Anti-reverse-bias protection diode vs. performance

[hollowman]

Due to the complexity of a CDP external PS project (and potential danger of accidentally reverse biasing ICs and ckts) I'm thinkig about placing diodes at the inputs of ckts and even IC PS pins.

How much performance loss (if any) can one expect in such an application if one uses, say, "high-quality" diodes (e.g., fast + soft recovery)?

[anatech]

Hi hollowman,
Normal robust diodes should work fine reverse biased across the supplies. Make sure you fuse the supply lines before your diodes.

I can't imagine what performance hit you could take from this, they represent a small capacitance across your supplies to common. Do make sure they return to a robust ground, right at the plug or socket.

-Chris

[hollowman]

Originally posted by anatech I can't imagine what performance hit you could take from this, they represent a small capacitance across your supplies to common. Do make sure they return to a robust ground, right at the plug or socket.

(Thx for your reply).
Agreed: a robust ground is important. However, many of the ICs or ckts I use are not connected to Earth (plug/socket) ground; rather, they are digital ground, analog ground and even virtual ground for a headphone amp that that runs on battery/AC. Dunno whether it will make any difference in topic of this discussion (i.e., "Anti-reverse-bias protection diode vs. performance").

[anatech]

Hi,
Okay, just return those diodes to whatever power supply ground you are using. Just avoid the signal ground for any data buss.

-Chris

[TechGuy]

The issue with using a blocking diode is when your using low input voltages, because of the voltage drop across the diode. A fast recovery diode may have a forward voltage drop of 1 V or even higher. This can be significant if your using low voltage batteries as your power source. The Soft-slow recovery (cheap diodes) would be better as blocking diodes because the usually have smaller forward voltage drops. In this situation reverse recover performance isn't an issue.

One option is to use a fuse and a blocking diode across the input voltage and input ground. The Diode will short and blow the fuse, but this solution isn't an option if your load requires high input current or you concerned about depleting your battery. If your fuse is set for a high current load, the blocking diode will blow before the fuse blows.

Another option would be a set of back to back N-Channel mosfets with low Rds. and use a small driver that is reverse voltage protected.
http://www.linear.com/pc/downloadDoc...41,P1385,D4142

You can also use off the shelf reverse recovery devices:
http://focus.ti.com/lit/ds/symlink/ucc3952-1.pdf

[hollowman]

My main interest in adding protection diodes is to safeguard $$ and difficult-to-find digital ICs: PMD-100 (digital filter/oversampler), TDA1541A double-crown, etc. Hence, I'm thinking of adding protection diodes very close to the supply pins of these ICs. These ICs are already well bypassed with HQ caps, and the supply lines also have in-line ferrites.

Any thoughts on whether adding a protection diodes just before ferrites will degrade performance?