Scratch Build – In Progress NUC NAS / HTPC / Desktop by Wolowitz

NUC NAS / HTPC / Desktop by Wolowitz​

For a quick access to my reports you can do a search within this thread using as search parameters my username and #Report_.
This tag #Report_ is generic. To gain in granularity and facilitate the search you could use the following tags or simply click on them.
Tags:
#Report_Build
#Report_Paint
#Report_Electro
#Report_Firmware
#Report_Software
#Report_Accident

Reports:

• Photo report 1: Manufacturing of the structural aluminum profiles and wooden feet
• Front panel electronic architecture
• Development boards
• Photo report 2: Base aluminum panel
• Photo report 3: Side wood panels and strips

​

1. Specifications:

• Module size
  140x140x45mm without wooden feet
  175x175x50mm with feet
  
  NUC NAS size - 2 modules (NUC module + HDD module)
  175x175x100mm with feet

• External components are in wood, except the touch buttons and logo which use black anodized aluminum.
  
• Internal components are mainly in aluminum to ensure a minimal EMI protection.
  
• Plexiglas + RGB LED are used to fill the gap between 2 modules.
  
• Metal fan mesh
  
• 2-3 internal Wifi antennas (optional)
  
• The HDD module is compatible with 2 x 2.5” HDD (15mm height max). It integrates a SATA port multiplier Raid 1/0/JBOD controller card. The HDD module may also contain a USB 2.0 hub if I find an appropriate one. For existing NUC motherboard an USB 2.0 <->SATA adapter is needed, since there are no small, compact Thunderbolt<->SATA adapter available. The next NUC based on Haswell will have an SATA port.
  
• The front panel contains an electronic circuit with probably an ARM microcontroller connected to the PC over USB. It drives the different LED used for the gauges, touch buttons, pictograms and text (everything is backlighted by RGB LEDs)
  The front panel is easily interchangeable, it slides vertically, you only have to loosen 2 or 4 screws (I may use a screwless mechanism). No need to connect/remove any cables, the front panel uses spring loaded contacts.

I will start with the NUC NAS - 2 modules version, other versions may fallow perpaps:
NUC HTPC (1 or 1.5 module)
NUC Desktop (1 module)



2. Concept photos

(Click on the low resolution thumbnail to get the picture in high resolution)




Side & back view


NUC & HDD modules



Disassembled
Only the NUC module is shown. The HDD module is the same except that it has a different backplate and bottom vent holes, and an additional HDD mounting cage.
The top panel is coated with an aluminum panel (not modelised yet), for EMI protection and screwless mechanism to hold the top panel in place.

​

3. Front panel electronic

The front panel may also include a light and proximity sensor, to adjust the LED brightness according to the surrounding light conditions to minimize the power consumption.
The gauges will have less segments (1 or 2 per gauge) than shown on the pictures.

1. Touch sensitive buttons, black anodized aluminum. Allow to trigger functions on the PC
   • Start/Close Webserver
   • Start/Close Media Streaming Server
   • Indicates if Windows Updates are available (solid yellow dot), if updates are being installed (blinking yellow dot), if the updates installation failed (red dot). Allow to start the installation and also to restart the PC if needed after the installation (switching green/yellow + user touch to restart the PC)
   • Launch Server backup (you need to connect the external backup HDD before)
   • 2 spare touch buttons available
   
   
2. 2 round gauges
   • The LAN pictogram can also indicate if the LAN connectivity is limited or if no internet connection is detected (red or orangedot pictogram)
   • Shows the LAN download and upload activity.
   • CPU core activity.
   • Or both, CPU and LAN activity at the same time
   
   
3. Gauge that could serve for
   • Indicates the storage amount used on the HDD
   • Indicates the RAM quantity used
   • As Progress bar, for instance when a client PC backup is occurring combined with corresponding client PC backup client LED dot. (see point 6)
   
   
4. Temperature indicators. They warns if the temperature exceeds predefined limits (only if I2C temperature sensor chips are installed)
   • Moves from solid green, to solid orange and blinking red according to threshold values defined in the software.
   • May also indicates (don't know how yet ) if the controller switched on one of the internal fans (The HDD module may include one, the NUC module maybe not)
   
   
5. Power touch button
   • Solid blue: PC running
   • Solid or blinking orange: WHS warning
   • Solid or blinking red: WHS error
   • Solid Yellow (or whatever color or effect): PC in sleep state
   
   
6. Touch buttons. It’s a list of PC on the local network
   • Text (PC name) indicates the PC state (on or off)
   • Contains a pictogram if there is an WHS warning or error for that PC.
   • Contains a pictogram that shows if a client PC backup is occurring (blinking blue), if the last one succeeded (solid green) or if the last one failed (blinking red).
   • The touch button allow to WOL the PC or to launch the client PC backup.
   
