dfrobot Gravity: I2C BME280 Environmental Sensor

AI context

Dense context for pasting into a model prompt — addresses, pinout, voltages, tolerances and the vendor's reference circuit.

PART: DFRobot Gravity: I2C BME280 Environmental Sensor, SKU SEN0236, 22 x 25 mm, 12 g. A Bosch Sensortec BME280 (U1 in DFRobot's schematic) on a Gravity carrier that measures temperature, humidity and barometric pressure in one package. part is null because the vendor's own product title already carries the chip name, so lifting it out would print BME280 twice.
I2C ADDRESSES: DFRobot publishes none on the product page or the wiki, so addresses is empty rather than filled with a remembered BME280 default - scan the bus. The board exposes exactly one I2C pair, SCL on pad 3 and SDA on pad 4, also available on the second Gravity connector.
PINOUT: 1 '+' = 3.3-5 V supply, 2 '-' = GND, 3 'C' = SCL, 4 'D' = SDA. DFRobot numbers these four and prints the single-letter labels, so the labels and the bus functions come from the same table.
SUPPLY: 3.3 V to 5.0 V, 2 mA working current. Logic is whatever the host rail is: DFRobot's schematic carries both a _03V3 and a _05V version of each of SCK, SDI and SDO (net names NLSCK03V3/NLSCK05V, NLSDI03V3/NLSDI05V, NLSPI0SCK), i.e. a level-shift stage (Q1, U2, U3) between the sensor's 3.3 V side and the header.
MEASUREMENT: temperature -40 to 85 C at 0.1 C resolution and +/-0.5 C deviation; humidity 0-100 %RH at 0.1 %RH resolution and +/-2 %RH deviation, 1 s sampling time; pressure 300-1100 hPa. Working temperature -40 to 85 C.
REFERENCE CIRCUIT: U1 BME280 with 1 uF bulk (C1, C3) and five 100 nF decoupling capacitors (C2, C4, C5, C6, C7); the I2C lines pass through the level-shift stage (Q1, U2, U3) rather than straight to the header. DFRobot's schematic does not put its resistor values in the PDF's text layer, so no pull-up value is recorded here - do not assume one.
TRAPS: (1) The chip's own limits, from the linked Bosch data sheet, are VDD 1.71-3.6 V and VDDIO 1.2-3.6 V, so the module's 5 V rating only holds because of its on-board stage - never feed 5 V to the BME280 package. (2) SPI is not broken out on this carrier even though the chip supports it - the four pads are I2C only. (3) The 1 s humidity sampling time is the vendor's figure; polling faster returns stale values.