smurfer77
Super Member
I have a Sansui AU-X11 and was interested in subs for some of the unobtainium FETs. However, very little info was available so I decided to do some measurement to share with everyone here. I also ordered some substitutes from Linear Integrated Systems (LIS) who happen to be just across the bay from here in SF, and checked out how they shape up. LIS are the only folks I know of that are currently manufacturing (or rather, having manufactured for them), dual N-JFETs and complementary dual P-JFETS. They have a few neat single JFETs too and are expanding there line-up. I saw Paul Norton and some of the guys (Kevin handles my orders) at Burning Amp 2016 a couple of weeks ago. Here is a video of Paul describing their product line.
I found LIS partly because of a nice Sansui AU-919 thread where @noveltone mentioned the LSK389 as a possible sub for the 2SK129A. This was pointed out to me by Hyperion and some supporting info came from Ronito over in my AU-X11 thread. The LSK389 is readily available and under production by LIS.
However, I wasn't happy with existing info to just go for the LSK389. And there are variants of the LSK389... which one to buy? Others have mentioned various options for sticking two single JFETs together but I don't have a reliable line on any of those mentioned in various threads (including that AU-919 thread)... I wanted to stick to something in production too! Oh, did I mention the LSK389 are monolithic? (the 2sk129A are two singles stuffed into one package) So the matching could be really good!
Here are a couple of pics. The single unit is the 2SK163 (which is a single N-JFET in the Sansui AU-X11).
Note the pinouts. GPS16 pointed out in my AU-X11 thread that it might be worth double-checking as if the same single transistor packages are stuffed into the can, but flipped around, you might think D and S should be reversed on one side, compared to the other. However, they are interchangeable as we will see later (not that uncommon for small signal JFETs)
In that AU-919 thread @Stene showed a graph for the 2SK129. From that graph I_dss is 2mA (@V_ds = 15 V) and V_gs_off is -.3 to -.4V. I've pulled 4 working (yeah, mine are all working, I don't need new ones, I'm just curious about the substitute for now!) 2SK129A and the characteristics are different. And then over in that AU-919 thread, @noveltone kindly measure for us about 7 to 8 mA for the I_dss of the 2SK129A. It turns out noveltone's I_dss measurements on hits one device, agree well with what I measured on four devices.
2SK129A I_d vs V_ds curves for different V_gs
I don't trust cheap curve tracers anymore so the measurements were made point by point manually, using lab power supplies. This curves are for a single channel of a single device but all four of my 2SK129A are really similar. I_dss (V_ds = 10 V, V_gs = 0 V) was between 7 and 8 mA for all 8 channels of my four 2SK129A.
2SK129A I_d vs V_gs (V_ds = 10 V)
V_gs_off was about -0.75 for the single channel plotted below. I.e. V_p = 0.75 V. This was pretty similar for all 8 channels, but you know V_gs_off can be a bit tricky to determine precisely... but you get the idea.
2SK129A Drain & Source symmetry
So back to the question of the pinouts and whether D and S are interchangeable on the 2SK129A. Well it turns out that, like some other small signal JFETs, the structure is such that D and S are symmetric and interchangeable. Below I took a single 2SK129A and you can see what happens when I flip around D and S on both channels... not much. As we will see later, LSK389 is also interchangeable. You may also notice from the I_dss that I used a different 2SK129A from the one in the graphs above. Here I_dss is about 7.8mA at V_ds = 10 V. Gate-source voltage here is 0 V of course.
Substitutes
If you go to the LIS website you can grab a spec sheet for the LSC389. V_P os meant to be 0.2 to 2 V. The LSK389C is meant to have I_dss between 10 mA and 20 mA (@10V_ds), but I grabbed some anyway to check them out. I got the TO-71 metal can package.
What I found is that they are not nearly as well matched as in the 2SK129A. But, pretty good.
For V_ds from 1 to 30 V, I_d goes from 12.6mA - 16 mA... but note that the upper V_ds measurements were tricky because the can really heats up and there is significant drift. The temperature response was negative... actually i got 15.93 mA at 20V and 15.2 mA at 30V steady-state... if I take measurements really quickly before it heats up, I can get something else. Case temperature went from 25 C at 1V, to 32 C at 20 V and 50 C at 30 V.... well we are past the 400mW Pd so can't really complain. But yeah, I got the feeling that the dissipation for the C version of the LSK389 is really a bit insufficient.
I_dss for the LSK389C at 10 V was for the 4 samples:
(I_dss1, Idss2)
16.0mA, 15.09mA
16.2mA, 15.1mA?
12.84mA, 13.04mA
13.10mA, 13.25mA
which gives Idss1/Idss2 within spec (10% variation), but not as good as the 2SK129A that I have at hand. I have a "?" beside one of the I_dss values because it kept jumping around between two values, almost like it was bistable somehow... it was not my setup. I have more detailed data that could be used to make curves if anyone wants, since curves are not given in the spec sheet.
