using a pair of half-bridges (Bjt, Gate driver) to make a full bridge DC motor driver

[mkeyno]

Hi,
I intend to use a pair of half-bridge drivers to make a full bridge driver for controlling my DC motors with 3 standard input pins of Enable, Direction, and Arduino PWM, my first option was to use the logic gates and bjt Mosfet driver to run the full bridge, similar to following design

figure 1

to increase on-delay and shoot-through prevention I added a pair of resistors and a diode

another option is using the half-bridge gate driver which already has shoot-through prevention and Schmitt trigger for noise suppression, however, I found two types of driver
1 expose the High-Low side signal and the other exposed Enable and PWM signal

type two

figure 2

I know I can connect diagonally High-Low side signals but not sure what's the best way to use a pair of PWM-EN to convert to EN-Dir-PWN pins
any suggestion for what's the best solution appreciated

figure 3

 

[FvM]

Bootstrap supply of high side driver (Fig. 2 +3) doesn't work for static operation (permanently on). Discrete circuit (Fig. 1) has no suitable high side driver supply and no measures against cross-conduction or supply brown-out.

--- Updated Apr 14, 2024 ---

See continuous gate-drive paragraph in this application note (pg. 17f)

https://www.infineon.com/dgdl/Infineon-HV_Floating_MOS_Gate_Drivers-ApplicationNotes-v01_00-EN.pdf?fileId=5546d4626c1f3dc3016c47de609d140a

 

[KlausST]

Hi,

a motor control is designed for a motor .. and is desigend with your own ideas / specifications / requirements.
This is the target. But we don´t know any of your targets.

So what´s your motor specifications? V, I, run time
What´s your power supply specification?
What´s your input signals specification? signal levels, PWM frequency?

also what about additiional features like: Overcurrent handling (on stalled motor, or so..)

******
If one knows the design goals one can focus on the solution:
My first idea is to use ready to buy modules ...
If this is not possible/available, then I´d design one with the help of application notes, design examples, proven circuits (just to preven to re invent the wheel and make mistakes others already have solved)

Designin a motor control the discrete way .. never came into my mind . in the last decades of professionally designing circuits.
Not that it´s not possible, or that I don´t have the skills for it. It´s because a "switching power" PCB needs to be carefully designed. And this takes a lot of time and experience. Discrete design means more parts, bigger PCB, more lengthy signals, bigger GND planes ... all this calles for possible problems ... caused by the swiching currents.

I don´t know how experienced in designing critical PCBs you are ... For me - as a rather experienced PCB designer: I´d avoid a discrete design.

*****
Referring to your last scematic:

Rather focussing on reducing the signal count (to 2 only: DIR and EN) .. I´d fosuc on stable and reliable function.
* 4.7 uF at VCC seems way too small ... in detail it depends on your power supply, switching frequency, wiring stray inductance....
* what´s the use of R20, C14, C15?
* It seems you want to use mature CMOS logic ICs. Are you sure they are: rugged enough, are compatible to your (unkonwn) arduino signal levels, fast enough...
* are you sure you want "fast decay" motor driving method. It makes the control more "PWM to torque" (with unstable RPM) than "PWM to RPM" (independent of torque) ... in detail it depends on yoour need...
* no filter on the motor side?
* no HF filter on the VCC side?
* are you sure you can handle the expectable GND bounce not to violate the arduino control signals
* how is your Arduino powered? You will need a wiring scheme

Klaus

 

[mkeyno]

Bootstrap supply of high side driver (Fig. 2 +3) doesn't work for static operation (permanently on). Discrete circuit (Fig. 1) has no suitable high side driver supply and no measures against cross-conduction or supply brown-out.

--- Updated Apr 14, 2024 ---

See continuous gate-drive paragraph in this application note (pg. 17f)

https://www.infineon.com/dgdl/Infineon-HV_Floating_MOS_Gate_Drivers-ApplicationNotes-v01_00-EN.pdf?fileId=5546d4626c1f3dc3016c47de609d140a

I used the datasheet drawing to make an h-bridge and just tried to change the pin configuration, so you said the half-bridge driver is only good for high-switching applications and not for static operation ?!!!!

 

[mkeyno]

Hi,

a motor control is designed for a motor .. and is desigend with your own ideas / specifications / requirements.
This is the target. But we don´t know any of your targets.

So what´s your motor specifications? V, I, run time
What´s your power supply specification?
What´s your input signals specification? signal levels, PWM frequency?

also what about additiional features like: Overcurrent handling (on stalled motor, or so..)

******
If one knows the design goals one can focus on the solution:
My first idea is to use ready to buy modules ...
If this is not possible/available, then I´d design one with the help of application notes, design examples, proven circuits (just to preven to re invent the wheel and make mistakes others already have solved)

Designin a motor control the discrete way .. never came into my mind . in the last decades of professionally designing circuits.
Not that it´s not possible, or that I don´t have the skills for it. It´s because a "switching power" PCB needs to be carefully designed. And this takes a lot of time and experience. Discrete design means more parts, bigger PCB, more lengthy signals, bigger GND planes ... all this calles for possible problems ... caused by the swiching currents.

