DAN A4-SFX: The smallest gaming case in the world

Would there be an option to buy both sets of panels (gray and black) as one package? (Something like the prototype package you got)
 
Prototype looks very nice, congrats.

I wonder if we can get a manufacturer like Silverstone or Cooler Master to make a low profile 120mm cooler that would fit in the A4.
I mean there's space over the PCI-E slot, which most "Mini-ITX" coolers have to avoid because it would hit the GPU. On the other hand, in the A4 there's a PCI-E riser there so they should be able to fit more fins in that area, and the area on the opposite side.
The problem is with the current CPU cooler height restriction in this case (48mm) there's not enough room for a 120mm fan+heatsink without violating the maximum component height of the motherboard.

Here's a drawing showing maximum allowable component heights on the motherboard:

HztPwn0h.png


In practice, the majority of motherboards keep the components at or under the 39mm limit, rather than the 57mm limit at the middle of the board. The ASUS boards with VRM daughterboard might be an exception to that. So:

  • 48mm max CPU cooler height
  • ~8mm added for the CPU socket
  • 39mm motherboard height restriction

48 + 8 - 39 = 17mm

So 17mm is what's available for the fan + heatsink fins. At best, you could have a slim 120mm fan (13-15mm thick), but the heatsink itself would be limited to the area of the socket (95x95mm for 1150/1155/1156), negating any of the benefit of the larger fan.

This is why low-profile 120mm CPU coolers have such thin fin stacks btw:

GUQaseth.jpg


That cooler is 10mm more than the max height in the A4, but you can see if the height were reduced by that much there would only be enough space for either the fan or heatsink fins, but not both.
 
After lurking this thread for a year I made an account just to tell you that your case is amazing and that myself and the thousands of other lurkers are very excited to buy it.

Been behind the scenes as well and I agree with you on this one. The case is amazing! I'd get a USB port cover though to mask that tinge of blue in the front.
Here's to hoping that the prize will be just as attractive upon the kickstart launch date.

And yes, this case is enticing not only for uSFF enthusiasts. The folks in the eGPU community are still waiting on this case as it is easily one of the most compact GPU+PSU+cooling cases you can find.
 
Last edited:
*snip*
That cooler is 10mm more than the max height in the A4, but you can see if the height were reduced by that much there would only be enough space for either the fan or heatsink fins, but not both.
First of all thanks for the detailed explanation :D
Secondly... I'm sad :(
It seems we'll just have to make do with NH-L9i or something similar.
I'm really interested in the Dynatron cooler with X99 though, I've been reading around. I read somewhere that it wouldn't work with a fan over it, because the fins/heatsink are designed in such a way that air is forced over it in one direction (like in server chassis). I figure X99 in the A4-SFX would be impossible if that were the case :( (at least for now)
 
Been behind the scenes as well and I agree with you on this one. The case is amazing! I'd get a USB port cover though to mask that tinge of blue in the front.
Here's to hoping that the prize will be just as attractive upon the kickstart launch date.

And yes, this case is enticing not only for uSFF enthusiasts. The folks in the eGPU community are still waiting on this case as it is easily one of the most compact GPU+PSU+cooling cases you can find.

Is there actually hype and discussion on this case in the eGPU community? That's awesome!
 
First of all thanks for the detailed explanation :D
Secondly... I'm sad :(
It seems we'll just have to make do with NH-L9i or something similar.
I'm really interested in the Dynatron cooler with X99 though, I've been reading around. I read somewhere that it wouldn't work with a fan over it, because the fins/heatsink are designed in such a way that air is forced over it in one direction (like in server chassis). I figure X99 in the A4-SFX would be impossible if that were the case :( (at least for now)
Not impossible, just impractical. Cooling 130-160W in this little space is going to be a challenge no matter what. A 1U copper finned heatsink + slim 92mm fan might work, though I'm skeptical something like the Noctua fan would have the static pressure to push enough air through the tight fins on those heatsinks. An alternative is a 1U blower cooler, but I can guarantee you wouldn't want to be in the same room with it if you value your hearing or sanity.
 
