Case Mod - In Progress ⭐ (In Win 901) Asteria II: Rearmoured

Many thanks to you! I'm so incredibly frustrated that many aspects of my life keep getting in the way of this it was over a year ago I was nominated for the Modding Update and, having come second, I wanted to keep the momentum going.

Still, work does continue. The stock portions of the case have been stripped down now with the final few mods, cuts and flush nuts being installed. I need to redo the acrylic work AGAIN and still faffing about with this DDC pump, but at the very least I'm happy that I'm working on it again.

 

A YEAR? Pfffffff. Hold my beer...

 

My birthday last month marked the 3rd anniversary of actually buying the case, and worklogs started 2 years ago this month. This better be bloody worth it

 

In life, always keep up and the only thing you wont go thru is death, so this case is to be finished sooner or later.keep up! there is many relevant information in this thread. And and don't worry, trends are coming back every 20 years, lol

 

Well that's good to know, it'll be 20 years by the time I've finished this

And all my future projects are based on older cases too. If I ever get the funds or the support to actually do them, we got

Dioxidane is a Fractal Node 202
Anesidora is a Bitfenix Pandora mATX
Jasmina Reprise is a Silverstone Sugo SG05
Akasha Renaissance is an Antec PlusView 1000 AMG

No doubt all of these will be considered retro by the time I get to do them lol

 

Greetings to all!

Blimey, it's been a while. 2018 wasn't the most fantastic year so many things were shelved, delayed and ignored but work does continue (albeit very slowly) and I thought I'd share some success with you.

I didn't cover in this particular log my pump mounting issues (but detailed here if you're interested), but I'm very happy to say a significant step has been made in resolving it: the DDC is finally in!

After all the toing and froing regarding measurements and poor accuracy, I confirmed that the distance from the edge of the GTX Titan and the glass side panel gave me an operating width of 43mm. DDC heatsink bodies are on average 20mm tall and my choice of pump top is a further 20mm. Factor in some manufacturing tolerance and the combined pump+top is 42mm tall. What that doesn't factor in though are the fins on the bottom of the DDC heatsink, which add 5mm, and DDC tops traditionally are fed from the top. All in and we're too big to fit. I've already discussed modding the Aquacomputer DDC top to have a side feed and using a custom stainless steel plate to seal off the top feed, so that keeps the top down to 20mm, but it was the heatsink that was the major issue.

The first plan was to just chop the fins off my Alphacool heatsink. Because Alphacool use a very thin thermal pad, removing the fins would reduce the heatsink to 19mm. Nice, but then put 1.5mm back on for my mounting bracket and we're up to the limit again. Plus, I didn't fancy the idea of getting nickel everywhere from cutting and filing, and I know I'd just want to strip it down and replate to keep things tidy. Same deal with the EK heatsink; easy enough to strip the anodising off the aluminium, but then it would need recolouring. But then I noticed Barrow take a slightly different approach.

Barrow's heatsinks are actually in 2 parts: a chunky piece of aluminium which does the actual heatsinking and a thin body ring. Could I combine then the actual heatsink portion and mounting bracket? That would save a lot of space. So I grabbed one and redesigned my mount. Believe it or not I actually cut this by hand!



Scruffy? Yes. But I'm stoked to get the end result that good with just a step drill, coping saw and a crap wrist

Quick assembly to ensure my mounting holes are correct...


...and then fold that bottom tab into place.





Nice job, but does it actually work?





So far so good, but there's supposed to be a computer in there...





Hot damn that's tight, but it's not actually touching the Titan. Major win (and relief) #1


Now, in a perfect world the entire pump assembly would fit within that 43mm operating space so I can put the side panels back on with no further mods, but if need be I can put in a 1mm spacer to just edge the panels out without issue. Do I need to?




Maybe. The pump mount is almost flush with the case tabs that hold the side panel mounts. If the mount is flush then the glass will press on the pump and we'll get vibration and noise. But 1mm in this build is acres of space so I'm not too concerned right now.

Now at this point I'm still just test fitting for viability. As a result I'm using standard dome head M4 screws to bolt everything together. The head height for those screws though is 2.3mm so of course they're not going to work. But the plan is to use countersink screws and get them flush with the mount.






