Rebuild worn CoreXY, what options should I consider?

Hi,

my home brew CoreXY machine has been used a lot since I first built it in august 2017. I changed a lot in electronics and I went to a bowden cable, removing the extruder from the carriage, I added bed heating, changed motors, but the mechanics remained the same. Now after a lot of kilos of PLA, the rods are rusted, the bearings are worn and squeeky, it is time to service the machine. Here it is in better days:
Link to Youtube clip
Another one

Now the easy route would be to buy new rods and bearings, and be done with it. But I also see the carriages running on wheels on the 2020 profiles, nowadays. My machine is already built from 2020 and 2040 profiles, I could go to the wheels that fit those profiles. I also found this clip of a sliding mechanism I don't yet quite understand, but it is interesting:
Digital dentist's sand table
It seems a simple teflon bearing block in the 2020 profile, does that suffice for a 3D printing CoreXY?

Please let me know what route you would go on this rebuild, prefferably based on your own experience.

Thanks,

Hugo

Edited 1 time(s). Last edit at 11/02/2021 09:07AM by HugoW.

After I built my first printer, MegaMax, I kept a running list of all the problems I ran into as I used it. One of the things I wanted to do was to print ABS, but at the time I built the printer, I didn't know I would need a warm enclosure for the machine. So armed with the list of all the problems, I started making a list of changes that would address all those problems. When I was done I started on a redesign that ended up as Son of MegaMax. I suggest you consider a similar approach. What doesn't the machine do that you want it to? What sort of changes will make it better at what it does?

The PTFE sliding bearings used in the sand table are not suitable for a 3D printer. They are fine for a sand table, but they do not provide the sort of precise, accurate motion that is needed to produce a high quality print with a nozzle that is a fraction of a mm in diameter. The best linear bearings are linear guides, but these days there are a lot of crappy ones being sold cheaply to 3D printer hobbyists. Don't buy them. Buy surplus industrial parts for about the same price via ebay.

I suggest that you redesign for fully enclosed printing. Even if you print with PLA, full enclosure will yield better prints. That means don't put the XY mechanism on top of the printer's frame, and don't use wheels running on the t-slot frame for linear bearings. 2020 t-slot is pretty flexible stuff, so you can make the frame more rigid by bolting panels- metal, sheet plastic like polycarbonate, or plywood- to it. If you use wheels running on the t-slot, you probably won't be able to bolt on panels to enclose the machine.

You used a bunch of toothed 3D printer pulleys. Those things have tiny bearings that don't last long. You might want to consider using stacked ball bearings because they will last longer. Also, it's a good idea to support both ends of the pulley axes instead of standing them up like fence posts. Belt tension will cause the post-type bearing axis to flex and that will lead to problems with the belts tracking on the pulleys. Stacked bearings means smooth surface for the belt, so unless you use larger bearings like F608, you'll probably want to put a twist in each belt so that the smooth side of the belt contacts the smooth pulley surfaces. F625 bearings are pretty good, smaller alternative to F608. They need a 5 mm axle instead of the heavier 8mm axle of the F608s.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Thanks for all the input. In the same order:

All I want to improve for future wishes are a full metal heatbreak (actually I am looking at the bimetal ones), allowing me to print NyLon. The only reason to change the concept would be to make it more accurate and/or lighten the X-axle and head for the same reason. So far not real changes have been necessary.

I was hoping I missed some important detail making them useable in 3D printing, but it is as I feared. OK, no PTFE slide bearings.

I have a (foldable) enclosure I can place around the printer, with a lid. It is separate from the printer, very cheap and easy, but it works. I used it when printing ABS, I might reinstate it for other use. It is fully separate from the printer frame.

Yeah, that area is certainly up for improvement. It was easy to make when I didn't have a 3D printer handy but now that I do, I can make better supports. The F625 bearing with printed flanges I will use and I am looking at corner pieces like the ones in your sand table. I might just stick to the current contruction, just buy better linear bearings and rods. These were all AliExpress, I am amazed they lasted as long as they did, but there is a lot of room for improvement. Just lowering the X-axle weight is something I want, despite the fact that I don't really have a reason to want it.

Cheers,

Hugo

In the sand table I used printed flanges to help keep the belts on the pulleys over the very large size of the table. The printed tires on the bearings were used to push the flanges outward against the bearing flanges. I don't think you need either the wider flanges or the printed tires for a 300x300 mm bed printer. I needed them in the sand table because of its large dimensions. In UMMD doesn't have any extra flanges or tires and the belts stay where they are supposed to.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Thoughts on your rebuild

1. I wouldn't use wheels on 2020 extrusion. Although OK at first, they wear a lot quicker than linear bearings. They also take up a lot of space so require compromises in the frame design.
2. Depending on how much you want to redesign/upgrade I'd consider using linear rails rather than linear rods for the CoreXY mechanism. The print quality from well-adjusted linear rods or linear rails is the same, but using linear rails makes a neater and more compact design - its easier to attach linear rails than rods and the X and Y carriages using linear rails are simpler. (Note for the Z-axis the tradeoffs aren't so clear).
3. I concur with the digital dentist that redesigning for a fully enclosed frame is a good idea. This is much more compact and convenient than placing an enclosure around the printer. It can also add rigidity to the printer.

