Banana BPI R3 3D Printed Case

There’s also a 2xUSB2.0 header (2x5 pins next to 2x13) worth to be noted. USB 2.0 sockets may be put “2nd floor” above the rest of stuff.

Also extensive cooling system may require some more space (btw - does this thing really need an extensive cooling?)

I would say that absence of ventilation holes on the right side, together with active cooling fan on the top cover, would allow to cool down the radiator on 4G card (if present) much better than if the holes on the right side woulld be present… it’s just a thought though :grinning:

With fan maybe,but for my last tests, fan is not required for normal usage,so i suggest air ventilation holes on both sides. If do not want them you can close them with duct tape from inside of case or remove them before printing case.

Finas told me he will design higher case to allow plugging debug-uart and gpio. And yes,for usb2 header i do not know any angled connectors and the one i have are 16mm in height.

1 Like

sorry guys, been out of town again. will do a taller version this week. will also make a version with ventilation on both sides.

Added case that is 8mm taller ( from top of board to top cover is 20mm. made all the variations possible, like with double vent with and without nvme slot.

I haven’t test printed the taller versions. please report and I will then upload it to printables.

3 Likes

That is an amazing piece of craftsmanship! I am trying to clarify whenever it is possible to operate two fans on one header. Guess it only depends on the power amps. If possible I had planned to reach out to someone with a water jet or laser cutter to make some cut outs on the metal case.

However your nice work seems much more feasible, though. I have never done some 3D modelling but will give it a try. Anyhow, I wanted to ask:

  1. In addition to the ventilation honeycomb holes in the side panels, which are really appealing, why not using this style element instead of slits on the left and right sides …maybe for structural strength?

  2. Which fans did you chose? I thought of using two Noctua NF-A4x10 5V PWM with a shallow depth of just 10 mm. One cooling the chips, the other the Mini PCIe but with the purpose of pushing more air over the chips with copper heat sinks with fins.

  3. Any 3D-Print service you would recommend (I live in Germany) or would you actually consider to offer that yourself?

Edit: Copper Heatsinks with fins like these

https://de.aliexpress.com/item/4000107291369.html

Best Mike

1 Like

I plan to use this fan: https://m.de.aliexpress.com/item/1005003663915828.html?spm=a2g0n.order_detail.order_detail_item.4.20bb6368c57Vk7&gatewayAdapt=glo2deu

For metal case i search for a way to make ventilation holes…

Need my friend for using local makerspace for printing or converting the 3d case to a laser cut variant.

Hi @finas,

I tried to modify the stl files but, assuming because they are volumetric, it seems impossible to me. Any suggestion what my mistake might be or how I could modify the design?

Many thanks in advance Mike

Adding a second fan, even a very low profile one like the NF-A4x10 5V PWM , is almost impossible due to the cables. So please discard my question about adding a second fan intake.

Even after some minor cable management, though. I tried the other cables, instead of the RG178, which are shipped by default. Since they are so short, they cannot be managed in any kind of way and even become stressed since they are rigid.

IMG_3954

With the 8mm raise of the case I opted for a copper fin heat sink, aluminum is said to be better for passive and copper for active cooling, with a height of just 10mm. That leaves some spare room for air circulation.

I will cut this into smaller pieces with a water jet or a good old metal saw. Searching for adhesive thermal paste or pad now.

https://de.aliexpress.com/item/1005004023326782.html

Instead of connecting all 2.4 GHz antennas on one side and all 5 GHz antennas on the other side you will get better results if you interleave them and thereby maximizing the distance between antennas which are used for the same band. 2~3 cm between two antennas of the same band will not allow for good results with MiMo, the more distance between them, the better…

3 Likes

I thought about that too but could not find any source testing this theory out. However, since MU-MIMO requires multiple antennas, pointing two in one and the other two in the opposite direction might not be favorable. Since this is going off topic, I guess creating a new post would be good.

Edit: There we go

Thanks Mike!