Now, the LSK389B though, is what we are really interested in. So on tho that next with some curves (which are not in the spec sheet). (sorry, I meant to finish this post including the single channel JFET data tonight, but I have an early start... so will resume this thread soon... but don't worry, it all works out and we have a nice sub for both!).
I found LIS partly because of a nice Sansui AU-919 thread where @noveltone mentioned the LSK389 as a possible sub for the 2SK129A. This was pointed out to me by Hyperion and some supporting info came from Ronito over in my AU-X11 thread. The LSK389 is readily available and under production by LIS.
However, I wasn't happy with existing info to just go for the LSK389. And there are variants of the LSK389... which one to buy? Others have mentioned various options for sticking two single JFETs together but I don't have a reliable line on any of those mentioned in various threads (including that AU-919 thread)... I wanted to stick to something in production too! Oh, did I mention the LSK389 are monolithic? (the 2sk129A are two singles stuffed into one package) So the matching could be really good!
Here are a couple of pics. The single unit is the 2SK163 (which is a single N-JFET in the Sansui AU-X11).
Note the pinouts. GPS16 pointed out in my AU-X11 thread that it might be worth double-checking as if the same single transistor packages are stuffed into the can, but flipped around, you might think D and S should be reversed on one side, compared to the other. However, they are interchangeable as we will see later (not that uncommon for small signal JFETs)
In that AU-919 thread @Stene showed a graph for the 2SK129. From that graph I_dss is 2mA (@V_ds = 15 V) and V_gs_off is -.3 to -.4V. I've pulled 4 working (yeah, mine are all working, I don't need new ones, I'm just curious about the substitute for now!) 2SK129A and the characteristics are different. And then over in that AU-919 thread, @noveltone kindly measure for us about 7 to 8 mA for the I_dss of the 2SK129A. It turns out noveltone's I_dss measurements on hits one device, agree well with what I measured on four devices.
2SK129A I_d vs V_ds curves for different V_gs
I don't trust cheap curve tracers anymore so the measurements were made point by point manually, using lab power supplies. This curves are for a single channel of a single device but all four of my 2SK129A are really similar. I_dss (V_ds = 10 V, V_gs = 0 V) was between 7 and 8 mA for all 8 channels of my four 2SK129A.
2SK129A I_d vs V_gs (V_ds = 10 V)
V_gs_off was about -0.75 for the single channel plotted below. I.e. V_p = 0.75 V. This was pretty similar for all 8 channels, but you know V_gs_off can be a bit tricky to determine precisely... but you get the idea.
2SK129A Drain & Source symmetry
So back to the question of the pinouts and whether D and S are interchangeable on the 2SK129A. Well it turns out that, like some other small signal JFETs, the structure is such that D and S are symmetric and interchangeable. Below I took a single 2SK129A and you can see what happens when I flip around D and S on both channels... not much. As we will see later, LSK389 is also interchangeable. You may also notice from the I_dss that I used a different 2SK129A from the one in the graphs above. Here I_dss is about 7.8mA at V_ds = 10 V. Gate-source voltage here is 0 V of course.
Substitutes
If you go to the LIS website you can grab a spec sheet for the LSC389. V_P os meant to be 0.2 to 2 V. The LSK389C is meant to have I_dss between 10 mA and 20 mA (@10V_ds), but I grabbed some anyway to check them out. I got the TO-71 metal can package.
What I found is that they are not nearly as well matched as in the 2SK129A. But, pretty good.
For V_ds from 1 to 30 V, I_d goes from 12.6mA - 16 mA... but note that the upper V_ds measurements were tricky because the can really heats up and there is significant drift. The temperature response was negative... actually i got 15.93 mA at 20V and 15.2 mA at 30V steady-state... if I take measurements really quickly before it heats up, I can get something else. Case temperature went from 25 C at 1V, to 32 C at 20 V and 50 C at 30 V.... well we are past the 400mW Pd so can't really complain. But yeah, I got the feeling that the dissipation for the C version of the LSK389 is really a bit insufficient.
I_dss for the LSK389C at 10 V was for the 4 samples:
(I_dss1, Idss2)
16.0mA, 15.09mA
16.2mA, 15.1mA?
12.84mA, 13.04mA
13.10mA, 13.25mA
which gives Idss1/Idss2 within spec (10% variation), but not as good as the 2SK129A that I have at hand. I have a "?" beside one of the I_dss values because it kept jumping around between two values, almost like it was bistable somehow... it was not my setup. I have more detailed data that could be used to make curves if anyone wants, since curves are not given in the spec sheet.
Now, the LSK389B though, is what we are really interested in. So on tho that next with some curves (which are not in the spec sheet). (sorry, I meant to finish this post including the single channel JFET data tonight, but I have an early start... so will resume this thread soon... but don't worry, it all works out and we have a nice sub for both!).