pulse metering: maximum size meterconstant?

by enervalis on

Hi,

In our final setup we need to make use of a quite big meterconstant (between 200 and 875).
I noticed in a test setup that with a meterconsant of 100 data was reported correctly, but when using 875 measured value is incorrect.
Could there be an overflow problem when using big meterconstants?

Kind Regards,

Frederik

‹ Number of clamps/phases can't get Flukso to work with smart meter and correct values with 2 x 50a clamps ›
gebhardm's picture
gebhardm |

The source provides the answer...
flm02/openwrt/package/flukso/luasrc/fsync.lua

  1. local meterconst = math.floor(real)
  2. local fraction = math.floor((real % 1) * 1000)

flm02/avr/a/ctrl.c
  1. ctrlWriteShortToTxBuffer(sensor[i].meterconst);
  2. [...]
  3. ctrlWriteShortToTxBuffer(sensor[i].fraction);

flm02/avr/a/main.c
  1. MacU16X16to32(state[sensor_id].nano, sensor[sensor_id].meterconst, ADC);

flm02/avr/a/main.h
  1. struct sensor_struct {
  2.         uint32_t counter;
  3.         uint16_t meterconst;
  4.         uint16_t fraction;      // expressed in milli
  5. };

So, if the multiplication runs into an overflow...

gebhardm's picture
gebhardm |

By the way: What do you mean by "measured"? Displayed in the dash or provided via REST or MQTT???
It maybe also the backward calculation of Wh into W with large time intervals...
You have to be a bit more precise in where you see an "error"...

enervalis's picture
enervalis |

Thanks

enervalis's picture
enervalis |

Related question: is there a way to reset the accumulated Wh (the counter value)?