JFET’s J113, 2SK170, 2N5457 and others

Hello,

I have been playing with the software and hardware. Here are some of the plots of the JFET(s) I have been looking at.

First 2SK170. It is interesting that the RIP old K170 get all the talk, people are looking for a substitute.

In future post I will take a look distortion and noise; you know things that would make a JFET less attractive. For this look at distortion I am thinking of looking at the Common Source Amplifier.

See the attached K170 plots;
Vg ID transfer curve with gm curve with IDSS Marker
Vg ID transfer curve with gm curve with CutOff Marker
Vg ID transfer curve with gm curve with Marker and tangent line AKA gm.
Vd vs. Drain current with 0.1V Gate Volt steps.

Next time similar curves for J113.

Thanks DT
 

Attachments

  • 2SK170 IDSS.png
    2SK170 IDSS.png
    53.3 KB · Views: 2,325
  • 2SK170 CutOff.png
    2SK170 CutOff.png
    53.9 KB · Views: 2,231
  • gm is Graident.png
    gm is Graident.png
    65.7 KB · Views: 2,204
  • Id Vd Plots.png
    Id Vd Plots.png
    53.2 KB · Views: 2,219
See the attached J113 plots

Hello,

Still playing with the software and hardware.

This time J113.

See the attached J113 plots;

The IDSS could be matched to be close to the K170.

The transconductance and CutOff are nowhere near.

Thinking Yup, the J113 is not a drop in replacement for the K170.

Thanks DT
 

Attachments

  • J113 GV steps.png
    J113 GV steps.png
    58.5 KB · Views: 464
  • J113 CutOff.png
    J113 CutOff.png
    59.5 KB · Views: 472
  • J113 IDSS.png
    J113 IDSS.png
    57.5 KB · Views: 2,077
Member
Joined 2010
Paid Member
There's little if anything worth considering any more as a replacement for 2SK170 among the standard industry JFETs. In some applications, 2N/PN4091 or 4391 can be adapted but genuine parts aren't available in TO92 or TO18 sizes any more so you're still relying on over-priced remnants of remaining stock, high-priced copies from Linear Systems or fake NOS parts.

If you're serious, tiny adaptor PCB adaptors for SOT23 to leaded format are available at low cost. At least some of the useful JFETs are still available in SMD format from reputable suppliers and this allows us to continue with standard types for the time being.
 
Okay.

I am sure that there is a NOS stash on the shelf at your house.

I want to look at the tools to optimize the JFET’s at hand and also stashed away.

In my bin there is a stash of Toshiba 2SK246’s.

Take a look at the attached Toshiba 2SK246 plots.

Thanks DT
 

Attachments

  • 2SK246 CutOff.png
    2SK246 CutOff.png
    51.2 KB · Views: 430
  • 2SK246 IDSS.png
    2SK246 IDSS.png
    51.1 KB · Views: 356
  • 2SK246 Steps.png
    2SK246 Steps.png
    59.2 KB · Views: 410
What's wrong with SMD ?
If anything, they make PCB design much simpler, and will allow cost-effective automated manufacture.
There is also no issue using them in DIY.
Only when you require power devices or high voltage that through hole still makes sense.

Here is a good example :
The Pioneer Super Linear Circuit

And as mentioned before, 2x 2SK209GR in parallel will be a equally worthy replacement for 2SK170BL.
NJFETs for Source Follower Applications
Some people suggested 2SK2145, but they are not always well matched.
~10% in a large batch will lie way outside the 3% Idss tolerance used by Toshiba for their matched devices.


Cheers,
Patrick
 
The little buggers (SMU's) try to run away when they see me coming

What's wrong with SMD ?
...
There is also no issue using them in DIY.
...

...

And as mentioned before, 2x 2SK209GR in parallel will be a equally worthy replacement for 2SK170BL.
...


