Transistor H bridge base resistor values

[Zero999]

Use an H-bridge IC. The cheapest, most common ones are the L298N and SN754410 but they do have fairly high on voltage losses, so aren't ideal for running the motor off 5V. There are more modern motor driver ICs available, with much lower voltage losses, but they come in SMT packages, with a thermal pad, which require a reflow oven to solder, but modules with the ICs already soldered in are available.

https://www.sparkfun.com/datasheets/Robotics/L298_H_Bridge.pdf
http://www.ti.com/lit/ds/symlink/sn754410.pdf

[danners430]

Use an H-bridge IC. The cheapest, most common ones are the L298N and SN754410 but they do have fairly high on voltage losses, so aren't ideal for running the motor off 5V. There are more modern motor driver ICs available, with much lower voltage losses, but they come in SMT packages, with a thermal pad, which require a reflow oven to solder, but modules with the ICs already soldered in are available.

https://www.sparkfun.com/datasheets/Robotics/L298_H_Bridge.pdf
http://www.ti.com/lit/ds/symlink/sn754410.pdf

Neither are really ideal... It's going to be driven using PWM (separate thread for that in this forum), and since the motors are actually 6V, we kinda need as much voltage as possible :-)

As for the modules, it sorta defeats the purpose of this STEM event which is to get the students to build their own circuitry...

Would a MOSFET H-Bridge work for driving a motor with PWM in either direction?

Cheers

Sent from my ONEPLUS A3003 using Tapatalk

[Zero999]

What's the difference between using a module and an IC, other than the latter provides a bit more soldering practise?

If you're not expecting them to design the circuit, then the only purpose of building it, rather than using a module, is practical skills and if that's the point, rather than teaching them to think, fair enough.

Yes, MOSFETs can work. The problem is shoot-through. Here's a circuit I designed awhile ago, which has minimal shoot-through and is protected from both transistors activating simultaneously, if both outputs are high.


https://www.eevblog.com/forum/projects/dc-to-ac-converter/msg16288/#msg16288

[danners430]

What's the difference between using a module and an IC, other than the latter provides a bit more soldering practise?

If you're not expecting them to design the circuit, then the only purpose of building it, rather than using a module, is practical skills and if that's the point, rather than teaching them to think, fair enough.

Yes, MOSFETs can work. The problem is shoot-through. Here's a circuit I designed awhile ago, which has minimal shoot-through and is protected from both transistors activating simultaneously, if both outputs are high.


https://www.eevblog.com/forum/projects/dc-to-ac-converter/msg16288/#msg16288

Unfortunately, this is pretty much it - we need to provide them with the circuit, teach them how it works, and still get them to physically assemble it.

The circuit looks ideal - would there be any changes necessary for 5V operation?

Cheers

Daniel

[Zero999]

Yes, just use logic level MOSFETs. Make sure they will pass the required current for the motor, when the gate voltage is only 5V.

[danners430]

Yes, just use logic level MOSFETs. Make sure they will pass the required current for the motor, when the gate voltage is only 5V.

Hi mate,

Two quick questions regarding your H-bridge circuit:

1. Would this circuit work using PWM?
2. Would an alternative circuit using four logic outputs from the MCU work too, where the MCU programming prevents shoot-through by enforcing a delay between switching the FETs?

With regards to the second question, I got that idea from another source when investigating the PWM control - would this also be easier to wire up if I have the spare GPIO?

Cheers

Daniel

[danners430]

Just finished a circuit diagram - the PNP transistors are there to default the P-channel FETs to off in case of MCU failure :-)

Daniel

[Zero999]

There's nothing to turn the upper MOSFETs off.

To answer your previous questions:

• The circuit I posted was not designed for PWM. The turn off time for M3 and M4 is quite slow. It's possible there will be some shoot-through, if the opposite side is turned on very soon after the other side is turned off. Although this isn't good, it won't destroy the MOSFETs, if it doesn't happen very often.
• Relying on the software to not destroy the MOSFETs is bad design practise. Logic gates could be used before the MOSFETs, to ensure they don't both turn on simultaneously.