..I'm really interested in the Dynatron cooler with X99 though, I've been reading around. I read somewhere that it wouldn't work with a fan over it, because the fins/heatsink are designed in such a way that air is forced over it in one direction (like in server chassis). I figure X99 in the A4-SFX would be impossible if that were the case :( (at least for now)

forced what? cant you take conclusions of how it works from the pics? jesus, for someone who claims to do x265 encoding you sound pretty dumb dude.
 
hmm.. Black5Lion: to give you an idea, a dynatron T318 + NF-A9x14 should work.
I think a 92x92x20mm fan will also work with that heatsink. We should find out soon what fan configs work the best, but i'm guessing fans that are close to the side panel will work better.
 
The silverstone AR05 should work with a total height, with fan, of 36.7mm.

Now with that said, I would prefer allowing for the more widely available 58mm HSF setups. Not sure how much would need to change and if that is a feasible change at this juncture.
 
Now with that said, I would prefer allowing for the more widely available 58mm HSF setups. Not sure how much would need to change and if that is a feasible change at this juncture.

Not likely; from the FAQ in the first post:

The case is too small to support tall CPU coolers and i’m am not willing to make the case larger to support coolers taller than 48mm.
The current offering of low profile CPU coolers are able to sufficiently cool a Core i7-4790K with good temps and noise levels.
 
forced what? cant you take conclusions of how it works from the pics? jesus, for someone who claims to do x265 encoding you sound pretty dumb dude.
I meant since the fins only go one way, the air has to go in one direction (or two if the fan is on top of the heatsink).
Also, you don't need a PhD. to encode stuff and just because I do some x265 encoding (through VidCoder/StaxRip mostly), doesn't mean I'm some nerd who knows everything there is to know about computers. I'm here to learn as much as I am here to give ideas/suggestions.
hmm.. Black5Lion: to give you an idea, a dynatron T318 + NF-A9x14 should work.
This is what I've been wondering about, should work is theoretical. I want actual tests that not only see if the CPU doesn't throttle, but what temperature it runs at and what noise levels.
The silverstone AR05 should work with a total height, with fan, of 36.7mm.

Now with that said, I would prefer allowing for the more widely available 58mm HSF setups. Not sure how much would need to change and if that is a feasible change at this juncture.

The Silverstone AR05 is an LGA115x cooler though :( It would work with 2011-3.
I doubt he could change the width of the case at this point, and even if he could I wouldn't recommend it. Since it would just make the case bigger, and the NH-L9i/AR05 are already good enough if you don't overclock. In addition to the fact that increasing the width to fit bigger CPU coolers would leave wasted space in front of the PSU.
 
The Silverstone AR05 is an LGA115x cooler though :( It would work with 2011-3.
I doubt he could change the width of the case at this point, and even if he could I wouldn't recommend it. Since it would just make the case bigger, and the NH-L9i/AR05 are already good enough if you don't overclock. In addition to the fact that increasing the width to fit bigger CPU coolers would leave wasted space in front of the PSU.

this is a very specific build oriented case, you cant stuff whatever you want into it, if you want x99 you might have to look for something else. i'll be doing mine on a ft03-mini for example, reasons are flexibility and because i find it cute :D
 
this is a very specific build oriented case, you cant stuff whatever you want into it, if you want x99 you might have to look for something else. i'll be doing mine on a ft03-mini for example, reasons are flexibility and because i find it cute :D

Yeah, I realize there are sacrifices to be made. We'll just have to wait and see how it performs :)
 
The Silverstone AR05 is an LGA115x cooler though :( It would work with 2011-3.
I doubt he could change the width of the case at this point, and even if he could I wouldn't recommend it. Since it would just make the case bigger, and the NH-L9i/AR05 are already good enough if you don't overclock. In addition to the fact that increasing the width to fit bigger CPU coolers would leave wasted space in front of the PSU.
I wonder if one could pair a 120mm fan with one of these heatsinks.
The fan would then cover the entire heatsink instead of just the circular inscribed region.
 
Maybe a slim 120/140mm fan could be mounted to the side panel instead of the heatsink.
 
