H-bridge using N channel FETs
Theory of Operation
by Eugene Blanchard
The problem that most people run into when using N channel MOSFETs for
H-bridges is that the MOSFET used to turn on and off the positive power supply
voltage, Vcc, will not work. This is the MOSFET which sits between the motor
The reason it does not work, is that a MOSFET is a voltage controlled device
(transconductance). A voltage on the input (gate-source) controls a current on
the output (drain-source). The voltage on the gate must be above the threshold
voltage (4.5 to 7 volts) of the MOSFET in order for the MOSFET to turn on.
The problems with N channel MOSFETs
Here lies the problem. The MOSFET's drain lead is connected to Vcc. The goal
is to turn the MOSFET on such that the top of the motor is at Vcc potential.
This means that the MOSFET's source lead will be raised to Vcc's potential. BUT
we need the gate to be 4.5 to 7V higher than the source to keep the MOSFET
turned on and we only have Vcc to work from! We need a gate voltage that is
higher than Vcc by 4.5 to 7 V!
A higher voltage can be accomplished by using a voltage doubler circuit to
generate a voltage that is larger than Vcc. The chip that I selected is the
Intersil 7662 negative voltage source which can be used as a voltage doubler
also. It is easy to use, small (8 pin DIP) and inexpensive. The ICL7662
about 5 mA of output which is more than sufficient to drive the MOSFETs. It is
good for a Vcc up to +18V or so while its sister chip the ICL7660 is only good
for a Vcc up to +10V.
How does it do that?
The components C2, D1, D2 and C3 of Figure 2 make up the voltage doubling
circuit. Basically, the ICL7662 voltage doubler is an oscillator that keeps
switching pin 2 from ground to Vcc and back again (See Figure 2). To start the
cycle, pin 2 is internally switched to ground. This charges up C2 by the
following path: Vcc to D1 to C2 to pin 2. C2 charges up to Vcc minus the
drop across D1 which is roughly 0.7V or so. Across C2 there will be (Vcc-0.7)
When ICL7662 switches again, pin 2 goes to Vcc but C2 has a full charge on
it (Vcc - 0.7V). This means that the anode of D2 is now Vcc from pin 2 PLUS the
charge of C2. The voltage on C2 is elevated above Vcc!
Diode D2 dumps the charge into C3. C3 now has a charge that is equal to Vcc
from pin 2 plus the charge on C2 MINUS the voltage drop across D2 (0.7V or so).
C3 has Vcc + Vcc-0.7V - 0.7V on it which is 2xVcc - 1.4V.
The voltage loss across the diodes can be minimized by using germanium
(0.3V) diodes or schottkey (0.25) diodes. The frequency of operation of the
ICL7662 is 10 kHz at room temperature.
Bonus - Negative voltage source!
In addition to being a voltage
doubler, the ICL7662 is a negative voltage source (actually that's its main
purpose - we're just not using it right ;-) . If you don't need a low-power
negative voltage source, you can leave capacitor C4 and diode D3 off the
The capacitors C1 and C4 and diode D3 make up the negative voltage source.
Basically, the ICL7662 voltage doubler is an oscillator that keeps switching
2 from ground to Vcc and back again and at the same time switches pin 4 from
ground to pin 5 (See Figure 2). To start the cycle, pin 2 is internally
switched to Vcc and pin 4 is switched to ground. This charges up C1 to Vcc with
the positive side of the charge connected to pin 2 (very important).
Next, pin 2 is internally connected to ground and pin 4 is internally
connected to pin 5. The positive charge on C1 is connected to ground (not
shorted - only connected). The capacitor has been electronically turned upside
down! This places the negative plate of C1 at pin 4 which results in a negative
Vcc onto pin 5.
Pin 5 is connected to diode D3 which dumps the negative Vcc into C4. The
charge on C4 is -Vcc minus the voltage drop across diode D3 which results in
-(Vcc-0.7V) across C4.
(Rev 01- 10k) is a schematic of the N channel H-bridge discussed here.
Figure 2- (2.6k) is a schematic of a simple
inexpensive voltage doubler circuit. Both are GIF files that can be opened by
any Web Browser.
