Cantilever printer design based on leftovers

Status update:

I bought some NEMA motor mounts with the idea that it would save me some time. I may have spent more time altering them to fit than I would have fabbing from scratch from square tube, and I forgot to adapt the model for the slightly thinner metal (about 1mm) and it left me with Z screw alignment issues that took a while, and some shimming, to resolve. I’ll know whether I got it perfect when I install the Z motor.

I needed a stop to make sure I don’t run the Z nut off the top of the Z screw, so I made a knurled knob that will let me adjust Z when the printer is off that also functions as a stop. :slight_smile:

I used a crosscut saw with a good carbide blade to cut the v-slot. It made nice clean square cuts, and the frame assembled square with no “convincing” or shimming; I can’t see light shining through when I hold my machinists square to the corners. My youngest drilled all the base frame holes. I showed him how to set up a jig on the mill, and the square fit shows that he did a good, careful job.

I should have included spacers in the design between the Y carriage and the “frog” to leave more room for bed adjustment screws. I’m using some 6mm openbuilds spacers right now to give myself a little more space. I’ll see if that’s enough; if it isn’t I’ll have to source longer screws as that has me using the longest flathead M3 screws I own, countersunk as deep as I dare.

I decided not to cut the tower piece down to 500mm. It’s tall, and I can hang the filament spools on it, above the print area. I gave myself about 15mm extra room on X and Y vs. what I modeled in case I forgot something, and it made it possible for me to put corner brackets inside the frame. I didn’t cut any of the 450mm MGN12 rail, so it sticks about 20mm behind the printer, so in retrospect I should just have designed it to be 450mm long. I also made the X beam longer than needed, because I used the offcut from cutting the shorter frame pieces out of a 1m section and didn’t cut it further. That way if I need to slide it around (say, to accommodate a second tower as discussed above) I don’t have to trash it and start over.

I didn’t model how to attach the tower to the side before I started. I used a piece of square bracket, screwed to t-nuts in the frame v-slot (the thick slide-in kind, not drop-in) and to a custom double t-nut in the 4040 t-slot that I’m using for the tower. Both use M5. I put a few extra T-nuts in the frame before putting loctite on the screws to hold it together. I’ll see whether that’s sturdy enough or not.

Not yet done, at least:

  • Electronics: nothing done yet; no wires, no routing, not even much planning, no electronics enclosure design
  • Mounting any motors
  • Fabrication:
    • Y idler pulley tensioner mount
    • Y belt clamp for saddle
    • X idler pulley rod
    • Extruder mount
    • Heated bed (planning to use an old heated bed as a heat source for a new cast aluminum bed plate)
3 Likes

My plan is to use the bed from my old hictop gantry printer as the heating element, at least at first. I want to hold it to the cast aluminum plate. That bed is marked for 12V and 24V but 24V is probably about not overwhelming the power supply, and of course the alternative is to use a a more powerful 24V supply and get 4x the power output.

I’ve been using that old gantry printer’s heat bed for mask sterilization. I set the bed to 75⁰ and put the mask on the bed, under insulation, for about an hour (Stanford protocol is 70⁰C for 30 minutes, and it takes a while to heat up, and the insulation isn’t perfect). But today, I used polyisocyanurate board to make a filament drying oven over the old bed and control board from my tronxy, so my heater is available.

It was bittersweet to tear down my first ever 3d printer for spare parts, but it’s done now, and no going back. I was a little low on M3 hardware, but not any more!

I finally realized that the heating element bed does not fit my plans for a kinematic mount as is; it is too big and I can’t mill through it and have it still work. So instead, I’ll mill kinematic mount features in small pieces of aluminum and glue then to the heating element with high temperature RTV. Then if I get a different heating element, I can switch to milling kinematic mount features in the plate itself.

Today I tried to buy RTV high temperature silicone because the tube I already had was dead, but the particular home depot I went to happened to be out. So I ordered from Amazon instead, but it will take a week to get here. Maybe I’ll get impatient and go shopping again.

I didn’t make much obvious progress on the printer itself today. I made the X idler bearing pin. That enabled me to discover that the X belt mount arrangement I had come up with will rub the belt, so I have to make a new part that’s a little more compact. Not hard to make, just a little frustrating to have to start over on it. My measurements were with a different belt mount (printed in ABS), that turned out not to be quite strong enough. I have yet another design, I just need to spend a little more time milling out a new part.