Maybe a slim 120/140mm fan could be mounted to the side panel instead of the heatsink.

That is whay I was planning to do with the deepcool gerstorm 120x20mm fan. I purchased the x99e-itx, dynatron t318 , fan, ram, 1tb ssd, psu and all I need is for my sister to stop by microcenter to get the 5820k. Once I have all the parts I will try and simulate the A4 to see how manageable the temps are. I'm really excited for this case and my build.
 
That is whay I was planning to do with the deepcool gerstorm 120x20mm fan. I purchased the x99e-itx, dynatron t318 , fan, ram, 1tb ssd, psu and all I need is for my sister to stop by microcenter to get the 5820k. Once I have all the parts I will try and simulate the A4 to see how manageable the temps are. I'm really excited for this case and my build.

Looking forward to seeing your results! :D
 
Is there actually hype and discussion on this case in the eGPU community? That's awesome!

Google "eGPU case" and see for yourself. There has been some discussion, but not much hype yet since the exact price hasn't been declared. That'll all change soon. :)

A little off-topic, the current eGPU setup comprises a Thunderbolt2 to PCIe adapter, a powered riser cable, the Desktop GPU, and a PSU (ATX or SFX).
c33a09f4d3792a6f52d62fbb81e12120.jpg


All these fit nicely into the A4:SFX (with additional space for fan), making it an ideal compact solution.
 
Last edited:
Yeah, I realize there are sacrifices to be made. We'll just have to wait and see how it performs :)

sacrifices meaning not being realistic to stuff a 120~140w multicore xeon into this case, not a big deal if you ask me, as most everything is good to go on it :)
 
Maybe a slim 120/140mm fan could be mounted to the side panel instead of the heatsink.

I thought about it, but if keeping the perforations as they come I don't think it will allow enough airflow or you will need a fan with a very high static pressure.
 
GUQaseth.jpg


That cooler is 10mm more than the max height in the A4, but you can see if the height were reduced by that much there would only be enough space for either the fan or heatsink fins, but not both.

that cooler is inefficiently designed for a height-constrained system. can you imagine video card with coolers like that? they'd all be 3 or 4 slots thick!

ITX motherboards need to be designed with low-height components around the socket, and the cooler heat pipes need to be replaced with a vapor chamber (equally effective, but lower height)
 
I actually testing all fan combinations. Noctua with default and 25mm FAN and Cooltek LP53 and Intel Box. The temps are very good but I will show you the results in the next days.
 
that cooler is inefficiently designed for a height-constrained system. can you imagine video card with coolers like that? they'd all be 3 or 4 slots thick!

ITX motherboards need to be designed with low-height components around the socket, and the cooler heat pipes need to be replaced with a vapor chamber (equally effective, but lower height)

It may be inefficient but (to restate what Necere said) it must be that way in order to conform to the Mini-ITX spec. What you're suggesting can be (and has been) done with 92mm heatsink/fans, but you won't see something much shorter than the cooler pictured with a 120mm fan for Mini-ITX.
 
You still have the backpanel with some ports and current boards tend to have more of them now rather than only two.

Also it might happen that oculus won't be the first one to get released and dominate the market so stay put and wait till something hits the stores.
 
You still have the backpanel with some ports and current boards tend to have more of them now rather than only two.

Yes I know you aren't limited to one total USB; it's would be more of a convenience thing since you're going to be walking around your room and will probably want as much cable as possible.

That said they might mean one cable for the HMD and one for the camera, so it may not even be an issue.
 
Yes I know you aren't limited to one total USB; it's would be more of a convenience thing since you're going to be walking around your room and will probably want as much cable as possible.

That said they might mean one cable for the HMD and one for the camera, so it may not even be an issue.

That's the beauty of the A4, fits in your fanny pack.
 
I don't really think that there'll be a lot of games with walking around with VR headset initially. Those type of games bring a risk of hitting something IRL and thus design

Guys at oculus are trying not to fail the launch of this tech like the last approach of VR did 30 ? years ago. So there's a lot of restrictions and guidelines for not making people sick and other design stuff.

