The next best thing to 2SK389 / 2SJ109 ?

Just as everyone else, I am running down on my reserve of 2SK389/2SJ109s, and am looking for the best possible substitute.

The LSK389 is only available in N-JFET, and is not pin compatible. So for me no solution.

The most obvious solution is to use still available 2SK170/2SJ74 matched pairs. If you have a large pool, you can get them even better matched than the 2SK389/2SJ74 (spec at 3% match).

http://www.diyaudio.com/forums/showthread.php?postid=1689279#post1689279

But how about thermal coupling ?

You can glue them on the flat surfaces with silver epoxy, but you mess up the pin compatiblity to the duals.

So this is my solution. ;)

It has a factor of 3 more thermal contact surfaces than just gluing on the flat surface, for much improved thermal coupling, and is still small enough to fit in most applications, with a footprint of 16x8mm.

The heatsink helps to increase thermal insertia and hence thermal stability, and keep the temperature a bit lower than without.

They are not extrusions (with a minimum order of 1 ton in weight), but rather CNC wire-cut from solid. This also explains the rough surface finish, which is an added advantage as it increase surface area for heat exchange.


Merry Christmas,
Patrick
 

Attachments

  • dual jfet.jpg
    dual jfet.jpg
    90.3 KB · Views: 9,565
And of course you can use the same heatsink for any TO92 devices -- MPSA18, 2SK369, ........

One more fantastic application I am using them for is the JohnCurl Complementary JFET Buffer with a Idss-matched pair of 2SK170 / 2SJ74. You will get next to zero thermal drift in DC offset even if you use the V grade at full dissipation.

:D


Patrick
 
My absolute offset is less than 10mv after warm up.

The relative offset remains -0.5 volt in warm up mode then0.15mv in normal operation.

So I devise warm up/stand by 1/2 bias current and let amp await in warm up mode like Pedator then full bias via DPDT switch...prepare for some excitement.... Big Bang.
 
EUVL said:
Doesn't look like people think thermal coupling is important enough.

If you are the only one, I can probably let you have some from my own reserve. Give it a day and I shall let you know.


Patrick


Hi Patrick!

very nice idea :) I'm sure it works very well

I would take some if available...are you planning to run a gb for it?

ciao
Vale
 
Group Buy ??

So there are some interest afterall. ;)

I could organise a production batch if someone would be kind enough to organise the GB as follows :

-- Minimum 500 pcs in total, no max but in multiples of 100 pcs.
-- Maximum 3 addresses to post to (e.g. one Europe, one Asia, one America).
-- USD 2.50 per piece including registered post to the said 3 maximum addresses, plus 4% Paypal fees.
-- As shown, wire cut aluminium, black anodised, with 0.1mm gap for silver epoxy.
-- Parts to be ordered when quota is reached. Payment before delivery.
-- The 3 local coordinators are responsible for payment (i.e. I shall get max 3 payments).

Local delivery I shall leave to you to arrange. That way I do not have all the work to do.

Fair deal ?

You can still get a pair of 2SK170s for less than 1USD if you buy enough, and a pair of 2SJ74s for less than 2 (while stock still last). A heatsink plus a drop of silver epoxy, and you have as close as substitute as you can get. For me a better deal than buying 2SK389 / 2SJ109 on Ebay for 10 Euros, with no better than 3% match in 60% of the cases (Believe me, I have measured enough of them).


Patrick
 
the 389/109 spec calls up Idss match of 0.9:1 That is <=11% matching guaranteed.
But, there is no specification for tracking of the pair, i.e. how transconductance matches (or doesn't) at various Id.

The only advantage the dual FET has over the individuals is thermal coupling. Much of the thermal coupling can be achieved using individuals. But, far better matching of both Idss and transconductance can only be achieved with individuals from large batches.
 
> The only advantage the dual FET has over the individuals is thermal coupling. Much of the thermal coupling can be achieved using individuals. But, far better matching of both Idss and transconductance can only be achieved with individuals from large batches.

I am not sure how many you have measured and whether they are Toshiba's or LSKs.

From my experience (with full range measurement from 0mA to Idss with 1000 measurement points per FET) over 200 Toshiba Duals and over 500 Toshiba Singles :

a) 60% of the Duals will be within 3%, less than 30% within 1%, and less than 10% within 0.5%. Those within 0.5% will normally be 0.5% over range and not just Idss.

b) Within the same bag (of 200, as from Toshiba), one would hardly be throwing away 20% even when looking for 1% match or better.

c) 0.5% Match over range from singles within a 200 pc bag is not difficult to find at all, just a lot of work in data processing.