In reply to your questions:

    1. My design started of from one I found on thingiverse that used slits on the side. I ended up designing it from scratch but kept that “slit” feature. So no real special reason. The honeycomb pattern would work just as well.
    1. I bought some cheap 40mm x 10mm 5v fans on amazon. I know there are 6 and 8mm versions but I wasn’t able to find any for sale. The 10mm one works well even with the heatsinks I used ( also bought from amazon, I ordered one of those boxes that have alot of them in different shapes and sizes ).
    1. no idea really, I print using the printer I have at the office. The only thing I would recommend would be using a material that is heat resistant, like ABS or PC or Nylon with carbon fibers.
    1. regarding the STL files, Have you tried copying the project I made on thinkercad and edit it yourself, instead of importing the stl files to whatever software you are using? this is the link: https://www.tinkercad.com/things/fqroXeQOz7a-banana-pi-rpi-3-case
    1. regarding adding a second fan, I added a taller version of the case that may allow for it. So there are basically two sized versions.
    1. I would be glad to print a case and send it to you. Currently the printer is printing stuff non stop that I need for some projects we are working on ( nothing related really, it’s printing some case parts for a visual AI sensor we are making ), but when this stops, probably next week, I will have print time to print a case for you if you want.
1 Like

You are most welcome and thanks for your comprehensive feedback! If my local 3D-Print-Service I found can’t help, I’d gladly accept you offer printing me a case … all expenses + tip will be covered :wink:

Does your 40mm fan by any chance have the same dimension as my nf-a4x10-5v-pwm:

  • Mounting holes 33mm apart
  • 37mm diameter intake

I place the cutout in the top case slightly off center (both 49mm form left and bottom) to the main CPU to cool the WiFi chips. The 4G Antennae, so the cables take the same space as those from the Wifi antennae, are placed:

  • 9mm respectively 29mm (centered above the SIM from the left
  • 16mm from the top

Just for reference in case any measure makes more sense to you compared to the current placement.

Cheers Mike

So I have two different fans ( in terms of physical aspect ), but they both have the same 33mm between mounting holes and 38mm diameter intake.

photo_2023-02-24_15-33-57

1 Like

Hey awesome work! But is the project still on tinkercard? I tried to follow the link but got an error

yep:

I missed sharing the promised 3D-Files. It is not fully production ready but is coming quite close. See my additions at the bottom.

All files provided in courtesy of Simon from 3D-Druck in Bensheim (no add, he’s a really nice guy!)

As follows the files with their respective modifications.

Case

  • Intake (hexagonal)
  • Quick access for m.2 Slot
  • Antenna mounting holes with flush site to prevent rotation upon tighten them

BPI-R3 Case 3D-Model.rar (662.6 KB)

Cover for SSD - Not available in metal case from sinovoip!

BPI-R3 Cover Bottom 3D-Model.rar (125.0 KB)

Top Cover

BPI-R3 Cover Top 3D-Model.rar (366.5 KB)

Final notes. There are a few modifications I, two them absolutely required, that are worth adding:

  1. Required - Wall thickness: In order to gain sufficient thread to mount the antenna, either add a cavity to fit the 8mm hex nut (basically flush mount) or simply decrease the wall thickness.
  2. Required - Reset buttons: Fine tune size to 4mm diameter and location of reset buttons
  3. Optional - Antenna holes in the top case: Align size and form to those Wifi-Antenna mounts of the main case
  4. Optional - Flush mount screws of the fan in the top case
  5. Optional - Jumper hole is quite tight

Thanks for working on this too

You can decrease the wall thickness and move screws onto extra areas (i guess these are the cause for wall thickness) and make wall same height as front for better alignment.

So from top view (front up):

_________
|•/   \•|

My not yet ready case:

That is quite a great idea relocating the top lid screws to attenuate the wall thickness. I believe a combination of both is favorable considering the longevity, i.e. when the threading get damaged, of the case.

I’d not make the walls too thin as, depending on the used filament, the mounting pressure and applied forces, i.e. when you accidentally bend a wifi antenna, could cause some structural damage. Using the 8 mm wrench was a bit annoying / wiggly but I am uncertain if, by flush mounting them, the filament can withstand the forces when tightening the antenna screws.

Maybe both are not necessary when the top lid designed to slide into the case entirely and only two screws, like in the metal case, are used for fastening on the bottom of the main case which should be sturdy enough.

This is my case, I 3D scanned the upper glove box in my Prius and not the R3 load balances three LTE modems :slight_smile: its battery powered now and the push button, well… its a push button. I need to add a few more fans though, the board been getting hot even with the little pwm fan.

anyone have a bright idea on the type of plastic i can use to route the lights on the board to this front area?