I had DK1 and DK2 in the office at my last job and I wouldn't say that I'd like to play a lot walking around with it in my room. It could be a great experience when the game arena is prepared like those recent ideas showing up but inconsistency between the game world and real world make standing and walking around a weird workout where whole body is trying not to fall over even when not needed.

It'll be awesome for games where you're in cockpit of a race car, space ship or mech etc and maybe those where you still use controller to move around but we're simply not there yet with the tech supporting walking around with this at home. Try to remember the wii era of breaking tv's with wiimote.
 
Yeah, the walking around stuff will be new to the consumer version of the Rift. DK1 and DK2 were officially "seated experiences", but the Vive allowed you to walk around a small room (it also knows where walls are and puts up a virtual grid to let you know you're getting close to a real world wall). In response to all of this Oculus said that the Rift CV will have better standing-up capabilities and will allow you to walk around.

I'm actually surprised at your guys' lack of enthusiasm for VR! Having tried it I'm 100% convinced it will be huge!

(Also if the :p face in my original post didn't make it clear I'm not actually expecting dondan to change the design at this point, was just joking around).
 
This forum is now a VR chat room...if usb length is an issue buy a usb extender or hub. Options are out there. Dondan has stated multiple times that the system is in it's final stages and doesn'twant to make changes if not needed. Usb placement and quantity have already been discussed and the majority vote was 1 front usb.
 
This forum is now a VR chat room...if usb length is an issue buy a usb extender or hub. Options are out there. Dondan has stated multiple times that the system is in it's final stages and doesn'twant to make changes if not needed. Usb placement and quantity have already been discussed and the majority vote was 1 front usb.

+1...I agree
 
Yeah, the walking around stuff will be new to the consumer version of the Rift. DK1 and DK2 were officially "seated experiences", but the Vive allowed you to walk around a small room (it also knows where walls are and puts up a virtual grid to let you know you're getting close to a real world wall). In response to all of this Oculus said that the Rift CV will have better standing-up capabilities and will allow you to walk around.
I was talking about AR as the M$ HoloLens. Anyway what you stated is a ton of hacks and stuff required to make it work that will make the experience worse or more expensive.

I'm actually surprised at your guys' lack of enthusiasm for VR! Having tried it I'm 100% convinced it will be huge!
It's not lack of enthusiasm for VR in general but being cautious with ideas of moving around with VR headset. We're excited about VR and there are already few studios around in my country that are making the VR dedicated games but at the same time there's also a ton of new other stuff that we're focusing on like physical based rendering and using 3D scans in games(Check out "the vanishing of ethan carter" and "get even").

The problem with VR headsets and dedicated games is that you won't know how big the market share will be initially when you ship your game so you might fail simply because coming out too early

(Also if the :p face in my original post didn't make it clear I'm not actually expecting dondan to change the design at this point, was just joking around).

But some other people lurking might have not realised that and might think they need the two front USB ports for VR.

Getting back to the case talk - I really like the idea of using this case as a eGPU housing. It would be nice if lurkers from eGPU community made a mITX form pcb dedicated for use with standard cases and power supplies.
 
It's not lack of enthusiasm for VR in general but being cautious with ideas of moving around with VR headset. We're excited about VR and there are already few studios around in my country that are making the VR dedicated games but at the same time there's also a ton of new other stuff that we're focusing on like physical based rendering and using 3D scans in games(Check out "the vanishing of ethan carter" and "get even").
I'm more excited for the "big screen" feel of VR rather than the "moving around in VR", also did you see Virtual Desktop (maybe pair it with Display Fusion to get 3+ virtual monitors)? That is something I could see replacing my monitor(s) once the resolution of VR headsets is good enough.
Getting back to the case talk - I really like the idea of using this case as a eGPU housing. It would be nice if lurkers from eGPU community made a mITX form pcb dedicated for use with standard cases and power supplies.
It seems like it would work great for an eGPU case, but I think what Silverstone showed off would work better, we just need a PCB/port that's more common and works better/similar to thunderbolt.
I remember reading somewheer that USB 3.1 could be used to transfer different standards in special caes, maybe someone could make it PCI-E?
 
