I_sub_dss zero gate drain current

Place it out on the midday hot pavement.

He'd have to shovel the snow to get to the pavement.

Cheers

Phil Hobbs

It's in the 70's here in LA. Beautiful! I've found where to hang out. The yard house. Must be at least 100 beers on tap. I'dpst

Have you ever looked at fet leakage?

Oops I'm still learning how to fly my tablet.

So when turned off does a fet look like a resistor? Or something else?

George H.

Above 0V, the drain current increases by factors of 10x for every 178mV increase in gate voltage, see the plots and discussion in AoE. When it reaches about 1mA, then it transitions to start acting more like classic MOSFETs are claimed to act.

I just measured a 2N7000 at about 180 pA at 20 volts on the drain, pretty flat with voltage.

Somebody, maybe Win, sent me this:

I do wish people would put their names and dates on things.

Thanks Win, So the max spec is 10^6 greater than the typical?

I'll have to measure.... that's always best.

George H.

Great,. Anything less than 10 nA should be fine. According to the anonymous graph making Vgs negative isn't going to help.

George H.

With 2N7002E's there isn't any. JFETs leak more at higher temperatures, but the effect is weaker at low V_DS, which is where you want to work anyway, so I generally don't worry about it.

Cheers

Phil Hobbs

George H.

The E version (now defunct, but I have a reel) is the one without gate protection. Their leakage is pretty hard to measure.

Cheers

Phil Hobbs

As some of us have mentioned many times, specs like that are based on the measurement limitations of ATE, semiconductor mass-production automatic test equipment. If they were to setup the ATE to measure lower maximum currents, they'd have to dramatically slow down the measurement time for each part. This assumes that their equipment allowed the choice. We know that there are very few power MOSFETs customers who would care. Generally it's not done. We do our own incoming inspection, or we will rely on common device physics.

NSC offers fA leakage-current op-amps, and Bob Pease made special test equipment, so they could test to those specs. See the AoE tables. Parts were accordingly priced a little more. LTC has done the same. So it certainly can be done, in production, if there's a good enough reason.

Yes, that was my graph. I was sorry to see I didn't label the two plots, or show Vds. I'm not sure about negative Vgs, but it probably doesn't help. My measurements were at lab temp.

Note, pA currents are 10^4 lower than 10nA.

The HP 4140A instruments do well to below 10^-12 A. Pretty cheap on eBay, or I'd be happy to send you one.

)The HP 4140A instruments do well to below 10^-12 A.

Thanks, Win, that's very kind of you. I actually have a few instruments tha t can do that. The best is probably a box I made long ago with some Russian 100G glass resistors and an LM660AN CMOS amp in PDIP, but the one I reach for is a Keithley 610C, which has a 10 fA FS range--completely useless, but still.

My sentimental fave is a Keithley 410, which takes a couple of hours' warmu p before it's 100 fA FS scale is useful, but then it's from 1963 (when I wa s 3) and uses a tube.

Then there are a few many-digit DMMS whose lowest voltage ranges make prett y decent picoammeters.

Cheers

Phil Hobbs

The 4104A lets you measure up to a +/-100V bias voltage, and it lets you do scans.

OK thanks, I'm thinking it's better if I fly with gate protection. :^)

There's a 2n7002e from vishay. Maybe different?

George H.

Same die.

Hi George,

I took a new-old-stock 2N7000 TO-92 made by ST circa 2006 from a fresh unopened old antistatic bag and shorted S-G and put in oven. Here is a photo (the blue wire is a thermocouple alligator clipped to hold it close to the DUT):

Current with Vgs=0 vs temp vs Vds was:

Room temp