The Si705x-EB has small through holes on the PCB. You can solder wires to the small through holes. There's a through hole for each pin. You'll need to run VDD and GND to your processor with the i2c lines for it to work.

The Si7210 is an I2C position sensor with a programmable output switch. By default, the Si7210-B-04-IV comes with an output programmed for Bop=1.1mT(max) and Brp=0.2mT(min). The typical values programmed are Bop=0.84mT and Brp=0.44mT. 

A host can reprogram the output switch points by changing the sw_op and hysteresis values. Section 4.1.3 Fields Associated with Configuring the Output Pin in the datasheet covers how to program those values. 

Simplicity Studio has an example driver file for the Si7210 I2C hall sensor in the driver folder:

C:\SiliconLabs\SimplicityStudio\v4\developer\sdks\gecko_sdk_suite\v2.7\hardware\kit\common\drivers

The driver files are: si72xx.c/.h

That is on the high side. I recommend you test the sensor in your environment. The sensor's polyimide should be able to detect water molecules in that environment. However, the sensor will still need adequate air flow to accurately detect the environment

That is on the high side. I'd recommend you test the sensor in your environment. The sensor's polyimide should be able to detect water molecules in that environment. However, the sensor will still need adequate air flow to accurately detect the environment

The Si7006, Si7013, and Si7021 all have the same I2C address, commands, and read/write sequence. The Si7013.c driver is compatible for the shared commands. Since the Si7013 is the only device with an additional external thermistor measurement, the voltage measurement commands are not compatible. Only the temperature and humidity commands are compatible.

Thank you for the scope capture. The jpg "when return 0" shows that the host is sending the command: 0xE3 after the initial St+Addr+W+ACK. The command 0xE3 is a temperature hold command which requires the host to allow the sensor to stretch the clock throughout the measurement duration. The scope capture doesn’t show a clock stretch event. Here’s how the i2c sequence should occur for the 0xE3 command:

 

The measurement duration should take at least 7msec typically so there’s no way the sensor is ready to output data after one half clock cycle.  I'm not sure why the SCL waveform doesn't show bus conflict. The SCL waveform should show some contention during the clock stretch portion of your waveform similar to the first SDA high pulse on the initial address.

If you can’t configure your host for clock stretching, I recommend you use No Hold Master Mode commands for issuing measurements.

 

 

I recommend using an oscilloscope. Monitor the initial address by the host and check for the sensor ACK. 

Well, first of all I suggest to operate the sensor standalone without any i2c programming and monitor the output pin. The only thing you need to take care is the output-pin-configuration of the sensor. If it is "open drain" type then you have to apply the pull-up-resistor of course.

When you play with the magnet (gently aproxximate and remove in z-direction), the output pin should go on and off as seen in Figure 4.1. of the Si7210 datasheet: https://www.silabs.com/documents/public/brochures/magnetic-sensors-selector-guide.pdf. Same with reversed polarity.

You could also have a look into AN0018 which is a good starting point (and also covers interrupts): https://www.silabs.com/products/development-tools/sensors/si72xx-magnetic-development-kit

Datasheet Link: Si7210 Datasheet

Table 1.5 has the magnetic sensor specifications. With a 5V supply, the typical noise performance is 30uT. As the supply is reduced, the noise is expected to increase.

You should be installing a supply decoupling cap near the sensor. This will help reduce noise.