Zener reverse leakage current question

xanatos

I have a 5.1 volt 1W Zener. When placed across the power rails of a supply delivering 4.98Vdc, I measure a whopping 38mA of reverse leakage current. Is this normal??? I tried it with two others, nearly identical. I know there is a knee on these things, but that's a lot of leakage for a Zener rated 0.12 volts higher than the supply voltage... I thought... On my datasheet, it lists Izt(mA) as 49mA - I thought that was what the diode could handle when in full-on Zener Mode at a voltage running over Vz of 5.1 (it's a 1N4733A). I guess I never thought that was what it drained just sitting there. Am I doing something wrong.

FYI, I'm not using this to regulate the incoming voltage, this is to shunt any voltage spikes that come through from my clamping diodes to the power rail, to ground, and those are current limited to 110mA max. My upstream regulation is very robust, these Zeners just keep the boards themselves in line.

Thanks for any thoughts/confirmations/advice.

Dave

Phil Pilgrim (PhiPi)

According to the Fairchild datasheet for their 1N4733A (5.1V nom.) zener, the zener voltage can range anywhere from 4.845V to 5.355V. So what you're seeing seems to be within spec. You could plot a current vs. voltage curve and see for yourself where the knee is.

-Phil

xanatos

Thanks Phil. I've got 200 of these, so I'll just go through the lot and select out the ones with the lowest leakage... :-) It just surprised me. Also - my datasheet sucks (NS version). I downloaded the Fairchild version - MUCH better quality data.

Thanks,

Dave

Tracy Allen

Dave, a sharper knee with lower leakage is possible by combining the zener with a transistor. A PNP transistor base connects to a zener biased ~0.3V below the Vdd supply. The emitter connects to the power supply output, and the PNP collector connects to ground or to a low value resistor that will absorb the transient. The transistor turns on only if and when the fault exceeds about 0.3V above Vdd, so it protects the regulator and other circuits but does not normally draw current. It is possible to use a synthesized zener such as the LM385 to get an extremely sharp knee in combination with less than 20µA of quiescent current to bias the transistor.

jmg

xanatos wrote: »

On my datasheet, it lists Izt(mA) as 49mA

izt usually means Current(i) thru Zener at test.

A 1W zener is quite grunty, & 49mA is 250mW which seems a reasonable test point.

Zener knee operation is not square, and varies with Zener Spec voltage. ~ 6.8V is sharpest and lower voltage ones like 3V9 are frankly, pretty lousy. 5v1 is only moderate.

xanatos

I needed a Grunty Zener!

Tracey- that transistor idea is fantastic, thanks! Plus I can get an even gruntier transistor that can shunt some serious current! :-)

Dave