LM2593HV: LM2593HV-ADJ FB input protection

Hello

The protection diode is used to clamp the FB voltage to a small negative value, if the output is shorted.

When using a CFF capacitor the voltage on the FB pin will spike negative when the output is shorted; the diode 

helps to clamp this voltage.  However, it does not depend on the input voltage; it depends on the output voltage.

I believe that is a misprint in the data sheet.  I don't think you need the diode for an output voltage of 12.7V.

If you wish to use a clamp, the BC junction of an NPN transistor will have much less leakage than a Schottky diode.

You can place a resistor in series with the CFF cap to limit the current.  However, this will affect the compensation of the loop.

You could also place a resistor in series with the FB pin.  However, if the value is too large, this will also affect the loop stability.

Thanks

Hi Frank,

Thank you for the fast response.

I understand the problem with the negative spike caused by the CFF cap during an output short circuit. I just wondered how it could possibly be related to the input voltage - it makes sense that the problem is related to Vo instead of Vin!

Do you think that I'll need to make short circuit tests to remove the diode - how certain are you :) ?

Will you (TI) possibly make a revision of the datasheet?

BR,

Lars Jespersen

It looks like when I answered the original E2E I was too quick to assume

that the data sheet contained an error in the wording.

In any application where the Cff is charged to a voltage greater than

a silicon diode drop there is the potential for causing an issue.

When the output is shorted, Cff will discharge into the ESD diode

connected to the FB pin.  When this occurs, a parasitic structure will

turn on and possibly cause incorrect behavior and/or damage.

The effects of this event will depend on the particular device, the nature of the parasitic, and the

silicon  process.

It will also depend on the size of Cff; since a larger capacitor will cause

more energy to be dumped into the parasitic structure.

Typically, when a parasitic transistor turns on, it will cause current to flow from the

input supply of the device and/or cause substrate current flow.  Larger input voltages

will create large power dissipation, in conjunction with this flow of parasitic current, thus

aggravating the issue.

For the LM259x family, with input voltages above 40V, this parasitic turn-on

will damage the device.

Therefore, the data sheet wording is correct.

Thank you

Hmm... I've already removed the diodes from the PCBA's... :-(
So I hope the risk of damage is small..