO (Amazon Affiliate)
3D printer:
- Kingroon KP3S 3.0 amzn.to/3N4fr96
Filaments:
- ANYCUBIC PLA amzn.to/3W3MmwE
- filalab PLA filalab.shop/en/shop/pla/ (non-aff) (-10% with DIYPERSPECTIVE code)
📷 VIDEO SHOT ON:
- Fuji X-S10 amzn.to/3Zar9BW
📢 OTHER MENTIONED THINGS:
- Z-HOP TORTURE TEST www.printables.com/model/4950...
- GJB Test Cube www.printables.com/model/2931...
🕗 TIMESTAMPS:
00:00 - What this is about
00:41 - The gantry drop
01:34 - Design choices and potential problems
02:27 - The belted Z mod
03:07 - Print Quality
04:12 - Got lucky?
04:44 - Dial indicator tests
05:14 - Concerning issues
06:14 - Final thoughts
🔗 YOU CAN FOLLOW ME:
Twitter: / diyperspective
Instagram: / diyperspective
#3dprinter #zaxis #test - วิทยาศาสตร์และเทคโนโลยี
I have a large DIY bed slinger and last year I converted it to dual drive belted Z axis. I designed a simple gearbox with a 5:1 reduction (using 80T and 16T pulleys). It all works great, but there is very little improvement in terms of print quality over leadscrews. I didn't lose quality, but I didn't gain much either. But there is one big advantage and that is that belts can move the gantry much faster than leadscrews. I have my gantry set to do all moves at 75mm/s and 1000mm/s2, and that makes Z-hop much gaster than with leadscrews, and on some prints (with lots of travel moves) that can make a huge difference in print times. The reason I did it was because I enjoy tinkering and modding my printer, so for me it was worth it, but if you are doing it to try and get better print quality, I wouldn't recommend it.
With regard to your print quality issues, did you try increasing the current to the Z stepper? If the current is too low, it's easy for the stepper to skep microsteps, and that could account for the artifcats you are seeing.
Thanks for the info from your side. The fast Z-axis speed is great, but you don't really need that on regular prints. You mostly need accel and I needed to tighten the belt a lot to minimize the wobble that occurs when moving up and down on z-hops. But again that can be minimized with more or wider belts or a lighter gantry.
The stepper wasn't skipping any steps at 750mm/s2 accel and 25mm/s travel speeds. And my prints were with 100mm/s2 accel and 8mm/s travel.
It improves quality by eliminating wobble and other issues due to a bent leadscrew and other issues with worn leadscrew brass nuts like z compression etc. If the printer isn't experiencing any of those issues then a belted Z will not show any improvements (when adequately implemented).
@@hd-be7di just do a wobble x upgrade then?
Nice video. There is also something to consider. More pulleys = more excentricity issue...unless very high quality and tigh tolerances ..but its hard to fnd. And excentricity on a z system will create artifacts a bit like z wobble
Thanks. That is most likely why I got worse results. But even if they would be better, I am not sure I like the idea of the belted z :)
@@DIYPERSPECTIVE I personnaly don't like the idea :) though some have good results with it.
@@DIYPERSPECTIVE look the mod kevinakasam did to the ender 3 dual z belt axis no more z binding from the cheap lead screw and like other said you can put key back mechanism on it but doesn't need; with that system the ony part z is controlled is by the v slots don't need to be parralel to the lead screw meaning that moviments on z axis are always fully vertical
I must admit that i always wondered why people aren't using belts for their Z-axis since it's something really accurate, so those results are extremely interesting for me. Nice videos in general dude, keep up the good work.
Thanks. I heard good things about belted z too but, I have to admit, I spent so much time analyzing what can be wrong when I got worse results. The more I thought about it the more things I discover that I am not a fan of. So I just ordered another ball screw for the printer :|
For the price they are good, but rather complex as you need a gear reduction. That means two tensioners and a good chunk of space. And you mustn't cheap out on pulleys as this would also result in z "wobble". Personally i prefer integrated leadscrew steppers and pom nuts. Silent, grease free, space efficient and precise. As long as you have the same z steps belts and screws are equally precise, have equal force and same top speed (not accounting for screw or belt system inertia). Consistency is dependent on the quality of the used components, a shitty leadscrew is just as bad as wobbly idlers or pulleys
Cetus (which the printer shown here is copied from) from 2018 era had the belt driven z axis. It had that gantry drop problem, the fix they added was to use a friction brake, so a plastic collar rubbing motor axis. It kind of worked. I remember someone in reddit did a servo mod (so a servo held the gantry at the end of the print).
