New Low Noise JFET from TI

[blackdog]

Hi,

Have you guys checked out the new JFEt from TI, it has pretty good specs.
I just think the 1/f noise is not good enough for reference measurements below 1Hz.

https://www.ti.com/product/JFE150

Kind regarts,
Bram

[Echo88]

Havent looked at it yet, but here a relevant infos that may be helpful for people interested in it:
https://www.diyaudio.com/forums/vendor-s-bazaar/375079-ultra-low-noise-jfets-texas-instruments.html
https://www.planetanalog.com/amplify-small-signals-in-low-noise-circuit-with-discrete-jfet/

[Kleinstein]

The low frequency noise still looks good, comparable to other low noise JFETs. Noise wise it looks close to the SK3557 - just with a higher maximum voltage and higher input capacitance.  I am not so sure why one would want a higher maximum voltage for a low noise amplifier though: for low noise and thus relatively high current, like 2-5 mA one does not want a high voltage, but more like 2-5 V to keep the heat low and avoid extra 1/f noise.  There is a general tendency to get more 1/f noise, when the DS voltage is higher than some 6 to 10 V.

Still good to have another source of low noise fets and new types.  The included diode and low gate current can be nice.

It gets interesting when they may offer pairs.

The LF noise may show popcorn noise. So the frequency domain description may be a little misleading and not follow the exact curve.
There is a chance that the LF performance does scatter between units too.

[David Hess]

It has better low frequency noise and gate leakage than an LSK170 but the high saturation current will make it more difficult to use in some circuits without compromising the noise.

The specifications and internal arrangement make me think that this is the integrated JFET used in TI's recent low noise JFET operational amplifiers like the OPA140 series.

[Kleinstein]

I am not that much  worried about the saturation current. The normal operation is at a much lower current, like 1 mA to 5 mA. So the relevant point to compare is more like the transconductance at a given current, like 1 mA.
For 2 mA the SK170 curve shows some 25 mS, the JEI150 gives 18 mS at 2 mA.  So one would expect a slightly lower noise for the SK170 in the medium frequency range.  The LF noise looks a little better, but the SK170 data-sheet is not very specific and the LF noise can scatter and may improve over time with process improvements. So later SK170 may be better than the DS.

The low gate leakage can be a factor for high impedance use.

[jbb]

Hey, TI just announced JFE2140 dual JFET.  Data sheet includes Vgs matching specs and curves but I’m not well placed to appreciate them.

[Kleinstein]

The VGS matching specs ( 1mV typ and 4 mV max.) are very good. For the older types in metal they usually have grades with 5 / 10 / 20 mV limits.

Good VGs matching means one can get away with little or no extra source resistor in a differential amplifier.

[David Hess]

The very best other dual JFETs (LS843)  are 1mV and 5uV/C maximum so 2 or 4 times better precision depending on how you squint.  4mV and 10uV/C are still very good and better than most.  I like the price if they ever become available and the noise is outstanding for such a low gate leakage current.  They are only good to about 120 MHz so slow but that is hardly unexpected.

If you have the time, matching discrete parts by hand and mounting them together can get into that range.  Dual monolithic transistors cannot be used in some applications because of coupling between the transistors on the same die.