Weather Station Pt 2

Today I made a couple of tweaks to the schematic to correct a few errors, and then completed a first pass at laying out the PCB for the weather station carrier board.

Tweaks Made#

The main schematic change was replacing the original 3.3V regulator, the AMS1117-3.3, with the AP2114D-3.3.

I wanted to move to a more efficient regulator because I may power this design from LiPo batteries in a future revision. The AP2114D-3.3 is not a drop-in replacement, but it offers a few advantages that made the change worthwhile.

First, it allowed me to remove the electrolytic capacitor and use ceramic capacitors instead. It also has a much lower dropout voltage, around 250–400 mV, compared with the 1.2–1.3 V typical of the AMS1117-3.3. On top of that, its quiescent current is in the microamp range, rather than the roughly 2–10 mA typically associated with the AMS1117.

The AMS1117 is certainly cheaper, but I do not expect to build these boards in large quantities. If the next revision ends up mounted at the top of a mast, and especially if it is battery-powered or supported by solar charging, I would rather prioritise efficiency over the small extra component cost.

I also removed a Schottky diode that I had originally connected to the 5V pin of the ESP32-C3 Supermini. I had mistakenly assumed this was an output pin (it was late — I should have gone to bed), when in reality it is a 5V input that can be used to power the board directly.

As a result, D2 and D3 are now used to OR the external power supplies feeding the regulator input. The regulator (U2) then feeds the 3.3V rail, which powers the Supermini and the rest of the circuit.

I have since increased the value of R3 to 1 kΩ (or higher) to reduce the current through the LED, assuming the brightness remains acceptable. That said, if this device is going to live at the top of a mast and power efficiency is critical, the LED can simply be omitted entirely by not populating it on the board.

I also added jumpers to the I²C pull-ups so they can be enabled or disabled as needed. This gives me a bit more flexibility, particularly if I decide to use a Qwiic (JST-SH) connector, which often already includes pull-ups. It is a small addition, but it avoids needing bodge wires later on.

One of the things I may consider changing in a future revision is the use of Schottky diodes for D2 and D3, possibly replacing them with an ideal diode arrangement. For now though, this is more than sufficient to get a first revision built while I consider the next steps.

I am also starting to think about EMC and how I might need to tackle that if I decide to use this in conjunction with amateur radio equipment. That will likely involve paying closer attention to grounding, decoupling, and possibly filtering in a future revision.

The updated schematic is shown below. The main changes are the revised power input stage, the switch to the AP2114 regulator, and the addition of configurable I²C pull-ups.

First pass at the PCB#

This is a first-pass layout, so there will almost certainly be refinements needed, particularly around RF considerations, grounding, and power routing. However, it should be more than sufficient to validate the design in hardware.

At this point, the next step is simple: get some boards manufactured and see how it performs in the real world.

If you are working on custom carrier boards, sensor integration, or ESP32 prototypes and need an experienced engineering partner, let’s discuss your project at Protoworks (NI).