djacobow / pidog2 Goto Github PK

Having enabled this feature on a pair of lab systems, I noted that within a 24 hour period each system would be reset. The serial monitor revealed no anomalous voltage readings and then the pidog would switch off.

[ 0x00000229, 0x00000110, 0x0000001E, 0x0000012C, 0x0000001E, 0x012603A8, 0x038E0000, 0x141703D8, 0x00000000, 0x70640207, 0x033C0000, 0x02870000, 0x602D3647, 0x74636800 ]
[ 0x00000229, 0x0000010F, 0x0000001E, 0x0000012C, 0x0000001E, 0x012603A8, 0x038E0000, 0x13D103D8, 0x00000000, 0x70640207, 0x033C0000, 0x02870000, 0x602D3647, 0x74636800 ]
[ 0x00000229, 0x0000010E, 0x0000001E, 0x0000012C, 0x0000001E, 0x012403A4, 0x03840000, 0x13D103D8, 0x00000000, 0x70640207, 0x033C0000, 0x02870000, 0x602D3647, 0x74636800 ]
[ 0x00000229, 0x0000010D, 0x0000001E, 0x0000012C, 0x0000001E, 0x012403A4, 0x03840003, 0x140003DC, 0x00000000, 0x70640207, 0x033C0000, 0x02870000, 0x602D3647, 0x74636800 ]
[ 0x00000229, 0x0000010C, 0x0000001E, 0x0000012C, 0x0000001E, 0x012403A8, 0x03890003, 0x140003DC, 0x00000000, 0x70640207, 0x033C0000, 0x02870000, 0x602D3647, 0x74636800 ]
Powering off pi.
[ 0x0000020B, 0x00000000, 0x0000001D, 0x0000012C, 0x0000001E, 0x012403A8, 0x03890000, 0x13BB03D8, 0x00000001, 0x70640207, 0x033C0000, 0x02870000, 0x602D3666, 0x74636800 ]
[ 0x0000024B, 0x00000000, 0x0000001C, 0x0000012C, 0x0000001E, 0x01230007, 0x03A80000, 0x13BB03D8, 0x00000001, 0x70640207, 0x033C0000, 0x02870000, 0x602D3666, 0x74636800 ]
[ 0x0000024B, 0x00000000, 0x0000001B, 0x0000012C, 0x0000001E, 0x01230007, 0x03A80007, 0x13D10000, 0x00000001, 0x70640207, 0x033C0000, 0x02870000, 0x602D3666, 0x74636800 ]
[ 0x0000024B, 0x00000000, 0x0000001A, 0x0000012C, 0x0000001E, 0x01280001, 0x03870007, 0x13D10000, 0x00000001, 0x70640207, 0x033C0000, 0x02870000, 0x602D3666, 0x74636800 ]
[ 0x0000024B, 0x00000000, 0x00000019, 0x0000012C, 0x0000001E, 0x01280001, 0x03870000, 0x13A40007, 0x00000001, 0x70640207, 0x033C0000, 0x02870000, 0x602D3666, 0x74636800 ]
[ 0x0000024B, 0x00000000, 0x00000018, 0x0000012C, 0x0000001E, 0x011F0000, 0x03810000, 0x13A40007, 0x00000001, 0x70640207, 0x033C0000, 0x02870000, 0x602D3666, 0x74636800 ]

Code inspection suggests that the reason for the sudden cutoff is that the reading of the ADC is sometimes generating garbage that allows the code path to execute a cutoff as though the monitored voltage is too low.

Discussion with @djacobow led to idea that an EMA filter should be applied to the ADC readings. I've implemented his suggestion and have not experienced any additional power cuts as a result. Because of the severity of this problem, I will submit these changes as a PR today, but note that additional changes might be warranted.

The convergence time of the EMA filter, based on a 1V change in the monitored voltage, is several minutes. This is fine when monitoring a battery, but probably not ideal when monitoring the output of something where the voltage changes are expected to be more rapid (e.g. a solar panel). As such, the code changes will not apply the EMA filter to VSENSB. As such, its recommended to NOT use VSENSB with the LVDC/HVRC feature.

it looks so great, exactly what i was looking for.
Could you please confirm if it will work with Pi 4 hardware and latest raspberry pi operating system? If anyone has tested it with Pi4, please let me know and confirm.