Microcontroller drives H bridge to power a permanent-magnet dc motor
Improved H-bridge requires only two driver signals.
Luca Bruno, ITIS Hensemberger Monza, Lissone, Italy; Edited by Brad Thompson and Charles H Small -- EDN, March 15, 2007
A traditional method of driving a low- to medium-power permanent-magnet dc motor involves using four MOSFET or bipolar transistors in an H-bridge configuration. For example, in Figure 1, the motor connects between collector pairs C1 and C2 and C3 and C4. Turning on diagonally opposite transistor pairs Q1 and Q3 or Q2 and Q4 steers current through the motor and allows for reversal of its direction of rotation. However, this method requires that each of the four transistors receive its own control input. Depending on the motor’s voltage requirements, the upper two drive signals may require electrical isolation or a level-shifter circuit to match the microcontroller’s output-voltage limitations.
This Design Idea describes an alternative circuit that drives only the H bridge’s two low-side switching transistors. In a standard bipolar-transistor H bridge for bidirectional motor control, Q1’s and Q4’s bases connect to Q3’s and Q2’s collectors through resistors R3 and R4 (Figure 2). Inputs VINA and VINB each control a pair of switches. When Q2 turns on, resistor R4 and diode D6 pull Q4’s base low, saturating Q4 and pulling current through the motor and Q2. Similarly, turning on Q3 pulls Q1 into saturation and drives the motor in the opposite direction. Diode D5 ensures that Q1 remains off when Q4 conducts, and D6 performs the same function for Q4 when Q1 conducts. Resistors R1, R2, R7, and R8 increase the switching speed of their associated transistors, and resistors R5 and R6 limit base-current drain from the microcontroller’s 5V high-logic-level outputs to approximately 15 to 20 mA. Resistors R3 and R4 set Q1’s and Q4’s saturation base currents. Their value depends on the motor-supply voltage and Q1’s and Q4’s dc current-gain according to the following equation: R3=R4≤[VCC–VBE(ON)(Q4) –VF(D6)–VCE(SAT)(Q2)]/[(IMOTOR)/hFE(MIN)(Q4)]. For best performance, select bipolar-junction transistors with low collector-emitter saturation voltages, VCE(SAT), and high values of dc-current gain, hFE. Currently available medium-power transistors compete favorably with MOSFETs by offering these characteristics in combinations that minimize collector-power dissipation and require little base drive.
Discrete devices such as On Semiconductor’s NSS40200LT1G PNP and NST489AMT1 NPN bipolar transistors work well in the circuit in Figure 1. For a more compact implementation, you can select an integrated H bridge, such as Zetex’s ZHB6790, which operates at power-supply voltages as high as 40V, with 2A continuous and 6A peak pulse-current collector ratings. Its minimum current gain of 500 at a collector current, IC, of 100 mA can decrease to 150 at IC of 2A. At a worst-case collector current of 2A in Q2 and Q3, achieving a saturation voltage of 0.35V or less requires a base current of 13 to 20 mA. Fortunately, many microcontrollers’ outputs can source or sink as much as 25 mA and thus directly drive the H bridge independently of the motor’s power-supply voltage. To further reduce drive current or to use a standard CMOS or TTL IC as a drive source, you can buffer Q2’s and Q3’s inputs with small-signal transistor inverters. As an option, you can connect fractional-ohm resistors between the emitters of Q2 and Q3 and ground. This approach can provide analog voltages proportional to motor current, allowing the microcontroller to detect a stalled or overloaded motor.
-
In searching for a web reference to an EDN Design Idea of mine whose acceptance for publication is shown in a
letter in my possession signed by Charles H. Small dated 29th January 1992 (#1104 Bridge Motor Drive) I was both disappointed in not finding a web reference (a lot more searching may have eventually yielded it) and dismayed to find the current idea which is fundamentally showing the same topology as mine, the only real difference being that the two lower devices in my design were N-ch MOSFETs (an even better choice for driving at moderate speeds from a uC port bit it might be added). The design was more than a theoretical whimsy - I used it on several occasions and it worked very well. At the time of its publication I was working in what was the Wellington Polytechnic, New Zealand.
Malcolm John Watts - 2007-3-9 18:26:00 PDT -
I believe this design has a problem with back-EMF unless the microcontroller waits for the motor to stop before trying to reverse it. This may not be possible if the load has a torque bias, and dynamic braking is ruled out in any case. Let''s say Q1 & Q3 are conducting (Q3 may be getting a PWM signal or continuous drive). The motor is spinning. Now the uC turns off Q3. The motor is now a generator, sourcing current through D1, D5, R3 and R1||Q1''s base. (D3 is not forward-biased because the motor''s EMF is less than the supply voltage and R3 is pulling the node up.) This will keep Q1 ON until the motor''s EMF drops below ((R1+R3)/R1 + 2)*0.6v. If the uC turns on Q2 any sooner than that, Q1 & Q2 will short out the supply at least momentarily, until Q4 turns on and diverts the drive from Q1. Since Q2 is also pulling the D3 side of the motor down in series with the motor''s EMF, Q1 is being driven hard until Q4 turns ON. Moreover, this happens to a lesser extent even if the motor is stationary. A means is needed to prevent this cross-conduction.
Dick Neubert - 2007-30-3 06:49:00 PDT
