View Single Post
  #42  
Old February 5th 17, 05:01 AM posted to rec.aviation.soaring
Paul[_10_]
external usenet poster
 
Posts: 10
Default Thoughts on panel upgrade?

On Friday, February 3, 2017 at 11:37:47 PM UTC-6, Surge wrote:
On Friday, 3 February 2017 07:57:40 UTC+2, Paul wrote:
That sounds pretty cool! Does it work well? Do you have any details/photos?


Some photos
https://s29.postimg.org/700iji3ev/pisoar1.jpg
https://s29.postimg.org/j35u72eh3/pisoar2.jpg
https://s29.postimg.org/f82g4hvbb/pisoar3.jpg
https://s29.postimg.org/wmr9qihtz/pisoar4.jpg
https://s29.postimg.org/pktc4be87/pisoar5.jpg
https://s29.postimg.org/76it0c1xj/pisoar6.jpg

I used single side, copper clad, FR4 PC board soldered together to make the case. Don't do this unless you like to itch and make all your tools blunt! In hindsight an aluminium case may cost more and be a bit more difficult to manufacture but it would be better than working with FR4 board and look more professional to boot.
I made an oversized case because I have plenty of space behind the panel and to leave space for additional circuity later.
The screen is a 1000 nit, 7" Xenarc resistive touch screen (works with gloves) with HDMI connection to the Pi. GPS source is from a FlarmMouse via an S80 vario into RS232 to USB cable (using RS232 levels - not TTL). I'm not a fan of wireless technologies such as Bluetooth since they both introduce RFI and are themselves prone to external interference so I stuck with shielded, wired connections.
The 40mm DC fan is at the bottom and a covered vent at the top rear.

The Raspberry Pi 2B runs XCSoar on framebuffer so there are no overheads from a window manager. I can still launch an LXDE window manager to use a web browser once I've quit XCSoar.

The main reasons why I built it is because XCSoar rocks, I like tinkering and I didn't want to spend $2000+ on a commercial solution that I can't customise later or find out X years down the line that the device is no longer supported (please upgrade).
The OpenVario project looks like a nice solution without requiring R&D. If I had to do it again I'd probably go that route than build my own solution.