Double-decker build

I’ve finally finished printing the stock or slightly-modified parts, and partly assembled some of the parts. I’ve also gathered electronics and thick-wall tube (1" O.D. and ~1/8" wall) for the gantry rails and z-axis.

[attachment file=105213]

I am finalizing the design for the corners which allow the legs to pass through vertically.

I still have to buy the EMT conduit, and I also have a lot of wood to buy for the underlying table and for side panels to stiffen the box.

The outside dimensions are going to be 3’ by 4’, so about 2’ by 3’ of working area, and it will have about 300 mm of working Z travel, but the overall Z height will be much taller. The goal is to have tall Z working volume, while still having stiffness that is as good as a small machine.

Hopefully I can get all the woodworking done within the next week or so and get the thing moving and measure the stiffness. I also need to move some crap around in my garage to make room, but I’m hoping this doesn’t take too much time.

I just discovered that my printer is probably under-extruding. I measured the weight of the printed parts and they are quite a bit less than the values posted on the parts page . I double checked that they had been printed at 55% infill. On the one hand this is troubling, because my parts might be low strength and low stiffness, but it also might explain why I had issues with the xyz part, when most people don’t have problems. I’m going to proceed with the parts I have, but I’ll probably be investigating the printer’s misbehavior and perhaps re-print the most critical parts.


Could your filament just not be as dense? I think you need to put a density number into the slicer, but it will estimate the weight, which might give you a clue. Perimeters also are probably a big part of the weight.

In principle I think there could be a small difference in density but I am seeing huge differences of like 30% to 50%. Now that I’m looking, it appears I have a partial clog of the nozzle.

I am glad to have finally discovered this, but holy cow, how long has it been like this and how many parts do I need to scrap? How much plastic and time have I wasted? :frowning:


Ah the joys of printing :slight_smile: I finally replaced my whole hotend about 4 months back after dealing with an unrecoverable clog. It was an old ubis ceramic (printrbot metal plus) so the whole thing was due for an upgrade.

Did you get clicking in your extruder? That was my heads up. Prints weren’t as badly under extruded as your picture, but still not great.

GL on the fix!

Oh bummer

So I’ll proceed in parallel, completing the machine with these iffy parts, and printing replacements. I think this is a good time to switch colors, or else I’ll go crazy wondering which parts are the bad ones.

Here is a further partial assembly, and also showing the mix-and-match rails that allow 3/4" EMT for side rails and vertical legs, while gantry (and Z) is all 1" tube. It’s fairly easy to modify the parts since the mating surface and screws between Roller and Mount are essentially the same between C and J sizes, so I just had to bore out a 1" cylinder on the C roller at the proper depth.

[attachment file=“blue parts.jpg”]

This also shows an attempt at the corners. Securing the ends of the belts will require some additional pieces, not yet designed, but probably resembling the after-market belt tensioners like this one:


Hi Jamie, how was the build? is you MPCNC working already?

I wonder if you have some inputs on how to setup the MKS GEN board in your machine, I have one like yours and have no idea where to begin.


I haven’t gotten it assembled yet or attempted to turn on the board, but I’m expecting the MKS Gen L to be equivalent to RAMPS as far as the firmware goes.

In case there was ever any doubt, I am not a carpenter. It has taken me forever to get this table assembled, partly because it’s hot and unpleasant in the garage, and also because I am not a fan of real work! :slight_smile: Where is the robot that can assemble this table for me??

The table top is 48" wide by 46" deep (largest that will fit in my car),with 2x4’s around the perimeter, and two 2x6’s diagonally underneath to give it stiffness in torsion. I wanted high stiffness in torsion because the floor is not particularly flat and I don’t want the table top to be compliant and follow the non-flatness of the floor. The table top has several coats of polyurethane in case of accidents.

The machine itself will be 48" wide by 36" deep (external dimensions), so there will be a 10" shelf in front for tools, maybe laptop, pen and paper, LCD, joystick…

In the picture are four X side rails (about 45"), four Y side rails (about 31") and four corner posts at 5’, all of which are 3/4" EMT. The two 1" X gantry rails and two 1" Y gantry rails are also shown. The 1" rails for the Z axis are off camera.

1 Like

Starting to look like a machine…

[attachment file=109581]

1 Like

I’m sorry but I don’t understand what the second level is going to be used for and why it is so tall overall.

I think there’s going to be another spoil board in between, so he can have two jobs going at once.

This was described in a separate thread, and in isolation I can see the motivation is not apparent.

Short answer is to try for the impossible: simultaneously achieve high stiffness and a large Z working depth (right now it’s about 11"). The idea is to use a very long Z axis and support the far end (top) with another copy of the MPCNC mechanism to prevent it from twisting away from vertical. It will use the same controller, but each X and Y stepper driver will drive four motors in series instead of two.

Here is the original explanation with more details:

Why do I want a large Z working depth? Partly for clearance for the tool changer system, and partly so I can experiment with a 4th (5th?) axis. And also to see if I can do it. The extra size in X and Y is room for the tool changer.

Oh, my mistake.

I’m sure that gap between gantries only needs to be about 1/4 the max sickout right? Too much longer or shorter and you loose effectiveness right?

I was thinking the other way around actually, where the vertical space between gantries should be 2x to 4x the max stickout. Larger is okay but doesn’t improve anything.

If the two gantries are too close vertically then the Z axis acts as a lever where horizontal load at the tool will be multiplied into larger horizontal forces on the two gantries and reduce their ability to hold the Z axis vertical.

With a large vertical space, it still acts as a lever but the horizontal loads on the upper gantry are less than the forces at the tool, and the horizontal force on the lower gantry is not much higher than the force at the tool.

Thats the idea anyway. In my mind it should be as stiff as an MPCNC with an extremely short Z. I’ll be measuring the deflection vs force and we’ll see if there is something I’m missing.

Boy, talk about making everything a bit harder to get square… Quad endstops anyone? :wink:

1 Like

Only nine motor drivers and eight endstops. Ryan can you whip up some firmware for me real quick? K thanks.


That has to be just a few copy pastes away from it working.

Gotcha, I guess I was thinking in terms of tube rigidity.

This is just me trying to work through this complicated problem in my head, no critiques or anything. Genuinely I have no idea, this is just really interesting to think about. I wish we could do an FEA on this…dam printed parts make that impossible.


Is this right? So think of it in terms of a standard build (2.5"Z) we assume the z tubes are good and deflections are coming from the gantry flexing and to some minor degree the XY rails, and lets use a generic load of 1kg@2". So if expanded from there to your larger stick out, the torque/force translated to the gantry is linear (T=F*L - double the distance double the force). We end up with ~5kg force at 10"…maybe

The sticky part is, now will the gantry nearest the force still be the pivot point or is it more of a distributed load between the two? Or the way I see this is going to be the top gantry will be more of the pivot point (less forces acting on it = more rigid) and the lower will still be doing more of the work. I think in either scenario what is actually happening is now the gantry twist is gone and what you are now fighting is the XY rails flexing. I see the lower XY rails being the new weak point and them flexing will be the limiting factor on Z rail stickout. Two Z rails (usually shorter) will always flex less than 1 X rail (usually very long).

No idea how to really test it but I am guessing you will be more limited by the lower XY rails flexing and most any dual gantry distance will stop lower gantry flex/twist. So maybe a purely X or Y move (single rail doing all the work) will flex more than say a diagonal (both rails splitting the load) when pushing your dual gantry with a long stickout where I don’t think that even comes into play on a short machine?

Can’t wait to see it in action…and no way, I really hope not to ever dive into marlin that deep again!

1 Like