...
| # | Title | User Story | Importance | Notes |
|---|---|---|---|---|
| 1 | Work with Nomura MPPTs | We're using 10 Nomura MPPTs, 5 in series for 2 strings total. They don't have any public way of interfacing with them, but we'd like to collect data on our array. | Must Have | Most likely using Using a master board for current sensing and I2C LIN slaves for each MPPT |
| 2 | Record voltage and temperature per-module | We'd like to know the voltage and temperature of each module so we can tell how each module is performing and any potential hotspots. | Must Have | |
| 3 | Record current per-string | We'd like to know the current provided by each string so we can calculate the overall power of the solar array. | Must Have | We will probably use a dedicated current slave to record the current of the string. |
| 4 | Isolated | We want to make sure that our ADCs are analog measurement is isolated from the rest of the boards and ADCscar. | Must Have | Look into analog isolation using a class D amplifier + digital isolator |
| 5 | Low-cost/Simple | We'd like to make these boards as simple as possible to reduce development time. Ideally we shouldn't need to write firmware for the slave boards.Must Have | Must Have | Once the LIN stack is written, we can use the built-in ADC to record voltage and temperature. |
Questions
Below is a list of questions to be addressed as a result of this requirements document:
| Question | Outcome |
|---|---|
| What kind of architecture are we using? | Master node with controller board living in the roof enclosure Slave nodes (no MCULIN) attached to each MPPT |
| How will our boards communicate? | We plan Originally, we were planning on using differential I2C. It allows us to communicate with multiple slaves and handles long wires pretty well.To improve our signal integrity, we're looking into using P82B96s, which are designed for I2C over long distance. This is a pretty good resource on I2C over long distancesHowever, after doing a cost analysis, we have decided to use LIN instead, with low-cost LIN slave controllers on each board. See Solar Slave - System Architectures for a comparison between the two. |
| What's powering the boards? | We will most likely run I2C SDA/SCL, power and ground to each of our slave nodes. Trying to power the nodes directly from the MPPTs would require a buck-boost for each boardThe MCUs and LIN transceiver will be powered from as 12V rail provided by the solar master node. To save on costs, we plan on using LDOs since our MCUs are relatively low-power. On the isolated side, we will likely use an LDO to convert the MPPT output to a voltage usable by our isolator. |