New CoreXY 400x400x560mm Build in Progress

Fickert

So as the title states I am in the works of building a design I have been working on for the past few months.

I first want to say I am typing this on a tablet so please bear with me if I have grammar errors.

Anywho, the start of this build was to convert my FT-5 to a corexy system but no design publically released addressed the issues I see in most corexy designs, Rigidity.

I also HATE the build plate platform design on the FT-5. Linear rods and bearings are used to keep it straight and flat, but fail doing so, as a matter of fact, I can tilt the bed with just light pressure on either the front or rear of the bed.

Lastly, I wanted a larger build area. 300 x 300 x 400 isn't enough, escpecially since the realistic build area I have found is 280 x 280mm.

The FT-5 did it right with using the mgn linear rails, and I am carrying that over to this design as well.

So Goals for this build:

Rigidity: Idler carriages and idler posts for the corexy system need to be solid. Will be solved with using aluminum square tubing and milled to size.
Lightweight: I want to create some inovation by making the gantry the lightest possible while being direct drive. Additionally, the idler carriages to also be light by using aluminum.
Bulletproof Build Platform: Having an unlevel build plate regardless of a BLTouch, is a pain. Aside from being level the build plate NEEDS to be repeatable.
Build Area: I wanted to get as close as possible to a TRUE 400mm x 400mm square size with at least a 500mm build height.

So to solve a these issues:

Rigidity: A lot of factors play into the rigidity of the design. One big factor is the aluminum tubing. Others are:
1. Aluminum is always going to beat any printed part in rigidity and weight.
2. Aluminumsquare tube will support the idler shafts for the bearings on both sides. A lot of designs only support the bearing shaft by one side, that can result in slop in belt tension during directional change.
3. Any printed part will be made from Carbon Fiber PC, which is a very rigid material; if you can print PC I highly recommend it for builds over ABS, PETG, Etc.
4. The frame is pretty similar to the FT-5, there are a few more extrusions but I think it will be plenty study, especially with using 90 degree cast aluminum braces and likely external braces.
Idler Bearing Size: I stumbled across an article from a builder that had a direct drive system printing at 150mm/s and higher https://drmrehorst.blogspot.com/2017/07/ummd-corexy-mechanism.html and he swore by using 608 bearings (dia of 22mm was optimal for skewing of the position of the head or something) so I wanted these to be used in the design.
*Can't remember now off the top of my head will get the link and post when I get to work tomorrow.
Lightweight: When I say lightweight I mainly mean mainly the gantry, carriage idlers, and x-axis.
1. Gantry: I am swapping to a nema 14 round stepper and a single 4010 blower fan (may opt for 2 fans, will test cooling after it is assembled), and the stepper motor alone shaves 100g from the standard nema 17 pancake stepper motor for the titan aero.
2. I opted for a MGN9N rail for the x-axis. This shaves off a fair amount of weight to throw back and forth on the y-axis. I do have concern of harmonics on the rail at high speeds but there isn't a really good way to check this until its up and running.[link text](link url)
Bulletproof Build Platform: I mentioned the FT-5 build platform in my opinion is garbage. There are quite a few key features in my build plate design that I feel will make this rock solid:
1. The base will be made of 2020 aluminum extrusion with 90 degree cast aluminum corner braces and Carbon Fiber PC braces for flatness and rigidity.
2. Linear motion will be guided by two 600mm MGN 12 rails, which GREATLY reduce the slop the stock bed has in the FT-5.
  The photo below does not have the lead screws in yet but you can see the three supports off the bed platform.
3. There will be 3 Leadscrews to fully fix this build plane, and these will all be tied to a single stepper motor.
4. The build plate I splurged on a bit. I ended up buying a 1/4in thick 6061 aluminum plate measuring 415mm x 415mm. The plate has been ground with a flatness tolerance of +- 0.38mm. Should be pretty good even if I decide to not use glass.
Build Area: The build area was a huge "Gotta have" thing for this build to help drive keeping everything to be at the absolute maxed position to keep clear from the gantry. My main focus like I stated was the square flat area of 400mm x 400mm, while having a direct drive extruder. To achieve this, I eliminated the BL Touch and will be using stall homing (on all axis actually) as well as a change in stepper motor. The round Nema 14 stepper motor's lightweight is one large positive, but its form factor was the other. The total footprint of the nema14 stepper and the 4010 fan are smaller in depth than the nema 17 pancake stepper, and still in the same 40mm x 40mm form factor.

