lowest leakage, lowest Vf diode?

I'd guess impossible. You'd need a schottky for the forward drop, but they leak a lot. You might be able to fake it with a small mosfet, depending.

John

Peltier cooled schottky?

Trouble is with leakage - the importance of sub-microamp drain leakage in mosfets not being specifiable.

RL

snipped-for-privacy@wwc.com a écrit :

You'll probably won't find one, but depending on what you're trying to do you might come up with something to do the trick.

What is it for?

I've used some that were consistantly good. When what you want to do is impossible with guaranteed specs, you may have to settle for actuals. Many semi "max" parameters are set where they are just to speed up testing, so can often be finessed if one is careful.

Anybody know typical Idss for a 2N7002? It's spec'd at 1 uA, but I suspect it's a lot less. I guess I could test a few. Gate leakage is spec'd in nA, but it should be spec'd in electrons per second.

John

Do you not mean "not in the datasheet"? The nanoamp region is very achieveable. Since the OP wanted a low Vf diode, the problem with FETS is that the gate control is sensitive to temperature, so a fixed bias scheme for enhancement *or* depletion mode FET is a bit iffy - especially in this case where the gate control voltage is going to change only a few hundred millivolts. Now that temperature sensitivity can be compensated some with a scheme similar to my first Codatron (TM) design; the patent for that has been released into the public domain - so feel free to use as you see fit. Be advised that there can be some mis-matching between FETS, so that for this "low Vf" application, that can have a large effect on the total operation if multiple composite devices are needed (read: volume production greater than 1000 per day). That can be mitigated to a degree by making it as IC, wafer testing offset before breakout and use. Shoot, one could even trim one of the FETs my having graded sizes for cutting out small areas; and the bias offset can also be trimmed (refer to the ALD110800 series).

Agree on the leakage spec; many devices are *at worst* orders of magnitude better than datasheet spec.

Last week's problem (the one MassivePratt squirmed out of addressing) was the opposite: nearly 200 uA of current coming out of the input of a cmos TinyLogic chip while the input voltage was well within the rails. We were impressed.

John

OK, Mr. Smartyboots, how many amps is(are?) one electron per second?

Please show your work. ;-)

Thanks! Rich

John

The mosfet sounds about right, but it'd have to be connected backwards on account of the source-drain diode, as in Bob Pease's polarity protection trick. Maybe the OP could use a lithium battery to get the gate bias right. It ought to be possible to power a comparator off wherever that 100 mA is coming from.

Cheers,

Phil Hobbs

That's a good point- I was playing with a an, um, I think it was BS171, a few weeks ago, in the trudge through my parts bin for an analog switch (for which I eventually settled on a 2SK30 or something).

After discovering it was in fact a MOSFET, I had it set up on the breadboard as a source follower into a 4.7kohm resistor from +12V or so. I connected the gate to a wire in air. While playing, I discovered that a plastic bag, charged with a stroke over my hair, switched it from two feet away, with no apparent drift in the potential (that is, the distance at which it saturates). Not bad at all. Probably, the adhesive label on the bottom of the protoboard has more leakage than that thing.

Tim

-- "Librarians are hiding something." - Steven Colbert Website @

This makes an interesting shot noise then! One shot per second, exactly :)

Pere

It is possible to produce electrical currents that have little or no random shot noise, to literally dispense exactly one (or more!) electrons a second, rigidly periodically. SETs (single electron transistors) are one way to do this.

But I guess that one electron a second does have a lot of shot noise in a 1 Hz bandwidth.

John

What's it for indeed! Some special circuit configuration may be able to handle the problem. For example, two 1n5822 diodes in series will drop about 450mV at 100mA, and the connection node between the two diodes can be bypassed with a resistor to ground for the purpose of diverting most of the leakage current of the first diode with 13 volts across it. If the resistor is a low enough value the voltage drop across it from the first diode's leakage will be low enough to avoid creating much leakage across the second diode.

(Single Ended Triodes? ;-) )

The gain on those things must be horrific. One electron admitted for -- how much capacitance to switch it?

Tim

-- "Librarians are hiding something." - Steven Colbert Website @

A Radio Frequency Single Electron Transistor (RFSet) can have a bandwidth greater than 100MHz and extreme sensitivity. It is described as a fraction of the charge on an electron.

Here's one that runs at 700MHz with a sensitivity of 3.63 * 10-5 e/RootHz. They show how it is measured:

Schoelkopf has a bunch of papers on them:

Here's one that operates at 1.7GHz with a sensitivity of 1.2 * 10-5 e/Roothertz. Fig. 1 shows a SEM photo of the device.

Fig. 3 shows the time-domain response for a large (~5.5 electrons peak-to-peak) signal, 10 kHz triangle-wave applied to the gate. The SNR looks very good:

These things fascinate me. I want one. No clue what I'd do with it, but it would be nice to have laying around just in case I found a use for it:)

Regards,

Mike Monett