beta at low currents

Looks like an 'opportunity' for a new product line to sell, after figuring out how to make parts for another.

Jon

Cool. How much current? Off leakage should be low.

How do you know?

John

We were in the 2-3 picoamp region. This was not servoed and just hand regulated for a proof of principal. It came about from another twisted use of a neon that needed a controlled emission spot on the electrode at a very low current. I had no way to see the popcorn noise , that might be a issue.

As for the sensor it was used in, I can't comment. Hint, the Keithley electrometer was in "Q" mode.

Just about any old phototube might do this, if you can find one. Burle is still making glass.

Wenzel has calculated the off resistance here:

Steve

Crazy, man! DigiKey's much cheaper:

Cheers, James Arthur

I've ordered some of the gigohm Vishay axial parts (to make a picoammeter) for about $2.75 or so. I was just impressed with the Ohmite prices. I wonder if they sell any.

John

There's probably a miltary spec electrometer somewhere that uses them and government contracts keep them alive.

We sell a modular widget where one of the modules is 90% power supplies and one small microcontroller board that accepts serial port commands and routes them around to other modules. On one particular contract, for each widget purchased the government also ordered not one but two spare power supply modules... our best guess is either (1) whoever's in charge of the project has had very serious reliability problems with power supply modules or (2) someone clever figured out that our power supply module can be used as a drop-in replacement for someone else's much-more-expensive power supply module (and just ignores the extra microcontroller board). Weird...

Given another decade these things will probably show up on eBay, since two date we've had exactly one power supply module fail in the field... so all those spares are sitting in a warehouse somewhere. Look for a 1U tall, black-anodized unit labeled "PCU" with two columns of 3 diagonally slanty LEDs by various rail voltage labels.

---Joel

?? Then Shoot Johnson quietly, and ditch the resistors??

On a sunny day (Wed, 22 Apr 2009 15:53:21 -0700 (PDT)) it happened snipped-for-privacy@uakron.edu wrote in :

Very nice, informative :-)

Cute.

I've used green LEDs shining on red LEDs to make adjustable current sources of a few picoamps, but all photocurrents have at least full shot noise. (Currents from photodiodes and vanilla phototubes [i.e. not PMTs] have exactly full shot noise.)

Cheers

Phil Hobbs

Why use a red LED and not just a Si photodiode? Do you get more compliance from the LED? I've a light source and photodiode to make

George Herold

That would be a cool component: two photoelectric plates, super-well insulated in vacuum, and two blue or UV led chips. It would be a near-perfect insulator that could be made to pump small currents in either direction. Three electrodes would be even better.

Like an old-fashioned phototube, but two cathodes:

--------( . )--------- | | |

I want one.

John

Johnson is immortal. I don't know if tohm resistors have shot noise or not; metal films don't. I asked IMS if their tohm resistors have shot noise, but they haven't replied.

John

One of my guys is testing various PIN photodiodes for dark current and such. We can get pd's with sub-pF capacitance; LEDs generally have a lot of capacitance.

John

Because I wanted nice repeatable picoamps, and Si PDs are so leaky by comparison. The very ordinary Chicago Miniature Lamp LEDs I used drew less than 100 fA for biases of -5 V to +0.5 V, and cost a few cents--far less than any PD I know about, even the Infineon ones.

The LEDs were being used as per-pixel switches for a low resolution pyroelectric camera, and the photocurrent supplied a bias charge on every readout cycle, so that the integrated charge was always positive regardless of the direction of the temperature change. Worked great--0.13 K NETD--and was very very cheap, about 2 orders of magnitude cheaper than the nearest competitor.

Didn't go anywhere, of course, because it was too cheap and too weird. I may resurrect it now that I'm out on my own.

Cheers

Phil Hobbs

ou

ut

g a

ic

like

en

Resistors don't have shot noise, or more accurately they have reduced shot noise. The shot noise is reduced by a factor given by the electron scattering length within the resistor material divided by the total length of the resistor. Each scattering can be thought of as an independent random event that transfers a fraction of the electron charge to the output of the resistor. This model is from Rolf Landauer and I have a reference if any are interested.

They may have 1/f noise.

George Herold

OK; thanks. :-)

I'm only a techie, and we never got into noise very much - that had been taken care of by the actual designers. In fact, the stuff I worked on in the USAF used noise to modulate the jamming transmitters, and the question was about the "quality" of the noise.

Cheers! Rich

True, though there are some really small ones. You can't make a sub-Poissonian current from any photodetector whatsoever, unfortunately.

There are high bandgap PDs e.g. GaAs, that have low capacitance and femtoamp leakage, but they're not particularly easy to get.

Cheers

Phil Hobbs

I just nabbed this:

so we'll be testing a lot of parts for obscure parameters.

It has a 1e-14 amp range, and a 1e14 ohms range. Does volts and charge, too. And it has a needle!

John

And you have the nerve to tell me that analog scopes are so last-century. ;)

Cheers

Phil Hobbs (Off to test out my new HP 3325A with my Tek 475.)