Golden! Major win (and relief) #2

What I've not shown you here is a small failure. I'm using 1.5mm aluminium and the countersink M4 head is 2.3-2.5mm, so I wasn't sure if I could just slap a countersink into the mounting bracket given I'm missing almost 1mm of material to cut into. The idea at this point was to JB Weld spacers to the inside surface and give me more material to form the countersink. These spacers would also serve as proper alignment feet for the Barrow heatsink ring given it's just 8mm through holes and slides around the screws as you tighten up.

The Mod Gods were merciful and I found some perfect spacers: M4 tight through holes (4.1mm) and 8mm diameter. Perfection Nickel-plated brass though, but they won't be visible so didn't matter if paint didn't take, but a quick rasping on the mating surface got rid of the nickel there ready for the JB Weld. The JB Weld didn't take though as a 2mm ring just wasn't enough surface to adhere to and take the pressure of grinding down a countersink cone. Plus it was possibly the hardest form of brass in the universe. So the spacers popped off, but wonderfully the countersink in the aluminium took perfectly without deforming! Nice bit of luck there.

For now I've slid the spacers inside the barrow body ring so they can still serve as alignment holes, but for the build proper I'm going to drop in some 16mm long nylon spacers to fill the gap from the pump top's alignment feet and the mount, hopefully providing a bit of support so I don't warp the countersinks as the entire assembly is tightened.

So moment of truth then. Does the glass panel actually fit back on?






Oh yeah Glass side panel perfectly screwed in. I can actually give the pump a push from the other side of the case and it doesn't knock against the glass. Major win #3

Speaking of the glass panels, I never actually told you about them either. Stay tuned...

 

OK, about those glass side panels.

The drive to retain as much of a stock look for the case as possible extends to keeping the same thumbscrews for the glass panels, so given I was replacing parts of the case which have panel mounts on them I needed to rebuild those mounts to fit the screws. Now although they're standard 6-32 threads, it took me a while to find brass standoffs that were 6-32 on both sides; usually they're 6-32 male and M3 female. Also they needed to be about 5mm in diameter too in order to fit the rubber grommets and mate with the glass.

Turns out I shouldn't have bothered

Eventually found some, marked up the new holes, tapped them into place and then realised they were too long. Not a problem, I'll just cut them down to fit, but then I realised a major error in judgement. Firstly, by chopping down the standoffs, there was very little female thread left to screw into, so there was no way the tempered glass could be supported safely even if the thumbscrews could actually go in. Then I realised that In Win had used through hole self-clinching threaded standoffs. This is how their thumbscrews would fully seat into the case and support the tempered glass.

So after carefully inspecting some of the original case parts, I tracked down the exact PEM fastener required, but the company in question no longer dealt with private customers, and I couldn't find any Chinese knockoffs (unlike my flush nuts). So I resolved this by doing something I didn't want to do: irreversibly cut up the original case

These particular standoffs fit into place by cold-forming sheet material into a little channel between its main body and a flange. As a result there's no welding or such required to get them in. So, by cutting into the aluminium either side of the standoff's main body, the material around it is actually no longer attached to anything. Gently pry it off and the standoff is released.




Since these things are steel pressed into aluminium, there's no deformation at all and can be re-used. So give them a quick sand down to remove the paint and get a proper through hole cut into my tabs (5.6mm in this instance). Then apply pressure.





And in they pop!

My half ton arbor press was sufficient for the pressure, but I had to use a chunky M8 spacer I had floating around as a makeshift anvil. Worked nicely though. Unfortunately I couldn't get the arbor press into position for the main body where we have this big B-shape series of folds. So I used my body weight instead; laying the fully folded, tapped and drilled final body work down onto a piece of MDF with a hole in it, putting a dome head screw into the backside of the standoff placing a couple sheets of material onto the screw head and then generating half a ton of pressure by balancing on one foot and doing squats is really, really not how I want to insert these things. There was this big pop and I did not know if the standoff had gone in or if I'd trashed the whole damn thing.

It was the former.



Beautifully stock.

 

Hey there. Might as well post some pretty renders of part 2 of the "seriously, haven't you fixed or discarded the DDC yet" saga.

