Stalled Freilite Brevis S: <5L Gaming Case with internal PSU

[CC Ricers]

Spaghetti cables aside (this is a just test rig after all) looks like all the parts will fit in nicely. It's also nice to see how much clearance the CPU cooler has, because my custom case would be about the same thickness, judging from the motherboard I/O opening on the back.

Yes they are indeed. Don't know the exact number off the top of my head, but they're from PEM.

I've gotten curious about those standoffs. I would really like standoffs that lie flush with the bottom of the tray, but having looked at how they are installed on their website, looks like you need precision tools... I don't have access to a pneumatic press or where to go for one. They don't look like something you can just put in with a hammer lol

 

[iFreilicht]

Thanks! Yeah the cables are long, but the ATX and EPS cables will probably be of similar length as they have to reach all possible positions of connectors. The SATA and PEG will be cut by a good margin, though.

If you have a way of getting a milled metal part from a shop in your general vicinity, you can insert these with a hand-operated press. Not sure what the correct terms is.
All you'd have to get the shop to make is the "anvil" as its called in the datasheets, and it's quite a simple part to mill.

 

[QinX]

Great progress on your project, it's always a good feeling when stuff works as you drew it.

That problem with the screen is not uncommon. My GTX970 had the same issue. It isn't PCIe CEM that is to blame but guess what, the industry just doing whatever it want.

Industry: Guys we need like 2x Display port, HDMI and DVI in 1 slot, but we would have to go out of spec.
DO IT! MOAR OUTPUTS IS MOAR BETTAH!

 

[Hahutzy]

The space from the bottom of the card-protrusion to the top of the bracket, is that 10.16mm (0.4")? As per Figure 6-2 of the v1.1 spec?

I did not refer to the specs when I drew that area of my case. Instead, I was referencing models of testbenches that people have put up around the internet. And from what I can gather, people have moved from 0.4" to 0.2" for that separation, because of reasons that QinX pointed out.

Also as a result of that, the cutout was also lengthened to around 95-96mm in height.

 

[jeshikat]

No idea how that could happen. I am 100% sure that the cutout is way larger than the PCIe CEM requires it to be, but maybe the size of it was changed after Rev1.1? Only two screens for now, bummer.

I have 7mm from the top of the PCI flange to the top of the cutout in Cerberus and haven't run into issues with video card connectors.

 

[Ceros_X]

Maybe a dumb question, but have you tried/thought about running the riser cable underneath the PSU? Would make for less clutter and let you use a shorter cable maybe?

 

[iFreilicht]

I'll have to re-check it, but I think there's about 10-9mm of space there, so if 7 are enough, that would be fine. I already made the cutout a bit wider to make sure the connectors of the R9 Nano would fit. Oh well. Thanks for the help!

Maybe a dumb question, but have you tried/thought about running the riser cable underneath the PSU? Would make for less clutter and let you use a shorter cable maybe?

That was the plan initially but running it above the PSU allows me to glue two strips of foam tape to it which will hold the HDD in place. If I run the riser underneath the PSU, I'd have to glue those to the HDD or PSU, both of which are components that are more likely to be replaced than the riser. It also allows me to use a thermal pad between PSU and the right side panel for a little bit of passive heat dissipation, but I'm not sure whether that would actually help.

 

[Ceros_X]

That was the plan initially but running it above the PSU allows me to glue two strips of foam tape to it which will hold the HDD in place. If I run the riser underneath the PSU, I'd have to glue those to the HDD or PSU, both of which are components that are more likely to be replaced than the riser. It also allows me to use a thermal pad between PSU and the right side panel for a little bit of passive heat dissipation, but I'm not sure whether that would actually help.

I used Velcro in my deployment Box to hold the SSD in place and it is rock solid, yet easily removable when needed. The installation also looks very clean. Foam tape would be better for platter drives, but Velcro just looks so good and is convenient.

 

[iFreilicht]

I completely agree with that, but I'm not talking about actually mounting them with double-sided foam tape.

The drives are held by a small mounting bracket, but there's too little clearance between the drive and the side panel and PSU, so I have to add single sided foam tape as padding to make sure it stays exactly where it's supposed to be.

If I were to use double sided tape, I would never mount the HDD on the riser, that would cause too much stress on the ribbon.

 

[iFreilicht]

I have 7mm from the top of the PCI flange to the top of the cutout in Cerberus and haven't run into issues with video card connectors.