Surely screws are more precise than belts though? The advantage of belts is that they're cheaper, lighter, and faster.
@@andybrice2711 of course they are more precise, exactly because they are "slower" 1 rotation of the motor only moves up the nut by 1 pitch, aka much greater and more precise control through reduction. A belt has a bunch of backlash and flexibility, even with a similar gear reduction a ball screw should be more precise (although can be more prone to certain specific resonances/vibrations, as belts are more likely to dampen them and stiff things tend to have more specific but more responsive resonances while flexibel things respond over a greater range but generally much more weakly).
The things is every extra gear for reduction also add extra inaccuracy due to tolerances that every product has (including bearings).
Belts can be good on systems where the Z axis might be over-constrained, or badly designed z-axises where they end up being disturbed by specific resonances that only can't get rid off.
The video did a decent job, but this comment by JimiDj is a bit off the mark.
Belts aren't all that accurate compared to many other systems but they are fast and generally cheaper. There's a reason they generally aren't used in a CNC machine ...
The original design of the belted-Z system you showed at 01:40 as an example to use 20T : 125T pulleys,
which you also showed at 01:02, is actually a 1:8 reduction ratio design.
To deal with the Z dropping issue, you can also use a counter-weight or a spring system to act against the weight, like the Key-bak design on Voron Switchwire.
Thank you man, i was considering this to remove uneven layer height but in the end you proved that it's not worth the hassle.
Glad to see your channel growing, count me in.
I'm using a belt z mod from kevinakasam on my ender 3 and my print quality improved, most noticably when z hopping was enabled. You do have to maintain pretty tight belt tension like you observed in your video. Also it enabled faster z speed so probing the bed is faster. I don't regret the mod
Thank you very much for taking the time to do these tests and comparisons. I am about to redesign my home made CoreXY printer to replace the use of screw/nut with pulleys/belt. With this information, I better analyze the possibilities before moving further. Very good videos! Thanks for sharing! 😁
Thank you for the nice comment, glad you liked it! :) Probably the easiest and most cost-effective way would be just using POM anti-backlash nuts on regular properly aligned lead screws. After this video, I will really need to test Voron 2.4 z-axis (I need to build it first lol).
@@DIYPERSPECTIVE Anxious to see it! I have to do more research about using the stationary bed and moving the rest. Good luck with the assembly!
Great videos as always!
Check the Cetus printer from 2018's time.. It didn't have any reduction for belted z axis. They just put a one way clutch on the z motor shaft running inside a fixed plastic collar. You tighten the collar just enough where the gantry won't drop without motor actively pulling it down. That works and stops the gantry from free falling. The rigidity of the axis is more on the linear rail on that printer. The belt was not causing/reducing the regidity. But given the y axis is fixed to z carrier with only three m3 kind of low on regid scale to start.
Somebody already said this but I'm glad somebody took this shower thought out of my head.
Furthermore, your video is really well presented and clear, with numbers and experiments and well!
I wish research mags would be open to making content like you, science research can be frustrating.
I think your findings make sense, but also I think there's some things you can do to improve results. Looking at your "Idler mount Belted Z" in the video I see it's plastic and even looks like it's bowing under the strain of the belt tension, I think that part needs to be re-though stronger or printed in metal. Also most belted Z I've seen are also dual-belted, so you're putting all that strain on a single belt now. I'd try using a thicker belt so the belt itself has less flex and more surface contact on the idler, also for the idler maybe you can sneak a larger diameter in there for that same surface contact reason. In the end the belt's surface is still rubber and the reinforced strength in it is length-wise, all the tension creates compression across the face of the idler, so it can still squish there if overly exerted upon, the randomness you are seeing might be the effect of it inconsistently squishing the surface as well as the idler mount bowing upwards in the middle. Also for your linear rails, I think you'd get more strength that resists play, by using a single rail and having the sliders one above each other, then having a nice triangle shape coming out from that that, it could maybe even go all the way to the middle of the X-axis beam? The distance you have the carriages spaced apart will directly correlate with how much force they exert on their bearings to resist flex, but having them too far apart eats into valuable Z height. At the very least try and aim for the pitch v-rollers use.
