ATMEGA328P Module with integrated LoRa and CAN-BUS

INTRODUCTION

In my quest to perfect my LoRa telemetry system, I have gone through quite a few prototypes by this time. This post will focus on the next node design. Due to the fact that the area where I will deploy the system is quite large, but with roughly square boundary fence-lines, I decided to try and reduce the number of LoRa Radio nodes needed to cover the entire area. This opened up the opportunity to utilise CAN-BUS to attach sensor-only nodes to a Radio node and have them report status on exception as well as on requests from the radio node.

The device will thus function as a LoRa-to-CAN-BUS Gateway, with some local automation to control the transmission of data to the master station. This concept can also be adapted for use in other areas, such as home -automation, or an industrial setting.

At the heart of the device, I have stuck with the versatile ATMEGA328P, which, current chip shortages excluded, and current high prices excluded, are a very inexpensive chip, with lots of well-tested libraries, and a relatively low learning curve, largely due to its very wide use in the Arduino ecosystem.

The LoRa component is handled by the RA-02 or even RA-01H module, from AI-Tinker (not sponsored). This device, as we have seen in the previous prototypes, requires the use of logic level converters, due to the fact that it only accepts 3.3v logic levels. While I could get rid of those if I powered the ATMEGA328P from 3.3v, it would give rise to two problems, of which one will still force the use of level converters…

I chose to run the ATMEGA328P at 16Mhz, which basically forces me to use 5v to power the chip. The second reason is not so obvious unless you read a few datasheets very thoroughly…

The CAN-Bus component is handled by the MCP2515 Standalone SPI-to-CAN Controller, as well as the TJA1050 CAN Bus transceiver.

This is where things get interesting… The MCP2515 can operate on 3.3v, but the TJA1050 is a 5v only device. I could thus in theory use logic level converters only between the MCP2515 and the TJA1050, while running the rest of the circuit on 3.3v…

Given that I would rather run the ATMEGA328P on 16Mhz, as well as the fact that my LoRa Radio Module circuit, with its logic level converter circuitry, works extremely well, I decided not to change that, and keep the CAN Bus running at 5v all the way through, as I would still have to use a 5v regulator on the PCB anyway just for that purpose.

IO connections for LoRa and CAN BUS modules

You can order this PCB directly from PCBWay, by clicking here

Both of the two integrated components ( Lora and CAN ) are SPI devices. This means that they share common SCK, MISO and MOSI lines ( provided on the ATMEGA328P by pins D13, D12 and D11 respectively. The individual SPI device is then further selected for operation by the use of a CE pin, one unique pin per device, which is pulled low by the MCU to indicate to the device that it should pay attention to the data being transmitted on the SPI bus…

Both LoRa and CAN makes use of other pins as well, LoRa needs a Reset pin, connected to D9, a CS/CE Pin on D10 as well as a hardware interrupt pin, connected to D2. ( Note that this is for use with Sandeep Mistry’s LoRa Library. The Radiolib library would require an additional pin, usually connected to DIO1 on the LoRa module. The device does not provide access to those pins in its current layout, so you can only use it with the Sandeep Mistry library, for now at least… )

The CAN module uses a CE/CS pin at D4, with an IRQ pin on D6, which, although not a hardware interrupt pin, does have PCINT functionality.

Pins D10, D9 and D2 are not broken out for user access. although I decided to give access to D4 and D6, as well as the SPI bus, D11, D12, D13, to allow interfacing with logic analysers, or adding other SPI devices to the bus…

This brings us to a very interesting point… Does the two SPI devices actually play nice together?...

Read more »