EE: Refresher & Catch-Up — Season 1 Curriculum Map

What this is: A graduate-level audio course that re-teaches core Electrical Engineering and pulls each topic forward to today's industry standard. It's meant to be listened to like a book — you are being taught the subject matter, not given a news roundup. Every episode increases both knowledge and usable skill.

Who it's for: An engineer fluent in Laplace/Fourier, circuit and semiconductor physics, control and digital fundamentals. We do not re-explain Ohm's law, KVL/KCL, or what a transistor is. We do re-derive the things worth re-deriving, then go past 2013.

Format - Voices: Elena — the instructor, who carries the teaching. Marcus — a co-host who asks the question you'd ask and keeps the pace. Explanation dominates; this is a lesson, not banter. - Length: 17–20 min (~2,300–2,600 spoken words). - Lesson shape (every episode): 1. Objective — what you'll understand and be able to do by the end. 2. The refresher — re-ground the fundamental at graduate level. 3. What changed — the genuine post-2013 shift. 4. The mechanism — actually teach how the modern thing works, with intuition + light math. 5. Industry standard today — where it lives in real practice and standards. 6. Self-test — a concrete exercise to convert listening into skill.

Episode 1 — Grid-Forming Inverters & Virtual Synchronous Machines

Pillar: Power & Energy Systems

• Refresher: The synchronous generator and the swing equation; inertia constant H, damping, droop; why grid frequency has historically been pinned to spinning mass; the power-angle relation P = (V₁V₂/X)·sin δ.
• What changed: Inverter-based resources (solar/wind/battery) carry no rotating mass. As penetration rises, system inertia falls, RoCoF climbs, frequency nadirs deepen.
• Mechanism: Grid-following (PLL + current source, needs a stiff grid) vs. grid-forming (voltage source behind an impedance; sets its own angle; can black-start and island). Virtual synchronous machines: implementing the swing equation in firmware so a battery emulates rotor inertia and damping.
• Standard today: IEEE 2800-2022, UNIFI specs, grid-forming BESS mandates.
• Self-test: Estimate RoCoF for a load step at a given system inertia; reason about why a PLL-based grid-follower destabilizes as short-circuit ratio drops.

Episode 2 — OFDM, the Cyclic Prefix, and 5G Numerology

Pillar: Signal Processing & Communications

• Refresher: Multipath → frequency-selective fading and ISI; why equalizing a wideband single-carrier channel is painful; convolution ↔ multiplication.
• What changed: OFDM made the channel diagonal; 5G NR made the waveform scalable.
• Mechanism: IFFT/FFT as the modulator/demodulator; orthogonality from 1/T spacing; the cyclic prefix doing two jobs — guard interval against delay spread, and turning linear convolution into circular convolution so each subcarrier is a one-tap equalize. PAPR as the price.
• Standard today: 5G NR scalable numerology (SCS = 15·2^μ kHz), CP-OFDM, DFT-s-OFDM.
• Self-test: Given a delay spread, pick a numerology whose CP covers it and compute the CP overhead.

Episode 3 — Field-Oriented Control of PMSMs

Pillar: Embedded & Control Systems

• Refresher: Why a 3-phase machine is hard to control in abc coordinates; why a DC machine is easy (orthogonal, decoupled flux and torque).
• What changed: Cheap fast MCUs + SiC inverters made high-bandwidth vector control and sensorless operation the default — it's how every EV traction drive runs.
• Mechanism: Clarke (abc→αβ) and Park (αβ→dq) transforms collapsing sinusoids to DC; two PI loops controlling i_d and i_q independently; torque ∝ i_q, i_d for field weakening; MTPA and reluctance torque in interior PMSMs; sensorless observers.
• Standard today: EV traction, industrial servo drives, sensorless FOC libraries.
• Self-test: Explain why i_d = 0 is suboptimal for an interior PMSM; relate electrical and mechanical frequency through pole pairs.

Episode 4 — From FinFET to Gate-All-Around Nanosheets

Pillar: Semiconductors & VLSI

• Refresher: Dennard scaling and its death; short-channel effects — DIBL, V_t roll-off, the ~60 mV/decade subthreshold-swing floor; why planar MOSFETs lost gate control.
• What changed: FinGate (3-sided) → Gate-All-Around nanosheets (4-sided) as the way to keep electrostatic control alive; backside power delivery; CFET on the horizon.
• Mechanism: How wrapping the gate restores channel control, steepens subthreshold slope, and cuts leakage; fin width quantization vs. nanosheet width tunability.
• Standard today: 2 nm-class nodes (Samsung MBCFET, TSMC N2, Intel 18A RibbonFET).
• Self-test: Tie subthreshold swing to off-current and explain how geometry beats DIBL; reason about drive-strength granularity, fin vs. sheet.

Episode 5 — Battery State of Charge: Coulomb Counting vs. Kalman Estimation

Pillar: Power & Energy Systems (with an estimation/Kalman skill spine)

• Refresher: SOC = ∫current/capacity; why open-loop coulomb counting drifts (sensor bias, unknown initial SOC, capacity fade); why open-circuit-voltage lookup is only valid at rest.
• What changed: Model-based estimation became standard in every serious BMS; joint state+parameter estimation now tracks aging (SOH), not just SOC.
• Mechanism: Equivalent-circuit battery model as a state space; the Kalman predict/update recursion as an optimal blend of model prediction and voltage measurement; EKF vs. UKF for the OCV nonlinearity; dual estimation for capacity/SOH.
• Standard today: EV and grid-BESS battery management systems.
• Self-test: Argue why pure coulomb counting drifts and pure OCV is unusable under load, and describe how the Kalman gain arbitrates between them.

Standing note: the recurring single-lesson rotation (Power → Signal → Embedded → Semiconductors) continues alongside this season. First rotation lesson (2026-05-15) was "Wide Bandgap Semiconductors and the Switching-Loss Equation" — not repeated here.