DRV8833

The die is the same but the thermal capability is not. PWP is a 16-pin HTSSOP with a PowerPAD and RTY is a 16-pin WQFN, both rated 1.5 A RMS and 2 A peak per bridge (3 A RMS paralleled). PW is a plain 16-pin TSSOP with no exposed pad and is rated only 500 mA RMS and 2 A peak per bridge (1 A RMS paralleled). HTSSOP and TSSOP share the same 5.00 × 4.40 mm body size, so never let procurement substitute PW for PWP without re-rating the design.

They overlap heavily. The DRV8833 runs down to 2.7 V and up to 10.8 V with 1.5 A RMS per bridge and 360 mΩ Ron, and gives an nFAULT pin plus explicit brake and coast states. The TB6612FNG runs from 2.5 V up to 13.5 V with 1.2 A continuous, 0.5 ohm Ron, built-in regenerative diodes and internal thermal shutdown, and it derates to 0.4 A below VM = 4.5 V. Choose the DRV8833 for maximum current at 5 V and for fault reporting; choose the TB6612FNG if you need a 12 V rail or want the integrated clamp diodes.

It drives a bipolar two-phase stepper: connect one winding across AOUT1 and AOUT2 and the other across BOUT1 and BOUT2, then sequence AIN1/AIN2 and BIN1/BIN2 from your MCU. There is no built-in indexer or microstepping DAC, so all phase generation and any microstepping must be done in firmware. A unipolar five- or six-wire motor is not a good fit — for those use a ULN2803 or ULN2003 Darlington array, or an L298N with the centre taps tied to VM. If you want on-chip microstepping, use the A4988 or DRV8825 instead.

In fast decay the bridge is disabled during the PWM off time and recirculating winding current flows through the MOSFET body diodes; you get this with the PWM pin driving one input and the other held low, or with both inputs low for coast. In slow decay the winding is shorted through both low-side FETs (both inputs high), which is more efficient and quieter but decays current more slowly. Fast decay gives finer current control at low speed and better response when reversing; slow decay is more efficient and reduces audible noise. TI's Table 2 lets you pick either by swapping which input carries the PWM.

Use the built-in fixed-frequency PWM current regulation. Insert a sense resistor in the low-side return path and compare the resulting voltage against the internal 200 mV reference; the DRV8833 then chops the bridge to hold the winding current at the programmed limit. That protects both the driver and a small motor from stall current, which otherwise rises to the full locked-rotor value. Alternatively drive the xIN pins with a PWM signal from your MCU and let the motor inductance average the current.

Yes, with margin to check. VIH is 2 V for AIN and BIN and 2.5 V for nSLEEP, with VIL at 0.7 V and 0.5 V respectively and 0.4 V of hysteresis, so a 3.3 V CMOS output comfortably exceeds both thresholds. Input high current at 3.3 V is 16.5 µA typical for the control pins and 6.6 µA for nSLEEP, which is negligible. Just do not drive the inputs from an open-drain output without a pull-up to at least 2 V.

The most common fraud is selling DRV8833PW (TSSOP, 500 mA RMS) remarked as PWP. Inspect the package for the exposed PowerPAD on the underside and the correct body size, then run a thermal test: a genuine PWP holds 1.5 A RMS per bridge at VM = 5 V with a reasonable copper pour, while a PW part will overtemperature and pull nFAULT low well before that. Also verify the 2.6 V undervoltage lockout threshold and the 1.6 µA sleep current. We sample-test lots on a load bank and provide date-code traceability.

MOQ is one reel of DRV8833PWPR (2000 pcs, HTSSOP-16 with PowerPAD) for volume, or 100 pcs loose for prototyping; WQFN RTY packaging is available on reel as well. Shenzhen stock ships in one to two business days. TI authorized-channel lead time is ten to eighteen weeks, and low-voltage motor drivers are frequently allocated during shortage cycles.