1
\$\begingroup\$

I am still on my struggle trying to design my own Brushed DC H-bridge with all N-MOS, still in the research stage and burning mosfets left and right.

I have come across this T.I design ,relatively recent. The full document about their design is here.

On page 12 and pictured below is the schematic of the H-bridge portion.

enter image description here

My question is as follows: I see no schottky diodes protecting it from the proverbial inductive spike. How come?

I do see the zener and resistor network but the paper says:

a Resistor-Zener network is provided on the switch nodes of the driver to protect it from overshoots and undershoots

I assume this "overshoot" refers to overshoot due to the bootstrap circuit and not inductive spiking?

Could it be that if you drive an H bridge in a complimentary manner, where the top and bottom mosfets are driven opposite of each other this gives the inductive spike a route to ground on off times.

For examplel: If you turn Q2 and Q3 ON the motor spins one way and when you turn them OFF there will be inductive spike so if you then turn Q1 and Q4 ON the spike has a route to ground and to Vbus.

that is the only explanation i can think of why they dont have protection diodes...

and yes I understand there has to be deadtime to protect from shoot through and all of that but my question is why there are no diodes?

One more question is what is the purpose of the resistor and capactors highlighted inside the bridge?? are these what they call snubber circuits?

\$\endgroup\$
  • \$\begingroup\$ they dont have protection diodes What about those diodes drawn in the MOSFET symbols? All MOSFETs have such diodes (it is an intrinsic element of a MOSFET's design) and some are suitable to act as protection diodes. But here D3 and D4 probably do most of the protection, they're not zeners but TVS diodes. \$\endgroup\$ – Bimpelrekkie Jul 13 '18 at 16:28
  • \$\begingroup\$ @Bimpelrekkie I thoguht the "intrinsic" diodes of mosfets were not something to substitute a dedicated schottky diode specially when dealing with high voltage spikes? In other words, yes its there but not to replace proper design with dedicated components.. unless those mosfets where made as such where that intrinsic diode can be taken "seriously"? \$\endgroup\$ – Edwin Fairchild Jul 13 '18 at 17:34
  • \$\begingroup\$ per the datasheet of the design they are zeners. \$\endgroup\$ – Edwin Fairchild Jul 13 '18 at 17:35
2
\$\begingroup\$

I see no schottky diodes protecting it from the proverbial inductive spike. How come?

The bulk diode that is pretty much a 100% feature of all modern enhancement mode MOSFETs can be relied upon to shunt any excess energy from inductive loads to the power rails. Most data sheets go to great length to describe the bulk diode's characterisitic so that the reader is under no-doubt that they can be relied upon. For the MOSFET in your design: -

enter image description here

  • Forward voltage is maximum of 1 volt at 10 amps
  • Reverse recovery time is typically 53 ns

In addition it has a couple of graphs that help the designer: -

enter image description here

a Resistor-Zener network is provided on the switch nodes of the driver to protect it from overshoots and undershoots

That's to protect the MOSFET driver ICs.

what is the purpose of the resistor and capactors highlighted inside the bridge??

R6 and R7 protect the zeners i.e. they limit current into the zeners that protect the driver ICs. The capacitors supplement the internal drain source capacitors of the MOSFETs (circa 300 pF) and although they represent a switching loss overall, probably are provided for EMC reasons.

\$\endgroup\$
1
\$\begingroup\$

There is already diodes built into the mosfets, and D3 and D4 zeners give added protection.

There are a few ways to burn out mosfets:

1) Over current, make sure that only one side of the bridge turns on at a time. Verify this with an oscilloscope. If you switch direction rapidly with your motor the back EMF can generate enough current to exceed the rating on the mosfet and burn it out. You'll need some time (built in to the software) to slow down the motor and then switch the polarity of the h-bridge.

2) Don't exceed Vgs which is ±20V in this case. In some cases I have used a 15V zener across the top FETs (in this case pin 4 of Q1 and HS_Phase_B for example, do the same for Q2) because if HS_Phase_B gets much higher than HS_Gate_B then you can exceed Vgs.

3) And of course don't go over the Vds of 100V

4) Don't ESD the gate, it can be really easy to do, the gate is only nm thick and on early FET's you could almost blow the gate out by waving your hand over it. On modern FET's they have more protection, but the gate can still be blown out.

5) Don't violate Id or the power dissipation. If the FETs are off the current from the motor can still travel through the diodes in the FETs and create problems, if you have a very large motor or load with a lot of energy to dissipate, this power can be dissipated in the mosfets.

Verify both of the Vgs and Vds with a scope when your running the motor, if you see spikes that get close to the Vgs rating of 20V (especially on the top FETs, the bottom FETs are grounded) then that might be your problem.

My question is as follows: I see no schottky diodes protecting it from the proverbial inductive spike. How come?

I don't know, not a super great design, if you really want to protect it, then add the protection. First scope it to see where the problem lies.

I assume this "overshoot" refers to overshoot due to the bootstrap circuit and not inductive spiking?

Not sure, You'll get overshoot from the inductive spiking so the mosfets need to be protected against that.

I understand there has to be deadtime to protect from shoot through and all of that but my question is why there are no diodes?

When all of the FETs are off the Back EMF from a spinning motor will start to create a large voltage the diode D4 and R7 are designed to ensure this voltage stays below the 100V Vds rating of the part, R5 and C9 are designed to shunt short high frequency overshoots, the D4 diode can take over after that.

One more question is what is the purpose of the resistor and capactors highlighted inside the bridge?? are these what they call snubber circuits?

See above, they are called snubber circuits because they 'snub' the voltage (socially ignore or turn away)

\$\endgroup\$
  • \$\begingroup\$ Tek FET probes were classic examples of waving hand over probe and damage the FETs \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jul 13 '18 at 17:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.