Very good news. Thank you for your time.
Best posts made by pdinoto
Latest posts made by pdinoto
Random fatal error without callstack
i'm using neonious-one to constantly read an analog value from a water-pressure sensor to monitor my water tank level.
At random i found the board blocked, and after login the popup says:
An fatal error has occured in the program: TypeError: undefined not callable Program cannot continue.
The main problem is that i don't know what cause the problem since into callstack tab i have:
Paused outside program code, while waiting for events
What i can do to undestand the root cause of the problem?
RE: ADC pins accurancy
The documentation says:
Analog measurements up to 1.5 V. For use with setAnalogAttenuation. Only available if not on neonious one, as neonious one always allows up to 3.3 V at industry-standard precision (uses LPC822 for ADC, not the ESP32).
I have the problem with neonious-one, not a generic ESP32 board, this flag seems not applicable.
ADC pins accurancy
i have to read a 24v - 4-20ma output sensor with the board. I have attached an ADC pin to a 4-20ma -> 0-3V volts converter output.
The output of the converter is very stable, measured by hands with my multimeter, but ADC readings are very unstable, for reference with an output value of 1150mV i have readings between 1000 and 1200mV.
RE: IDE memory leak?
My first tests yesterday actually show that a small mem increase is normal, because additional resources are taken for sockets and so on which are not freed. However, after long usage it should stop increase.
The problem is, with the npm MQTT needing so much RAM, it is just difficult to check if it is a mem leak or a fragmentation problem.
And even if a possible error is fixed, you will always be near the limit.
Yes, it is from Espruino, but its only requirement is the net module, and this module seems to use the net module very standard-Node.JS like. So it might just work.
I will try, any suggestion to reduce RAM usage are very welcome!
RE: IDE memory leak?
As a workaround, I would suggest adding a lowsys.restart(true) periodically whenever it is OK for the device to reboot. That reboots completly including low.js.
In the meantime I will scan for memory leaks on my own...
Thank you for the suggestion, yes as a workaround i can save the current state of outputs (i'm currently testing the board controlling some high power led strip of my house), perform a reboot at late night, and restore previous state.
Memory fragmentation is becoming a problem