Monolith Maslow 1.0 (refined gantry style stiffened Maslow, back to linear bearings)

Hardware
1

Files:

TLDR: refinement & simplification of my stiffened linear rail design - only adds 1kg rather than 2, uses the original linear bearings, uses a variation of my Turtle Clamps, fewer bits to print and adds a few expansion options, uses 25mm rather than 30mm bolts in most places. Still retains the significant stiffness boost over the stock Maslow4.

1920×2560 654 KB

My linear rail conversion (I linear rails-ed my Maslow) did what I wanted - upped the stiffness a lot. But it was at a cost of quite a bit of weight gain, and complexity of assembly. I’ve spent a bit of time refining that as I knew it was overkill really, and there was a lot of scope for simplifying and saving weight.

The first thing I wanted to fix was the clamps - hence the Turtle Clamps (Turtle Clamp (non-split clamp design)). But further than that, I wanted to see if I could refine it a bit from over-engineered back to just…engineered. Although I like the linear rails, the linear bearings a stock Maslow comes with are sufficient for a router based CNC and the loads we’re likely to put on it. In time I aim to get a fat spindle going, and I suspect I’ll need the linear rails at that point, but for now this gives me everything I want for a trim-router based Maslow.

Although you can use three bits of 2040 with printed brackets:

1920×2590 812 KB

I don’t think this is necessary so i’ve just been using it with a single 2020 cross piece and some alloy brackets, which saves a few hundred grams.

Refinements and comments

I can do a step by step build guide if needed, but largely it will be the same steps as the linear rail one . So I’ll just do a list of points about this:

  • The 2040-to-sled printed brackets are now single-piece and sit lower. Which means…
  • It uses 25mm bolts where the liner rail version used 30mm (they seem easier to get hold of).
  • Although in the photos you can see I’ve used 300mm rods for the linear bearings, the stock ~200mm ones work fine.
  • The rod holders are design so the 2 through-bolts can be used to set the exact spacing when assembling, but then the brackets can be slid out and off via the t-nuts for maintenance without having to reset the spacing each time.
  • Similarly, the stock posts with the bearings are used (i’ll include variations of my own design in 1.1)
  • With the Turtle clamp the router-collet-to-base-gap now sits at ~8mm.
  • Maintenance - one aim was a bit of design-for-maintenance. It’s still a bit of a faff to get an arm out for example, but I have put some though into several of the common things I do.
    • Router out - as with the linear rail design, using 350mm 2040 sides means removing the router is a case of:
      • Make sure the cage is at the lowest Z-travel.
      • Unplug the control board.
      • Loosen the Turtle clamps.
      • Slide the router up and out.
    • Note the Turtle clamps have cutouts for the button pusher anyway.
    • Taking the whole cage assembly out is a case of:
      • Loosen the side t-nuts of the cross-piece, and slide off.
      • Loosen the t-nuts on the rod holders, and the rod-clamp-bolt
      • Slide the rod-holders up and out.
      • Do Z-travel up until the cage is free and lift out.
      • When re-assembling, you shouldn’t need to re-adjust and spacing this way, just tighten the loose nuts.
    • And to just remove one arm it’s a case of:
      • Router out as above.
      • Raise the cage / Z-travel to the top but not off the nut.
      • Unscrew either the top or bottom post depending which arm it is you’re removing.
      • Slide the arm out - watch out as you often get both the arms on that side slide out and you need to remember the ordering.
    • As you can see, the aim is that nothing that needs spacing-adjusting during assembly should need loosening for general maintenance.
3 Likes
2

This is so fun to follow along with :smiley:

1 Like
3

Thanks, I’m really enjoying the hardware hacking as a change from software stuff!

4

how much additional stiffness do you get from the top crossmember?

my 3d printed towers are very stiff, but I haven’t tried putting a crossmember on them yet.

at some point I’d love to compare the two. I think the printed towers may end up being stiffer (the geometry gives them better triangulation to the base)

1 Like
5

Files added:

I need to break out a dial gauge and some weights and measure it precisely so I can say more than ‘a bit’. It’s going from being able to flex the base still, to not. But very much it’s less and less returns replacing Clamps > Towers > Crossbrace.

Triangulation? I am guessing you mean forming a triangle from top of the tower to the sled?

Yeah, I need to print your towers and compare too - I (personally) am not a fan of nuts inserted mid-print, so I was going to try and find the time to add tunnels for the nuts before I try.

It’ll be interesting for sure to compare, and I suspect I can re-design the sled-to-upright clamps to get better triangulation while still not fouling the belts, it’ll be interesting to try and find the best trade-off of stiffness / fouling / weight!

There is a design logic for me going with the gantry design (I did look at fully printed towers to start with), which may not hold for you. I was mostly focused on x-shear (as in the whole thing flexing as a parallelogram in the X-axis), as my initial experiments showed me a decently stiff clamp more or less eliminates Y-shear as the clamps rest on the Z-motors. It’s also the reason I use the 2040 aligned the way it is, and not with all the 2040 pieces rotated 90 degrees around their long axis which would be helpful in other ways.

In essence:

  • Towers and a crossbrace is needed to get the least X-shear because the sled can flex, and the linear bearings don’t have z-motor support directly.
  • To get usable clearance for the crossbrace it needs to be about 350mm higher than the sled (to minimise the disassembly required for various maintenance tasks / not foul the control board / give yourself good levels of Z travel for functionality / etc.
    • This is the point that you might not agree with - you can definitely get away with a 300mm crossbrace and it’s just a pain to do maintenance, and you can probably limit your Z-travel (but still keep a practical cutting depth) and go lower. Or you can relocate the control board and add complexity. But it feels like limiting yourself.
  • If you are looking at a crossbrace at 350mm or higher, designing a fully-printed tower meant at least printing it in 2 parts (for me anyway), and keeping the weight down but the stiffness up while minimising fouling…using 2040 profile just worked better as far as I could see.

I am happy to be shown differently though!