Half Bridge Motor controller with 60% duty cycle

[Anna]

Hi everyone ,

I tried to simulate a circuit of a half bridge motor controller with 60% of pwm duty cycle, but the output voltage and current waveforms in the coil do not appear ideal. I added a freewheeling diode parallel to the load and some resistors on the gates to discharge the gate capacitance, but it seems insufficient. I am a student and would greatly appreciate your explanations and support regarding this issue. If you have any solutions or recommendations, please feel free to share them with me. Thank you so much! Attached are the circuit and simulation.

[Siwastaja]

The load (motor) is wired wrong, in series between the MOSFETs.

You should just connect the two mosfets in series (drain and source together). That middle point becomes your Switch Node. Motor is connected between Switch Node and GND (or alternatively, Vcc). In this case, the body diodes of the MOSFETs also act as freewheeling diodes.

[Benta]

Which raises the question: what is the prupose of H-bridge drive? Single-ended drive would achieve the same. Un less dynamic braking is needed, but that would require separate drive for the brake MOSFET.

[Siwastaja]

Which raises the question: what is the prupose of H-bridge drive? Single-ended drive would achieve the same.

Even that "single-ended" would have a diode; it's still a switch, just an automagic one:
 * Turned on when the MOSFET turns off;
 * Turned off when the MOSFET turns on;
 * Also turned off when motor current reaches zero in low-load conditions (this is what prevents braking); using a MOSFET you can keep it on for negative current.

Typical purposes for using a MOSFET switch are, as you say, (regenerative) braking, but also just decreasing the diode Vf loss, which can be quite significant in low-voltage converters which run in low RPM, high torque conditions (low duty cycle, so diode conducts most of the time, at high current).

[Anna]

Thank you so much for your support! However, I did not connect the load as you mentioned because I need to follow the datasheet from Infineon. You will find it attached. It describes another method of connecting the MOSFETs to the load. I believe the connection is correct, but it may need some refinements.

[Benta]

I understand what you mean.
That data sheet drawing is misleading. "TO LOAD" does not mean that you must connect a load between those two points.
Each driver can be connected to it's own load. One ground-referenced, one HV-referenced. Can also be combined.
But not the way you did it.

[PCB.Wiz]

Hi everyone ,

I tried to simulate a circuit of a half bridge motor controller with 60% of pwm duty cycle, but the output voltage and current waveforms in the coil do not appear ideal.

What did you expect ?

I added a freewheeling diode parallel to the load and some resistors on the gates to discharge the gate capacitance, but it seems insufficient.

You can monitor the current in the gate drive resistors to check if you additions are needed.

I am a student and would greatly appreciate your explanations and support regarding this issue. If you have any solutions or recommendations, please feel free to share them with me. Thank you so much! Attached are the circuit and simulation.

You should always label nodes on the SCH that you will graph. Spice node numbers ( 10? 17? 5?) mean nothing to readers.

You start the plot at 82ms, and what is a bit strange is the sub-period variations every ~ 10ms ?
I would zoom out a bit and look for nodes that correlate to that ~ 10ms

[ArdWar]

Incorrect half bridge wiring aside, that's one huge bootstrap capacitor. You rarely need more than 1uF.
Can it even start up properly with that size?

[Siwastaja]

Thank you so much for your support! However, I did not connect the load as you mentioned because I need to follow the datasheet from Infineon. You will find it attached. It describes another method of connecting the MOSFETs to the load. I believe the connection is correct, but it may need some refinements.

This datasheet is just wrong. I don't know what they have been smoking, but that definitely isn't a "typical connection". Either it's just an accidental mishap, or as Benta suggests, it's intentionally so, which would mean there is some mental acrobatics going on there, but I don't think it's worth of anyone's time trying to analyze what Infineon tried to say with this figure. Instead, just google the typical half-bridge motor control circuit. First result on Google image search: https://4qd.co.uk/wp-content/uploads/hbr.gif . Note that the motor can be as well connected between the switch node and GND instead.

Connection between Switch Node and VCC makes sense when you use only one MOSFET, and diode as another switch. In such case the diode must go over the motor; so connecting the motor on the high side allows you to place the diode on the high side, allowing you to place mosfet on the low side, allowing you to use a simple N-channel MOSFET and drive the gate ground-referenced. So essentially the only reason you see the motor connected between Vcc and SW (instead of GND and SW) is easier gate drive circuitry.

[ArdWar]

It isn't that uncommon to find errors in datasheets. Especially this one where the "Typical Connection" section is used for illustration on necessary connections and components around the part only and not meant to be the guaranteed "data" point nor supposed to be a working application circuit. Common sense and knowledge of circuit working principle is still needed to complete the design.

On another note, OP's mistake of connecting the bridge and load all in series is actually kind of fortunate. The whole circuit is so wrong that it circles back to become actually somewhat workable. With HIN and LIN driven together in-phase like that it usually ends up with the transistors blowing up due to shoot thru had the bridge and load connection were designed properly. Instead with this connection it ends up working as on-off modulation closer to what Benta proposed at reply #2. But of course since the HI FET is in phase with LO FET, the bootstrap cap will not get pumped and HI gate drive won't work properly.

[Siwastaja]

With HIN and LIN driven together in-phase like that it usually ends up with the transistors blowing up due to shoot thru had the bridge and load connection were designed properly.

Yeah. Some bootstrap gate driver ICs (apparently not this one, or it does not say) have internal logic which prevents shoot through as part of adding deadtime. For the OP, another type of gate driver which has (EN, IN) inputs instead of (HIN, LIN) would be easier to use - you just drive square wave into IN, and use the EN pin for enabling/disabling the motor completely.

Even then, there's still a long way for an actual motor controller. For example, there is no current sense, which is pretty much mandatory unless the OP plans to only drive very small motors.