Have a strange situation. One standard OEM wired temperature sensor connected to a 6-port breakout board is not giving correct temperature data. The sensor input seems to be being treated as cumulative data as I can see increases in the cumulative total when other linked temperatures are increasing. The firmware was updated to allow 12 sensors to be connected via the published process and don’t have same problem on a similar test setup. Other identical sensors connected to the same board are OK.
For scaling, the value has increased from 0 to 60 in about 3 weeks.
I can give some more detail on the specific setup but first wanted to just checked if there was any obvious cause we were missing.
Connected to what? emonTX 3, 4?
On the emoncms inputs page, is it showing the right temperature?
Connected to the RJ45 slot on an Emonpi. No, the Inputs page shows the slowly cumulating value.
@Robert or @TrystanLea will be able to help.
Probably also worth saying that sensor was replaced and breakout slot was changed around, but response is the same. Disconnecting the sensor does not reset the value. However, the new sensor started from 0 with the same type of response.
What about the Emonhub log? This is about as close as you can easily get to the One-wire input itself, without delving into the front end software in the ‘emon’ part of the emonPi.
How and by whom?
Is this the temperature reported on the Inputs page, in degrees Celsius? In 3 weeks, this is about 180,000 readings, so it’s acquiring one three-thousandth of a degree per reading?
I can’t think of a mechanism that would do this, either in the DS18B20 itself, in the emonLib or emonLibCM software or in the standard front end sketch. I think it must be in the “updated” software that a mistake has been made.