Overclock.net › Forums › Specialty Builds › Small Form Factor Systems › [Build Log] Freilite Brevis S - GTX 1070 and FlexATX in a <5L package
New Posts  All Forums:Forum Nav:

[Build Log] Freilite Brevis S - GTX 1070 and FlexATX in a <5L package - Page 2

post #11 of 86
Thread Starter 
Quote:
Originally Posted by ccRicers View Post

The layout looks nice. Looks like it will fit in a standard 1U rack. My Slim Machine build is aimed to be more like the PS4 in shape, it has a more square profile.

Also, having a small access "door" on the side which you can unscrew to install or replace hard drives would be nice from a user friendliness point of view, in the similar way that modern game consoles do. This might be more difficult to do with the way you have your hard drives arranged. But I want to do something like this in my case. Instead of fiddling with cables, the hard drives attach directly to SATA adapters that are fixed inside the case.

I suggest making a angled stand with a bit of style. There is a SFF build about 3-4L in size that I saw on ExtremeSystems but I can't remember the name. It had a stand that was angled diagonally when you looked at it from the front, which added some asymmetry to the design.

Thanks! Nope, too thick for 1U biggrin.gif

Yeah, yours does fit a larger GPU as well, which is nice, but I assessed that a mITX 970 is enough for me.

Well you can just unscrew the top panel, that will give you access to every component in the case. I could make those panel into three separate pieces, but I have no incentive to do that, really. And yes, that would be very difficult when using two HDDs. When just using one that is oriented in parallel with the PSU, it would be quite a bit easier.

Angled could mean pretty much anything, I'm not sure what you exactly have in mind there.
post #12 of 86
For my case, I'll be putting the hard drives stacked vertically, in the empty space in front of the GPU. There will be a small side panel to unscrew to access those hard drives, so no need to take the entire lid off the case.

What I mean by angled is to have an asymmetric angle look when viewed from the front. I found the case I was talking about here. The angled stand provides balance as well of breaking the boxy look.
post #13 of 86
Thread Starter 
Quote:
Originally Posted by ccRicers View Post

For my case, I'll be putting the hard drives stacked vertically, in the empty space in front of the GPU. There will be a small side panel to unscrew to access those hard drives, so no need to take the entire lid off the case.

What I mean by angled is to have an asymmetric angle look when viewed from the front. I found the case I was talking about here. The angled stand provides balance as well of breaking the boxy look.

Hijacking my own buildlog here, but did you think about using a few of these guys? You can unscrew the PCIe bracket which makes them barely thicker than a 9.5mm HDD, I'd say about 11mm. You could stack three of those on top of each other and hide them behind the panel you're talking about.
I thought about using one of these myself and just have one HDD, but I'd have to make the case a bit thicker for that, which I don't really want to do.

Ah I see. I wanted to make the case look like it's floating anyway. That was a very good inspiration, thank you! Oh and that case looks pretty cool as well, reminds me of LittleMac and Neutronium V3.
post #14 of 86
Hot swappable drive bays are a good idea. They'll still be inconspicuous being on the side of the case, and won't interfere with the stand if placed vertically. I'll probably get one of the Icy Dock ones that holds dual 2.5" drives.
post #15 of 86
Thread Starter 
Small Update:

Today's work so far was designing new intake holes for the top panel, which are, let's say, inspired by the PC-Q12:



I may make a few other designs as well and then do a poll to decide which is received the best smile.gif

And I also made the reversible PSU mount, which I earlier assumed to not be possible. I surprised myself.



This now allows to install FlexATX PSUs in any orientation, so no matter which PSU I choose to install or if I later want to upgrade, I can always install it in such an orientation that the hotter side is farther away from the HDDs. Good stuff.

As you can see, one screw is only holding on to the edges of the mount, but that's fine as the PSU won't have much room to move anyway, so I don't see any stability issues there. You could probably use just two screws and be fine. Maybe there are screws with an extended flat head that grip more of the metal there.

EDIT:

One last thing I did today: (Audio for the win)

3VTsnTE.png?1
Edited by iFreilicht - 5/17/15 at 3:36pm
post #16 of 86
Sub sub biggrin.gif
Lori Lemaris 023
(21 items)
 
Felicity Harmon
(23 items)
 
Moneypenny
(12 items)
 
  hide details  
Reply
Lori Lemaris 023
(21 items)
 
Felicity Harmon
(23 items)
 
Moneypenny
(12 items)
 
  hide details  
Reply
post #17 of 86
Thread Starter 
Quote:
Originally Posted by WiSK View Post

Sub sub biggrin.gif

Hi WiSK, great to have you on board! thumb.gif

Sorry that I don't have any large updates, there is a major Development Issue I encountered.

Front audio

There are quite a few ways I can do this. The more complicated, the higher quality the final solution will be. The problem is that the different solutions may require different sizes for the PCB for the front audio connection.

The pin header in the following schematics is the HD_AUDIO connector.

1. Easy, dirty way:

So the most easy way would be to just use a TRRS 3.5mm jack without a switch, connect that directly to the audio header and shorten the two sense inputs to the sense signal pin like so.



Advantages:
  • Easy to make
  • Few components
  • Small size
  • Low cost

Disadvantages:
  • No EMI filter
  • Noisy, bad quality audio output
  • Mainboard and OS always think a mic and headphone is plugged in

Maybe I'll still make this one first to just have something that works, but it's not desirable in the end, having unfiltered audio on the front ports would result in high noise, so I wouldn't ever use it.

