NJFETs for Source Follower Applications

The one and only
Joined 2001
Paid Member
Post #1, "Each of the N-JFET pairs used in the testing was matched to <10μA in Idss."
So DC was < 1mV in all cases.

But yes, if one can measure upfront and use the JFET with higher Idss for the CCS, a small source resistor will allow easy trimming of DC offset.
I normally do not use source resistor for the upper JFET.
Somewhere I read before it is supposed to create more H3.

2SK117 can still be bought, 2SK209 for sure.
So no need to be desperate, for now at least. ;)

I have been working with J113's in an effort to get good results with
cheap and available parts, and I continue to make some progress.

I have a follower and CCS operated 6 to 12 mA, and with part selection
by Vp and appropriate Source resistors I am currently getting less than .002%
at 1V (-94 dB). If the devices are matched Vp, then equal value Source
resistors on both Fets gives consistently good results.

It gets better with a little more supply voltage, as high as +/-15V.

Of course, the part is too noisy for the input of an MC phono stage...

:cheers:
 
Member
Joined 2004
Paid Member
Is part of the choice of J113 the fact that it's available in TO-92 to make it more DIY friendly for most people vs parts that are only available as SMD?


I have been working with J113's in an effort to get good results with
cheap and available parts, and I continue to make some progress.

I have a follower and CCS operated 6 to 12 mA, and with part selection
by Vp and appropriate Source resistors I am currently getting less than .002%
at 1V (-94 dB). If the devices are matched Vp, then equal value Source
resistors on both Fets gives consistently good results.

It gets better with a little more supply voltage, as high as +/-15V.

Of course, the part is too noisy for the input of an MC phono stage...

:cheers:
 
Replacement JFET

Hello All,


It seems that everyone is looking for a replacement JFET. From my view they are all on the endangered list.


I like the idea presented of a standard performance test. In this case a source follower. Plus the idea of a small source tuning resistor seems to be becoming to be fairly common. We do not want a JFET that measures close to go unwanted.


I think of these JFETS as rapidly becoming extinct in the wild. I have a zoo of these things in my cage. I also have a bunch of out of print tubes. For commercial or mass DIY use I am glad to see the J113 remaining on the shelfs for sale.


I have access to an APx555 to also test and compare distortion numbers with variable circuit voltage and current. Most interesting is noise performance.


What I do with tubes: I hook up the power supply and audio analyzer; tweak the power supply voltage and dial the bias, most often there is a null or sweet spot where THD and Noise find a minimum.

Thanks DT
 
2SK2145-BL is in stock at Digikey for USD $0.28 at 1,000 pieces. $285 gives you a lifetime supply if you are a hobbyist. You have to deal with surface mount. Not insurmountable.

You could parallel a GR grade part and have a 3-terminal BL part.

A PCB adapter that pinned out to TO-94 would be very nice.

Or just buy $200 worth of genuine 2SK170 from AMPSLAB for $2 each.
 
The one and only
Joined 2001
Paid Member
When I started getting serious about replacing the Toshibas as followers
I did pc art for a jig that used two Jfets and had variable resistance for
the Source of each and also the load. This allows you to search for the
sweet spot without a lot of effort. After I did this one I have figured out
some improvements...
 

Attachments

  • jfet buffer.jpg
    jfet buffer.jpg
    990.2 KB · Views: 1,527
Hello All,

For grins I took the first J113 out of the bag and cut the first J176 off the tape. By chance the IDSS for the J113 (10.1557mA) and J176 (-10.0588mA) are pretty close to matching. The J113 (16.3870mS) and J176 (-6.6730mS) Transconductance are nowhere near matching, they are exactly what you would expect from a look at the datasheets.

Vgate Vdrain Idrain gm
V V mA mS
J113
1/1 0.00000 15.0000 10.1557 16.3870
1/2 0.10000 15.0000 11.7944 16.3870
1/3 0.00000 15.1000 10.0898 16.2490
1/4 0.10000 15.1000 11.7147 16.2490

J176
2/1 0.00000 15.0000 -10.0588 -6.6730
2/2 0.10000 15.0000 -9.3915 -6.6730
2/3 0.00000 15.1000 -10.0546 -7.4890
2/4 0.10000 15.1000 -9.3057 -7.4890


Most of the formatting disappeared when copied to DIY.


Before connecting audio analyzer, has anyone tried picking a higher IDSS J113 and dialing back both the Drain current and gm with a source resistor?

Seems that if we can adjust the J113 and J176 Drain currents and gm’s closer to matching we could get better distortion performance.

Anyone tried?

Thanks DT
 
The one and only
Joined 2001
Paid Member
Yes, of course. But now you can see why the Toshiba's are prized....

You would want to try J112's, J113's, J175's and J176 with different
Idss figures, different Source resistors, different loads, different supplies.

I pulled up some notes indicating that if you juggle these you can get some
0.00X% distortion numbers at 1V into as low as 1.5 Kohm.

That said, it's better and easier with matched J113's where one is the
follower and the other one is a CCS.

:cheers:
 
Member
Joined 2011
Paid Member
If the source follower application has |total supply voltage| > (max supported peak to peak output swing) - 12 volts, you could consider cascoding each J113 with a J107. The J107s need not be matched (but of course YOU would match them anyway, because that's so YOU) Then you'd have constant voltage AND constant current AND the sweet spot. Geepers what's not to like.
 
Yes, of course. But now you can see why the Toshiba's are prized....

You would want to try J112's, J113's, J175's and J176 with different
Idss figures, different Source resistors, different loads, different supplies.

I pulled up some notes indicating that if you juggle these you can get some
0.00X% distortion numbers at 1V into as low as 1.5 Kohm.

That said, it's better and easier with matched J113's where one is the
follower and the other one is a CCS.

:cheers:

Nelson,

Have you tested the J230 and 2N3819 with your jig? If you can post a schematic, I can try it on my end.

Cheers!
 
Hello All,

There for a minute I failed to recall where I learned to dial in the sweet spot. I clicked on PassDiy and it all came rushing back at me. Back when, I rushed out and purchased a used AP2522 audio analyzer. You can adjust the power supply voltage plus adjust the multi-turn variable resistors up a little or down a little while you watch the Harmonics waltz in or out on your computer monitor.

We are talking simple buffers with matched n channel pairs maybe later get fancy and try complementary push pull n p pairs.

This is not so much about finding the perfect replacement JFET as developing the methods to select from the JFETs that you have dusty on the shelf. It is much easier to match n channel pairs than complementary push-pull n p parts.

I have some K170, K246, 2N5457, PF5102 and a butt load of J113 to test.

Thanks DT

I am looking at sample buffers in the LX-mini Analog Crossover.