Electronics: Obviously this will be using the Duet Wifi, but additionaly it will have all the following:

Controller: Duet Wifi
PSU: MeanWell SE24 600Watt
Stepper Motors: Standard nema17 stepper motors from the FT-5
Heat Bed: 120V AC 1000W 400mm x 400mm silcone bed heater controlled by a SSR with heatsink
Extruder: E3D Titan Aero; not sure if I will be using the volcano or v6 hotend yet.

Still to do:

I still have to design a few parts yet, but with my parts coming in the mail and seeing my frame take shape I wanted to start to document the build for others.

fcwilt

@fickert said in New CoreXY 400x400x560mm Build in Progress:

+- 0.38mm

Hi,

Thanks for taking the time to post your progress.

About the tolerance on the plate - is +/- 0.38mm correct? That's 2 to 4 layers for the slicer settings I commonly use.

That seems a lot although I have nothing to reference it against.

Frederick

Fickert

Yes that is correct. I say that because a lot of the aluminum plate that is used for head beds I believe has this tolerance or more.

I beleive this is still considered more flat than most large pcb beds too.

And you could get a more fine tolerane but that would have made the tolerance skyrocket into the $200+ price range.

I haven't had the chance to check the flatness on our granite block at work. Will look into that tomorrow.

fcwilt

@fickert said in New CoreXY 400x400x560mm Build in Progress:

I haven't had the chance to check the flatness on our granite block at work. Will look into that tomorrow.

That would be interesting.

Frederick

Stephen6309

@fickert You may want to put some diagonal bracing on the side to keep the printer from twisting. On my corexy i made, I drilled access holes and tapped the extrusion ends to screw them together to not use any angle brackets.

Fickert

@stephen6309 said in New CoreXY 400x400x560mm Build in Progress:

@fickert You may want to put some diagonal bracing on the side to keep the printer from twisting. On my corexy i made, I drilled access holes and tapped the extrusion ends to screw them together to not use any angle brackets.

I did think about this. Probably not a bad call at all. I was even thinking of doing some diagnal cable bracing as an alternative. Similar to pull barn structuring. Figure it's something I can add down the line.

Fickert

@fcwilt said in New CoreXY 400x400x560mm Build in Progress:

@fickert said in New CoreXY 400x400x560mm Build in Progress:

I haven't had the chance to check the flatness on our granite block at work. Will look into that tomorrow.

That would be interesting.

Frederick

Just an update Frederick,

was able to check the flatness of the plate on our granite. About the flattest possible surface to check on (hence it was 30k for us to get made) and attached are my results with our feeler gauges.

from my findings, it looks like the flatness tolerance is between 0.10 - 0.15mm (0.10mm was slightly loose, and 0.15mm was snug) from each furthest corner. Considering thats spanning over 415mm square, I am very pleased with this tolerance. I am curious how close a sheet of glass would be to this.

Anywho, this helps me relieve that this is pretty darn flat

timcurtis67

The piece of jig plate i used on my printer was out .007" total for flatness. I bought it from the same place as you did a few years ago. It is an 18" square plate.

I would say .004"-.006" of flatness is great. You won't need to use any kind of bed mess compensation. Just get it indicated nice a flat with respect to the travels of the hotend.

It should stay pretty stable when heated too.

Fickert

@timcurtis67 said in New CoreXY 400x400x560mm Build in Progress:

The piece of jig plate i used on my printer was out .007" total for flatness. I bought it from the same place as you did a few years ago. It is an 18" square plate.

