sm7tix wrote:Network issue solved!!
I don't know what was wrong, but i rewrote the /etc/network/interfaces file and now it works like a sharm! I can run graph, manual and cam at the same time. GREAT work!
I get some error on read and error on write from cloud. Do you know how to debug that more?
Do you know how to set static values in the QT window? Volt for example could be between 2 and 4 volt. Easier to the trend when draining the battery.
Is there a way to get the graph-window without X?
I have setup a "mission control" on a headless virtual machine and it should be nice to have one console with the graph in in it.
When i run the manual rover code it picks up keystrokes from all my terminal windows. I will see if i find anything on the net on this.
The pixel lib, is it something done just for the spirit or is it a common lib?
Keep up the good work!
Very good questions but I don't know the answers to most of them
Ignore the read/write errors appearing on the clouddemo.py. They are actually generated by the rover.py so my guess is that it somestimes has trouble talking to the PIC. At worst you may lose a couple of sensor measurements per second which is nothing since you are sending over 20 samples per second.
The scale on the y-axis for the graphs should be possible to be locked to a certain range but I never checked how. It was the first time i use python to make any graphs and most of the code is copied and modified from an example. Speaking of seeing the trend, you can go in the graph script and modify the size of the arrays with the number of elements you want. I currently fill them with 100 samples of 0. You can make them 1000 if you want the graph to display a longer duration. I don't know what the max array size is in python. Maybe it is worth checking out if you want to see a trend. Otherwise, you can write a subroutine that will take multiple volt readings and average them and only update the volt graph once every few minutes.
graphs without X means curses on terminal really. I have never programmed anything with curses but it sounds like an interesting approach.
Good thing that your manual control seems to pick up the keyboard from inside terminal windows. I just confirmed again today on my laptop that it won't do that and I must open another application and select that in order for my keystrokes to register in the manualcontrol.py. No idea why this is happening on that laptop.
Today was the rover's first trip over the prototype 5G network and it performed beautifully.