   
7. 2 Touch buttons to modify some settings of the gauges
   

​

4. Current status of the project
(Will be updated regularly)
Last update: 23rd August

​

Stay tuned


(Sorry for my bad English, if something isn't clear in the description, or if there are too many grammatical/orthographic mistakes, don't hesitate to inform me with a private message. )

 

Photo report 1: Manufacturing of the structural aluminum profiles and wooden feet
Tags: #Report_Build

Plan​

​

1. Aluminum structural profiles

1. Cutting L-aluminum profiles, 10x10x1mm (1.5mm thickness would have been a better idea)
   
   
   
2. Marking the different parts that need to be cut off
   
   
   
3. The cutting is done with a jigsaw (Yes, still have all my fingers )
   
   
   
4. Cutting the extremities in an angle of 45°. Had first to find a way to keep those profiles in place for a precise cutting, I used a metallic squared profile and a screw to keep the element to cut at the right distance.
   
   
   
5. And the result was this. Enough structural aluminum profiles for 5 modules. (some are missing on the picture, like the vertical corner profiles, they look like those on the left on the picture except that they are 1mm less wider and have M2 threads to hold all those horizontal profiles that we see on the photo.
   
   

2. Now the wooden feet

• Cutting the wooden feet at the right length with a manual miter saw. Sanding the extremities and adding marking for the next step.
  
  
  
• Drilling some areas unreachable by Dremel grinder discs (on the right).
  Cutting with the grinder discs and finish the work with chisels.
  
  
  
• On the left the expected result. On the right the unexpected result
  Pine wood, it's a softwood easy to work with, maybe too much
  
  
  
• The 8 feet drilled
  
  
  
• And here are the 8 cylinders drilled although not being totally finished.
  I drilled all the aluminum profiles to pass the screw through and did the first assembly.
  

 

reserved 2

 

This looks like it is going to be good I'm Subbed.

 

definitely Subbed.
This is gonna be good, really interested in seeing the front panel take shape.
Very cool

 

Amazing!!

The design is superb, but the front panel is absolutely a master piece

 

@Mechh69, LegendaryScott, longyny

Thank you!




I just edited the second post, and added the first photo report

If someone has any improvement proposals don't hesitate to tell me.

 

This looks really excellent Wolowitz, subbed.

 

your planing looks excellent. The front panel is an amazing idea and can wait to see it come alive.

 

Looking good Wolowitz! Intrigued to see how you're going to make the front panel.

 

Nice build in general but that front panel will be awesome. Look forward to seeing how you realise the concept.

 

Front panel electronic architecture
Tags: #Report_Electro, #Report_Firmware

I have that picture of a fan controller prototype I did several years ago, it gives you a rough idea of the visual effect it will have. Don't pay attention to all those light leaks, I will use a different technic this time. (Sorry for the bad image quality)



The gauges PCB will look similar to this one. It was a prototype for a previous Mini-ITX project before switching to the NUC one. There are a bunch of traces missing, the purpose of this PCB was only to check if there were enough space to place all those LEDs and how to place the thermal pads, before designing a clean one from scratch.



I will do the firmware/software development in several steps, and it may last years . I'm a .Net developer, I only have basic electronic knowledge.

1. First I will develop a .Net PC application that will directly manage the different communication buses on the different PCBs. No microcontroller connected to the PCB, only a basic I2C<->USB adapter.
   
   
2. Next I will add a STM32F4 (ARM Cortex M4) controller card. The next step will be to transfer the business logic into the ARM firmware, so that the PC application only needs to send raw data to be displayed, all the visual effects and bus communication would be managed by the controller. The PC application will only be limited to a bunch of plugins that collect data on the PC to send to the ARM controller, nothing more. I will have several firmware development choices.
   
   
   • I will use either the .NetMF (Micro framework ) from Microsoft that allows me to program the controller in .Net so that I may reuse parts of the PC application source code on the ARM without big modifications to do. But it seems that the .NetMF is too power hungry.
     