With the pump assembly perfectly in place, it now transpires that I can't make the tube routing from pump to radiator through the floor to work without a stupid, ugly amount of adapters, and it's too tight for soft tube too and just kinks. So as suspected it's time to return to the custom pump top idea. Completely redesigned this time though. Enjoy the renders






I've extended the body width by 19mm to cover the distance between the pump body and the radiator holes in the case floor, so now an EK 90 degree rotary fitting will align directly above the radiator port's centre line, with sufficient vertical space to get a compression fitting on and angle some tube. That extension means I've gone down a distro plate approach to go from the volute outlet to the port, which has of course necessitated o-rings and a sealing plate.

And yes, the strong resemblance to Aquacomputer's DDC top is intentional; wanting to keep the look, I've drawn on many design cues from their dual DDC top to create mine as a homage to the top I've had to discard. I don't think I can get in trouble for that

To allow for measuring and manufacturing tolerances I will be using Mayhem's ultra clear soft tube for this part of the loop, but it's very clear for soft tube, 13mm OD is indistinguishable from 12mm OD acrylic when piped in, and once I have Oil Black Pastel coolant running you'd be hard pressed to tell the difference between the soft and hard tubes I think.

And I'm also very excited to say I'm discussing manufacturing of this right now. CAD models are being verified as I type

 

Minor change of plan already

As I was fine-tuning the CAD work, I realised that I'd completely forgotten a fundamental point in this entire build: clearance spaces. And specifically, the entire reason I've been working and fighting with side ports is because I have zero space on the top face. So why in the hell did I put cap head screws along the horizontal centre line? They're going to foul the bloody graphics card! What a wally!

So as I was reworking things, it also occurred to me that as I'd gotten a bit carried away with replicating Aquacomputer's aesthetic, it would actually look out of place with my aesthetic. Countersunk screws solve both of these issues! So, the portion of the pump top that is visible peeking over the graphics card will have countersunk screws just like the GPU backplate. Also, my logo would be totally obscured by the GPU too.

So new model and new renders...




Now I'm not so sure about 6 screws so I modelled a 4-screw version. The top plate is 1.2mm stainless steel at only 77x58mm in size so it's not going to flex with only 4 screws, but there's something about the symmetry of the 6 that I like, despite it feeling a little cluttered. 4 screws is nicely spacious without being structurally unsound, but has a weird diagonal slant. I'll see if I can get a render out to show you all.

But faffing or not with 4 or 6 holes, this IS getting made in the week

 

OK, because it's a slow day at work and I'm procrastinating, the 4 or 6 screw thing is bugging me. So I've done 2 quick layouts to illustrate (Fusion 360 in a browser is awesome!).





Looking at these without any lighting or textures and fancy projections, I think I'm actually satisfied with the 4-screw variation, it feels a lot more "open". The 6-screw symmetry is very nice but does feel a little cramped when I apply the graphics to the plate.

Any thoughts?

 

Just commented in your O-ring thread. -Screw the O-rings. With acetal and stainless, you could just use silicone glue.
Also, with cord stock O-ring, instead of two rings, you could use one ring with a single divider cord. - like a zero with a strike through the middle.

 

Cheers dude, we'll see how I get on!

 

My guy did some CAM tests in acrylic for the internal structure.





Hot damn that looks good! So good in fact there's a small part of me regretting going black acetal for this. But, just like the cap head screws, however good this looks it doesn't fit with my overall theme.

Devil says "Except for the fact the SSD covers are a brushed metal plate with an acrylic light ring, and with the brushed stainless plate on top of this, it would actually look like the SSD covers if you light it up."
Angel says "Except the corners then wouldn't match because these are angled with a parallelogram motif and the SSDs are rounded rectangles."

NO! I shall be strong and stand by my convictions! And stop talking to myself!

 

Until you buff it out and decide to change your theme. Looook at it. -It's sexy plexi.

 

Let's just say I'm going to ask him to send that acrylic test over too just in case

 

Clear with light Cheesecake dude.

 

Ooh! My pump top is in somebody else's video!



Many thanks to Lucas at Mp5works for his help, incredible work.

 

OOoooh! Lovely.
The coolant was probably unnecessary since acetal is like milling a stick of butter. Since it's really expensive butter, I can understand not taking chances.

 

Lucas had never worked with acetal before so I guess he wasn't taking chances with the fluid assist. When I did my forstner bit experiment I found just a bit of air was sufficient, but then I was pecking at it and not doing continuous contact milling out channels.

Still, looks like it turned out well. He's doing a couple of other little bits for me so he's not posted everything over yet. Can't wait to get the thing in my hands!

 

Not too shabby!