You might want to take a look at MaybeCube, a 3D printer I have designed and built, for ideas. The description of the printer and its design goals are here: [github.com]
The assembly instructions (so you can see how it all fits together) is here: [github.com]

Quote
the_digital_dentist
The PTFE sliding bearings used in the sand table are not suitable for a 3D printer. They are fine for a sand table, but they do not provide the sort of precise, accurate motion that is needed to produce a high quality print with a nozzle that is a fraction of a mm in diameter. The best linear bearings are linear guides, but these days there are a lot of crappy ones being sold cheaply to 3D printer hobbyists. Don't buy them. Buy surplus industrial parts for about the same price via ebay.

Yeah, the great thing about one-off machine designs is that you can use whatever rails you can find cheap. My mini CNC mill has mostly $10 Chinese rails, but I did get one high quality wide rail on ebay. The two 300mm cheap rails are bowed, which would be a problem for a 3D printer where the rail itself is used as the X gantry, but when screwed to a stiff frame it's no problem. The six 200-250mm rails were all perfectly straight, with the only problem being preload consistency. But after mixing and matching the slider blocks to whichever rail they fit best, they're just as good as the fancy rail.

SBR rails (round rod screwed to aluminum extrusion) are great too, because you can fiddle the set screws in the slider blocks to get the bearing preload just right, unlike full round bearings which have to be manufactured perfectly to begin with. Smart way to get high precision without high cost. The slider blocks should work fine on floating rails too. But 10mm is the smallest size available, so not the common 8mm rods used for 3D printers.

Thanks for all the input. I am seriously considering to use real good linear rails, not the cheap Chinese ones, but searching ebay doesn't ensure me the quality. I can buy them locally from a reputable dealer, but that would set me back hundreds of euro's. My printer is quite big (400 x 400 bed). I will look into the linked Maybe Cube, the SBR rails I am not a big fan of. Just personal prefference.

Or I could just rebuild with the same cheap Chinese stuff and print for another 3 years. But somehow this rebuild feels like I need to improve.

Cheers,

Hugo

When you're buying linear guides via ebay, look for quality brands Thomson, THK, IKO, Bosch, etc. I avoid parts with "MGN" or "MGH" part numbers- the Taiwanese company (HIWIN) that started (?) that numbering scheme makes quality guides, but the vast majority are cheap knock-offs. Don't worry about the exact length of the rails- you can cut them down using a cutoff wheel on a grinder. Make sure you get enough bearing blocks with them- buying blocks without the rails is expensive and may be difficult to fit properly. It's probably best to get guides with light or no preload.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Look for name brand rails like IKO, THK, NSK, Misumi. You may not be able to find any long enough for a 400mm bed, though. Especially not under $100 each.

A 400 x 400 bed is very large. Unless you really need that size, I'd consider a smaller printer. A smaller bed heats up more quickly, and all other things being equal a smaller printer will be more rigid. I have 3 printers in regular use,
the smallest has a build volume of 90 x 80 x 60mm and I print most of my stuff on that.

If you really need the size, then I would consider doing the minimum update of your current printer for large prints, and building a new smaller printer.

The thing about cheap chinese linear rails is that they are cheap - much much cheaper than name brands. Part of the reason they are cheap is the finishing - they are not cut precisely to length (it looks like they are cut by hand using an angle grinder), and they often are covered in swarf and then bathed in oil. But if you take the trouble to clean them up, re-oil them, and clean and re-oil the carriages they are good enough for 3D printing.

Of course it is better to use name brands, but if budget it tight cheap linear rails are adequate.

Thanks for the input. There is a reason I built the large bed, I use it. I'll take the cheap rails plus effort option into consideration.

Cheers,

Hugo

One option would be to keep the Y rods like you have them but switch to SBR blocks for the adjustability, and buy a single high quality linear rail for the X axis since that's the one that really makes a difference in terms of moving weight.

Though it should be noted that rails rust much more easily than chrome plated rods, so if you're already having rust problems then you may be better off rebuilding it as-is.

Another nice thing about floating rods is that since the bearing balls run in straight lines along them, if the rod begins to wear, you can rotate it a bit to present fresh surface to the bearings.