Small test print for my architecture model, before I decide

[a-max 0]

Hello,

I've been struggling with the idea if I should invest in a 3D printer for my graduation architecture physical model and after lengthy discussion with some brilliant guys in the Questions & Answers area of the forum they pointed out how it could be printed, as well as some advices.

However the windows and roof part of my architecture model, are quite tricky so I was wondering if anyone could do a test print for me, in order to help me decide if I shall or not go the 3D printed route.

I'll be willing to pay for your efforts via Paypal, if needed, cause I need to see some images by Friday the latest, since I'm a little behind schedule and need to know as soon as possible if I shall or not buy an Ultimaker.

The STL files are found here:

https://www.dropbox.com/sh/56qwkmmh7e3oogk/hRRn_UTZ5d

Please note that you need to scale them to 1.5 in all directions due to a bug in my CAD software.

 

The 3d printed parts:

My full architecture model:

Thank you in advanced!

 

[illuminarti 17]

I've looked at these some more, and I'm actually running some test prints at the moment. But even in these simplified forms, they are going to be very hard to print.

Printing with support doesn't seem to be an option - the parts that need support are very thin - such as the supports under the roof - and so very hard to separate from any support that gets printed under them (I tried). Perhaps it would work in a dual nozzle set up with a soluble support material such as PVA in the second extruder. But using just one material, mostly seems to give a mess if the print relies on support.

(Another problem is with the model - it looks like the struts that make up the roof section are separate objects that intersect one another; Cura can fuse these together ok, but it's support generation is a poor at present. On the other hand, Kisslicer seems to do a better job of generating support that might be removable - but it can't deal with intersecting struts, and generates the nodes where they intersect as empty, so the struts don't actually join up).

But without support, the problem is that there are almost no flat surfaces from which to build. The window section has a very slight curve to it, so it can't be laid flat on the bed. If it was flat, on the back, such that the lower 'wall' and the ladder part could both touch the bed, it would print great. But it isn't, and they don't. So the best hope of printing it looks to be to stand it up on the longest edge, so that the only bridging to be done is the two long strips - and those are laid over the top of the smaller slats, so they don't have to stretch over as much empty space to print. Even then, its a bit of a stretch.

The Roof has a curve and twist on the outside that prevents you from laying it on that side. That might be a candidate for some support, as it would be relatively easier to remove from a continuous surface, but you'd end up damaging the most obvious visible surface quite badly that way.

In addition to the support struts having modeling issues, they're also very thin, even printed at 150% scale. The printer doesn't do particuarly well at printing very small cross-section parts like that - especially when those parts are suspended in mid air, with little support. Perhaps printing the whole thing at an even larger scale might help. Or deconstructing the support lattice into simpler flat parts, and then assembling them afterwards?

All of that said, I'm making some sample prints now - I'll upload them a bit later.

 

[illuminarti 17]

I gave up on printing the roof when it was about half way done - after about 3 hours. I printed it at 60mm/s, 0.06mm layers (to try to give it as good a chance as possible of printing the almost horizontal struts in the scaffolding), with a single perimeter wall (0.65mm as my nozzle is larger than standard).

I printed it vertically, in the end, oriented as shown in this picture: cropping off 1mm or so at the bottom to get a flat edge on the print.

The separate scaffold structure on the left section of roof came loose during the print as it was barely attached to the main roof panel - it wobbled a lot, and just turned into a blobby mess. So I removed that. The rest I cleaned up a little by using a scalpel to remove some of the worst stringing. But it's still pretty rough - you could maybe dremel it a bit - its pretty sturdy once it's all connected up.

You might get better results printing slower, or cooler - you'll need to experiment, as we've already said. Finding ways to slice the model up into flat pieces that are better suited to printing will be critical - especially for complicated and fiddly parts like this scaffold under the roof.

 

 

 

[a-max 0]

Thank you once again illuminarti for the in depth reply, I must say ouch... these seams quite troublesome to print and they are actually only small portions of the model and are probably the "easiest" since the other parts are whole lot bigger, like 20cmx20cm, with even more complex curvatures & twists. Unfortunately I can't go bigger then that since already the model would be around 2.1m x 1.5m and it would require around 6-8 rolls of PLA as long as higher print times, which I estimated to being around 10days for now, printing almost non stop.

Regarding the model intersections, unfortunately I can't fix those since that's how my CAD software generates them and deconstructing the roof truss into other parts and then assemble them later is again problematic due to having a large area of them, aproximately 10 parts with a size of 20cmx20cm, the number of pieces that it will require assembling would surpass 1000 and that's extremely time consuming.

Thank you once again for all the effort you put into helping me decide, i'm extremely grateful for all the info you've provided.

 

[illuminarti 17]

The windows I printed at the same speed and layer height - 0.06mm layers, 60mm/s. I printed it oriented vertically, with a raft in the hope that it would keep it upright - the panel is curved, so it leans backwards quite a bit.

The retraction left some interesting organic patterns within the window panels, but hopefully those will clean off pretty easily once the print finishes....

 

[a-max 0]

I've just seen the images after I've posted and I must say that although the prints seam relatively ok, it requires quite a lot of post production, as you mentioned and I must agree, the windows do look quite nice with that organic pattern.

Unfortunately the time needed to print these are a lot bigger then expected, since I've estimated around 5 hours/module which could have helped me reach the deadline but after reading that it took 3 hours, the time needed for a whole module would be around 15 hours or so which is too time consuming and unfortunately I don't have that time at the moment.

Thank you once again for the effort and tremendous help, you finally helped me decide and also save quite a lot of time,headaches and money. Please let me know your Paypal address as well as the costs in order to pay you back for your efforts.

Thank you !

 

[illuminarti 17]

Well it finished printing in about an hour and a half. The stray bits of plastic cleaned up pretty easily, but the whole thing is just too weak... it basically fell apart as I tried to remove it from the bed and raft. the junction between the vertical and horizontal bars was surprisingly weak - I wonder if there wasn't some weirdness in the model for those too that was causing them to be poorly connected... mind you, this isn't the best filament either, and it was a pretty rushed print, so that could be part of it. Overall, it printed better than the roof, but too weak to withstand clean up, or be useable for much I think.

Which brings me back to my original observation that while it may not be impossible to get good results with FFF printers like the Ultimaker - it certainly won't be easy, especially not at this scale.

 

[valcrow 144]

I Imported the file to take a look, and it's what I expected. The geometry is interpenetrating which is causing cura to put a small gap between one object and the other (support struts and Wall panel) Which might be what's causing it to be super weak. This is an issue not only where it connects to the wall, but each support interpenetrates each other as well, causing them to be weak to each other.

What you need to do on the 3D end, is clean the tiny gaps between the walls. And then boolean everything together so that they are one single mesh and ensure there are no interpenetration of geometry.

Slice after that and you'll have a vastly improved solution from Cura. And I'm pretty sure you'll get a much sturdier part.

Here's an example: your cylinders are like the ones on the left.

You might be able to use Daid's Fix horrible options, or netfabb if you don't want to clean your geo up.

The window pattern is really cool btw.