diode leakage, input protection

May 22, 2014 7 Replies
J

Does anybody know how power diodes (rectifiers, zeners, TVSs) are fabricated? Do these use patterned lithography, or do people just fab big unfeatured wafers and dice them up? I wonder if chip edge damage contributes to leakage, or maybe the sheer combination of size and doping makes power diodes leaky.



The context is that I have a test/cal connector that I want to protect from overload, at least +-20 volts and preferably more. I can't add a lot of series resistance.



A polyfuse and clamp might work:



formatting link



but I need fairly high-current, low leakage clamp diodes, and most official diodes are leaky, even typ values at room temp, much less max at 75C. It seems that transistors are better diodes than diodes.



Analog Devices just EOL'd the neat ADG465. Bummer.



John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Vote
Twitter Facebook LinkedIn
G

Yeah use a transistor maybe... the BC junction for higher voltage (I think) (I've got a nice B. Pease paper on clamping and bounding) ... or maybe even a diode connected transistor?

George H.

Vote
Twitter Facebook LinkedIn
J

Once-upon-a-time wafers were indeed unpatterned, and the die were edge-etched after sawing to clean up.

I think everything modern is now patterned diffusion or implantation.

Leakage _is_ proportional to size and crud content.

...Jim Thompson

| James E.Thompson | mens | | Analog Innovations | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | San Tan Valley, AZ 85142 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at http://www.analog-innovations.com | 1962 | I love to cook with wine. Sometimes I even put it in the food.
Vote
Twitter Facebook LinkedIn
W

So, it's time to design in a protective relay. High current: check Low leakage: check

Vote
Twitter Facebook LinkedIn
J

We could sense the overload and drop out the test relay, but that would take a millisecond or two. We need to protect the ICs meanwhile.

John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Vote
Twitter Facebook LinkedIn
L

Den torsdag den 22. maj 2014 19.31.10 UTC+2 skrev John Larkin:

a relay? either short or disconnect

I looked at a random smd reed datasheet, max 1ms open/close time I seen some RF reeds say 0.1ms

-Lasse

Vote
Twitter Facebook LinkedIn
C

I suggest two diodes in series, in place of each of your existing diodes. By itself this doesn't help much with the leakage, but if you then get an op-amp voltage follower with a gain of 1, and drive the junction between the two diodes to the same voltage as the low-leakage node that you are trying to clamp, then the diode leakage is much reduced, as provided they are in the dark, diodes leak very little when there is no voltage across them.

e.g. +-----|\ | | >---+ | +-|/ | | | | | +-------+ | | INPUT--*--*-|>|--*-*--|>|--Vpos | | +--|

Vote
Twitter Facebook LinkedIn
W

So, you provide a driven shield, protect the shield with leaky diodes (but it's driven, so no big deal), and clamp the high-Z node to the shield, with diodes that (in normal operation) never see over a few millivolts. Excellent! As I read the original problem, though, there was high current (i.e. low impedance) involved. The second part of the problem, is to keep the high (allowed!) normal currents while suppressing, somehow, overrhigh (fault) currents. Fuses, polyfuses, and relays are capable of this, but are all kinda slow. Maybe use a choke and a transzorb, to blow the fuse before the sensitive gizmo (on the other side of the choke) gags?

Vote
Twitter Facebook LinkedIn

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required