3 Likes

Turns out that moving the motor back a bit for the Z lift plate fix also threw off my belt alignment calculations, which were imperfect to start with, so I had to take 2.5mm off the bottom front of the X carriage mounting block to avoid rubbing the belt. It’s not perfectly aligned end-to-end so it won’t run true, but since the carriage won’t get closer than about 40mm to the motor pulley and it’s off by about 1mm, the cosine error won’t be significant. I think.

I attached the Z stage motor, which finally set my minimum Z height. I have about 35mm between the top of the “frog” and the end of the nozzle. If I end up wanting a little more, I can slide the Z stage higher.

2 Likes

Finally came up with a design for a Y idler tensioner that I’m happy with; a relatively simple hinge.

  • The block at the top is screwed into a T-nut in the top slot with an M5 screw
  • An M3 screw forms the hinge pin through the top of the arms and the block
  • An M5 screw through the middle of the arms holds the idler pulley
  • An M5 screw in the bottom of the fork frame screws into a T-nut in the bottom slot and tensions the idler pulley.

It’s sized so that I can upgrade from the 6mm belt I have now to proper 9mm Gates belt now that I know where I can get that conveniently (Belt, for $1.50/100mm idler for $7 and pulley for $6) — 6mm belt for the X is fine, that doesn’t have much mass, but a thicker belt for the bed might turn out to be a good idea so I’d like to leave room for it.

I think I’ll fab that tensioner, then cut the belt, and drill the saddle (red, next to it in the model) wherever makes the belt run flat without trying to model it more thoroughly.

1 Like

Of course, this other project used up the 24V power supply that I had intended to use for this printer, and my remaining supplies are 12, 36, or 48V, so I had to order another 24V power supply. I’ll just consider power supplies fungible and think of the one I ordered today as having been bought for my drying oven, not as breaking my “leftovers” design constraint further. :smiling_face: Sadly, the MeanWell 24V 350W supplies don’t seem to have a thermostatically-controlled fan.

1 Like

@Eclsnowman since you are using an SKR mini E3 v1.2 with at least one printer, did you get the TFT35-E3 as well? If so, did you design an enclosure for the pair? Right now, I have the pair sitting on the table next to my printer and it’s kinda ghetto, and while I could design a case I figured if I can cadge a case design off a friend why not? :wink:

1 Like

I have an enclosure design for the board that could be modified pretty easily. I have one of the minis on my ender 3, and I thought I messed up the boot loader so I bought a replacement board but then figured out the issue was with flashing it a firmware above 256k without using the 512k hack. So I ended up with a second one and replaced the controller board on my Talos3D Spartan. So that’s the one I designed the electronics enclosure for, cuz the under 3 already has the standard enclosure.

I don’t have that particular TFT, I just did the wiring hack to hook the smaller tft control to it. But to be honest I like the interface better of the Marlin non-touch screen. So I almost always run the screen in Marlin emulation mode. Instead of the standard touch interface.

1 Like

Give me some time here tonight and I’ll get the files together for you, what format would you prefer them in for modifying?

As usual, you do beautiful designs. :slight_smile:

I also use marlin mode instead of touch mode for pretty much everything except typing gcode into the integrated terminal, thought I might be the only one…

That design won’t work for my corexy, but it sure will for the new cantilever printer which is nearing functionality :crossed_fingers: so that would be awesome! I would have to modify for the TFT35.

As of recently, I ended up installing Fusion360 on the windows machine, and I’m an EAA member, so SolidWorks is also free for, me so I have that installed too. So probably whichever is native is most likely to be useful. I keep working in openscad and FreeCAD because I hate lugging out the windows machine when I’m overall so comfortable in Linux. :slight_smile: So I’m pretty much the confused user in Fusion360 and SolidWorks still. Therefore, I might have some learning to do to modify the front for the TFT35.

1 Like

Here is a google drive download for the files. I have them in Soldiworks 2018, as well as Step, and X_T. I also had been looking at the TFT35 V3 before, so I had a model of that and chucked it in there for good measure. Should make designing a case pretty easy. If you want any help on that front let me know because I am pretty quick at knocking things like that out. Only thing to watch for is I didn’t make that TFT35 model so an enclosure for it might take a test run to confirm all that modelers positions were correct.