I think there are enough people here who buy 100 pcs at a time. And there are also enough offering for matched pairs. That is why I am not going to offer anything unless they are unique.


Patrick
 
Re: Group Buy ??

EUVL said:
......For me a better deal than buying 2SK389 / 2SJ109 on Ebay for 10 Euros, with no better than 3% match in 60% of the cases (Believe me, I have measured enough of them).


EUVL said:
................From my experience (with full range measurement from 0mA to Idss with 1000 measurement points per FET) over 200 Toshiba Duals and over 500 Toshiba Singles :

a) 90% of the Duals will be within 3%, less than 30% within 1%, and less than 10% within 0.5%. Those within 0.5% will normally be 0.5% over range and not just Idss.

b) Within the same bag (of 200, as from Toshiba), one would hardly be throwing away 20% even when looking for 1% match or better.
c) 0.5% Match over range from singles within a 200 pc bag is not difficult to find at all,

your yield went from 60% passing the 3% Idss match to 90% passing the 3% match. Are the Ebay ones comparable to the genuine Toshiba?

I only checked the tracking over the range 20%Idss to 100%Idss and never more than @ 5points, after finding close (1%) matches @ 3points.

0.5% Idss matching of BL grade sk170 from a 200batch is not common.
Then compare transconductance (as a thermally coupled pair) and the apparent matches are seriously not matched. The yield of 0.5% Idss and at lower currents is nearer a 2% yield from a 200batch and yield virtually NIL from a 100batch.
I have only measured 1000FETs, so my experience is limited.
 
> your yield went from 60% passing the 3% Idss match to 90% passing the 3% match.

Typing mistake corrected.

> Are the Ebay ones comparable to the genuine Toshiba?

No idea. I never bought any from Ebay. But at 10 Euros a piece, they better be genuine.
I bought mine ages ago from a reliable German supplier.

>0.5% Idss matching of BL grade sk170 from a 200batch is not common.
> Then compare transconductance (as a thermally coupled pair) and the apparent matches are seriously not matched. The yield of 0.5% Idss and at lower currents is nearer a 2% yield from a 200batch and yield virtually NIL from a 100batch.

I just measured a batch of 200 pcs 2SK170BLs a week ago (rough screening with one measurement point at Idss), and have no problems finding Idss matches to 20uA, which is as accurate as my multimeter will do over an hour, even with calibration against a current reference every 15 minutes.

Low range (<7mA Idss) is difficult. But anything from 7.5mA to 10.5mA, I have no problems finding excellent matches. Maybe I just have luck.

;)


Patrick
 
EUVL said:
............. have no problems finding Idss matches to 20uA, which is as accurate as my multimeter will do over an hour, even with calibration against a current reference every 15 minutes.

Low range (<7mA Idss) is difficult. But anything from 7.5mA to 10.5mA, I have no problems finding excellent matches. Maybe I just have luck.
Much more likely you have better techniques, I can't achieve that.
Yes, the gr grade are much more difficult to match due the wide spread of the Idss range and the temperature sensitivity since we are working at below the optimum Tcomp current.
The V grade are easier to match in % terms but Vds has to be reduced below the 10Vgs standard to keep temperatures sensible using my slow methods at the high end of the range (20mA @ 10V takes the device to 50% of absolute max Tc=25degC limit and should be de-rated as Tc rises).
 
I use 9V Vds, which is about the optimal working point.

Low Idss take ages to stabilise thermally, as the current keeps drifting. One breath of air, and it changes by 20uA easily.

Anything above 9mA is easy. They warm up fast and remains stable over minutes.

My Idss measurements were all taken at steady state. A thermal shield from the environment will help.

And you need to do that at the same room temperature. It is therefore easier in Winter than summer, but you need to measure room temperature.

Of course if you have done a rough sreening first, and you can just match a pair right after each other to reduce those variables.


Patrick
 
EUVL said:
Of course if you have done a rough sreening first, and you can just match a pair right after each other to reduce those variables.
rough screening into batches of ~1% increments.
Then pair matching in a thermally coupled LTP to find Idss matches, followed by selected pair matching in that same LTP thermally coupled jig.

I do not try to compare two samples by reading numbers. I select by reading numbers and then I pair match. I did BJTs this way and modified the method to allow FETs to be done. It's the enormous spread in parameters of FETs that makes matching difficult. BJTs are so much easier.

If you are comparing numbers from separate test runs then I am beginning to doubt the results of the yield of matching pairs.