TB6612FNG motor driver

The Toshiba TB6612FNG dual motor driver on its breakout board: two H bridges for two DC motors. Aliases: TB6612, SparkFun motor driver, dual H-bridge.

It is the driver in the SparkFun Inventor's Kit and most small robot kits. Its bridges are MOSFETs rather than the L293D's Darlingtons, so it keeps a fraction of a volt where an L293D keeps more than two, and its logic runs on 3.3 V as well as 5 V.

Pins

Logic side on the left, power side on the right, in the order the SparkFun board prints them:

Pin	What it does
PWMA, PWMB	Each channel's speed: analogWrite here.
AIN1, AIN2	Channel A's direction.
BIN1, BIN2	Channel B's direction.
STBY	Must be HIGH, or both channels are off.
VCC	Logic supply, 2.7 V to 5.5 V.
VM	Motor supply, 2.5 V to 13.5 V (15 V absolute maximum).
AO1, AO2	Motor A.
BO1, BO2	Motor B.
GND (three)	One ground, strapped on the board.

The truth table

For each channel, with STBY high:

IN1	IN2	PWM	Outputs	Mode
H	L	H	OUT1 high, OUT2 low	clockwise
L	H	H	OUT1 low, OUT2 high	counterclockwise
H or L	the other	L	both low	short brake
H	H	any	both low	short brake
L	L	any	both off	stop, the motor coasts

PWM low is a short brake, not a coast. So at half duty the motor is driven and braked in turn, and it slows down firmly rather than drifting.

While it runs

The face shows what each channel is doing: off, CW, CCW or brake.

The library

SparkFun_TB6612.h works as the Inventor's Kit uses it:

#include <SparkFun_TB6612.h>

Motor motor1 = Motor(2, 4, 5, 1, 9);   // AIN1, AIN2, PWMA, offset, STBY
Motor motor2 = Motor(7, 8, 6, 1, 9);   // BIN1, BIN2, PWMB, offset, STBY

void loop() {
  motor1.drive(255, 1000);   // forward at full speed for a second
  motor1.drive(-255, 1000);  // and back
  motor1.brake();
  forward(motor1, motor2, 150);
  delay(1000);
  brake(motor1, motor2);
  delay(1000);
}

drive() takes -255 to 255 and sets STBY high itself; brake(), standby(), forward(), back(), left() and right() do what upstream's do. The offset argument, 1 or -1, flips a motor wired the other way around.

What the model gets right

From Toshiba's datasheet:

The truth table, row for row, including the brake on PWM low.

Internal pull-downs. Every input has 200 k to ground inside, so an unwired STBY reads LOW and the chip does nothing at all. That is the first fault on every forum thread about this board.

The volts it keeps. 0.5 V at 1 A across the two transistors of a bridge, which is 0.25 ohm a side.

Where the current comes from. With a battery or a bench supply on VM, the bridge is part of the analog solve, so the motors' current comes out of the motor rail and a battery sags under it.

The supplies. Under 2.7 V of logic, under 2.5 V of motor supply or over the 15 V absolute maximum, the outputs are dead.

What it does not model

The 1.2 A average and 3.2 A peak current limits and the thermal shutdown behind them, the dead time, the 100 kHz PWM ceiling, and the flyback diodes.

Common mistakes

Leaving STBY unwired. Its pull-down holds the chip in standby.

Expecting PWM to coast. On this chip it brakes.

Running the motors from the logic supply. Give VM its own supply, and join the grounds.

See it in action

TB6612 robot drive runs two gearbox motors from a four-AA pack. Open it at /templates.