Hi All, Ive been using the Linear Systems LSK 389A dual JFET for a while, it works brillinatly in my low noise audio amp.
Unfortunately I have just been informed that they are having problems fabricating all but the LSK389C version which will draw much more current in my design.
Does anyone know of any other low noise dual jfets that are comparable?
Unfortunately I have just been informed that they are having problems fabricating all but the LSK389C version which will draw much more current in my design.
Does anyone know of any other low noise dual jfets that are comparable?
2sk389, but you will have to pay to buy some after a long search for the little available stock.
If I wait long enough, my small stock might make a millionaire of me
If I wait long enough, my small stock might make a millionaire of me
Somewhat lower transconductance, and common source. If you can live with that - and actually find a seller, Toshiba 2SK2145 might fit your bill: MOSFETs Products - Toshiba America Electronic Components, Inc.
2x Toshiba 2SK170 with matched Idss and thermally coupled.
http://www.diyaudio.com/forums/solid-state/135154-next-best-thing-2sk389-2sj109.html
Patrick
http://www.diyaudio.com/forums/solid-state/135154-next-best-thing-2sk389-2sj109.html
Patrick
Mouser has 2SK170GR in stock -- these have Idss of 2.6 to 6.5mA -- if you purchase about 20 you should get several matches +/- 0.2mA. The distribution of parts isn't "normal" and seems to "mode" around 4 to 5 mA. What you don't use as matched can be used for CCS. Let me know if you have a problem.
Thanks for the suggestions, Unfortunately I cannot afford the time to perform such a select on test procedure (to match IDSS) though with over 100 audio channels. The Toshiba part would be usefull if the two source terminals were seperate, but as they are, it would require a major redesign to use it.
Does nobody make discreet dual JFET's anymore?
Does nobody make discreet dual JFET's anymore?
IMO, you're lucky to find discrete anything anymore. A few years ago we were warned that discrete parts, especially leaded ones, were simply going away. It hasn't happened, but I find the specialty parts are getting tough to find. Also, the actual stock on the shelves at places like Digikey seems to be less than it used to be.
You may well have to make the time to match some parts.
You may well have to make the time to match some parts.
What's a good price for the 2SK389? I know a parts supplier that has some stock of these. Maybe I can join that millionaires club with you.
Found this source for the LSK 389A dual JFET: N-Channel: SMSLSK389A LOW NOISE MONOLITHIC DUAL N-CHANNEL JFET same as Linear Integrated Systems LSK389A, Linear Integrated Systems LSK389A TO71 manufactured by Semiconix Semiconductor - Gold chip technology for known good N-Channel die, N-Channel fl
However they have a minimum order of 550 USD.
However they have a minimum order of 550 USD.
Try approaching Linear directly.
They have supplied some members with low quantities in the past.
We organised a group buy some years ago where we approached 10k pieces, but mixed over their range of the sk170abc & sk389abc.
They have supplied some members with low quantities in the past.
We organised a group buy some years ago where we approached 10k pieces, but mixed over their range of the sk170abc & sk389abc.
I made a MC pre-amp with a 2SK389 dual JFET. Rsource=47 ohm; Rdrain=3k3 (while Rg=82, Rin=68ohm), so classical Le Pacific.
One channel takes more current (5 mA noisy one against 3.8 normal one) while amplification is less (-3dB). I measure a high gate current.
And it is noisy. The bandwidth is more than double (450 kHz against 180 kHz).
I tested with input 10 mV into 1000ohm against the two inputs which // 34 ohms.Output good channel = same level, = what I wanted for SPU
One channel takes more current (5 mA noisy one against 3.8 normal one) while amplification is less (-3dB). I measure a high gate current.
And it is noisy. The bandwidth is more than double (450 kHz against 180 kHz).
- Just bad luck?
- Or did I do something wrong - to much heated when soldering for instance --> just to learn something
- or is it oscillating ... 😱
I tested with input 10 mV into 1000ohm against the two inputs which // 34 ohms.Output good channel = same level, = what I wanted for SPU
Counterfeit perhaps? Soldering doesn't normally damage silicon devices if you iron is temperature controlled.
For MC you need low voltage noise, not low current noise, even the phenomenal 2SK170 JFET is struggling at 10 ohms source impedance (2SK389 is worse), low noise bipolars like the ZTX851 have considerably less voltage noise.
[ 0.25nV/√Hz is reported - Ultra low noise amplifiers ]
For MC you need low voltage noise, not low current noise, even the phenomenal 2SK170 JFET is struggling at 10 ohms source impedance (2SK389 is worse), low noise bipolars like the ZTX851 have considerably less voltage noise.
[ 0.25nV/√Hz is reported - Ultra low noise amplifiers ]
I bought them from a reliable hi-end manufacturer ( as surplus stock) but as I should have expected, that stock included rejected stock. This part could be relabeled. What is not?
I had 2SK240; but they are two discretes in a common alu-case. One half started to drift from 3 to 6 mA . . and having it on the cathode of a tube, the whole DC amp went beserk.
I bought four 2SK389, and have one over, so I'll replace.
Indeed, I have not seen components fail at large from soldering; I often use LMP solder at 305 degrees.
As regards gate current,
I had 2SK240; but they are two discretes in a common alu-case. One half started to drift from 3 to 6 mA . . and having it on the cathode of a tube, the whole DC amp went beserk.
I bought four 2SK389, and have one over, so I'll replace.
Indeed, I have not seen components fail at large from soldering; I often use LMP solder at 305 degrees.
As regards gate current,
- I have two pars of 2SK170/2SJ74 from the very reliable Al Weit, they are originals, perfectly matched; but with a high gate impedance above 200K - both pairs starts to drift to +2 mV common source output, so the 2SK170 from those pairs leak just a tiny bit. It is also sensitive to the voltage, increasing at higher voltage.
- very old 2SK30 AGR has a negative temperature coefficient, new relabeled 2SK30s have positive tempco, fake. .
Last edited:
LSK389, LSK489 are available from NacSemi (Linear Systems is a big supporter of the DIY community).
I have some NSVJ5908 which have a 1dB NF but the two devices have somewhat dissimilar Idss.
I have some NSVJ5908 which have a 1dB NF but the two devices have somewhat dissimilar Idss.