Just my thoughts on things that might help!
i build the KevinakaSam dual belted Z, and its Awesome.
I’m using the Kevin aka sam belted z on my e3v2neo, print quality noticeably improved, I think the original z screw might have been slightly bent or damaged, which is why the print quality improvement was so large
Maybe consider steppers with electromagnetic brake or friction brake then lower the gear reduction. Adding a counterbalance to make the gantry light enough to be held by detent would also be a interesting idea.
There are a fair amount of machines used to manufacture medical catheters that use belt drives for longer distance transverse or z axis movements. I think the belts have a pretension of something like 500-900lbs.
Love this video! Not sure about montage music tbh
I’m in the process of designing a belted z-axis upgrade for my printer with a 30:1 reduction using a worm gear, which allows a high reduction in a smaller footprint and prevents back driving, so no gantry drop. I use linear rails on the back sides of the extrusions and recently added the rollers back to the outsides (no eccentric nut). The rollers added a significant amount of stability. I want to do belts with reductions for the increased resolution, which should help with ABL
Suggestion for using less pulleys whilst achieving the large reduction and avoiding the lack of meshing teeth.
Use two smooth tensioner pulleys to tension the timing belt around the driven gear, such that more teeth engage.
Then you don't need a staged system.
You actually don’t need that much of a gear reduction for a belted Z axis. The key to using a belt on the Z-axis is to use a counterbalance. Uses far less torque this way. Rigidness can also be accomplished with linear rails.
Apparently a counterweight can have inertia issues on heavier z axis (e.g. beds) and you'd want to use a constant force spring instead. (Count around 20-30 bucks for a pair of good quality 50k cycles ones.)
@@satibel Yup, it would increase the inertia, but I don’t think inertia is as critical on the Z compared to the X,Y. This is why Bambu Studio uses carbon fiber rods on their X,Y. My guess is that he believes the lead screw on the Z axis might be inducing some amount of play. Interesting experiment but the X,Y play a far greater roll in print quality and any play that a lead screw could induce on the Z is probably negligible. As long as your steps are the correct increment and there is not a lot of play on the Z stage or whatever, it shouldn’t make that much of a difference.
I also turned my KP3 into a belted Z drive using a geared stepper driver, think it was 8:1. Was an old Kysan that I used on my 2014 Mini Kossel builds extruder.
Have you ever tried using a counterbalance setup to keep the z axis from crashing down? Or a servo lock that triggers once the print is complete?
The artefact magnitude seems similar but the artefact quality is different. It occures to me that your belted Z suffers from random layer to layer misalignment while your lead screw build has a repeatable pattern. It occurs to me that the results would be improved if you had a direcrional preload of some kind, and odds are, the compliance of belted Z can take it just fine, while fitted nut screw might become overconstrained and bind up. I think dual lead screw Prusa Mendel style builds are really overconstrained that way and belted Z is a good design trait for them.
Interesting. As far as I remember your cantilever video, being one is a single biggest limitation of KP3S. I'm not sure why I was expecting belted Z on that modified de-cantilevered printer, I'd expect it to work better there.
How's the noise, is it quiter?
Maybe it's worth checking if the titan extruder is the source of those artefacts. After all it's a cheap titan clone. Was actually considering switching to a bowden extruder to reduce weight of the toolhead. Cool project for sure, thank you for all your hard work. My KP3S suffers from Z wobble and I'm still undecided what should I use to fix it, but right now I'm leaning towards wobble rings designed by EvoMotors (thingiverse) or WobbleX sold by Mellow
I think most of the artifact that he's seeing can be explained by miragec inconsistent extension video
Btw wobble X and ring probably won't work, the gantry of the kp3 is not heavy enough and wil cause the z to jitter since the wobbleX is not being compress
When mine was stock it suffered too. But with some drilling and trial and error, I got it to a perfect lead screw alignment. The prints look fine not under the worst lighting conditions. I can show them to look flawless too with the right lighting angles like all people on social media do, just saying. :) BTW make sure it is not the X-axis wobble as it is a cantilever 3d printer.
