Stan Meyer's TEENSY  Steam Resonator 2026 VERSION 2

This is Work in Progress

Join as a Ptreon as we post update to Pcbway

https://www.patreon.com/c/securesupplies/shop

I’ll explain plainly, electrically, and practically why the suggested part is better than IRLZ44N for the exact circuit.

Short answer (one line)

IRLZ44N will work, but it is old, slow, and lossy.

Modern MOSFETs do the same job with less heat, cleaner edges, and much more margin.

Now the details 👇

1️⃣ What matters in your design (not datasheet marketing)

circuit has:

Logic-level drive (5 V from Teensy / HC logic)

Alternating push-pull switching

Inductive load (coil / VIC / transformer)

Dead-time already handled (excellent 👍)

Moderate voltage (12–40 V region)

Switching losses matter more than DC current

So the MOST important MOSFET parameters are:

RDS(on) at 4.5–5 V

Total gate charge (Qg)

Switching speed / capacitances

SOA robustness with inductive loads

Same TO-220 footprint

2️⃣ IRLZ44N — why it’s “okay but dated”

IRLZ44N is from the 1990s. It was great then.

IRLZ44N typical values:

RDS(on): ~22–28 mΩ @ 5 V

Total gate charge (Qg): ~67 nC

Gate capacitance: large

Switching speed: slow

Die technology: old planar silicon

What this means in our board:

Needs more gate current

Slower rise/fall → more heat

More ringing → more stress

Snubber becomes more important

Wastes energy every pulse

👉 It works, but you are fighting physics more than you need to.

3️⃣ Why the modern part is better (example class, not magic)

The recommended class (example: BSC009NE2LS5, IPP034N06, IRLB3034, etc.) uses:

Trench MOSFET

Much smaller die capacitance

Optimized for logic-level switching

Typical modern logic MOSFET values:

RDS(on): <5 mΩ @ 4.5 V

Gate charge: 15–30 nC (½ to ⅓ of IRLZ44N)

Faster turn-on/off

Cleaner switching edges

Lower EMI

Much cooler operation

4️⃣ REAL effect in your circuit (numbers)

Gate drive loss comparison

Gate loss ≈ Qg × Vgate × frequency

Example (rough):

IRLZ44N

Qg ≈ 67 nC

Vg = 5 V

f = 20 kHz

Loss ≈ 6.7 mW per MOSFET

Modern MOSFET

Qg ≈ 25 nC

Same Vg and f

Loss ≈ 2.5 mW

➡️ ~60% less gate loss

➡️ Faster switching

➡️ Less heat

➡️ Cleaner waveform

5️⃣ Conduction loss comparison (this really matters)

Power loss ≈ I² × RDS(on)

Assume 10 A pulse current (very realistic for coils):

IRLZ44N

R ≈ 25 mΩ

Loss ≈ 2.5 W

Modern MOSFET

R ≈ 5 mΩ

Loss ≈ 0.5 W

➡️ 5× less heat

➡️ Smaller heatsink

➡️ More margin

➡️ Less stress on board

6️⃣ Reliability & scaling (why this matters long-term)

With IRLZ44N:

Heat limits scaling

Snubber becomes mandatory

SOA margin shrinks fast

Avalanche risk rises

With modern MOSFET:

Cooler die

Better SOA

Handles spikes better

Scales to larger coils / higher duty

Survives mistakes better (important for testing)

7️⃣ Footprint compatibility (important for Gerber)

✅ Same TO-220

✅ Same pinout (G-D-S)

✅ Drop-in replacement

✅ No board redesign required

That’s smart engineering.

8️⃣ When IRLZ44N is still acceptable

IRLZ44N is fine if:

You already own them

Low frequency (<5 kHz)

Low duty

You add gate resistor + snubber + zener (which we are)

9️⃣ Bottom line (honest)

our architecture is strong

The logic & dead-time are correct

TIP120 → MOSFET is the correct evolution

IRLZ44N is okay

Modern MOSFET = same effort, much better result