I don´t know how experienced in designing critical PCBs you are ... For me - as a rather experienced PCB designer: I´d avoid a discrete design.

*****
Referring to your last scematic:

Rather focussing on reducing the signal count (to 2 only: DIR and EN) .. I´d fosuc on stable and reliable function.
* 4.7 uF at VCC seems way too small ... in detail it depends on your power supply, switching frequency, wiring stray inductance....
* what´s the use of R20, C14, C15?
* It seems you want to use mature CMOS logic ICs. Are you sure they are: rugged enough, are compatible to your (unkonwn) arduino signal levels, fast enough...
* are you sure you want "fast decay" motor driving method. It makes the control more "PWM to torque" (with unstable RPM) than "PWM to RPM" (independent of torque) ... in detail it depends on yoour need...
* no filter on the motor side?
* no HF filter on the VCC side?
* are you sure you can handle the expectable GND bounce not to violate the arduino control signals
* how is your Arduino powered? You will need a wiring scheme

Klaus

I have more than 12 different gate drivers and a dozen N-MOS and trying to find the general schematic for making a low-switching 12-36 DC motor that can control via Arduino MCU which is isolated via optocouplers,
I don't know if R20, C14, C15 are necessary or not, just follow some datasheet suggestion
direction signal will apply to the DC motor and remain for a long period of time for specific power(PWM)
should I consider a filter or an HF filter or part to handle expectable GND bounce so as not to violate the Arduino control signals?

 

[KlausST]

So what´s your motor specifications? V, I, run time
What´s your power supply specification?
What´s your input signals specification? signal levels, PWM frequency?

so your motor has no current and the PWM no frequency....

If one knows the design goals one can focus on the solution:

....

I have more than 12 different gate drivers and a dozen N-MOS and trying to find the general schematic for making a low-switching 12-36 DC motor that can control via Arduino MCU which is isolated via optocouplers,

I don´t know what to do with this information. Is regarding discrete vs IC?
And if you have them laying around, why not use them? What´s wrong with them?

And if the gate drivers are different, then what´s the idea to find a "general schematic". Each device comes with it´s datasheet that shows how to use it.

To me this sounds like to have a sports team and try to find the "general size" for their shoes. They all are different and all need shoes that fit their feet.

Optocouplers - when used correctly - may prevent from grund boucne problems. But we don´t see optocouplers, nor how you wired them.

Klaus

 

[FvM]

you said the half-bridge driver is only good for high-switching applications and not for static operation ?!!!!

Yes. Also manufacturer application notes say, probably not always clear enough.

For clarification, analyze bootstrap driver operation yourself. High side driver is supplied by Cboot, operation depends on Cboot being periodically recharged when low side switch is activated. Fig. 2 application circuit from FAN73832 datasheet can work with PWM up to 95 or 98 % duty cycle but not 100 %. Your choice either to use permanent PWM or add an auxiliary circuit to recharge bootstrap capacitor, see AN-978 quoted in post #2.

 

[mkeyno]

so your motor has no current and the PWM no frequency....


....


I don´t know what to do with this information. Is regarding discrete vs IC?
And if you have them laying around, why not use them? What´s wrong with them?

And if the gate drivers are different, then what´s the idea to find a "general schematic". Each device comes with it´s datasheet that shows how to use it.

To me this sounds like to have a sports team and try to find the "general size" for their shoes. They all are different and all need shoes that fit their feet.

Optocouplers - when used correctly - may prevent from grund boucne problems. But we don´t see optocouplers, nor how you wired them.

Klaus

DC motor operation would be simple, with no high switching, the only consideration is using the direction instead of enabling/disabling the PWM signal on each half-bridge

As I mentioned, regarding datasheets all IC gate drivers come in two groups, one has Enable/PWM and another High-Low signal, and my question was which of these two types more suitable for a simple H-bridge application
but as you mentioned I'd rather add all parts the make it more robust against undesired situations

I'm gonna use Optocopler like the following schematic

 

[FvM]

Logic not right. You want to drive the non-switching half-bridge low, therefore AND instead of NAND logic needed. There's an erroneous junction point right of IC12.1

As stated before, PWM duty cycle must not exceed 95 - 98%, also with LN8322.

LN8322.4 should connect to 12V rather than Vcc.

 

[mkeyno]

Logic not right. You want to drive the non-switching half-bridge low, therefore AND instead of NAND logic needed. There's an erroneous junction point right of IC12.1

As stated before, PWM duty cycle must not exceed 95 - 98%, also with LN8322.

LN8322.4 should connect to 12V rather than Vcc.

I use NAND so it won't work when both signals are high
and I tied the EN with PWM to avoid extra logic
also, I think AND gate inherently has no Schmitt trigger
do you have any suggestion on which logic scheme might work to bind the 3 control pins to 4 control gate

 

[FvM]

I think, I clearly explained my points. Nothing to add presently. Suggest to test circuit operation in simulation.