@@DIYPERSPECTIVE My reasoning for considering bowden is: 1) less weight = higher accel and less ghosting, 2) as X value changes during printing the nozzle's distance to the print also changes, because only one side of the X axis is supported
@@twanheijkoop6753 There was a remix with 10x2mm (N52?) magnets that fix this issue, but not sure with how many other printers it was tested
My experience with Ender 3 clone:
- brass lead screw nuts are OK-ish in single direction
- belted Z and POM nuts behave better with Z-hop (there is no slack, so they can pull X gantry down. POM nut can wiggle a bit, but usually won't have unrotated movement along lead screw)
- Z wheels must move really smoothly (as in each wheel "cage" must fall freely without any other weight, e.g. stepper or gantry). If there is only single lead screw, the "free" end wheels should be really smooth movement (fairly loose) or it will start to get behind and gradually catch up with driven side?
- counter weight prevents belted Z from falling (belt from gantry, around idler or bearing on top) but may create new problems with needed space and swinging
- dual steppers with separate belts or lead screws better be identical
After all that, it's running dual lead screws synchronized with belt, POM nuts, loose wheels and single larger-than-normal stepper motor.
If I would add linear rails somewhere, it would be Z 😅
Well, standard stepper motor with standard lead screw and half-stepping gives you a 0,01mm resolution. The only problem: most 3D printers and cheap CNC machines don't include backlash cancellation on Z axis, relying instead on gravity. A simple fix would be a backlash cancelling nut, a spring and 3D printed slot to keep nut from turning. This adds little to mass but spring not only would cancel backlash, but would also act as dampener. I'm planning to add this when I'll be adding a second Z-axis screw and stepper to my Ender 3 V2...
Another option that won't be as bulky or have as many concentricity issues would be a 6:1 planetary gearbox on the stepper , then just a normal 20T pulley. might be worth a try
I have a belted Z on my Pro 3D V-King 400 and I just have end g-code to lower the bed after a print has finished, a gantry could use the same method although filament would need to be retracted and the bed and nozzle would need to be cooled before lowering.
And how is your z doing with belts? After this, I am really interested in voron 2.4 z axis and how well it works.
How about a NEMA 17 motor brake? Power off condition engages the brake. Power on releases for normal stepper function.
You could have bought the relay module from the Lulzbot Mini 2. It crosses the stepper motor windings when the power is turned off.
Thank you for investing such amount of time and effort. I still believe people underestimate the power of software. If you know how to tune your parameters in the slicer, plus having a good firmware (stable, updated, etc), you can reach incredible results compensating these hardware issues/limitations.
I think the elasticity of the belts undermine every addicional decimal of precision that you might get from that sistem.
On the other hand, threaded rods are usually not perfectly straight but this issiue can be compensated with mechanical decouplers so you can still achive very high precision and rigidity.
Would be interesting to see a direct drive 0.9degree stepper rather than the complicated pulley gear box (which probably is introducing noise) driving the Z. You would need a one-way clutch or counterbalance though on the gantry
I just printed a cali dragon on a slightly bent dual z-axis lead screw printer and got a print almost like it was injection molded i cant find any problem even if i wanted to. Printer: crealiy ender 3 max neo with sprite extruder and klipper
Is your bed PID tuned? It can reduce bed flexing and z artifacts quite a bit
Yes, it is. Before and after results were with the same PID tuning. If you think prints look horrible that is because of the lighting angle. They look fine otherwise. :)
What if instead of the gear ratio you use a counterweight so the the printhead floats even without power? Of course double the weight to move = slower z hops but it would make the design much simpler and include a belt as you intended. Would love to see that :)
I thik you should try using a lighter extruder like an Orbiter V2 or the Sherpa Mini. That have great pushing force despite their size and will help reduce the weight of the toolhead as well as reduce reduce resonances.
Great video. I've thought about converting my super tall printer to belt Z due to leadscrew wobble, but I wouldn't be surprised if belt elasticity causes problems of its own. Rack&pinion may be best for it.
