Protecting an H-Bridge

Thread Starter

cmartinez

Joined Jan 17, 2007
7,879
I already have some experience working with H-Bridges, although I'm far from being an expert in the subject.

My question has to do with how to protect them from inductive loads. I can tell from the following diagram of a conventional low-side nFets and high-side pFets that their intrinsic diodes work together to protect the components from voltage spikes when the load is switched off. The circuit in question is using DMN2056U nFets and DMP2045UQ pFets. And both are rated at 20V drain to source, and 8V gate to source. Both transistors have a capacity of about 4 amps, continuous. The solenoid's impedance is 4.2 ohms.

1668370057280.png

I normally place a TVS diode of an adequate value (in this case, about 18V) in parallel with the solenoid coil, and it's been working fine for years now. But is it really necessary? Aren't the transistors' intrinsic diodes enough to protect the circuit? ...

And by the way, the solenoid's activation/deactivation period is about 5 seconds.
 

Thread Starter

cmartinez

Joined Jan 17, 2007
7,879
The body diode has always worked for me. I don't understand the concern.
My concern has to do with the body diode maximum forward voltage. I don't know what the proper term for it is, but it's the voltage at which the diode will fail if exceeded. And I can't find that info anywhere in the transistors' datasheets.
 

MrSalts

Joined Apr 2, 2020
2,620
My concern has to do with the body diode maximum forward voltage. I don't know what the proper term for it is, but it's the voltage at which the diode will fail if exceeded. And I can't find that info anywhere in the transistors' datasheets.
The body diode can handle the rated current of the mosfet - the body diode has a vf of a normal diode (about 0.6v). The diode will make sure the inductive kick from the diode doesn't exceed v(f).
 

Thread Starter

cmartinez

Joined Jan 17, 2007
7,879
The body diode can handle the rated current of the mosfet - the body diode has a vf of a normal diode (about 0.6v).
perhaps I used the wrong term, as I said ... so here goes another try: What's the maximum voltage that the fet's diode can withstand before it fails?
 

MrSalts

Joined Apr 2, 2020
2,620
perhaps I used the wrong term, as I said ... so here goes another try: What's the maximum voltage that the fet's diode can withstand before it fails?
The reverse voltage (the concerning voltage of any diode) is the voltage rating of the MOSFET. But your solenoid creates an inductive kick (a reverse current). Ohms law will determine the voltage spike in most cases but, you are shorting the current to the positive rail with a diode. A diode has a forward voltage of 0.6v. There is no possible way the voltage will exceed 0.7v - for thst very reason, it is called a "protection diode" on any inductive circuit (external diode is just not needed on a Mosfet H-bridge).

The same diode protection circuitry is often used on the input of an IC or even microcontrollers to insure no negative inputs.
 

Thread Starter

cmartinez

Joined Jan 17, 2007
7,879
The reverse voltage (the concerning voltage of any diode) is the voltage rating of the MOSFET. But your solenoid creates an inductive kick (a reverse current). Ohms law will determine the voltage spike in most cases but, you are shorting the current to the positive rail with a diode. A diode has a forward voltage of 0.6v. There is no possible way the voltage will exceed 0.7v - for thst very reason, it is called a "protection diode" on any inductive circuit (external diode is just not needed on a Mosfet H-bridge).
Thanks for the explanation ... I gather then that my question should rather be "how much current can said diode withstand?", which you have already answered ... but one doubt remains ... kickback inductive voltages are normally quite high (about up to 10 times of the normal applied voltage, if I'm not mistaken), and since V=IR according to Mr Ohm, then the surge in current will also be affected proportionally... I'm guessing that the transistor's rated "avalanche current" has to do with that?
 

MrSalts

Joined Apr 2, 2020
2,620
Thanks for the explanation ... I gather then that my question should rather be "how much current can said diode withstand?", which you have already answered ... but one doubt remains ... kickback inductive voltages are normally quite high (about up to 10 times of the normal applied voltage, if I'm not mistaken), and since V=IR according to Mr Ohm, then the surge in current will also be affected proportionally... I'm guessing that the transistor's rated "avalanche current" has to do with that?
The collapsing magnetic field of the solenoid causes current flow. The voltage can only get high if you have a large resistance in the circuit (I*R=V). If you have low impedence paths to the positive rail, which the diode does provide, you never get to large voltages. Protection diodes work (whether in the Mosfet package or not).
 

nsaspook

Joined Aug 27, 2009
10,698
Another consideration is you normally want to dump the stored inductive energy as close to the source as possible. All wiring and conductors from points A (solenoid) to B (H-bridge) are possible sources of problematic energy induction into a circuit. So even if the circuit can handle the current, it's wise to cut it off at the head as a precaution when practical.
 

Thread Starter

cmartinez

Joined Jan 17, 2007
7,879
Another consideration is you normally want to dump the stored inductive energy as close to the source as possible. All wiring and conductors from points A (solenoid) to B (H-bridge) are possible sources of problematic energy induction into a circuit. So even if the circuit can handle the current, it's wise to cut it off at the head as a precaution when practical.
The head being the solenoid coil itself? ... so what I mentioned in my first post, the use of a TVS in parallel to the coil, is not overkill?
 

crutschow

Joined Mar 14, 2008
31,126
As has been stated, the intrinsic MOSFET substrate diodes should be adequate to handle the inductive current from the solenoid without a problem.
Then the reverse inductive voltage across the MOSFETs is no more than the drop across the substrate diodes when the are forward biased with a current equal to the solenoid current.
I see no reason for a TVS diode across the solenoid.
And I wouldn't be too concerned about the length of the wires to the solenoid, since any added inductance from the wire will likely be quite small as compared to the solenoid inductance.
Basically the solenoid inductance swamps the wire inductance.
 

crutschow

Joined Mar 14, 2008
31,126
it's about the wire as an source antenna loop area for EMI/RFI that can interfere with circuit operation in places other than the MOSFET.
Your reference is for using a TVS diode to suppress the inductive transient, which still generates significant voltage ringing that can certainly generate some EMI.
Here the transient is suppressed by a (substrate) diode, which will generate a <2V transient with likely only slight ringing, so I wouldn't expect significant radiated EMI, even from a long wire.

But if you want to be on the safe side, connecting the solenoid with a twisted-pair wire will significantly reduce any possible EMI.
 
Last edited:
Top