I've seen a reply where someone said it's okay to replace a 2SK107 with 2SK389 (same for 2x 2SJ74 in place of 2SJ109) but I want to confirm.
Also why is there a need to glue them? I saw this photo: http://www.diyaudio.com/forums/attachment.php?s=&postid=1133174&stamp=1171282640
The transistors are facing each other. But since I want to replace a 2SK389 with 2SK170, I cannot have them facing together unless I bend the pins to make it work?
thank you for the help
Also why is there a need to glue them? I saw this photo: http://www.diyaudio.com/forums/attachment.php?s=&postid=1133174&stamp=1171282640
The transistors are facing each other. But since I want to replace a 2SK389 with 2SK170, I cannot have them facing together unless I bend the pins to make it work?
thank you for the help
yes, it's okay to use the single versions.
The devices are glued together for thermal coupling, which improves the degree to which the can be matched (not the best choice of words, but you get my drift).
I'd recommend about 15 of each single version to find 1 closely matched pair, as in my experience these can vary widely even with the same grade.
The devices are glued together for thermal coupling, which improves the degree to which the can be matched (not the best choice of words, but you get my drift).
I'd recommend about 15 of each single version to find 1 closely matched pair, as in my experience these can vary widely even with the same grade.
C'mon! You need thermal coupling, otherwise your matched fets are going to drift thermally -> they're no longer tightly matched.
That's the reason the 2SK389 is a double package, tight thermal coupling and best matching since both fets are sitting next to each other on the wafer.
Don't worry about glueing them together. It would be nice todo if possible, but if you don't need the last bit of performance (then the 2SK389 would anyway the only choice) put them in and that's it.
Cheers, Hannes
That's the reason the 2SK389 is a double package, tight thermal coupling and best matching since both fets are sitting next to each other on the wafer.
Don't worry about glueing them together. It would be nice todo if possible, but if you don't need the last bit of performance (then the 2SK389 would anyway the only choice) put them in and that's it.
Cheers, Hannes
The moment i see an auction of Chateau d'Yquem bottles with a single digit vintage label of 7 i'll let you know.
Who knows, you might hit the '67 jackpot.
Just thought it amusing, someone selling this stuff at such a ridiculous rate isn't likely to call them NOS parts.
(Japanese manufacturers, all except NEC, use a year and month manufacturing code, goes back to the seventees. Januari is A, Februari is B, E translates to May. The J109 was not manufactured in 1987 and not anymore in 2007, so it's a 1997 part)
Here's a nicer vintage : 1L, December 19-?-1
Who knows, you might hit the '67 jackpot.
Just thought it amusing, someone selling this stuff at such a ridiculous rate isn't likely to call them NOS parts.
(Japanese manufacturers, all except NEC, use a year and month manufacturing code, goes back to the seventees. Januari is A, Februari is B, E translates to May. The J109 was not manufactured in 1987 and not anymore in 2007, so it's a 1997 part)
Here's a nicer vintage : 1L, December 19-?-1
Attachments
This isn't true and you need to look at more 1970s and 1980s components.
Manufacturers work production in weeks, not months.
eg: 79R corresponds to 1979, 18th fortnight. (Sanken etc)
8822 corresponds to the 22nd week of 1988 (Toshiba on ICs)
Most of the Japanese (ICs, caps, transistors etc) use a single (or two) digit year and the letter (A-Z) corresponds to the fortnight (1-26) of manufacture or if in two digits, the week (up to 52).
Last edited:
I think you are talking about the 2SK240/2SJ75 which if I remember correctly are 2 matched 2SK170/2SJ74 inside a metal can.
No. One 2SK389 = two matched 2SK170.
I still don't understand why, on the 389, the gate of one side reacts - multimeter in diode test - with the drain/source of opposite side....
Read the first post of this thread here and look at the pictureshere and you'll see what I'm talking about.
When testing FETs, discharge all residual charges that was left behind by the last measurement by sticking all the pins on a conductive foam (the black sponge where ESD devices are placed). Remember you are applying voltage across the pins and when you measure the next pair, the device will store some of that charge and screw up your next measurements. Also test FETs with the DMM in resistance mode, not diode. Wakarimasu?
Yes, you are right, the 2SK240 is exactly two 2SK170 units in a metal can.
But, at least as I read everywhere (even in this thread), the 2SK389 is a dual version of 2SK170, with matching advantages over two 2SK170 matched units, as the 2SK240, isn't? A half 2SK389 in place of one 2SK170 will sound exactly the same or will be different?
Last edited:
Remember the 2SK389 is a 50V part while the 2SK170 is 40V. They are different devices that share the same DNA hence some common traits.
The question of both devices sounding exactly the same or different lies on the listener and that is a purely subjective matter. It's the type of discussion I don't want to be part of.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.