FS 2SK170 2SJ74 matched sets

[tony399]

I have got just over 230 of each 2sk170 and 2sj74 unused jfets left over from previous projects.

They have all been initially measured for Idss at 10v and are BL version so there is a spread from around 6.5mA to 11.5mA.





To start with I will only supply NNNN-PPPP (4 x K170 and 4 x J74 matched sets). When these run out I will go to NNN-PPP then NN-PP then N-P. N-N and P-P sets may follow.

They will be remeasured (to reduce temperature, supply voltage and instrument drifts) and set matched using a meter with 1uA resolution, you can request;

Tolerance
Vds*
Preferred Idss

*The Vds of the specific circuit they will be used in will affect matching tolerance given the different transconductance curves of individual devices (a 2% match at 10Vds certainly doesn't guarantee a 2% match at 5Vds or 20Vds).

As supplies dwindle request options will be removed but I can then give more detail for the remaining jfets.

If you PM me with requirements I will let you know what I can supply and cost. Actual measured values will be given for each jfet of the set prior to purchase.

 

[zoe]

Hello Tony
how much do you want for a set of around Idss 10.

 

[MASantos]

Where did you source them? Do you have any assurance these are genuine? I might be interested.

 

[tony399]

Thanks to all for the inquiries so far.

At the moment I am looking at £30 to £40 per matched 8x set depending on tolerance and Vds requirements plus p&p which I hope is considered 'reasonable' for what they are. Paypal is my preferred method of payment.

To be able to supply a good number of people a restriction on quantity per person will be imposed, I guess four sets (16 x K170 and 16 x J74) will allow enough for personal use so I will say this to begin with.

These are genuine Toshiba as far as I can be sure. A lot of them were sourced from Ampslab and I will try to find out where the rest came from (was a while ago).

All the markings and measurements are consistent with Toshiba and the circuits I have built with them all work fine. I have read that fake ones don't even measure correctly.

Please bear with me and I will try to answer PMs over this weekend but I do have work on Saturday.

 

[tony399]

Fakes?

Speaking of fakes, I thought I had found more supplies

But when I looked closer the lettering is not the same as mine

Or Amplabs (where the majority of mine came from)

The lettering on all my devices is the same style which anyone with genuine Toshiba parts will recognise.

 

[tony399]

Rethink

I might rethink the custom options and pricing.

If I just match them all at 10Vds, group them in sets of 8, 4 and pairs at different tolerances then publish the results for people to choose what they want it will be far less work for me and price can come down.

Please leave feedback for me on this thread with your thoughts..

Thanks, Tony.

 

[Geografica]

I think that's sounds good.

 

[qusp]

it does make them less useful though unless everyone is using them at 10vds... I understand you probably underestimated the time customizing might take, so not a problem at all. I think the ones from analog metric look fine, there were a few different types of lettering and I have seen legit J74 like those pictured

 

[CaféNoir]

I agree too

 

[bear]

Please shoot a close up shot of the round side of the jfets?

I'd like to see what is in the backside.
Some of the fakes I have in hand are easy to identify by that...

Otherwise, should they prove genuine, I'd be interested in more or less random IDSS 2SJ74s... I found that a quick look on the curve tracer revealed fake J74s very very quickly - they're not even close. The K170s are harder to spot because the fakes may not be fake, since they are pretty close in terms of gain and linearity. Noise and a few other not quite as easy to measure specs might spot them.

_-_-bear

 

[madisonears]

If you're taking the time to match parts and charging a premium for it (which is certainly fair), perhaps you could accommodate different builders and applications by matching some at one Vds, then another separate batch at a different Vds. That way, buyers could identify which sets can be used with confidence at or near the Vds they require, perhaps 10 and 15, which seem to be pretty common. Sort of a semi-custom specification with meaningful flexibility yet simplified, and less expensive, work on your part.

Peace,
Tom E

 

[tony399]

I will take a close up of the backside and post it tonight if I get a chance, there is a round indent which has no letter in it.

So far I have remeasured around 40 to 50 of each type which is a range from minimum Idss up to 8mA. This plus the initial range sorting measuring has taken a couple of days work.

The first ten or so I did a measurement at 10Vds and then at 20Vds and again I will post the results so people can see how matching at the correct Vds is important.

If further Vds points are necessary to prove transconductance curve I'm happy to do this on a few samples.

 

[tony399]

By the way has anyone investigated diytrade.com, they are showing J74s which look ok to me..

 

[tony399]

Funnily enough I had a picture of the backs already for this very reason bear. I hadn't posted it because it was taken on the works camera, my computer at home doesn't take SD cards and I didn't have a usb cable to fit the camera.

 

[tony399]

Example one;
K170 @10Vds = 6.65mA, @20Vds = 7.10mA
J74 @10Vds = 6.65mA, @20Vds = 6.99mA

Example two;
K170 @10Vds = 6.82mA, @20Vds = 7.28mA
J74 @10Vds = 6.81mA, @20Vds = 7.14mA

Example three;
K170 @10Vds = 6.88mA, @20Vds = 7.34mA
J74 @10Vds = 6.88mA, @20Vds = 7.23mA

This is the level of matching I can achieve on some pairs, seems a pity not to do it at the ideal Vds..

 

[tony399]

Due to the amount of counterfeit devices around at the moment I am going to work with another diyaudio member to get my devices measured on a curve tracer. Apparently this a fairly good method of spotting some of the better fakes.
Please no more PMs for the time being.
Thanks,
Tony

 

[NicMac]

Your measurements nicely indicate that Idss does not change dramatically with Vds
Personally, I would prefer knowing both Idss and Yfs at a single Vds rather than Idss at multiple Vds.
Tracing will show you that your J74 are genuine - and that they will never trace K170 well unless you degenerate

 

[tony399]

curve tracer

I have been told I can get a curve tracer from ebay to attach to a scope for $20, the only one I can find is;

Power Transistor Curve Tracer adapter XY Oscilloscopes | eBay

Obviously this is designed for power transistors not fets, does anyone know of one for fets that is cheap and works well?

I've had a quick look at octopus but need to do some more research.

Also, on the question of matching, the circuits that I have seen are designed to cope with a relatively loose tolerance.

This would help me as I could match more sets and not worry about custom Vds, in fact all the ones I have can be done at 10Vds with 5% tolerance at worse.

Anyway, need to get them verified first.

 

[Salas]

Get a Peak DCA75 made in the UK.

 

[tony399]

Get a Peak DCA75 made in the UK.

Nice bit of kit but not cheap at £115!