Base-collector junction as analog input stage protection

Try comparing guaranteed leakage currents. You can buy low leakage diodes, but they tend to be more expensive than jelly-bean transistors

-- Bill Sloman, Nijmegen

Yeah, much lower leakage currents. This is a nice article,

George H. (If the link doesn't work, then search for EDN +bounding and clamping)

The c-b junction of a BFT25 leaks roughly 20 fA at room temp. It's tricky to measure.

John

You probably want to do the EB short. Floating anything is a bad idea, and high fields lead to junction damage which will surely leak.

Generally when you ESD test a chip, you keep zapping until you see junction damage. This sets a "known quantity" on how much the device can take. [It isn't exactly kosher since you do multiple tests on the same pin and often go test another pin after zapping an adjacent pin.] However, you also zap a rail at your factory spec and run them through ATE, just to insure nothing bad happened internally.

Hi George,

Many thanks.

It is exactly this article I search for.

Regards,

Thoma

What is far better, is to measure the NF before and after ONE test. Betcha the results will raise your eyebrows and swear off multiple testing like that..

I found something like a few pA for a 2n3904. But the opamp leakage was also ~ 10pA so that was about my limit.

George H.

Hmm, I left the emitter floating. (That's the way Bob P. drew it.) Does shorting EB change anything?

I tested the ESD protection on our diode laser by first scuffing up a big charge and touching things.

I then got this piezo sparker from a butane lighter and zapped it with that. (both polarities.) Watching the laser diode output you could see little spikes in the light if any charged leaked in.

ESD damage also tended to 'wonk' the wavelength a little.

George H.