Follow up "LTC4367 gate problem"

PeterG. said:

In case of reverse polarity, the LTC4367 ties VIN to GATE internally.
Wouldn't that cause an uncontrolled high current through the LTC4367 if the diode D12 forward voltage is lower than the body diode forward voltage of Mosfet T2?

Where would the D12 current come from?
When GATE follows VIN as it goes negative, T2 and T3 are turned OFF.

Here's a simulation showing D12 and T2 have no current flow when V_IN goes negative.


-Aaron

Hi,

Thanks for your answer.

Ashapiro said:

When GATE follows VIN as it goes negative, T2 and T3 are turned OFF.

Yes Of course T2 and T3 are OFF when VIN goes negative!

Ashapiro said:

Where would the D12 current come from?

T3 can still avalanche, especially with large and low impedance output capacitances (6000µF @ 3mΩ in our case). 

Which could lead to a current flow throw D12 (if D12 forward voltage is lower than the body diode forward voltage of Mosfet T2)

Ashapiro said:

Here's a simulation showing D12 and T2 have no current flow when V_IN goes negative.

Thanks! We have done several Simulations our self as well and they all show current throw D12. (up to several amps.)

Wouldn't it be better to use a bidirectional diode instead?
Or is that just a simulation problem?

Many Thanks, 

Hi Peter,
My responses below:

PeterG. said:

T3 can still avalanche, especially with large and low impedance output capacitances (6000µF @ 3mΩ in our case). 

Which could lead to a current flow throw D12 (if D12 forward voltage is lower than the body diode forward voltage of Mosfet T2)

I see.
To prevent this scenario from happening, choose a FET where: BVDSS > (V_IN,MAX + 40V).

PeterG. said:

Wouldn't it be better to use a bidirectional diode instead?

I don't think it's necessary with back-to-back FETs, but this would guarantee the zener never gets forward biased.

PeterG. said:

Thanks! We have done several Simulations our self as well and they all show current throw D12. (up to several amps.)

PeterG. said:

Or is that just a simulation problem?

Check if your back-to-back FETs are installed backwards in the simulation. In this case, reverse current would flow through T3's body diode.

Have a good day,
-Aaron

Hi Peter,
My responses below:

PeterG. said:

T3 can still avalanche, especially with large and low impedance output capacitances (6000µF @ 3mΩ in our case). 

Which could lead to a current flow throw D12 (if D12 forward voltage is lower than the body diode forward voltage of Mosfet T2)

I see.
To prevent this scenario from happening, choose a FET where: BVDSS > (V_IN,MAX + 40V).

PeterG. said:

Wouldn't it be better to use a bidirectional diode instead?

I don't think it's necessary with back-to-back FETs, but this would guarantee the zener never gets forward biased.

PeterG. said:

Thanks! We have done several Simulations our self as well and they all show current throw D12. (up to several amps.)

PeterG. said:

Or is that just a simulation problem?

Check if your back-to-back FETs are installed backwards in the simulation. In this case, reverse current would flow through T3's body diode.

Have a good day,
-Aaron