After digging into my 2 i3's pretty deeply over the last month I've come to the conclusion that one of the biggest generators of artifacts, be they banding, salmon skin, irregularities, in the X and Z directions is the wheels shipped with these kits. They are unfortunately cheap and not manufactured to the highest standards. About what you'd expect given the price. I've actually had to get FLSun to ship me a bag of replacements because of this, but only a few in the entire bag were near being 'true'.
Has anyone had any luck with replacing these from a 3rd party? I've seen Delrin wheels (well, supposedly Delrin) on AliExpress, but has anyone tried them? Any good?
I have a redesigned carriage system I'd like to publish that greatly improves X and Z rigidity, but the wheels are doing the same thing they have on every carriage I've tried and caused it to move in ways it shouldn't.
I got around to replacing my X carriage with a linear rail. Cost of the rail was $20 canadian and I had to buy a longer 2020 rail which was maybe another $10. I cut an idler and nema plate out of aluminum but they are openbuilds I've seen them online for maybe $8 apiece. I am able to print as fast as the extruder can melt plastic with only some slight ringing from the high speeds... very happy with it.
Looking at my old V-wheels they had some ridges indented in them but the FLSUN rounded wheels still looked like brand new so it seems like they are not bad compared to third party stuff. Leaving the wheeled Z plates on since they run really nice and I'm not seeing any problems from them. The Y carriage I got high quality linear bearings and pressed them into 3d printed pillow blocks and those are working great.
Nice. Agree on the wheeled Z. I'm getting zero wobble and the new carriage design will definitely make its way to the HEVO (especially since its going to cost me much less to do).
Knocked out some more prints this week and am pretty happy with the build. Finished up the last of the X-Y prints for my HEVO in that black PETG. Little more dialed in but still need to find the sweet spot for extruder temp on this machine. Switched over to some promo clear PLA I had lying around and did some filament container parts and a ThreeA-inspired robot. Looked pretty damn good.
The Y-axis is the weak point now. On my little i3 I did block bearings and printed holders, but I'm thinking if I keep this thing as is for a while it'll need something more drastic. I've got more Z travel than I'll ever need to redesigning the lower chassis to use a wheel setup isn't out of the question. Think in the vein of a Tevo or Ender, but with two rails instead of one 2040. I could potentially have enough extrusion left after cutting the Z's for the HEVO to do it without completely redesigning the lower chassis, but we'll see. I'll post a screen cap of the CAD model once its done.
For anyone still watching this I have a couple of updates.
First off the PA wheels I ordered. They do appear to have tighter tolerances according to my measuring jig, but are far from perfect. I don't have them on the current build, but will give them a test when I pull apart the Z carriages on the i3 Plus to swap the leadscrew and z-motor mounts around. The biggest issue I see with these, and the same would go for V-slot wheels made of harder materials, is that they are much more at the mercy of the quality of extrusion. They pick up every little bump and wave.
The new carriage design is working pretty good. I'll admit there were some oversights though, particularly with respect to cable management. If I can find the time to fix it I will and post the final draft to Thingiverse. The X and Z carriage assemblies are attached.
The attached test cube is print zero. Spool3D black PETG printed at 230/60, 0.2mm layer height, 75 mm/s for all. As you can see I have some adjustments to make... Print settings and the Klipper config file were copied over from my little i3, maybe a little aggressive. Accel is currently at 3000 which is just too high for how I have the Y set up. Very happy with the X though.
Your observations regarding the quality of the extrusions is spot-on. FWIW I compared V-slot extrusions from 2 different sources and the difference was enough to convince me to just pay more and move on. The same can be said for wheels no matter the style.
Everything we do with our own built printers are trade-offs and you need to choose how far to pursue in order to get where we want to be.
"Everything we do with our own built printers are trade-offs and you need to choose how far to pursue in order to get where we want to be"
Words to live by, with any design. I try to impress this on my engineers with every project we work on. Surprising how it doesn't always resonate.
Debating right now whether or not to print out these carriages for the little i3 or not. It'll be donating its parts to a corexy soon as make a decision on using this Z carriage design on it, or sticking with smooth rod/linear bearings for now.
So happy this thing is up and running. Been a frustrating build.
General update for anyone curious:
I built a little test rig to run each wheel under a dial indicator to check for flat spots or uneven POM material. They were actually better than I thought they'd be with only 5 of the 21 wheels I have on hand being out enough to avoid using. Interestingly I did discover some of the slots on the aluminum extrusion to be pretty bad. I know there is actually a pretty wide tolerance range for these things, but didn't think it would be as bad as it was (one had a bump of nearly 0.5mm right in the middle).
I finished up the design of and printed a new Z-X carrier system that also allows for mounting of the extruder assembly as a bonus. New X-carriage that is vertically mounted as well. Each carrier relies primarily on 2 wheels to provide a motion guide, with the 3rd opposing wheel under very little compression. Every seems to move pretty well, but haven't had enough time to finish assembly yet. Hopefully in the next couple of days I can get it together and knock out some testing. Really a matter of wiring at this point, and getting some mounts done for a MOSFET and some TL-smoothers.
