This FET is testing weird

[Ricardus]

Question. In my recent preamp builds, the DI board in this kit uses a 2SK170 FET. These haven't been made for a while, so I'm wondering where the kit seller is getting them. I've been testing them and they all test like this with my DCA75. But they test as a FET with one of those cheaper multi function testers like you can get on eBay for 10 or 15 dollars.

I happen to have some NOS Toshiba 2SK170s in my stock so I am installing those instead, but any idea why one tester would say dual diode like this one (supposedly the better tester) and the other would say it's fine?

I mean, I know it's the software, but I have yet to see a FET test like this in the DCA75 until the FETs in these kits.

As I said, I've been installing some NOS 170s I already had instead, so I have no idea if these would even work in circuit. That said, this is a popular kit from a reputable company, and if the FET wasn't working seems like they would know it.

 

[pucho812]

Maybe the are linear systems repros, they make the lsk170 which is a drop in pound for pound replacement

 

[Ricardus]

Maybe the are linear systems repros, they make the lsk170 which is a drop in pound for pound replacement

Yes. I've used their LSK389 dual JFETs in some cloudlifter clone builds that Abbey designed.

As I said, every transistor I've ever tested in the DCA75 tested fine, unless it was a failed part. But everyone of the 2SK170s I've received so far (4 in total) test out as dual diodes, while the cheaper tester says they're fine.

It's just oddly curious.

 

[gyraf]

The market for fake jfets is huge. Recently it's more common to get fakes than not, even from real retailers

Could it be that your jfet's are kinda real, but marginal: de-selected parts that failed post-packaging tests, perhaps?

/Jakob E.

 

[Gus]

The gate does form a diode with the rest of the substrate

So I would look into how the Peak tests devices vs the cheaper tester things like the currents and voltages used.

I would test them with one of the DIY IDSS, Vgs testers that are a fast build

Then I would build a simple gain stage to compare the devices.

There is more to think about like Vgd breakdown

 

[JohnRoberts]

+1 Perhaps body diodes to substrate.

This is usually detailed in data sheets, but who knows what actual silicon device you got.

JR

 

[Ricardus]

Could it be that your jfet's are kinda real, but marginal: de-selected parts that failed post-packaging tests, perhaps?

/Jakob E.

Seems possible!

 

[dom]

Assuming your gate is on the center, it looks like your meter is correctly showing you that you have a JFET connected. I don't see any problem here. Gate-source and Gate-drain of a JFET just looks like PN junctions.

Your meter has some software bundled with it and can actually do I-V sweeps, which is pretty neat. It's not just a simple diode tester. Have you tried installing the app to play with that functionality? It might offer better diagnostics than what you get through just the meter hardware interface.

 

[Ricardus]

Assuming your gate is on the center, it looks like your meter is correctly showing you that you have a JFET connected. I don't see any problem here. Gate-source and Gate-drain of a JFET just looks like PN junctions.

Your meter has some software bundled with it and can actually do I-V sweeps, which is pretty neat. It's not just a simple diode tester. Have you tried installing the app to play with that functionality? It might offer better diagnostics than what you get through just the meter hardware interface.

Every other JFET I've ever tested on it has identified as a JFET.

I haven't done any sweeps with this particular part.

 

[ruffrecords]

I have exactly the same tester. If you want to lend me one of the FETs I can check it on mine.

Cheers

Ian

 

[Winston OBoogie]

+1 on all the above's

Most of the K170's listed on Ebay are fake (I know of one seller there that I'm positive the devices are bona fide). so it's certainly a possibility that the kit seller bought some fakes in good faith.

However, it's not evident to me that these are indeed fakes. Are the kit J-Fets the same IDSS grade as your NOS Toshiba's? G, BL, V?

As Gus mentioned above, the best way to test these is with an easy to build IDSS measurement circuit.
Nothing beats also measuring the device at the operating point used in your circuit or, in situ. Compare with your known good K170's.

What Dom suggested regarding software sweeps on your DCA75 seems cool

 

[rackmonkey]

This issue is actually mentioned in the DCA75 manual. Faulty transistors or transistors with “very low gain” will often be identified as 2 diode junctions. It’s a limitation with the instrument where “low gain” is concerned. With JFETs, that means that devices with low Yfs (usually also in combination with low Idss - but not always) will be identified as such and you won’t be able to get the data you’re after for matching, selecting etc. This always happens with devices like 2N4338, which i’ve never been able to measure successfully.

It should NOT happen with 2SK170. Yfs is unusually high with that device and i have never seen a real 2SK170 measure as 2 diode junctions. The kit maker may have gotten some fakes or out of spec parts in a batch, or the FET may simply be damaged from static or something. Regardless of what the issue is, I can assure you that the DCA doesn‘t have an issue testing good 170s.

Link to manual - see page 20:

https://www.electrokit.com/uploads/productfile/41015/dca75.pdf

 

[Ricardus]

This issue is actually mentioned in the DCA75 manual. Faulty transistors or transistors with “very low gain” will often be identified as 2 diode junctions. It’s a limitation with the instrument where “low gain” is concerned. With JFETs, that means that devices with low Yfs (usually also in combination with low Idss - but not always) will be identified as such and you won’t be able to get the data you’re after for matching, selecting etc. This always happens with devices like 2N4338, which i’ve never been able to measure successfully.

It should NOT happen with 2SK170. Yfs is unusually high with that device and i have never seen a real 2SK170 measure as 2 diode junctions. The kit maker may have gotten some fakes or out of spec parts in a batch or something, or the FET may simply be damaged from static or something. Regardless of what the issue is, I can assure you that the DCA doesn‘t have an issue testing 170s.

Link to manual - see page 20:

https://www.electrokit.com/uploads/productfile/41015/dca75.pdf

Oh wow. Thank you! Shame on me for not having seen that in the manual in the first place.

 

[rackmonkey]

Oh wow. Thank you! Shame on me for not having seen that in the manual in the first place.

I share your shame . I initially went through the same mental gyrations as you did before I figured out what was really going on. RTFM!!