[danners430]

There's nothing to turn the upper MOSFETs off.

To answer your previous questions:

• The circuit I posted was not designed for PWM. The turn off time for M3 and M4 is quite slow. It's possible there will be some shoot-through, if the opposite side is turned on very soon after the other side is turned off. Although this isn't good, it won't destroy the MOSFETs, if it doesn't happen very often.
• Relying on the software to not destroy the MOSFETs is bad design practise. Logic gates could be used before the MOSFETs, to ensure they don't both turn on simultaneously.

The idea was that the PNP transistors act like a NOT gate to hold the FETs off unless they're powered by the MCU.

The reason I decided to go with this design is that it's reasonably simple, so easy for students to build on a breadboard - the MCU will be pre-programmed by ourselves, and we'll add in checks to ensure a good few ms delay between switching the FETs to prevent shoot-through.

Cheers

Sent from my ONEPLUS A3003 using Tapatalk

[Zero999]

There's nothing to turn the upper MOSFETs off.

To answer your previous questions:

• The circuit I posted was not designed for PWM. The turn off time for M3 and M4 is quite slow. It's possible there will be some shoot-through, if the opposite side is turned on very soon after the other side is turned off. Although this isn't good, it won't destroy the MOSFETs, if it doesn't happen very often.
• Relying on the software to not destroy the MOSFETs is bad design practise. Logic gates could be used before the MOSFETs, to ensure they don't both turn on simultaneously.

The idea was that the PNP transistors act like a NOT gate to hold the FETs off unless they're powered by the MCU.

Sorry, I should have said, how are the tp MOSFETs supposed to turn on?

The reason I decided to go with this design is that it's reasonably simple, so easy for students to build on a breadboard - the MCU will be pre-programmed by ourselves, and we'll add in checks to ensure a good few ms delay between switching the FETs to prevent shoot-through.

Cheers

That's a shame. You could have used it as an opportunity to teach logic gates, although I accept it would make it more complex and could take too long.

[danners430]

The idea behind the top FETs was to use them as digital inputs, PWMing the N-channel FETs. So a delay in the top ones actuating isn't a problem. By powering the PNP transistor, it would switch off, making the P-channel FET switch on... Or am I talking rubbish?

Sent from my ONEPLUS A3003 using Tapatalk

[Zero999]

The P-channel MOSFET's gate needs to go to +V to turn it off and 0V to turn it on and must be connected to either power supply rail. The gate must never be left disconnected from anything, i.e. just left floating.

[danners430]

The P-channel MOSFET's gate needs to go to +V to turn it off and 0V to turn it on and must be connected to either power supply rail. The gate must never be left disconnected from anything, i.e. just left floating.

So it needs pull-down resistors?

Sent from my ONEPLUS A3003 using Tapatalk

[Zero999]

Yes, pull-down resistors are required to turn the P-MOSFETs on.

[danners430]

Yes, pull-down resistors are required to turn the P-MOSFETs on.

Perfect. I'm guessing simple 1k resistors would do the job?

Cheers

Sent from my ONEPLUS A3003 using Tapatalk

[Zero999]

Yes, pull-down resistors are required to turn the P-MOSFETs on.

Perfect. I'm guessing simple 1k resistors would do the job?

Cheers

Sent from my ONEPLUS A3003 using Tapatalk

This really isn't the right way to do it.

I'd recommend not driving the MOSFETs directly from the MCU. If a MOSFET dies, it will take out the MCU. There are special MOSFET driver ICs available, but just for low speed/power/frequency applications, they're not worth it. A hex Schmitt trigger IC can be used, with RC circuits to delay the turning on of the top MOSFETs, to avoid shoot-through. The capacitors will charge, slowly via the resistors, delaying turn on, but discharge quickly through the diodes.