I would say .004"-.006" of flatness is great. You won't need to use any kind of bed mess compensation. Just get it indicated nice a flat with respect to the travels of the hotend.

It should stay pretty stable when heated too.

Thanks, that is good to know!

I was pretty impressed with this company, especially since this plate only ran me 62 bucks shipped. Shipping alone was 18 dollars.

fcwilt

@fickert said in New CoreXY 400x400x560mm Build in Progress:

I was pretty impressed with this company, especially since this plate only ran me 62 bucks shipped. Shipping alone was 18 dollars.

Did you say that for more money they could provide a piece with close tolerances?

Thanks.

Frederick

timcurtis67

@fcwilt said in New CoreXY 400x400x560mm Build in Progress:

@fickert said in New CoreXY 400x400x560mm Build in Progress:

I was pretty impressed with this company, especially since this plate only ran me 62 bucks shipped. Shipping alone was 18 dollars.

Did you say that for more money they could provide a piece with close tolerances?

Thanks.

Frederick

I don't think they have any better pieces for sale. To get the piece flatter you would have to either machine it which wouldn't be practical or mount it in such a way that you could dial the flatness into it. maybe with screws pushing it from the back. That would make mounting a heater difficult though. .006" over 18" is plenty good enough for a 3D printer.

Fickert

@fcwilt said in New CoreXY 400x400x560mm Build in Progress:

@fickert said in New CoreXY 400x400x560mm Build in Progress:

I was pretty impressed with this company, especially since this plate only ran me 62 bucks shipped. Shipping alone was 18 dollars.

Did you say that for more money they could provide a piece with close tolerances?

Thanks.

Frederick

Correct, I do not think they offer anything with higher tolerance. I was just impressed as similar cuts, I was seeing quotes of $200+ so this was very easy to justify in comparison.

Phaedrux

200$ sounds about right for a large plate of that size with tight tolerances.

One note with regards to your intention of using stall detection for homing on all axis. Don't do it. Your planned setup checks all the boxes for it to not work reliably. CoreXY is tricky because of two motors working at the same time. And on the Z axis you'll likely end up using some gearing between the motor and lead screws to increase torque. This will make it hard to tune the stall detection limit. The repeatability will also vary with motor tempm which can become maddening.

It will be far easier and reliable to just use simple switches. And on the Z axis having a probe to get a height map to verify the flatness and do compensation if necessary will be almost necessary given the large size of the bed and unknown tolerances.

Given the size of your build area I would recommend going with the volcano and a larger nozzle since if you're planning on printing large things you probably don't want them to take forever.

The Titan Aero is a great performing compact direct drive extruder, but it is also a tad finicky. Do a search on this forum to get a sense of what I mean.

But that's just my take on it. Your build is looking great so far. Can't wait to see how it progresses.

Fickert

@phaedrux said in New CoreXY 400x400x560mm Build in Progress:

200$ sounds about right for a large plate of that size with tight tolerances.

One note with regards to your intention of using stall detection for homing on all axis. Don't do it. Your planned setup checks all the boxes for it to not work reliably. CoreXY is tricky because of two motors working at the same time. And on the Z axis you'll likely end up using some gearing between the motor and lead screws to increase torque. This will make it hard to tune the stall detection limit. The repeatability will also vary with motor tempm which can become maddening.

It will be far easier and reliable to just use simple switches. And on the Z axis having a probe to get a height map to verify the flatness and do compensation if necessary will be almost necessary given the large size of the bed and unknown tolerances.

Given the size of your build area I would recommend going with the volcano and a larger nozzle since if you're planning on printing large things you probably don't want them to take forever.

The Titan Aero is a great performing compact direct drive extruder, but it is also a tad finicky. Do a search on this forum to get a sense of what I mean.

But that's just my take on it. Your build is looking great so far. Can't wait to see how it progresses.

