Low noise N-JFET ?

We needed low noise in the sub-Hz region, too. The Interfet IF9030 I mentioned in another reply is a full mm square device, so that brings down the noise a lot. (Obviously their processing is also optimized to lower the noise.)

Jon

Because the news is very bad down there, they don't WANT you to know. Try to find op-amp specs for very low F noise, it is practically absent. One op amp with really GOOD LF noise, and tey SPEC it too in the data sheets, it the AD706. It got us out of a bind where we had a LOT of extremely low-F noise. Our circuit would sit there runnign perfectly stable for minutes, then suddenly it would jump millivolts. The AD706 was the only thing in the entire book that had a guaranteed spec for .001 - 1 Hz. The spec wasn't that astonishingly low, but the fact they were willing to guarantee the 1/f curve out to .001 Hz was big news.

Jon

Le Tue, 06 Nov 2007 16:57:33 -0600, Jon Elson a écrit:

Noise wise the AD706 is an ordinary one and many others outperform them by far. Their virtue is more in the low bias current over temperature range and I wouldn't use them for anything else.

This one is also quite good even though the graph ends at 10Hz. But I never had to go lower than 5Hz or so:

That is surprising, although 17nV/rtHz at 10Hz is indeed not much to write home about. But it does have the knee nice and low.

I was planning to use the AD797 or LT1028.

Pieter

Hi,

The website is still there, and yes, the specs are exceptional. Only the capacitance is high, that that does make sens for a FET optimized for noise.

Pieter

Specs need to be tested. A noise spec at 0.001 Hz needs a at least 1000 seconds to measure. This will make the test very expensive, as the tester can only test 86 chips per day.