With the design for the printer mostly finished, I’ve been going through and double checking all of what I’ve done before making the big order.
Two of the really critical parts I’ve designed are the pulleys and the cage. As you saw in yesterday’s post, the pulleys worked out quite well. After I completed those, I started printing the cage.
The cage – aptly named due to its appearance as a cubic kind of shape that holds all kinds of stuff within it – is the joint, if you will, between the extruder / hot end and the cross gantry linear rails.
The engineering that went into this required a fair deal greater precision than a lot of other parts. There were some tight clearances to negotiate around. The goal in designing it was to give it as small a footprint as possible. The reason being that any unnecessary width or depth added to the necessary length of the cross rails in order to deliver on a 400×400 XY build area.
When you design a cross gantry machine, the size of the cross rails is determined by the print volume in a given direction, plus the width of the whole hot end / extruder assembly, however much extra you need to mount the rails or rods into the side carriages, and any additional clearance space. Suffice to say, it can add up rather quickly.
Design Method
I designed it by laying out all the parts, and essentially designing the cage around the parts. I put the rails as close as I felt I could together, while still giving enough breathing room to the extruder / hot end, and ensuring that there was enough material between all these things to hold it together.
I drew out a top-down sketch, creating what I felt was adequate clearance around each part. I knew the blocks would be inside, so I created a little “wiggle” room to allow for fine adjustment, expansion / contraction and such. Did the same thing around the rails, and the hot end / extruder assembly.
I started at the bottom, and created an extruded “slice” if you will that went up to the first part interference, and then created another slice that negotiated around that part allowing for proper clearance, and then once I got above that part, the next slice returned back over that part and negotiated its way around whatever other parts it needed to.
The end result is pretty cool.
Probably the most frustrating part was the lack of attention I’d paid to support, and where it would print. There were some pieces that knocked out easy, and some that took some rather long needle nose pliers and a lot of twisting and scraping to fully remove. I will address that in a revision.
All the holes were perfectly placed. Due to the design, it is necessary to take the bearing blocks off the rails, and then reinsert the rails once the bearing blocks are installed. Last night, first time I did it, I inadvertently lose a few ball bearings. I learned the secret is to slide the bearing block into the slot in such a direction that it is upside down, and before putting the M3 screws in, gently slide the rail, wiggling it slightly until it catches just right and then very carefully continue progressing it through until it comes out the other end.
The mounts for the extruder work reasonably well. The bottom holes are flush with the extruder. I left some space on the side ones for some breathing room and a washer. I will probably take those in flush on a revision to further tighten things up. I may add an additional set of holes to match the ones up a bit higher on the Bondtech, however I really don’t care for the placement of them so close to the cleaning ports.
The whole assembly – if we’re being honest – is rather heavy. I weighed each part individually before putting it all together, and then all the parts together. It feels much heavier than it is.
But what you have to consider is that this is all being split four ways.
There are four support points on each of the XY carriages holding it all up – front, back, left, and right. That’s 455 grams a piece.
There are also four pretty decently powered motors (right now I’ve got NEMA 17s spec’d, OMC 17HS19-2004S1, with 59Ncm, or 84oz.in of holding power), each driving 25% of the entire assembly.
All of this is being driven by 10mm GT2 belt, with 30 tooth pulleys. That’s a decently wide belt, with 150-200% the tooth engagement of the size of pulleys typically used on these machines.
Contingency
At this point, I’m about ready to take the plunge and place my order.
There will be a lot of trial and error testing going forward, which has led me to consider contingency plans in the event some or all of this engineering doesn’t work out the way I need it to.
Basically, if this doesn’t work, am I going to be saddled with all kinds of parts that aren’t going to work for something else? I think that’s a universal concern, before any large investment is made. And luckily, I think we’re good in this case. All parts should translate well between different approaches.
If this cross gantry doesn’t perform up to what I need it to, the next step I think would be to switch to a slightly lighter cross gantry setup with linear rods (8mm, since I already have them) and test that out.
While the present MGN12 driven design is supremely rigid, it is also rather heavy, and who knows where the exact sweet spot is between the two, and if one is much closer than the other? Such a design change would require a new, but arguably simpler cage to be printed, along with slightly different carriage blocks to accept an 8mm rod instead of 12mm linear rails.
A rather easy modification.
If that doesn’t work, I would be open to entertaining a CoreXY setup. It would definitely take some extra work, because the sheer size of this build volume is something that CoreXY struggles with. The long gantry span, the belt lengths of over a meter, controlling ringing; there’s a lot to deal with.
But, all of these designs have so much in common with one another parts wise that I don’t see an issue with any of it.
Time to “do the thing” as they say. More later!