2. Sophisticated solution with slight problems:

This requires a TRRS 3.5mm audio jack with an isolated, normally-open switch that will shorten the two sense inputs to the sense signal only when an audio jack is plugged in and it requires the use of three to four capacitors and inductors for the EMI filter.



Advantages:
  • Good quality audio
  • OS and board know when the front audio isn't used
  • Reasonably easy to make

Disadvantages:
  • Can't differentiate between regular headphones and headset with mic
  • 3.5mm TRRS with isolated normally-open switch is hard to find
  • No idea what inductors to use

The second issue can be worked around. I found a jack with a normally-closed switch that could be inverted by using a MOSFET, but I'll be searching one that works with this circuit out of the box.

The last issue is the most problematic right now. The Intel I/O connectivity design guide merely annotates the inductors with: "ZL should be 600Ω or greater @ 100MHz with a low Q (broad Impedance curve over frequency)". I have no idea what to do there, what kinds of components I need. That's why I can't really say anything about whether this circuit would even fit inside the case or not, that entirely depends on the size of the inductors.
If anyone could help me out here, I would highly appreciate it.

The capacitors are very small, though. Maybe I can eliminate one of the inductor-capacitor-pair because the microphone effectively only has one channel, not two.

Incidentally, it seems like most PC cases don't implement this EMI filter, that's why front audio is so shoddy on most of them. There's also the option of using just the isolated switch mechanic and leaving the EMI filter out.

2.5. Super sophisticated high-end solution.

In order to eliminate the differentiation problem of 2. between headphones and headset with mic, one can use a Texas Instruments TS3A225E, which can sense whether a mic is connected to the second ring of the jack and switch the sense signal accordingly.
This would make the whole circuit more complicated, but it would make it work out of the box with every mainboard out there, all adapters from TRRS to mic+audio and all headphones and even regular microphones that connect via the 3.5mm jack. I would really like to employ this Solution, but it could take quite a while to make it happen.

I know that laptops have different ways of solving this, their drivers recognise that the audio port is a TRRS so it can ask you what kind of device you plugged in every time, but I don't think modding drivers is a solution for me.
post #18 of 86
Thread Starter 
Small updates:

1. HDD mounting:

I finally decided to stick with adhesive tape. Designing a sophisticated mounting system has proven to be quite complicated and make installation of the HDDs more complicated.
The tape is not only a cheap solution, but it's also decoupling the drives and allows multiple orientations and positions for the HDDs, so no matter what components are used, at least one HDD will always fit.

Also, I decided against making the case a little bit larger to make two HDDs work in every configuration. Not only do I hate making anything larger, but every mm I add to the height of this case will decrease the compatibility with bags, and that's one of the main features I am going for.

2. Front Audio:

So this again. I finally found out how I can find the inductors I need.

An example on digikey: You don't go to the "Inductors" section, but the "Ferrite beads and chips" section, and there you can sort or filter them by "Impedance @ Frequency"! So, looking for >600Ohms @ 100MHz, I have quite a lot of choices. there even are SMD ferrite chips. So that's solved, I can easily fit those components on the front Audio PCB.
I'll look into the TI mic sensing IC as well. I really want this to be a plug and play solution, so if I made that work, that would be awesome.

3. PCIe Riser:

After talking to LiHeat and even getting a few pictures, I can now say for sure that the solution I envisioned for the riser will work. The riser is about 2.2mm thick and has a bend radius small enough to fit into the cavity the PCIe bracket encloses. They also made the surface non-conductive, so I won't need an additional shield underneath the GPU.
I got a quote for a somewhat shorter sample that I will use to try out how large the gap between the PSU and the mainboard really has to be and what kind of optimisations I can do there.



So yeah, things are happening. I wanted to ask metalworking shops for quotes this week, but I really want to have the parts finalised first.
Edited by iFreilicht - 5/28/15 at 6:39am
post #19 of 86
Great project, looking forward to seeing the actual case!

I initially didn't get what the difference between inductor and ferrite is. but I've found this short post on another forum:
Quote:
Inductor opposes current change & block AC signals. Ferrite beads suppress high frequency signals. Normally you use inductor in any power output especially switching regulators to separate the AC & DC signals then take advantage of opposing the current change in the circuit. you use ferrite bead when you want to separate two circuitry especially when application of certain circuit is high frequency. Hope this helps.

So the same function it seems but Inductors are meant for higher currents/lower frequency and ferrites are meant for lower current/higher frequency is seems. You learn something everyday biggrin.gif
post #20 of 86
Thread Starter 
Quote:
Originally Posted by QinX View Post

Great project, looking forward to seeing the actual case!

I initially didn't get what the difference between inductor and ferrite is. but I've found this short post on another forum:
So the same function it seems but Inductors are meant for higher currents/lower frequency and ferrites are meant for lower current/higher frequency is seems. You learn something everyday biggrin.gif

Thank you very much! Yeah me too, I can't wait to get the current stuff finished and get it made biggrin.gif

Yup, that's what I found as well. The confusing thing is that Intel calls the components "Inductors" in their Design Guide, not "Ferrite chips/beads" and ferrite chips and inductors both are inductive components in a way, at least that's how it sounds to me.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Small Form Factor Systems
Overclock.net › Forums › Specialty Builds › Small Form Factor Systems › [Build Log] Freilite Brevis S - GTX 1070 and FlexATX in a <5L package