TB6612FNG

Almost always, yes, if your rail is under 13.5 V and your current is under 1.2 A. The TB6612FNG bridge drops 0.5 V typical at 1 A versus about 3.7 V for the L298N, so it runs cool with no heatsink, wastes roughly a seventh of the power, and needs no external flyback diodes because regenerative diodes are built in. It also draws 1.1 mA instead of the L298N's tens of milliamps of quiescent current. Keep the L298N only when you need the 46 V rail or more than 1.2 A per channel.

12 V is fine — the operating VM range is 2.5 V to 13.5 V. 24 V is not: the absolute maximum VM is 15 V and the device will fail, often destructively. For 24 V systems use the DRV8825 (8.2 V to 45 V) for steppers or a DRV8871 / L298N for brushed motors. Also remember VM and VCC must be sequenced or at least present together, and a 100 nF ceramic close to each VM pin plus bulk capacitance is required.

Not for normal small DC motors. Toshiba specifies internal upper and lower regenerative diodes in the electrical characteristics table with a forward voltage between 1.0 V and 1.1 V at 1 A, and they carry the back-EMF current when the bridge is turned off. Add external Schottky diodes only for high-inductance loads, long motor cables, or when the supply cannot absorb regenerated energy — and always add a bulk capacitor on VM, because short-brake mode dumps winding energy back into the rail.

That is a datasheet operating-range limit, not a thermal one. Toshiba rates 1.0 A maximum output current for VM at or above 4.5 V, and only 0.4 A for 4.5 V down to 2.5 V without PWM operation. The lower rail reduces the gate overdrive available to the LD-MOS outputs, so Ron rises. If you must run from a single Li-ion cell, stay under 0.4 A or add PWM so the average current is inside the limit.

It has no built-in indexer, so your MCU must generate the phase sequence. Drive channel A from IN1/IN2 and channel B from IN3/IN4 with a two-phase bipolar stepping table, and use PWMA/PWMB for current or speed control. It suits a four-wire bipolar motor up to about 1 A per winding. If you want one STEP pulse per microstep with no phase tables, use the A4988 or DRV8825 instead — those have an on-chip translator and adjustable chopping current. A unipolar five- or six-wire motor is a poor fit here; use a ULN2803 or L298N for that.

Short brake (IN1 = IN2 = high with PWM high) turns on both low-side FETs and shorts the winding, so the motor stops quickly by dissipating its kinetic energy in the bridge. Stop (IN1 = IN2 = low with PWM high) puts both outputs in high impedance, so the motor coasts. Standby (STBY low) shuts the whole control block down and pulls both outputs to high impedance with under 1 µA of supply current. Use short brake when you need to hold position against a load, coast when you want freewheeling, and standby only in sleep modes.

Measure three parameters on a sample. Total output saturation voltage at 1 A should be 0.7 V maximum — clones with higher-Ron bridges read 1.2 V or more and get noticeably warm at 800 mA. Standby current with STBY low should be under 1 µA; remarked parts often leak 20 to 100 µA. And check the SSOP-24 body width, the 0.65 mm lead pitch, laser-etched Toshiba marking and a legible date code. We electrically sample-test each reel and can provide lot traceability on request.

MOQ is one reel of TB6612FNG(EL) in SSOP-24 for volume orders, or 100 pcs loose for prototyping; breakout modules are stocked in packs of 10. Shenzhen stock ships in one to two business days. Toshiba authorized-channel replenishment is eight to sixteen weeks, and the TB6612FNG has periodically been allocated, so keep a qualified alternate such as the DRV8833 on your AVL.