MOSFETs are extremely static sensitive but more important is that if the
Gate is left open (no connection), the MOSFET can self- destruct. The Gate is a
very high impedance device (10+ megohms) and noise can trigger the MOSFET.
Resistors R4, R6, R8 & R13 of Figure 1 have been added specifically to stop
the MOSFET from self destructing. It is very important to install these
resistors FIRST before installing the MOSFETs. You will find that after these
resistors are installed that the MOSFETs are quite stable devices. The
pull-down the Gates and turn off the MOSFETs, not to mention add some static
Back EMF protection
D1 to D4 route back EMF from the motor back to the power supply. Some MOSFETs
(actually most) have these diodes built-in, so they may not be necessary.
Q1 & Q7 are NPN transistors that invert the control signals to Q2 and
Q8. Q2 & Q8 are PNP transistors that control Q3 and Q5 respectively.
Looking at Q1 and Q2 operation only, we can see that a High at A will turn
on Q1. This allows current to flow from Q2's base. This turns on Q2 and raises
the Gate voltage of Q3 to +24V. Q3 turns on.
When point A goes low, Q1 turns off, which turns off Q2 and Q3's gate is
pulled to ground by R4. Q3 turns off. Q7, Q8 and Q5 operation is similar.
The circuit could of been made simpler by by using only a NPN transistor to
pull down Q3's gate except that should something happen to Q1 or Q7 (failure),
the FETs would be turned on. The extra transistors provide a
mode of operation.
When A=0 and B=0, the motor is stopped. R4 and R13 pull down the Gates of Q3
and Q5 respectively and turn off the MOSFETs.
When A=0 and B=1 (+5V), the motor is in reverse. Q1 is turned off, Q2 is
off and Q3 is turned off due to R4.
Q7 is turned on by the voltage at B. Q7's collector pulls Q8's Base to
ground. This turns on Q8 which raises Q5's gate to +24V. This turns on Q5. The
-ve side of the motor is raised to +12V. R5 raises Q4'ss Gate to +11V or so
(voltage divider) which turns on Q4. Q4's Drain goes to ground which makes the
+ve side of the motor go to ground. R7 is also connected to the +ve side of
motor which pulls down Q6's Gate and makes sure that it is turned off. The
current path for the motor is from +12V to Q5 to -ve contact to +ve contact to
Q4 to ground.
When A=1 and B=0, the motor is in forward. Q7 is turned off, Q8 is turned off
and Q5 is turned off due to R13.
Q1 is turned on due to the voltage at A and Q1's collector goes to ground.
This turns on Q2. Q2's collector raises the Q3's gate to +24V which turns on
Q3 raises the motor's +ve side to +12V. R7 raises Q6's Gate voltage and turns
it on. When Q6 turns on, R5 makes sure that Q4 remains off. The current path
for the motor is from +12V to Q3 to +ve contact to -ve contact to Q6 to
NOT ALLOWED Mode (or fuse test
IF A=1 and B=1 then all MOSFETs turn on which shorts out the power supply
among other things - Not recommended.
The tricolor LED allows you to test the circuit without connecting the motor.
The LED will be green for one direction and red for the other. Handy test.
Motors make a lot of electrical noise from the brushes when running and
huge electrical spikes when stopping, starting and especially changing
direction. C1 and C2 try to suppress the noise spikes. Negative spikes are
shorted to either ground or the power supply by D1 to D4. Z1 tries to clip the
MOSFETs turn on very fast, if you have problems with noise, you may want to
put 0.1 uF capacitors in parallel with R4 and R13 to slow the turn-on time.
will reduce the EMF generated by the motors. If the rise time is too slow the
MOSFETs may heat up excessively!
Try to keep the motor supply separate from the logic supply if possible or
go to extreme filtering techniques using coils, diodes and capacitors to
out the motor noise.
This site is sponsered by
Cat's Ass Kustoms
Parts, Performance, Passion
News for Motorheads
Disclaimer: The circuits described here are for reference purposes only.
There is no guarantee in any way shape or form that they will work for your specific application. Use them at your own risk.
Please send any corrections, suggestions or errors that you may have
caught to me.
Copyright Eugene Blanchard Jan 2007