Fully 3D printed custom case (update)

[HgArg]

Hi comunity of SFF.
I designed and made my first SFF case.

In this case I used AutoCAD to designed because I needed to see if all the parts will fit on the case, I don't have any idea about components.
I had two first conditions: use a MIni-ITX motherboard and a low profile graphic card (for graphic matters, no for game, I know the low profiles graphic cards are low end but PNY sells Nividia Quadro graphic cards with good specs: K420, K620 and K1200).

PC specs:

Motherboard: AsRock H110M-HDS, is not a Mini-ITX mobo but has 2 cm more with 2 PCI-Express expansion port (X16 and X1). Ideal to put a graphic card and a Wi-Fi card.
Graphic card: PNY Nvidia Quadro K1200 4GB GDDR5.
CPU: Intel i5 6400 Skylake.
PSU: SFX Sentey STY45-PMX.
RAM: 2x16 GB Crucial Ballistix Sport LT, are not the best RAM but the price was convenient.
Storage: Kingston SSD UV400 120GB (enough for the moment).
No DVD/BD/CD drive.

This is my first desing:

Layout:

It measures 216 mm in height because of the motherboard (has 19.1x18.8 cm, "Micro ATX" as AsRock call it). But the Mini-ITX version has only 196 mm.

I had two ideas:

1. 3D printed case, using a Prusa i3 (building platform 200x200 mm), this is the reason of these connecting pieces:

2. Put some holes to change the covers:

The upper holes allows to make two different covers and change the location.

To start, I made a paperboard mock-up to adjust dimensions:

Then I printed the parts and glued them together:

This is for now, next time I'll show the covers and the front with the USB ports and the switch power button.
I'll upload the files to Thingiverse.

>> Go to covers.

Bye!

 

[Phuncz]

Welcome and a nice project to start off with !

How is the strength and torsion of the case being 3D-printed ? Are you going to be using metal covers on the outside ?
Something you may have thought about but if your top panel (covering the PCI brackets) could have pins on the panel, where the screws of the PCI brackets would be. That way you could solve the need to fasten them down and make it invisible from the outside.

Interesting project, keep us in the loop !

 

[HgArg]

Welcome and a nice project to start off with !

How is the strength and torsion of the case being 3D-printed ? Are you going to be using metal covers on the outside ?
Something you may have thought about but if your top panel (covering the PCI brackets) could have pins on the panel, where the screws of the PCI brackets would be. That way you could solve the need to fasten them down and make it invisible from the outside.

Interesting project, keep us in the loop !

Hi Phuncz, thanks for your reply!

Surprisingly the case is strong enough to manipulate without problem because I made a full flange on the back and the top of the case to reinfor the structure (my English is not enough good, sorry ).
The covers are made of plastic too, with 4 and 5 mm of thickness.

I tought something similar about you told me, but the top cover has a cavity to solve the problem.

Bye!

 

[iFreilicht]

What a cool way to make a case, 3D printing separate parts that then glue together. How long did it take to print all of that?

 

[HgArg]

What a cool way to make a case, 3D printing separate parts that then glue together. How long did it take to print all of that?

Hi iFreilicht.
I don't know, maybe 10 hrs for the structure and 10 hrs more for the covers, are in different colors.
I'm making another design where the covers are the structure, like the cases of the market. I estimate it will take 12 hrs to have a complete case.

 

[BirdofPrey]

What material did you use?

 

[HgArg]

What material did you use?

Hi BirdofPrey. I used PLA, the only reason is has better adherence than the ABS.
But is not the best material to glue with cyanocrylate (Super Glue and similars), it's needed more time.

 

[ColorZepppelin]

I'll upload the files to Thingiverse.

Hope we get to see your design on Thingiverse. That would be awesome! Good job, quite innovative and interesting solution!

 

[Hahutzy]

Hi BirdofPrey. I used PLA, the only reason is has better adherence than the ABS.
But is not the best material to glue with cyanocrylate (Super Glue and similars), it's needed more time.

