I just wanted to share something I encountered yesterday and what I discovered on this topic. My A30 has been printing flawlessly for over a year now, the whole time controlled by Octopi, including running some pretty long prints (longest was 54 hours). Yesterday, I loaded up a print I wanted to do of a low-poly tree stump shaped tablet holder. Octoprint estimated it at about a 27 hour print. The base of it covers a pretty good portion of the 320x320 print bed. Before it would finish laying down the first layer, I would get the dreaded communication error and Octoprint would disconnect from the printer. Never saw this before - EVER. So after reading the threads in here and elsewhere on the topic and trying a couple of things without success, I ran out to the local computer store, bought all new hardware (new Pi 3 B+, short printer cable, ferrite ring, 3 amp power supply). It was too late when I got home to replace it all and try it, so this morning, I had a thought before I replaced the hardware. I cued up another 11.5 hour print and Vola! No Comms errors. The conclusion I'm left with, is there is SOMETHING in the Gcode of that tree stump file that is causing the printer to stop answering pings. I have always used Simplify 3D as my slicer and uploaded files directly to Pi for print. I tweaked some settings in that file (sped things up a bit), so we'll see if that makes a difference, though don't know why it would.
I'd guess that your A30 has an 8-bit controller board. This—combined with small line segments for curves—can fill up the receive buffer at which point it might get unresponsive.
Well, I do believe that is the chip on the controller board. Any way around it you know of?
Slice with bigger line segments.
In slic3r:
I don't know which value this is in Cura; I've never had this problem myself.
I will note that there is a regression bug in Cura related to someone thinking that setting the default z speed to light speed might be a great idea. In hindsight, it's causing problems.
OK - I managed to compensate for it. In the first file, the print speed in Simplify 3D for the one that was failing for the first layer was 50% of 1800mm/m. I sped the first layer up to 65%, and it made it over the hump - Been running for 2.5 hours now without a hiccup.