So I just checked the GPU and yes, 7mm would've been enough, just about. I was at 9.5, but that's changed now.

For my current prototype, I crudely corrected it:

I also mounted the SSP-300SUG and as I expected, it not adhering to the FlexATX Standard by makes it fit awkwardly and it can only be mounted with two screws.. I'll have to correct the cutout to be a bit taller. @Hahutzy make sure you do that as well with Hutzy XS!

Either way, the SSP-300SUG is wonderfully quiet in idle.

Right now I'm running it without a dGPU, but it makes for a lovely compact rig, too:

Which makes me think whether this could be made modular. Like having the frame as three separate pieces (MB, PSU, GPU) and being able to combine them however you like. Another idea would be to have a way of making the case thicker with an additional piece to give space for taller CPU coolers and more drives if desired.

 

[Hahutzy]

re cutout: Hmm can you measure the width/height of the C14 connector? I added some tolerance to my cutout when I sent it to manufacturers, but it looks like the connector is a lot bigger than normal..

re modularity: I thought about doing this with XS too. Your implementation is cleaner, but still has a similar deterrent. The top/bottom panels will have to be completely different SKUs than the normal one. Makes manufacturing a little more complex.

re PSU: Really interested in noise test at load. The spec for the fan model you snapped a pic of says it goes up to 42dBA, which is the same rating as the stock fan in the FSP 400W and the FSP 500W



For your motherboard standoffs, did you follow the specifications for the hole distance 100%? In my 3d printed prototypes, I have consistently needed to add a bit more distance for the mobo holes to be centered with the standoff holes..

 

[iFreilicht]

re cutout: Hmm can you measure the width/height of the C14 connector? I added some tolerance to my cutout when I sent it to manufacturers, but it looks like the connector is a lot bigger than normal..

Will do tomorrow. It's a little wider, but that's not the big problem. The problem is that its center is about 1-2mm higher up than it should be.

re modularity: I thought about doing this with XS too. Your implementation is cleaner, but still has a similar deterrent. The top/bottom panels will have to be completely different SKUs than the normal one. Makes manufacturing a little more complex.

Well you could have side-panels that were made up of multiple parts as well, but that would negatively affect the stiffness of the case.

re PSU: Really interested in noise test at load. The spec for the fan model you snapped a pic of says it goes up to 42dBA, which is the same rating as the stock fan in the FSP 400W and the FSP 500W

You and me both. It will probably be louder than those two units under the same load, though.

For your motherboard standoffs, did you follow the specifications for the hole distance 100%? In my 3d printed prototypes, I have consistently needed to add a bit more distance for the mobo holes to be centered with the standoff holes..

Yes, but I didn't check whether all holes were centered at the same time.

 

[iFreilicht]

Teaser update:

Just finished the cable harness and it lives!

I'll be heading home in a few minutes and write-up an update there.

 

[Ceros_X]

Looking good!

 

[iFreilicht]

Update time!

The cable harness for the SSP-300SUG was completed in two sittings. I used pins and connectors from Icemodz, which are very, very nice connectors BTW. The 6+2 pin PEG and 4+4 pin EPS12V have a particularly nice feeling to them. They just fit together so very well. Here's how everything went:

First off, all the tools I needed for crimping were at work. While I don't trust the GPU construction yet, the main frame was just put inside my messenger bag with no further protection. Perfect plan. At least it showed how much easier this case will fit than Freilite Alpha once it's completed.

Progress after day 1:

Transportation went flawless, I got my wiring chart setup, but I wasn't really focusing on making it beautiful. Just a somewhat good-looking proof of concept that I could use to test every component combination, no calculation of wire bends and stuff like that.

After day 1, I already had the 4+4 pin EPS 12V, 6 pin PEG and 6+2 pin PEG cables done.

Of course the PEG connectors are not planned to be used in full configuration, no ITX GPU has an 8-pin and a 6-pin connector. But both two 6-pin and a single 8-pin connector are common and able to deliver the same wattage, so I had to accommodate that somehow. Here's my solution:

Looks a bit dodgy at first, but the PSU has very few pins available, the 12V for the PEG connectors are just on two pins. The pins should be rated for 13A, so at a maximum rating of 312W I'm way above the 225W an 8-pin connector (or two 6-pins) is rated for.