Interesting take on homing with the stall detection. Even if I decide to home X and Y axis one at a time?

Z axis I am willing to give it a shot, I have been following the stall detection on a few threads and seems that once it is dialed right, the repeat ability is well within reason to use it. But I could run into issues using 3 lead screws and a heavy build platform. Worst Case I could try to setup a 3 - endstop system to level the bed and hope to never need mesh leveling.

We just got in the Raise3D Pro2 Plus and it comes with a pre-calibrated build plate and does not have a probe. So it is possible, but I am not sure if it is doable.

Thanks!

Phaedrux

@fickert I didn't have great success with stall detection on the Z axis with 3x 1mm pitch lead screws and a single 1.8 deg motor with 1:1 gearing, but you may have better luck with your setup. To help improve the resolution of the stall I switched to 0.9 deg motor but ended up losing too much torque to reliably lift my heavy bed platform. So I switched to 2:1 gearing, which got me more torque, but now made the sensitivity too weird and would seem to swing wildly with motor temp. In the end I switched back to the 1.8 deg motor, kept the 2:1 gearing and gave up on stall detection and use a BLTouch instead.

Even on a very flat plate you can find some warping. You won't know how it will turn out until you try and print on it.

Regardless, I think it would be something to setup after everything is already up and running as an improvement since the setup and tuning and testing is very time consuming and will likely involve a few head crashes.

Stall detection as endstop on X and Y in a CoreXY has been done (do a search to see what others have gone through to get there), but it is highly dependent on the motor.

I don't mean to sound down on stall detection as an endstop. It can be made to work, but when a simple switch is so cheap and reliable it's hard to see why you'd want to use anything else. Either way, stall detection while printing is totally separate and can work quite well, though it is still subject to the same motor sensitivity requirements, so your mileage may vary.

Fickert

This post is deleted!

Fickert

Quick update, got the two front lead screws installed. The KP08's for the rear lead screw are still in transit, but that rear lead screw does not need cut down. The front two needed cut to 575mm to slip into place. Sadly I forgot to loop in my closed loop gt2 belt so I will need to uninstall them just to get the belt in there.

Hoping to finish machining the Carriage Idlers and finish the rear idler design to print by next Monday.

mrehorstdmd

A few comments...

you checked bed flatness against your granite slab by slipping a feeler gauge under the edge of the plate while it sat on the slab. Was the protective plastic peeled off the side of the plate that was on the granite? By laying it on the granite, you "fully supported" the plate which is not what you'll be doing in your printer, so your measurement doesn't account for sag in the plate when it's supported by 3 levelers. If you really want to measure it in a meaningful way, put the plate on the bed support in the printer, tram it, and mount a gauge on the extruder carriage. Sweep the gauge over the surface and see what sort of range you get. That measurement will take into account sag in the X and Y axes, sag in the bed, thickness variations in the print surface, etc., all of which will determine whether you're going to get the first print layer to stick.
2020 extrusions seems pretty light duty to me for a printer this size. The machine is going to be throwing the X axis back and forth at pretty high speeds/accelerations and you can expect there to be some fore/aft movement in the frame. Bolting on rigid side panels should help. This matters because at the start of the print the bed will tend to move back and forth with the printer's frame, but as the print gets taller and the bed moves down, it will move less, so the print will move less, and the print quality may degrade.
The 9 mm X axis guide rail will be light, but over the long span you may have problems with it sagging or wobbling when you start and stop motion in the Y direction.
The extruder stepper needs to have enough torque to push the filament. You can find lots of tiny steppers, but without adequate torque, your light weight extruder won't work. If you gear the motor down to multiply torque, you limit speed. Light weight and speed are overrated. Even if you get the mechanism to move at 500 mm/sec, the extruder (any extruder) won't be able to deliver the molten plastic the way you want. My printer, UMMD, is not designed to be particularly light weight, yet I have been able to run it with acceleration at 10k and speed at 200 mm/sec. The X axis weighs about 1.5 kg. It did a fair job of printing, but how often is a fair job good enough? The faster it runs, the more it shakes and the noisier it gets.
It's not a big deal, but you shouldn't need antibacklash nuts on the Z axis screws. Gravity will keep the nuts firmly engaged with the screw threads at all times. Antibacklash nuts increase friction which may require more torque from the Z axis motor, especially since you're turning 3 screws with one motor.
If you enclose the machine you might be able to print ABS and other plastics that are more useful than PLA. An enclosure will also help contain the particles released when printing.