To get good belted z results you need wide belts (how wide and how many depends on the weight) and high tension. Voron 2.4 does it right and results are pretty good. Though I doubt anything beats a ballscrew precision wise.
@@DIYPERSPECTIVE Ballscrews are annoyingly complicated though, since you have to have a coupling that allows lateral motion. They're very stiff and usually not quite concentric, so unless the frame is super stiff like a milling machine, they cause worse wobble than a leadscrew if the nut is rigidly mounted.
I converted my ender 3pro with dual lead screws to Kevin's dual belted z and the z banding on my prints totally disappeared. Also I was now able to do much faster printing at 120mm/s and faster z hops too. So I do not agree with your conclusion that the belted z does not help. Also I am running a direct drive on my ender and despite the additional weight on the x gantry of the extruder and motor my gantry does not fall when the motors are off. One more thing, I am able to do 4x faster z homing and bed mesh with the belted z and that too without any noise which the lead screw would make. So overall my results are totally opposite to yours 😊
I never said it never helps. In my scenario, it didn't. But I had a perfectly aligned lead screw and had 0 problems before. I also tested at 0.12mm layer heights which are very demanding. If in your scenario it did, I am glad it worked out for you! :) I personally just find the mod way too complex with a few potential issues. Cheers!
same here
Did you also upgrade the roller wheels to linear rails? I think that's where there may be a lot of the quality improvement, as those roller wheels can develop flat spots or get debris caught under them and cause pretty bad Z banding as a result.
Nice video. I did the Kevinakasam dual belt z mod as well on my Ender 3 pro. In retrospect, I should have done my CR-10 instead as the gantry is much wider with less support with just a single z screw. But all in all, I'm quite happy with the mod on my ender 3 and will do the same to my cr-10 when I get a chance. I personally haven't found a need to install a key bak mod on my lightweight gantry with bowden and minimal hotend.
@@fredmertz4819 My enders are all direct drive with the mod from Thingiverse where the stock extruder directly mounts on top of the hot end and even without a keybak the gantry does not fall even when the printer is off, might be useful to you
Overall the best inexpensive method for the Z-axis for a bed slinger is double THICK lead screws. You will see a significant quality change by going from 8mm DIA to 12mm DIA with lead screws.
(you can go larger but check your step per rotation and it will start to get expensive)
the hob gear not being perfectly centred on the extruder motor shaft is what will be having the most impact on your surface quality.
Hi! Nice video. A single wormgear would solve your issue. Wormgears are not transferring power backwards, it cannot drop. of course it would not mitigate the flexing of the belt, just an idea.
Also also, Try a block and tackle belted Z axis next!
Have you thought of using a counter weight that hangs over the top of the Z-axis, like the ones used in normal elevators or lifts. This will double the weight on the Z-axis, but it should not be a problem. With a weight balanced Z-axis, you won't need such a lot of gearing.
Great video! Good for you for exploring other options. Funny, logic eould suggest that using belts should work since X and Y use them and the quality is Ok. Heck, corexy uses a very long belts and print Ok too. So, it could be your design (too many pulleys with errors) and "faulty" implemention (no disrespect) that could cause worse outcome than screw.
Thanks. The z-axis is not really comparable to X or Y as we have gravity pushing down the gantry by a lot (with this printer in a stationary position belt experience tension like during ~10k accel if it was on a X or Y axis). Meanwhile, X and Y moves are basically frictionless and equal in both directions.
I don't think it is a wrong implementation, it is just when you reduce the ratio so much with a lot of pulleys those inaccuracies multiply. And I used quite decent pulleys. Others who use cheap ones would probably get the same effect with one reduction belt. But because most probably make a mod like this on cheap Enders with leadscrew problems obviously results will be better, but none of them check actual Z movement accuracy with a dial indicator. Maybe I will change my mind about belted z after I build Voron 2.4. :)
Those pulleys are crazy overkill. The simplest and most compact way is to use a 5:1 worm gear. That maintains the 400 steps/mm on a standard T8 lead screw, and is also a self locking mechanism. The gantry will never drop.
all those wondering why belts are not used: a normal 8mm ball screw gives you 2mm per revolution, a 16t gt2 pulley has a circumference of roughly 40 mm. now if you rotate that small pulley one full rotation, you have 40 mm per revolution the belt is traveling. so instead of having a 20 :1 transmission on top of the belt its way easier to have a screw.