Cheers,
Patrick

Hello,

What qualities of the 2SK170 are we reaching for? Distortion? Noise? Input impedance? Or a drop in replacement?

The little buggers (SMU's) try to run away when they see me coming after them with a pair of pointy tweezers and a hot soldering tool.

When I do catch em I do have a cute little, clamp em in, test socket.

Do you have any samples to test?

Thanks DT
 
Hello,

Today I received a package of 2SK117GR's.

See the attached curves.

Note there are excel data files for each of the curves posted. I do not know how to attach them or DIYaudio does not like the data format?

Thanks DT
 

Attachments

  • 2SK117 Steps.png
    2SK117 Steps.png
    52.8 KB · Views: 1,041
  • 2SK117 IDSS.png
    2SK117 IDSS.png
    53.6 KB · Views: 1,025
  • 2SK117 CutOff.png
    2SK117 CutOff.png
    56.6 KB · Views: 1,082
DT, what equipment did you use to collect this data?

Jan

Jan,

The instrument is a Keysight B2912A, I purchased it used from Keysight on Ebay. The software is also Keysight, EasyEXPERT. The instrument arrived with a 30 day warrantee. When I received the instrument I took it to the Keysight service facility near me in northern California for calibration. So a new to me instrument, a fresh calibration certificate and the warrantee extended to 90 days.

https://literature.cdn.keysight.com/litweb/pdf/5990-7009EN.pdf?id=2035016

This instrument is also a precision dual power supply, up to +- 210VDC. Next step is to use it to test push/pull JFET amplifiers for distortion and noise on the audio analyzer.

The next little project is a JFET test bread board with clamp in sockets and DIP switching to select exact source, drain and load resistors. The parts were ordered yesterday.

I saw Nelson Pass post a photo of something similar he uses to match JFET’s and select precise resistor values to dial in resistors to minimize JFET distortion.

Thanks DT
 

Attachments

  • Pass Test Board.jpg
    Pass Test Board.jpg
    959.3 KB · Views: 477
  • 2N5457 Idss .jpg
    2N5457 Idss .jpg
    712.2 KB · Views: 416
Check out the contact resistors of DIP switches over lifetime.


Patrick

Thank you Patrick,

I did look at the life cycles of the selected switch, the datasheet says 10000 cycles. I expect that the JFET breadboard will be an iterative process. If I wear it out that will be an indicator. This is the selected switch.

https://www.mouser.com/datasheet/2/60/bd-1382595.pdf


The switches and resistors will be decade boxish; 1’s, 10’s, 100’s … so that you can switch in 482R or any other desired value. Whatever value gets you to the sweet spot.

I picked resistor values near the decade values, but not exact. Samples; 1% 10.0, 20.0, 30.0, 40.2, 49.9, 60.4, 69.8, 80.6, and 90.9, close but no awards.

MF1/4DCT52R60R4F KOA Speer | Mouser

The plan is to select resistor values under test to maximize performance then switch the “decade box” out of the loop and measure the overall resistance including DIP switches with the bench digital multimeter.

This test point 5006 Keystone Electronics | Mouser is used to attach the multimeter minigrabbers. Maybe even calculate the resistance from current value provided by the power supply and the voltage drop across the “decade box”.

Thanks DT
 
Last edited:
PF5102

Hello,

PF5102’s.

They are what looks like a different IDSS range of the J111 family of N JFET’s.

ON has notified End of Life for the PF5102 JFET.

See the plots.

PDF’s of the data this time.

Thanks DT

I received the switches and resistors to begin fabing the little decade PCB's.

Thanks DT
 

Attachments

  • PF5102 CutOff.png
    PF5102 CutOff.png
    43.2 KB · Views: 408
  • PF5102 IDSS.png
    PF5102 IDSS.png
    42.9 KB · Views: 210
  • PF5102 Steps.png
    PF5102 Steps.png
    50 KB · Views: 219
  • PF5102 Transfer Data.pdf
    258.3 KB · Views: 109