Fickert

@mrehorstdmd

Thank you for getting to review my build.

I want to at least clear up what I can and hopefully pick at your brain a bit more.

the aluminum bed was measured on a granite slab, and the film was removed. I even deburred the edges to be sure no aluminum burr was left behind when they cut it (wasn't much if at all, but did it to be safe). I think your perspective on the plate goes for the whole assembly, as I can't recall now, but the be spanning 4 points in the corners it sags very little. Enough to disregard it, I can run it though simulation again in SW to confirm that. Partly why I went with 1/4in thick plate.
On another note that did give me a good idea, I should assemble the 2020 extrusion bed ontop the granite block to confirm it has complete flatness.
I have wondered if the 2020 would be too little for this build, I pretty much just based it off of the FT-5 design, assuming it was enough. I didn't think about it like you mentioned when the bed moves downward and the bed quits moving with the frame. I will be enclosing the printer eventually, hopefully can reuse the enclosure from my FT-5. If not then I will get some alumium corner braces.
The 9mm rail is a concern for my I agree, and I still have an issue with it, BUT I am wanting to see how much of an issue. I probably should tackle it now, but am willing to see what I will be sacrificing. Maybe I am too knuckle-headed to changed my design at this time.
I check the stepper motor with a titan aero and it can extrude up to 20mm/s and still can retract 40mm/s so I believe it will be fine, but we will see. I don't see anyone doing it and so I wanted to pioneer a bit. If I steer away from the titan, than yes I will for sure have to ditch the stepper.
Thats a good thought. I just put them in there as habit, I have always been told to use them. But I see your logic, and with this weight of this build platform I doubt it will move, even with a z-hop. Especially if I try to use stall homing.

Again thanks for coming over to help me out, I realize now I was mainly focused on your build using the sqare aluminum tubing, I thought it was genius, but kinda missed all the extra rigidy you did to the frame.

mrehorstdmd

I didn't see the XY stage in the CAD model, so it seems like you're working from the outside in. You may find that the XY mechanism that will fit in the space you've allowed for it won't reach you print envelope goal. When you have a print size goal, it's better to start by designing the mechanism, then build the enclosure around it.

If your goal to to minimize moving mass, using the F608 bearings I used may not be optimal. Those bearings are heavy, and have 8mm bore, so their axles are heavy, too. If you look around you might be able to find plastic or aluminum flanged pulleys that use smaller bearings and smaller axle hardware, while maintaining the relatively large diameter. A lot of printers use smaller diameter pulleys and seem to have adequate belt life and print quality.

Those tiny corner brackets don't really do much for rigidity compared to putting a sheet of plywood (ugh) or polycarbonate on the sides of the printer.

You might want to put leveling feet on the bottom of the machine. If the machine sits on an uneven surface the frame will twist to conform to that surface, especially if the machine is heavy. That twisting will distort the XY plane and could make it hard to tram the bed. UMMD usually sits on uneven concrete floors. When I move it to a new location I push on the frame and see if it rocks a little and if it does (almost always) I adjust the leveling feet to minimize the rocking.

UMMD's frame would have been a lot more rigid if I had screwed PC sheets for the side walls instead of fitting the dual layer PC into the frame's slots. I wasn't too concerned about rigidity (the 4040 extrusion is pretty good) and was more concerned with thermal insulation and appearance (I take it to makerfaires, etc).