Here is the link for the downloads: https://drive.google.com/drive/folders/1vlIBAw07zbvf0dH1GCBH8gQeaQIjFoda?usp=sharing

And here are some pictures of the model I had found for the TFT35, it looks like good work by who did it:

1 Like

Thank you!

I am discovering in practice what I knew in theory: That some OpenSCAD and FreeCAD experience doesn’t make SolidWorks obvious to me. I’m guessing I’d need to go a long way down this rabbit hole to learn what I’m doing. I’ve literally only opened a few files and poked around before, never even tried to edit anything.

I want to get there, but it’s slow going. I figured out how to the the TFT35 inserted into the front cover in place of the the smaller display. But lots of “The selected component is fixed, it cannot be moved” when I try to move features on the enclosure to make them match the TFT35 display. I’m guessing there’s a way to select features and move one to match another quickly.

I feel like “I’m not a doctor, I just play one on TV”, or maybe waking up unexpectedly in Portugal and discovering that a smattering of French and less of Spanish isn’t the same as being able to speak Portuguese. I’m clearly wielding a power tool while unsure which end is the business end! :rofl:

1 Like

Several of the features of the housing are designed using “in context” constraints to the assembly. So the features are actually tied to each other and their respective models within the assembly. It’s a top-down vs bottom-up modeling question. If you give me an quick sketch of where you want the display mounted on your printer I can wipe up a quick concept. I like these sort of projects.

That would be very kind!

What you have designed, without the power supply holder, and accommodating the larger TFT35 display, would be perfect for the cantilever printer! (I’ll tuck the power supply away somewhere else; it would get in the way of the X motor on the Z tower for my design.)

I hope to be less helpless when I have a chance to learn more. :slight_smile:

Do you have your current printer design in a form of Cad I could pull into SolidWorks? Like step or something similar? If so I could model the mounts for the LCD and the controller around that to make sure I’m not obstructing anything. Either that or even a dimensioned 2D front view and side view so I knew what heights to stay below (for example so I don’t have the LCD sticking above the bottom of the bed where it would come in contact with it.)

Oh, the model is only in OpenSCAD, so I’m not aware of a good way to do any import that would pull into SolidWorks. It can produce AMF, but I don’t see importing AMF in the solidworks help. I tried using FreeCAD to convert the OpenSCAD to STEP but no luck. If I’d designed it in SolidWorks I could probably just tell it to make drawings for me, but then I would also know how to modify it myself. :rofl:

I think I should just make a design that attaches to the side and doesn’t wrap around the front. It would give me more freedom to modify the printer later. :slight_smile:

1 Like

Fabbed this idler tensioner tonight. I ended up making it a bit bigger so that I could get more M3 screw engagement in the hinge block at the top. I used two M3 screws, one from each side, so that I can adjust the location of the hinge block without removing the screw, and I don’t think I’d have a long enough M3 anyway to use just one. I ended up with 20mm between the arms, far more than I need otherwise but also not a problem, I expect.

Silicone heated bed arrived earlier than previously predicted, so I won’t end up using the old aluminum heat bed as a heater after all. Not sure what I’ll do with it now, though! Maybe an enclosure heater for the corexy printer? I think I’ll still fab the kinematic mount features in small bits of aluminum and attach the to the bed with high-temp RTV as I’d been thinking of using for the heated bed, because that lets me replace them easily if I screw up.

Still remaining:

  • Electronics: still pretty much nothing done, but the new 24V supply arrived to replace the one I am using for the drying oven.
  • Heated bed (120V 500W silicone heater, high-temp RTV, SSR, TCO, cast aluminum plate, kinematic mount fixtures)
  • Y belt clamp for saddle
  • Extruder mount
2 Likes

This morning I got the Y belt clamp set up. I had intended to attach the top side of the belt to the middle of the saddle, but then realized that the adjustment screw would be inaccessible. I then discovered that with a clamp piece I had already made (a test piece for the X axis clamp) it was aligned already with the bottom of the saddle, so all I needed was to screw an extension to the bottom of the saddle and fix the belt to that on the bottom side. It will reverse motor direction for Y, which is fine.

The tensioner worked great. I was able to slide everything around so that everything is square and it runs true. The clamps are wide enough to move to 9mm belt later if this 6mm belt isn’t sufficient.