     
   • Or I will use an ARM real time operating system like ChibiOS, but it has to be programmed in C or C++.

But the main problem will be the costs due to

• the limited volume, I only need one of it.
• the huge amount of RGB LEDs, LED drivers, voltage regulators, temperature sensors (for a minimal security, I don't want the house to catch fire ), voltage/current monitors (and some of them are only available in the USA, I can't find some of them here in Europe)
• the assembly costs since the components need to be soldered in a solder owen because of their tiny packages (QFN,...).

But I want it!

 

DAT FRONT PANEL!

This looks amazing just from the renders! I'll be following this closely.

 

Have 3 questions.

1. Which version do you prefer?
   (Technically only the first seems to be possible for me I think. Visually I prefer the second one.)
   
   
   
2. The illuminated plexiglass is thicker on the sides (5mm) than on the front (3mm), does it disturb you?
   
   
3. On the back side there is no illuminated plexiglass between the modules. The reason is that there are breathing holes on the bottom panel of the upper module (the Nuc module) to cool down the motherboard and mSata in the case of the lateral breathing holes were not enough. Does it disturb you? (I may intervert the NUC with the HDD module, technically there is no modification to do, I only need to add the plexiglass piece)

Thank you


EDIT:
Search tags are now clickable

 

I agree, visually the second one is the most pleasing to me. I think its definitely doable and if I understand your construction correctly I think it could be done fairly simply

 

Well the problem is that I want an uniform illumination of the plexi and avoid bright dots where we distinguish where the LED are placed behind.

http://www.overclock.net/t/1187138/...ing-testing-murdermod-luminous-panel-analysis

I want this:


Not that:


It is only possible if there are engouh distance between the LEDs and the plexi surface (~15-20mm).

The front panel has a depth of 15mm (could have been 12mm but I prefer to place the voltage regulators on a seperate PCB and fix them with a thermal pad on the alluminum panel serving as heatsink. If one VR fails, it could easily be remplaced).
That means that the front plexiglass strip will go through the whole front panel, the LEDs, like the others for the side plexiglass strips would be placed between the modules, not on the front planel electronic.
The reason? I want the same visuel effect on the front plexiglass strip than on the side ones.
I can't do the same for the round plexi strip that we see on the 2nd variant for 2 reasons:

1. I need to pass Sata, USB and power cables from the NUC module to the HDD module, and this happen in the center part of the gap between the front panel and the modules.

2. The PCB of the gauges will be big, and I also need to pass some traces from the upper PCB (the gauges, buttons, VR) to the bottom PCB containing the other buttons and ARM. It is impossible to make that the plexiglass reaches the other end of the front planel where the LEDs are placed.
​

There may be 2 solutions, if a majority of people really prefers the 2nd variant.
I could place the LED of the plexi strip on the same PCB like those for the gauges and boutons.
But as the PCB is only 6.5mm from the surface away combined with the large surface of the plexi, it won't produce the same visual effect as the plexiglass strips placed on the left and right side of the case. But I could try it

The alternative would be to use EL tubes, but there aren't RGB versions available, and I find that they are not bright enough.


I will publish an disassembled model of the front panel when it's done (it may take a while).

 

CoooooooL! Crazy surface mount PCB coolness too!
Don't forget to use something for size reference. You really don't convey how awesomely small everything is until you show it in the palm of your hand or something.

 

yep, I still think its very doable. You just need to think outside the box a little

 

We will see, if it work it would be great.

I tried to render 3-module NAS to see if I could reuse the electronic PCBs without changes.
Well, the result is not very pleasant to see unfortunately :-/








The radius of the round gauges needs to be increased like this, and the bouton moved elsewhere:


I think that version will never exist.


I will try to get further with the build this weekend and make the internal aluminum cage of the modules, even if I have I little technical problem to solve first.



It would have been interesting to be able to use the USB 3.0 front connectors of the upcoming Nuc boards with a Delock Adapter USB 3.0 (male-female angled 270°).
I first thought it wasn't that deep, but I only received the exact dimensions from Delock recently, and it is a little bit too late to add 5mm to the structure .

 

I like how your building this even though it's not in the competition. I see why it was so difficult to pick just five.

Looking forward to seeing how your front panel turns out, really clever stuff