I bet you could also use a worm gear to reduce backdriving
The PCBWAY thing, that ties everything together is most likely a 150-200€ part. If you get it right in the first attempt.
That's just for the ad. The part doesn't need to be made out of metal, normally you'd just print it.
Belts for speed, ballscrew for precision. All industrial CNC equipment is made that way for a reason.
This is the way
I've fixed that artifacts in your video by changing my stock extruder to BMG Original. Trust me, it's not about Z rod or Z issues. Is the extrusion inconsistency and the extruder gear issue.
You should have watched the full video before posting a "trust me bro" comment lol
Are you releasing any 3D models or parts that you used?
I said myself if anyone asks I will publish them :D so I guess I am doing that right now
I updated the description with the link.
@@DIYPERSPECTIVE oh great! Thank you :)
Well, I've answer all those questions a décade ago juste with an inexpensive planetary geared NEMA17 motor and quality belt and pulleys.
Could you use a finer lead screw instead?
I actually tested that, but I couldn't see a difference on my modded kp3s pro, so I didn't make a video. The best way is just to ensure that the leadscrews are properly aligned and don't wobble when attached to the motor's shaft. A lot of cheap 3D printers have problems with all this.
Great!
Try shut down heatbed that was my problem for Bad z
The thing is, a good lead screw might be better than a good belt, but even a cheap belt is going to be way better than a cheap lead screw. And we all know what kind of lead screws our 3d printers tend to have.
Also, belt usually means dual z, which is a massive improvement. You can do dual screws, but that means either driving them from both sides (major pita unless you have a board that can G34) or syncing them with - guess what - a belt.
Still don't get why not use a worm gear? Reduction as well as not being backdrivable so no loss in Z-hight and is less complex than all that reduction gearing
Yeah, I can't see why a z-belt would offer any advantage. Belts are lighter, faster, cheaper, and flexible. Screws are more rigid and precise. XY Belts + Z Screws already seems like the optimal design.
This test is flawed because in order to see the effects of the Z axis you must first remove the biggest source of inconsistencies: the extruder. The simplest way to do this is to get a bowden extruder. The bowden will even out any inconsistencies. The other way is to get a direct drive extruder with RIDGA gears and replace the idler gear with a bearing. Or get a CNC vz-hextrudort with twill gears.
CoreXZ bed slinger FTW
How about using a large gear extruder for these tests? That should be more consistent.
Yea, it would, but even ignoring the print quality the movement consistency of the z-axis wasn't as good as with a ball screw.
@@DIYPERSPECTIVE So if ball screws are better couldn't you build a 3d printer with ball screws for X and Y? It is probably really impractical but combine that with a large gear extruder and you might make the ultimate quality 3d printer.
Yes, you are talking about the Pantheon 3d printer :D go look for the price of those, as you need really good quality ball screws :)
I would like to see someone attempt a linear stepper z-drive, with an elevator safety brake style gantry drop mechanism.
Sounds like something you are better equipped to attempt than I am........ ;)
Maybe, maybe...
Why though? The Z-Axis makes hardly any movement. Surely it's the X and Y axes where you could achieve the most speed gains?
@@andybrice2711 Right. Hardly any movement for the printing of say a full Mandalorian helmet. 1) I am not interested in quality of print, rather interested in zero post-processing. 2) This video is not about speed performance. 3) I would like to see whether micro-stepping on a linear rail has the granularity a standard brass nut and rail has. 3) Because I am curious and proposed the challenge - ie why not?
@@truegret7778 I mean, sure, all those of experiments would be interesting. But my prediction is it would add major complexity and cost with no advantage.
Whereas linear steppers on the X and Y axes, I reckon that could be revolutionary!
could a counterweight be effective?
yes
Awesome thou im not down taking ur genius i bow down to u ceaser. A hundred percent positive ….
CoreXZ?
Just use a constant tension spring to counter the weight on the z axis...
