This is my 1st post here and as an added caveat.....I'm a software engineer not a trained electronics engineer (just lining up the excuses in advance!). My day job has been varied including being a PM in an Investment Bank but thankfully now back to writing software to control a Radiotherapy Linear Accelerator. So this is partly my way of gaining a new skillset along the way.
The seed of this idea came about when my Ham Radio clubs' 3KW generator was stolen. I'd already been looking into LoraWan and the suggestion was to build a tracking device for the replacement. So the design began....and grew....and grew, but always supporting the original requirement.
The board contains the STM32L4 (centre right) and a SX1262 (inside E22 module). The 2 sets of expansion connectors allow a number of optional boards to be fitted and are supplied by SPI, I2C, I/O pins and 2*3.3v lines (one switchable). Expansion board ideas (actually mostly built) so far are GPS, Leds, ePaper display, Air Quality sensor etc. This is where my imagination got away with me. The main thing is that the system is geared towards LoraWan data communication and the STM32L4. I did start with the Murata ABZ board but it didn't have enough I/O and my ideas for expansion outstripped it. I got to the point of connecting and sending data to TTN using mbed which was certainly useful.
There is a 18650 rechargable battery underneath the main board feeding to a TPS63031 buck convertor producing 3.3v. A mcp73831 Li-ion charge controller to allow USB charging.
I quickly realised that I needed to reflow it. The pain in trying to hot-air solder the TPS63031 sold me on a toaster oven. The board above is my first attempt with the oven. It works but need to get more experince with the solder mask (too much solder sometimes). Fromthe STM32L4 I can talk to the SX1262 chip and shut it down. Talk via I2C to a LED board (see below) and also via SPI to the ePaper Board. And I can shut everything down and go to SHUTDOWN state in the STM32L4. That's when the debug really started.
Tracking down the elusive 1mA
Another piece of equipment I realised very early on that I needed was something to measure very low levels of current. I wanted this device to last many months, if not years on a single battery charge. While watching a youtube video by Andreas Spiess (the guy with the Swiss accent) I saw he was using an Otii. Next day I had one. So back to how much current my board was using......With everything turned off (SX1262, Flash chip, both expansion connectors) I just can't get it down below 1.42mA in shutdown mode whereas it should be down around 40uA. All IO pins set to analog, no pull up/down. Desoldered the flash, then the SX1262 module, then any leds on the 3.3v side, removed the expansion boards...no change. I even changed the BOOT0 resistor to 1M. If I disable the 3.3v buck convertor (see SW1) it drops to 0uA so I know its the 3.3v side that is the problem.
My soldering of the STML476RCT6 (possibly) ?..... I have a few semi-populated boards coming from JLCPCB so watch this space!
What have I learnt so far
You can do a reasonable job of soldering 0603, WSON and LQFP parts in your kitchen
KiCAD is well worth the money and then some!
Powering on for the first time took 'courage'
The low profile female pin headers are stupidly expensive compared to the rest of the components.
The time it takes for the wall clock to move from 6pm to 3am while working on a pcb board layout is frightening.
Paying work gets in the way
Reduce power consumption on shutdown
Test the GPS board
Build and test the Air Quality board
Build and test the CAN board
Write a load of software!!