I did order some PA (Nylon) wheels of AliExpress as well, see if they turn out to be any better. I still feel like these are a viable option, especially when you factor in cost, as everything has weaknesses and issues with manufacturing tolerances. If I can get things working well enough there may be a QuadRap design using wheels and t-slot posted here in the coming months.
I put an MGN12 on the X axis. Honestly, the difference wasn't so huge. It was a much better bump going from 4 wheels to 3. When I attached the rail to the aluminum extrusion, I noticed that it wasn't super straight either. Cheap printer is cheap. Next time, core-xy with all linear rails.
That's where I'm leaning more and more. I'm not keen on the idea of having to rip apart a HEVO every few months for linear bearings. I've read of people getting bad MGN linear guides, but I feel like there's some preventative maintenance that can be done before hand to address the issue. The price difference for the X won't be a big deal, but the goes up on the Y and Z.
I've been using the rail at least 2 months and greased it, so far 0 maintenance. Only issue from the factory was the M3 screw holes.
That's what I was thinking. You never know what they're using for assembly lube on the factory floor. Best to grease then go. Do that with linear bearings as well.
What was up with the screw holes? Just need a chase?
I drilled them out to M4 because I didn't have M3 T-nuts.
The wheels used in that kit is perfect for shower sliding doors, not precision work. The only way to fix this is by replacing the extrusions with V-slot and using V-slot wheels. The gotcha with doing this, other than the expense, is that none of the acrylic parts will fit as the spacing for the wheels are different so it is not a trivial change.
And not a cheap one either. I've done the math and sourcing parts locally (Canada) the cost difference between v-slot and MGN rails isn't that great. When you factor in v-slot having issues similar to these "R" wheels I don't see that as even being an option to consider.
Reading around on the Open Builds forums I see they recommend shims to get the right amount of compression on the bearing. Incidentally the magic preload seems to be zero for v-slot which makes me wonder if allowing these R wheels just the slightest amount of play might solve the issue.
Hoping I can sort this out as the z rails on a corexy cube build would be vastly cheaper going the wheeled route over even rod and bearings. Eventually both these i3's are going to be cannibalized for cube parts.
I replaced the X carriage with V-Slot and V-Wheels and the openrail gantry. It was all great at first but now I'm having issues. Maybe I overtightened the wheels. I would go with rails if I had it to do all over again but I was seeing some people had to wind up replacing all the bearings on the cheap rails.
I've read of both issues elsewhere. Sucks because there aren't great options no matter how you cut it unless you're really willing to dump some cash.
What I find most odd is that when I finished the x-carriage prints I test fit it with wheels that didn't come with the i3 Plus, they were a set that FLSun had sent me separately for my other i3, and it seemed pretty smooth. I ended up just putting those wheels with the others for assembly and now have carriages that aren't a smooth as before.
Going to have to build a little test rig and just select the best 9 out of what I have and hope for the best. Ordered a lot of PA wheels as well, see if they're any better. Not ready to give up on this, just need to design around it.
Yeah.. managed to get my homebrew CNC machine up and running so I am planning on milling an aluminum all in one openrail gantry and titan aero mount... if I still have issues at that point I can pretty much guarantee it's the wheels. I like the idea of a test rig.. maybe something as simple as securely attaching a mechanical pencil or awl to the hot end and scratching/drawing a line on something. At least then you're eliminating anything going on with the extruder and filament. From my experience even if the wheels are good to begin with they don't always stay that way or you can still have problem with good wheels if the pressure on the rail is not adjusted correctly.
Been doing a little research on CR-10's/Ender 3/Tevo's since they are common and am seeing quite a few people complaining about V-slot wheels. One material is too soft and develops flat spots, the other too brittle. I think the best route is to design around it.
Yeah I think I overtightened mine and got flat spots so now I get repetitiive waves in the horizontal. The wheels that came with the printer actually seemed to work better for me. I ordered a linear rail from aliiexpress for $20 so maybe I will give that a shot. I think it will require a longer 2020 rail so I ordered one of those too. I like this design it could be adapted to our printer: https://www.thingiverse.com/thing:1733006
What I'm thinking is a dial indicator and mount. Won't give me the complete picture, but at least an idea.
Pressure on my carriages is being provided by a printed-in spring arm, similar to how some delta carriages are designed. There's just enough pressure on their own they don't need the tension screws for now.
FYI for anyone running into this issue.. spool3d.ca sells polycarbonate V-wheels you can use with skate bearings. I think they would probably solve the issue. I've already started off on the linear rail route though. I'm thinking cutting the bearing and motor mounts out of sheet aluminum or buying the openrail plates will be a better solution than anything I can print.