I would love to see you repeat these experiments after you get a better extruder. The imprecision of the extrusion gear is far worse then any runout on the belt pulleys.
Yeah, it would be interesting to see (I written down as a quick video idea). But also dual-drive extruders can have issues of their own. Even if we invalidate the print results, the test with a dial indicator shows that at least in this specific scenario, belted z wasn't as precise as I was hoped. I get more or less similar results on printers with a regular lead screw.
What about a counterweight?
That could also work
I am pretty much set on the saying if it is not broke don't fix it. However, I firmly also believe that there is always room for improvement regardless of what it is. This keeps humanity / civilizations moving forward the very thing that puts rockets into space and man on the moon. Thomas Edison found 10, 000 ways how not to make a light bulb it only took one that made it work. never give up my friend you may just be the one that changes everything and discovers the ultimate in 3D printer design making it faster and more reliable then ever. Keep posting Thx
Yup, it is a great take. :)
that would explain why majority of the voron 2.4's in the community have really bad layer qualities
IMO You can't use belts on the Z stage it can't maintain the increasing loads and the deflection or tensions for any real sized parts or weight
Make an elevator style contra weight that matches the weight of the arm.
I wonder what a 12mm belt will be like.
that ratio is absolutely rediculous.
i have a single reduction and i haven't turned my machine on in almost a year now.
and it's STILL in the air. in the exact same position i left it in.
can someone explain why gandry drop is a problem? i mean you level it at each print and all these powerloss safety things are just stupid
Yeah, power loss would usually drop it into the print, which would usually not be recoverable anyway due to the nozzle oozing onto it as it cools down. Bed Z is more of a problem, since it could have a long way to fall and a hard stop at the bottom.
@@dekutree64 and i never had a problem with power loss i dont know how is it in the usa but atleast where i live in germany never happend
Seems way over complicated. You could probably just use some kind of a spring tension brake system that holds the position instead of all that stuff. Maybe even a counterweight system, but a brake pulley system would probably make more sense.
Das ist nichts neues. Gab es bei einem sehr etablierten Hersteller schon vor über zehn Jahren. Ich habe einen 3D-Drucker mit Riemen an der Z-Achse und bin damit ganz zufrieden.
Very interesting, as always. Some kind of brake to stop the gantry from falling might work out better. th-cam.com/video/-39GtYVHS6M/w-d-xo.html
My current problem with my KP3S is that my linear rails seem to be worn and need to be "re-balled". It really is cheaply made, and there's a limit to the quality you can expect from such a machine.
This sucks. I am in the process of converting my CR-10 to Kevin AKA Sams dual belted Z conversion. The who time I wondered if belts were as accurate as the lead screw. Everyone kept saying this was so great and belts were better. I had my doubts but since everyone claims that it was better I proceeded. Sounds like they just kept all excited because it is a modification and looks cool. They are not very knowledgeable apparently. I wish people would think logically about stuff, and not just get all excited for modding for the sake of modding. This is like the guys who don't know what they are doing modifying cars and end up making their car slower and worse in every way with their modding.
Try corex-z
A belted Z direct drive on a cantilever printer... ... sure if you like to suffer go right ahead lol
Not worth it. Just buy a delta printer.
Why is the point of all this ? I tis useful to do that "hack" for people to see how belts in this application absolutely don´t make any sense. You should always prefer design slimplicity over overengineering.
I can ask you the same about your comment lol
@@DIYPERSPECTIVE Why would anybody choose a more complex more bulky technical solution with similar results ?
The point is experimentation. Testing an idea to see if it works. So far, the answer is "no". Which is still useful information, even if it might feel like a failure.
@@andybrice2711 experimentation makes sense if you don´t the outcome. Everybody should have known that a complex multi part array is not as good as a simple rod.
Where experimentation would make sense is comparing timing bets vs fishing line or POM wheels vs linear rails etc. This was basically a waste of time for clickbait.
@@sierraecho884 Yeah, that was my prediction too. But a fair few people are claiming they get better results with a Z-Belt. I'm very skeptical about that. But I guess it's worth testing, to try and work out what's going on there.
Just print a stopper lol 😂😂 over engineering to be slow lol .. stay simple .
consider, if you will, COREXZ