I installed X and Y stops. The Y stops are 9mm openbuilds spacers fixed with M5 screws and drop-in T-nuts that the saddle hits at the travel limits. The X stops are more interesting, M5x30 screws with locknuts screwed on to leave 7mm of travel, then an M5 washer, then a slide-in T-nut screwed into a slot and which the hotend mount hits at the limits of travel.

I’m so glad I cut the frame a little oversize just in case I hadn’t thought things through well enough. I’d missed something in my calculations and padding Y by 15mm gives me about 2mm extra room for bed travel in Y. That came close to being really annoying.

I am also making some feet, because I realized I needed to hold the frame off the table to run wires under it. I lathed them out of 20mm PE round bar stock; 20mm long, 5.5mm through-hole for M5 bolt, 10mm countersink for M5 bolt head 14mm deep, M5x12 bolt into slide-in T-nut in frame. After a day of drying out my ninjaflex, I will print TPU inserts (“shoes”?) that fit into the 10mm counter-sunk holes as the bottoms of the feet. Without the “shoes,” here they are on the corner. I like the flush frame mount that works from a good square cut!

I think I’m going to run electronics and try the printer out with wires run through wire loom before I commit to drag chain installation parameters; I’m not 100% on routing yet.

This means that the remaining tasks before first print are

  • Design, fab, and attach an extruder mount.
  • (Re)-design bed mounts; fab, assemble, and mount heated bed
  • Design and fab limit switch mounting blocks and flags; attach them
  • Wire up all the electronics

Before that’s all done, I should have the ninjaflex dried and the “shoes” printed and ready to use.

Here’s what it looks like now:

4 Likes

Okay, now that I see the layout and the surrounding equipment I can make up and LCD mount and control box. I didn’t see the pictures earlier when I saw this latest post. Just to confirm, the base is 2060 extrusion, and you’d like the LCD to the left instead of in front of the front extrusion, and the control box along the left hand side? Also would you like the electronics closer to the z pillar?

2 Likes

Yeah, the frame is 2060 extrusion mounted on 20mm feet (which will soon have another 4mm or so of TPU under them, in a few hours after my marathon print of fittings for the enclosure for the corexy printer finishes, which will be the tallest print I’ve ever made).

There weren’t pictures earlier; I am so ashamed of my disaster of a shop, where I am halfway through so many projects that I sometimes can’t find the tool I’m looking for, that I couldn’t bring myself to take pictures of the work I was doing in there. I back-filled pictures last night. :blush:

I finally got far enough in SolidWorks this morning, after hooking up my space mouse to that computer, to find the actual part (rather than an assembly into which it was linked), remove the smaller display, import the model of the TFT35, and start to try to figure out how to re-link constraints to the features of that model. (Got far enough to think that if I could just take a week or so to get past the initial ramp of the learning curve, I’d probably enjoy using SolidWorks enough to get over needing to run Windows to do so.)

Then I looked again at the display I have and discovered that it isn’t the part that was modeled. What I have is the TFT35-E3 V3.0 which, I think, re-packaged up the TFT35 in a form factor that is an exact replacement for the Ender 3 display board. The dimensions are on github but I don’t know of a 3D model for it. They have a 3D PDF but not quite clear to me how useful that would be; Linux support for that seems to be slim pickin’s…

I was thinking of an arrangement that exposes the microsd card and micro-USB slot in the mini-e3, because on-board SD has lowest potential for print artifacts, and the USB port is useful, with the board set vertically for better thermal management. I want to put the 24V power supply remote under the desk to reduce noise from the fan, but make the mains power switch accessible. I need 120VAC to feed the heated bed from the SSR anyway. I like having the screen right out front, easy to get to, and I have plenty of desk space for it.

I’m attaching pictures here roughly mocking up the layout that I have in mind:

This doesn’t show the mini-e3 board as high as I would make it actually sit in order to have access to the USB port. There is plenty of room between the frame and the bed, so the case can stick up above the frame without a problem. (I just ran out of pieces of foam nearby to prop it up. :slight_smile:)

The AC inlet could be on its side with the C19 receptacle toward the back, instead of oriented vertically. That would probably make more sense than the mockup.

Here’s a top view:

Thank you! :slight_smile:

2 Likes

Nice thing is it looks like there’s a model for the E3 version as well:

https://grabcad.com/library/bigtreetech-tft35-e3-v3-display-1

later on tonight once the kids get to bed or maybe tomorrow I’ll start playing around with some layout ideas.

1 Like