It gets a bit dodgier when we look at the EPS12V cable:

This is just connected via a single pin, so the maximum power rating for that connection would be 156W. In theory, an 8-pin EPS12V connector is rated for more, but you don't want to use any CPU over 88W if the biggest cooler you can fit is an NH-L9i. This would be different for APU builds that utilise an AIO in the GPU compartment, so those would have to use a PEG-to-EPS12V adapter or get a different cable harness.

Progress after day 2:

Everything's done! I did put a little bit of effort into the ATX cable to make sure it would at least somewhat stay in this shape, but you could certainly do better. Either way, it's a large improvement over the previous situation.

In this picture, you can already see the PC light up, but it wouldn't actually POST at this stage. No display signal, and no speaker header for troubleshooting. The problem was that I didn't make a connection for the -12V rail. This was deliberate, as the FSP500-50FSPT was once considered as a bundled PSU with this case and it lacks the -12V rail, but apparently some Mainboards are checking every voltage before POSTing properly, so that option is pretty much out the window now. After crimping the last wire, everything worked out fine.

I also removed the blue LEDs cables from the power button while I was at it.

And now it's sitting on my desk, looking a lot nicer already:

I tried to resolve the slanted GPU by chaining the HDPLEX and LiHeat risers together, but I didn't get a display signal with that. Next thing to try is testing two HDPLEX risers chained.

First tests with the new PSU:

So as a simple first test, I just ran Unigine Heaven for 20 minutes or so to see what the new PSU could do. It's actually pretty darn quiet for something with a wimpy 40mm fan in the back. It didn't get any louder than the GPU itself, as long as the exhaust was unobstructed. The fan is not exhibiting any high pitched whining, just the whooshing noise of air turbulence. In idle the GPUs fans are much more audible than the PSU, so that's certainly a good thing.

Interestingly, it took the PSU a lot longer (about 5 minutes) to ramp up its fan to maximum speed when running the benchmark than the GPU (about 25 seconds). With a highly professional temperature difference measuring device called "finger", I also confirmed that the side of the PSU close to the GPU was much hotter than the one close to the CPU. I'll have to see whether that is actually caused by the GPU exhausting or whether the PSU has a natural hotspot in that area.

I've also taken measurements on the PSUs non-standard-compliant power connector and will post the results of that in its respective review thread.

As a bonus, here's a picture of my dog sleeping in a chair. Thanks for reading!

 

[Hahutzy]

Are the taps in your flanges just straight taps? Any plans on pems or extruded taps?

I made a change for my flanges to use pems.

 

[iFreilicht]

Right now they are. Maybe I'll go for extruded taps in the future, but clinched nuts aren't possible. The frame is made from 1mm thick steel, though, not aluminium as it was initially planned, so there's less risk of stripping them.

 

[iFreilicht]

Updated the first post to be an overview of the current specs and status of the project, as well as the second one now being an accumulation of all updates in short form. I also updated the compatibility spreadsheet to include all variants of R9 Nanos (spoiler: they all fit), some updated GTX 960s, all the 950s that were previously missing and socket 1151 boards. It was a pleasant surprise to see that pretty much all 1151 boards fit without restrictions, except for the ASUS ROG Maximus VIII Impact with its huge daughterboard. It fits, but it pretty much blocks all exhaust the mainboard has.

 

[CC Ricers]

The new test setup looks great. Much easier to see what's going on lol. I do think you should consider the idea of making the case modular as well. Thinking about transporting it, if it's too long to fit in your back pack you can split it up and put the GPU section in one of the smaller pouches.

Also given the proportions, are you sticking with the stand ideas from your first designs? As I'm building a slim case myself that I want to stand vertically, I am a bit concerned about balance and easily knocking it over. Your case is even taller and a lot of the weight would be centered in the middle because that's where the PSU is. Without a proper stand could it stay upright without wobbling?

 

[iFreilicht]

Thanks!

Making the case modular that way would be a huge engineering endeavour I am not ready to take on, even if it is very tempting. I was more thinking about a way to reuse parts to offer a case without a PSU compartment for PicoPSU users, or one for APU builds with integrated PSU.

Without a proper stand could it stay upright without wobbling?

As long as nothing touches it But seriously, it will need a stand for the vertical position, no doubt about that. I actually made a concept for one that can be installed without tools and takes up very little space, but I didn't share it yet. I'll model it once I find the time.