For me that is strange. If you want to spend a raspberry - why not drive the display and the buttons directly?Mausnoab Williamston wrote: ↑Fri Dec 31, 2021 8:29 pmSeems like you are focusing currently on the hardware side.
Yes an no. I think in such project you have to keep a eye on the both sides. I learned programming in the times of magnetic core memory () and the "Wang 2000". So I keep every time an eye of the capacity of the hardware. And so I asked myself which capacity abuse the detour if you can´t make the thing direct. In the side of programming the detour is very much easier. Perhaps. But for an Arduino it´s very easy take out the serial information directly. Okay let´s see:
Oolite - MQTT - MQTT - Raspberry - (Arduino) - display
This - in my perhaps false opinion - costs capacity cause of the detour.
A raspberry has enough power to run a full linux distribution so it is the second computer you mention.
Otherwise you have a full PC in front of you with more compute power than Oolite requires. Another process is required anyway to get data in/out of Oolite, so this process could drive your addons via USB directly.
Less components usually mean less room for failure.
So you want to optimize the solution. For a custom build I believe the main cost are development cost. The one-time hardware purchase is neglectible.Mausnoab Williamston wrote: ↑Fri Dec 31, 2021 8:29 pmThe capacity of Mousenoabs-Micro controllers.It is economical thinking. If you have a big realistic project with realistic landscape and more. You will need a second computer to make it. So I think with a game like OOlite and some funny looking displays depends what do you want but one or two Arduinos are mostly enough. This is the combination of an Mega and a Leonardo/Micro(Pro). If you have a typical project one display and a button box mostly the Arduino pro micro seems the best. (It is to build a Joystick with 56 buttons).(And this are not the costs, calculate one button with 50 cents (50 x 56 buttons (28$), the Arduino Pro I bought for 5 Dollars!)
My concrete questions:
1) How many capacity costs the detour with using the debug-file and MTT in the computer? (I have two Raspberrys here.)
2) Is there a way with f.e. javascript and OXP to write the variables I need directly to the serial port (like to the monitor)?
Happy New Year!
So I'd throw in some processing capacity to keep development cost low. By using an ESP32 controller (e.g. integrated into a D1 mini board) I know I have WiFi, TCP/IP and can run MQTT directly. Communication to a PC is really easy, and the gateway process for Oolite already exists as Cholmondely had found out. If you like you can still use serial ports and USB.
What I still see as a challenge: I have no clue whether that ESP32 has enough power to drive a display, especially if you want touch capability. And I would not know how to wire the pins to support a joystick (well that would be a few analog pins) with 56 buttons (that exceeds the number of available pins so some better logic is necessary).
If I were to reuse available Raspberry's that would even ease development further as you do not have to code in C but can choose the language. Some time ago I bought the [https://www.raspberrypi.com/products/ra ... h-display/]official touch display[/url] to build a nice little appliance. The touch is a bit odd when programming yourself. But as soon as you run X everything comes in naturally and you can focus on the application displaying your gauges and receiving touch events. It could be a nice way to add features later to your cockpit.
But also the Raspberry does not provide enough pins to connect joystick and 56 buttons directly, so the wiring problem remains. But solutions exist anyway.
