Forums

reverse voltage protection with parallel diode

Started by Unknown July 21, 2012
On Saturday, July 21, 2012 8:24:17 AM UTC-4, markj...@gmail.com wrote:
> I have a diode 1N5402 connected in parallel reverse biased across the pow=
er supply for reverse polarity protection. The maximum absolute ratings of= the load IC says -0.3V to +30V. If reverse power is applied, anything abov= e diode's forward drop it will short circuit and break the fuse. No iss= ue here. If, let's say, 0.5V is applied in reverse, the diode will not = conduct and reverse supply of -0.5V will go directly to the load IC. Since = it is below the min -0.3V, will it damage the IC?=20
>=20 > -Markj
http://www.ti.com/lit/an/slva139/slva139.pdf
If you can tolerate schottky reverse bias leakage, that would provide a 
better clamp.

You don't indicate if the user is plugging in an external supply, or the 
supply is hard wired. I'm not sure why you would expect a hard wired 
supply to go negative. [OK, maybe from one of those home brew switcher 
designs, which is of course why you should always buy a controller IC.]

I have a break out box I use for mobile applications to distribute 12vdc 
from a marine type battery (AGM). I use a fuse and parallel Schottky in 
the box. Actually a few Schottky diodes. Due to current hogging, I 
expect the diodes to turn on individually, so paralleling them doesn't 
do much in terms of current handling. My guess was the first diodes to 
turn on might fail before the fuse pops, so as each diode fails, a new 
one will take over. I added some junk box standard silicon diodes as 
well. Needless to say I haven't tested this! I just tossed in the parts 
because I had them handy and I couldn't see how it would hurt.

Regarding damage to ICs, you would really have to reach a diode turn on 
voltage to get the chip to draw significant current. The manufacturer is 
just guardbanding this by specing 0.3V.

With static DC, there are two way to damage an IC. Overvoltage will 
break down diode junctions. Reverse voltage will "melt" the metal trace 
in the path of the current.

This leads to a classic problem in IC metal fuse popping. If upon 
popping the fuse the voltage on the test pad spikes upwards, as can 
happen with a sudden di/dt as the fuse pops, the overvoltage can break 
down an internal junction and cause it to leak. Generally is it better 
to set up the fuse popper so it forward biases a junction after the fuse 
is popped. Given that the energy to pop the fuse is stored in a 
capacitor, you can predict how quickly the cap will discharge and thus 
predict the time the metal trace will have to conduct high current. 
Basically you can engineer around the potentially damaging situation. 
Add protection diodes and insure a sufficient metal width to them.