Output to input protection

[Marco Trapanese]

Hello,

reading the ds I'm not sure if this two ICs:

http://www.farnell.com/datasheets/808903.pdf
http://www.farnell.com/datasheets/1610860.pdf

when turned of using the enable pin can withstand a voltage (3.3V) on
their output. The input voltage may change from 0V to 4.7V.

Thanks
Marco

 

[Marco Trapanese]

Il 17/03/2013 19:51, Jim Thompson ha scritto:

The way I read the data, VSHDN > 45% of VIN to activate the circuit.

Hi Jim!
Let's suppose the VSHDN is pulled down to 0V. The circuit is off.

As for "withstanding", I'd hazard a guess that the input has a series
resistor built-in, poly on oxide, so no junction forwards to fret
over.

Well, I was afraid about the mosfet body diode: if I have a voltage at
output and 0V on the input will it turn on? Even the IC is in off condition.

Marco

 

[Marco Trapanese]

Il 17/03/2013 20:15, Jim Thompson ha scritto:

Aha! You're asking a different question than I assumed. You wonder
about the case where you have commanded shutdown (\SHDN\=0), and
VIN=0, but there's a voltage _forced_ at the output?

Exactly!
I'm sorry again for my poor English!

The block diagram of your first device implies a body-diode problem...
though that doesn't necessarily have to be the case... I've designed
"body drivers" to cope with such events (sort of like bus drivers when
powered down.

The second chip, a switch, implies an N-channel switch.

Who knows? Test yourself for current flow and the magic smoke ;-)

LOL

In the first case I solved placing another diode at the input so there
is no return path.

For the second chip.... I found another one:

http://www.farnell.com/datasheets/608802.pdf

which is designed to avoid the problem.

Marco

 

[Marco Trapanese]

Il 17/03/2013 20:34, Jim Thompson ha scritto:

On most processes, use of a P-channel pass device makes it easy...
since it has its own tub, you drive the tub with the higher of VIN or
VOUT.... so you can survive an output voltage even when the chip is
unpowered.

Good point. Anyway I changed a bit the connections and I avoided the
problem at all.

Thank you again, Jim.

Marco