In various of his columns. He and his technician developed the test setups for these opamps. I imagine he was involved in their design as well. Didn't he have a few columns along the lines of, What's all this Teflon stuff?
What/where are BP's remarks about these ICs?
Thanks,
robert
...and had *no* triboelectric, accumulated surface charge orother teflon-related problems?
If I pull a random 2N7000 out of stock (smallish mosfet, TO-92 package), and connect a supply and LED in its source-drain path, I can tease the floating gate to barely turn on the led, then let it sit with the drain lead hanging. The led stays part-on for days, and the current must be in the 1e-17 amp range, ballpark 100 electrons/second.
John
I've done the same experiment, only with a 5.5-digit meter measuring the drain current. It's clear the plastic packages of these parts are essentially pefect when it comes to leakage.
Is that the one where test time was limited by cosmic rays passing through the air volume by the inputs? He said something about the ionization left by them created enough current to bugger the low input current test.
IIRC they had to make a little enclosure around the inputs to limit the volume of air to be ionized by the stray whatzit and after that they could reliably test the parts.
Robert
[snip]
You disconnect the drain "...drain lead hanging" supply?
Or you disconnect the Gate lead?
Robert
Gate of course.
John
Now *that* seems to be a good starting point for a rather sebsitive electrometer!
That also is the way i remember what ws said.
On 9 Oct 2005 05:35:51 -0700, Winfield Hill wrote in Msg.
Ah that. Yes, I remember. Thanks.
robert
Hello,
are you shure this is not caused by drain-gate leakage?
Daniel
Whew!
Thought I was going to have to relearn some electronics.
Robert
Could be, but two leakages cancelling each other out with such precision would be rather improbable.
robert
There was also at least one "What's all this femtoamp stuff?" IIRC
As matter of comment:- I believe Analog Device or Burr Brown make better device for picoamp meter, where I use T network in the transimpedance feedback circuit, using 10MOhm resistor (cheaper than single feedback using 1GOhm) but I have watch out for offset drift due to voltage drift. I was surprised Kiethley used NSC device. In the end I able to build reasonably accurate picoamp meter where performance are in par with significantly more expensive instrument.
I'm not aware that they do. What's your story on this point?
The T feedback-resistor setup, while cute up to a point, has a significant performance cost from the offset-voltage magnification issue. Keep in mind low-current MOSFET opamps are already worse than their more expensive JFET brothers. The "significantly more expensive instruments" typically have 0.01% performance, etc., even at a fraction of each full-scale range. This requires the use of high-value resistors, no short-cuts allowed. These resistors are not especially hard to get.
One way to get a nice set of resistors, plus an elegant unobtanium teflon-insulated switch is to purchase an old Keithley instrument on eBay. Nobody wants a vacuum-tube clunker, you can get 'em for $10,
Have something to add? Share your thoughts — no account required.
Ask the community — no account required