When I was prototyping the middle board of the build (the Hutzy XS and its big brother, the Hutzy HS) I used my own 3D printer and printed the boards in PLA.

However, I noticed that after extended use (around a month or so?), the board had warped significantly when I detached the motherboard and GPU from it.

As you may know, in the context of a computer heat output, PLA has a faily low glass-transition temperature of 60-65°C. This means that as the PLA reaches and goes above this range of temperatures, it goes from its glassy, solid state into a rubbery, flexible state.

In my case, this phenomenon together with the weight of the motherboard and GPU caused the middleboard to start sagging and warping.

In your case, if you stand it up like you do in that last picture after mounting the covers, warping might also start happening. So my advice is be mindful of that.

Other than that, I applaud your effort in 3D printing and assembling the whole frame. My build area is also 200mm x 200mm so I know how annoying it is to not be able to print the thing you want to make in one piece.

 

[HgArg]

Hi Hahutzy.

Thanks for your advise. I printed it one month ago and today I haven't any problem, only the covers make some noise because the desk isn't to steady ....
The motherboard is not fixed to the frame, I put 4 cylinders below the motherboard to screw it and glue after that to the frame, bad idea....

I think the best choice to make a 3D printed case is divide the model and glued the parts, it's stronger than a unique piece.
I used Cura and put 25% of Fill Density, it's strong enough .

 

[BirdofPrey]

Hi BirdofPrey. I used PLA, the only reason is has better adherence than the ABS.
But is not the best material to glue with cyanocrylate (Super Glue and similars), it's needed more time.

Well at least I see you gave some thought to gluing surface area with the extra tabs and some lap joints at the rear. If you need something more secure, you could also try dovetail or box joints in a future revision, and you could also try using epoxy next time (cyanoacrylate glues have low shear strength)

At the very least, if you want a strong bond, you should probably clamp the parts together and leave it overnight as well as using a glue with gap filling properties.

 

[HgArg]

The dovetail option looks interesting! I will try it in my new model (a thin HTPC with standard components)..
Yes, I used clamps, but with 1 hore is enough, with ABS is instantaneous.

 

[HgArg]

If you need something more secure, you could also try dovetail or box joints in a future revision....

Check out:

 

[3lfk1ng]

Not sure how much clearance you need but something like this would allow you to press the joint into place without it going through. The offjet joint on the side would also add some lateral rigidity.

 

[HgArg]

Not sure how much clearance you need but something like this would allow you to press the joint into place without it going through. The offjet joint on the side would also add some lateral rigidity.

Hi 3lfk1ng.

The main problem with your proposal is because is difficult to clean after the print (it happened!), with the dovetails option no need to clean and no need to clamp, just with glue on a flat surface (on a glass for example). Then the side covers will reinforce the case.

A mixed solution:

 

[3lfk1ng]

Ah, yes. Something like that would work.

I have a 3D printer myself (Replicator 2) so I know that pain. I'll be watching your progress, keep up the good work!

 

[EdZ]

One outside-the-box option for joining hard-to-glue thermoplastics is to use Friction Stir Welding: put a piece of filament in a rotary tool, spin it up, then apply to the seam. The joint created is excellent (it's a fairly homogeneous weld), but the downside is this is not really a technique suitable to serial production.

 

[HgArg]

One outside-the-box option for joining hard-to-glue thermoplastics is to use Friction Stir Welding: put a piece of filament in a rotary tool, spin it up, then apply to the seam. The joint created is excellent (it's a fairly homogeneous weld), but the downside is this is not really a technique suitable to serial production.

HI, I think ABS is able to do what you are saying, but for PLA I don't really know ...

 

[iFreilicht]

HI, I think ABS is able to do what you are saying, but for PLA I don't really know ...

Shouldn't it be easier with PLA because the melting point is lower? So you'd need lower RPM to get the same result. Just speculating.

 

[HgArg]

Shouldn't it be easier with PLA because the melting point is lower? So you'd need lower RPM to get the same result. Just speculating.

Interesting point, maybe works on both cases!