What silverstone showed off? You mean that last year's CES ASUS/Silverstone XG Station 2? The thing about toys dedicated to e-gpu is they are seriously expensive because they are dedicated.

Making a mITX form factor compatible eGPU board is a far better idea than leaving that to laptop manufacturers that will most likely make a proprietary tech fitting only high end/expensive ultra thin laptops. with mITX form factor you get to choose all other parts from anywhere you want.

As for the interface - I think it's doable but probably would be slow or unstable. If I'm not wrong USB, pci-e and pci are the same familiy of pci devices seen by the operating system.

Someone explained its problems here:
http://electronics.stackexchange.com/questions/141579/how-to-make-an-external-pci-adapter-that-runs-on-usb3-0
There are two challenges:

Latency

USB has two orders of magnitude higher latency than PCI. Many PCI drivers assume memory mapped access to registers on the PCI card with the CPU blocking on the bus transaction. Unmodified PCI drivers over a PCI-over-USB bridge are likely to fail because of wrong latency assumptions.

Check this study on PCI latency: PCI mean latency is in the order of 400ns, PCIe mean latency is in the order of 2000ns. USB 3.0 latency according to the article you linked is in the order of 30us.

Bus vs. Packet Network

While PCI is a bus, PCIe and USB are packet based networks. The PCIe root bridge hides the complexity of a packet network and presents the CPU with something that looks like a bus. Doing the same thing over USB is certainly possible, but if you want binary compatible drivers your hardware interface towards the CPU has to look more or less like PCI and you will end up with something like a PCIe root bridge. This is probably not something that can be achieved by common USB chipsets with *HCI interfaces.

Anyway I think we won't get proper eGPU support unless we drop the idea of having video output connectors on the card bracket rather than forwarding the output back through the motherboard and its connectors which would also benefit the console like form factor cases.

Sorry for going off-topic once again :(
 
What silverstone showed off? You mean that last year's CES ASUS/Silverstone XG Station 2? The thing about toys dedicated to e-gpu is they are seriously expensive because they are dedicated.

Making a mITX form factor compatible eGPU board is a far better idea than leaving that to laptop manufacturers that will most likely make a proprietary tech fitting only high end/expensive ultra thin laptops. with mITX form factor you get to choose all other parts from anywhere you want.

As for the interface - I think it's doable but probably would be slow or unstable. If I'm not wrong USB, pci-e and pci are the same familiy of pci devices seen by the operating system.

Someone explained its problems here:
http://electronics.stackexchange.com/questions/141579/how-to-make-an-external-pci-adapter-that-runs-on-usb3-0

Anyway I think we won't get proper eGPU support unless we drop the idea of having video output connectors on the card bracket rather than forwarding the output back through the motherboard and its connectors which would also benefit the console like form factor cases.

Sorry for going off-topic once again :(

Still OT:

Agreed, an mITX board that solely acts as an eGPU would be a much better idea than a proprietary solution.

As for the interface: Thunderbolt has the ability to transmit PCIe signals, USB C-Type could theoretically do that as well. Not sure whether the bandwidth would be enough, but those seem like the most probable solutions for notebooks.

Having all display connectors on the mainboard has down- and upsides:

As you said, console-like formfactors would highly benefit. Not only would the rear bracket be used exclusively for exhaust, it would also limit confusion for the customer as to which HDMI port he has to plug his TV into and it would give more freedom with placement of the GPU.
 
As you said, console-like formfactors would highly benefit. Not only would the rear bracket be used exclusively for exhaust, it would also limit confusion for the customer as to which HDMI port he has to plug his TV into and it would give more freedom with placement of the GPU.
Its not only good for console like form factors but mostly for the laptop eGPU's - currently you ether have to have proprietary transmission implemented in the notebook or you use the external monitor with the eGPU as far as I know. Which means your're converting your notebook in desktop workstation with screen, keyboard and mouse rather than simply attaching accelerator with cable while still chilling out on the couch :)
 
Back
Top