Controlling 3.6kW of Solar EV Charging with an Arduino GIGA R1 WiFi

Original Article ↗ Hacker News Discussion ↗

Solar hardware & sourcing

  • Few >1 kW charge controllers are visible in consumer channels; users say current scales well via multiple smaller MPPTs, so the market targets higher string voltages instead of massive current.
  • Higher-end stand-alone MPPTs (e.g., Victron, EG4) and DC optimizers (Tigo, SolarEdge) are recommended over generic Amazon products for safety, reliability, and per‑module MPPT/shutdown.
  • Many larger inverters are sold through solar specialists, not consumer retailers; some argue most US homeowners can self‑install, others note states like CA/AZ require licensed electricians for grid‑tie work.
  • Amazon is widely viewed as the wrong place to shop; electronics distributors and dedicated solar suppliers are suggested instead.

Panel tilt, orientation, and cleaning

  • Multiple commenters argue tilting/tracking rarely pays off for small systems; added cost can often buy more fixed panels with better net yield.
  • For roofs, suboptimal angles and even non‑ideal azimuths can still be economic, especially against retail electricity rates.
  • At grid scale, a few percent matters: east/west and carefully tuned angles can improve revenue by shifting output into higher‑value morning/evening periods.
  • Vertical and bifacial setups (fences, walls, agrivoltaics) are highlighted for better winter/snow performance, improved daily production profile, and dual land use.
  • Flat panels may accumulate dirt/moss and lose substantial output if not cleaned; adequate tilt for snow‑shedding is important in snowy regions.

Solar costs, tariffs, and grid economics

  • Reported new‑panel prices range roughly from ~$0.30–0.64/W in the US, lower in high‑volume/wholesale and some countries (e.g., India, Australia); used panels can be “dirt cheap.”
  • Several note that racking, inverters, and batteries can approach or exceed panel cost.
  • One view: over 25 years, home solar kWh can be ~¼ the grid price, raising questions about why grid power is so expensive.
  • Responses cite transmission/distribution build‑out, maintenance, reliability, night‑time supply, peak coverage, and utility profit margins as major contributors.
  • There’s disagreement on battery economics: some claim adding batteries for overnight use only increases system cost about 20% and still pays off; others see battery costs nearer 50% of install and not yet compelling, especially for full off‑grid sizing.

Grid-scale storage vs nuclear (unresolved)

  • One side asserts grid‑scale storage is orders of magnitude more expensive than generation and that nuclear plants provide cheaper continuous power.
  • The other uses specific US and large‑battery project numbers to argue that, on a lifetime cost per delivered kWh basis, big batteries plus overbuilt solar can approach nuclear costs, not differ by orders of magnitude.
  • They trade calculations using recent nuclear and battery projects, but dispute each other’s assumptions (construction overruns, required redundancy, degradation, power vs energy confusion).
  • No consensus emerges; both demand citations, and the debate ends with persistent disagreement.

Home vs grid, usage patterns

  • Some argue that pairing solar with batteries can cover nights cheaply enough to largely abandon grid power.
  • Others note that US “typical” all‑electric loads and winter conditions can require very large arrays (20–30 kW) and sizable storage (50–150 kWh) to bridge multi‑day low‑sun periods.
  • Contrasting examples from Australia with modest systems and low grid imports illustrate how regional climate and consumption shape viability.

Arduino vs other platforms

  • Question raised: has Arduino become suitable for “serious” or industrial use?
  • One perspective: Arduino’s newer pro/DIN products aim at industry but often lack industrial must‑haves (input protection, wide supply range, industrial buses), so are not widely used professionally.
  • Another: Arduino excels at rapid prototyping and simple control; small tasks (e.g., pump control with basic logic) can be implemented in about an hour.
  • Many hobbyists now favor ESP32 or RP2040/Raspberry Pi Pico: similar or better MCU capabilities, lower cost, built‑in Wi‑Fi/Bluetooth, and richer options.
  • Raspberry Pi (Zero etc.) is praised where a full OS simplifies development, updates, logging, remote debugging, and “self‑hosted” toolchains.
  • There’s a philosophical split between valuing tight, resource‑efficient MCU code vs valuing ease of development and maintenance.
  • One report of a dead‑on‑arrival Arduino “pro” unit is balanced by praise for responsive support.
  • A question about why the GIGA R1 uses a more expensive Murata Wi‑Fi module instead of Espressif is raised but not answered.

EV charging control setups

  • Several commenters describe integrating solar with EV charging using home automation.
  • One setup uses Home Assistant on a Raspberry Pi to modulate EV charging current based on real‑time solar production, aiming for nearly 100% solar charging when daytime charging is possible.
  • The “evcc” project is recommended: a standalone server with web UI that coordinates between inverters, wallboxes, car APIs, and smart meters.
    • Modes include solar‑only, minimum‑current plus solar boost, and solar‑independent charging.
    • It can enforce state‑of‑charge targets (e.g., stop at 80%).
    • Core code is MIT‑licensed, but some integration interfaces require a sponsorship token or patching around license checks.

Article and EV framing

  • Some question why a derivative Arduino blog post is front‑linked instead of the more detailed original Hackster.io project write‑up.
  • One commenter criticizes the article’s framing of the EV vs ICE discussion as vaguely negative toward EVs.
  • Others acknowledge that, pure ROI aside, the project is technically interesting and well documented, making it a useful reference.