I found some "Toshiba" 2SK389 SOIC packaged dual transistors on Ebay. They do NOT have the same pin-out as the Linear Systems LSK389.
They are marked K389BL OF
I soldered them onto some DIP adapters to test them. (I checked the adapters for correct pin translations)
The pin-out appears (and works) as:
pin2 S1 or D1
pin 3 gate 1
pin 4 D1 or S1
pin 5 S2 or D2
pin 6 gate 2
pin 7 D2 or S2
No apparent connections to pin 1 or pin 8.
The FETs trace on the curve tracer just like a 2SK170 and are well matched for gm. However the current offset ( Idss ? ) can be mis-matched by up to 25mV gate V.
(half a step division offset with 15 traces and 50 mV steps, so 25mV mis-match) One device did trace perfectly matched.
I would think that adjacent dies would be better matched on Idss. Could these be matched 2SK170 dies in an SOIC package?
Looks like these SOIC parts have surfaced before:
https://www.diyaudio.com/community/threads/help-2sk389-soic-packaged.100856/
but no resolution. No Toshiba SOIC 2SK389 apparently made.
Any way to tell which pin is Source versus Drain. They (and 2SK170) trace exactly the same way by swapping S and D around. Does it matter?
And a small detail: The SOIC to DIP adapters have a foil plane on the back side below the chip, and a big plated thru hole to a smaller foil plane just below the chip package.
I assume this is intended for some heatsinkng effect. To get good thermal contact with the chip package should this plate thru hole be solder filled or maybe just some glue fill to avoid over heating the chip. Since these surface mount things get wave soldered, it would seem that momentary solder heating should be tolerable.
I was planning to try out the 2SK389 for a Mimesis type board in place of the 2N5565 that Goldmund used. Also planning to try an Analog Devices MAT02. Those MAT02 devices DO match perfectly on the curve tracer. (TEK576) I even adjusted the beam focus to try and see any difference. Perfect match.
They are marked K389BL OF
I soldered them onto some DIP adapters to test them. (I checked the adapters for correct pin translations)
The pin-out appears (and works) as:
pin2 S1 or D1
pin 3 gate 1
pin 4 D1 or S1
pin 5 S2 or D2
pin 6 gate 2
pin 7 D2 or S2
No apparent connections to pin 1 or pin 8.
The FETs trace on the curve tracer just like a 2SK170 and are well matched for gm. However the current offset ( Idss ? ) can be mis-matched by up to 25mV gate V.
(half a step division offset with 15 traces and 50 mV steps, so 25mV mis-match) One device did trace perfectly matched.
I would think that adjacent dies would be better matched on Idss. Could these be matched 2SK170 dies in an SOIC package?
Looks like these SOIC parts have surfaced before:
https://www.diyaudio.com/community/threads/help-2sk389-soic-packaged.100856/
but no resolution. No Toshiba SOIC 2SK389 apparently made.
Any way to tell which pin is Source versus Drain. They (and 2SK170) trace exactly the same way by swapping S and D around. Does it matter?
And a small detail: The SOIC to DIP adapters have a foil plane on the back side below the chip, and a big plated thru hole to a smaller foil plane just below the chip package.
I assume this is intended for some heatsinkng effect. To get good thermal contact with the chip package should this plate thru hole be solder filled or maybe just some glue fill to avoid over heating the chip. Since these surface mount things get wave soldered, it would seem that momentary solder heating should be tolerable.
I was planning to try out the 2SK389 for a Mimesis type board in place of the 2N5565 that Goldmund used. Also planning to try an Analog Devices MAT02. Those MAT02 devices DO match perfectly on the curve tracer. (TEK576) I even adjusted the beam focus to try and see any difference. Perfect match.
Last edited:
The parts are symmetric so you can swap D for S. I recall matching of originals was .5 ma idss, so .25v max Vgs is realistic.
There is no records of Toshiba making this smd chip 2SK389.
Even if they ever did, shouldn't they have used the different part# for the different looking part,
just like they did for the 2SK117 dual (5 pins)?
Of course LS did use the same part# for the through-hole and smd versions of LSK389, but the
smd pinout is different from the "mysterious" part.
Yours:
pin 2 S1
pin 3 G1
pin 4 D1
pin 5 D2
pin 6 G2
pin 7 S2
The pics attached demonstrate that the boards were made to produce the same ZIP pinout as the
original 2SK389.
If your chip has the transconductance of 2SK170, that is higher numbers than the 2SK389 jfets
should have...
The pinout is also different from the new JFE2140 dual...
The first guess will be for chinese production...
Even if they ever did, shouldn't they have used the different part# for the different looking part,
just like they did for the 2SK117 dual (5 pins)?
Of course LS did use the same part# for the through-hole and smd versions of LSK389, but the
smd pinout is different from the "mysterious" part.
Yours:
pin 2 S1
pin 3 G1
pin 4 D1
pin 5 D2
pin 6 G2
pin 7 S2
The pics attached demonstrate that the boards were made to produce the same ZIP pinout as the
original 2SK389.
If your chip has the transconductance of 2SK170, that is higher numbers than the 2SK389 jfets
should have...
The pinout is also different from the new JFE2140 dual...
The first guess will be for chinese production...
Attachments
Wow, you hit the Rosetta stone there. Clearly the same part on that board as I have here. I wonder who makes the part in China. Probably less than a $1.00 from the manufacturer. I paid $9.50 for it on Ebay. "Original Toshiba" $$$, yeah right. Ha!
That JFE2140 is even better at half the price. Will have to get some of those.
I see MAT02 parts online for around $7, but many come from China, so I wonder it they are real and up to spec. I have a whole sleeve of MAT02 and MAT03 from Analog Devices from 15 years ago through a major distributor. (were cheap then) I put some into Randy Slone's symmetric Optimos amplifiers and they worked great. No undefined VAS current with those. And I put them into my old 1970s symmetric Tiger 01 (SWTP 207B) Amps by Daniel Meyers, along with 2SK1058 and 2SJ162 outputs, Still working good.
That JFE2140 is even better at half the price. Will have to get some of those.
I see MAT02 parts online for around $7, but many come from China, so I wonder it they are real and up to spec. I have a whole sleeve of MAT02 and MAT03 from Analog Devices from 15 years ago through a major distributor. (were cheap then) I put some into Randy Slone's symmetric Optimos amplifiers and they worked great. No undefined VAS current with those. And I put them into my old 1970s symmetric Tiger 01 (SWTP 207B) Amps by Daniel Meyers, along with 2SK1058 and 2SJ162 outputs, Still working good.
Last edited: