Beware!!!!! Fake Sankens

[mcp]

This is a much sought after part from Sanken. One is an original, the other a FAKE. The complementary part 2SC2922 looks very close like the original too. Information on how to tell them apart can be found at

http://www.ampslab.com/projects.htm

Regards
Mike

 

[Piersma]

FAKE

Micheal,

I have taken a look of some 2SA/SC transistors on my shelf and i noticed the following. The lettering "SANKEN" seems original to me but the production run number is in the middle.

I am confused, are they fake or not ?

Greetings from Holland
Piersma

 

[Elso Kwak]

Re: FAKE

Micheal,

I have taken a look of some 2SA/SC transistors on my shelf and i noticed the following. The lettering "SANKEN" seems original to me but the production run number is in the middle.

I am confused, are they fake or not ?

Greetings from Holland
Piersma

Hi, Yes I am extremely confused too. Maybe they are both real!
Below a picture from www.schuro.de that I believe is the genuine article. Mike do you have measurements or listening tests to prove which one is real?

 

[Netlist]

What about those?
They are the original transistors from an old Crest 5000.

/Hugo

 

[mcp]

Hi everyone,

The one on the right is the real one.

Piersma, Elso....look like your Sankens are OK . If you have the time, check out my site or Rod's. The visual differences are explained more in detail.

orginally posted by Elso
you have measurements or listening tests to prove which one is real?

I don't need to. Let's just put it as "supplier & customer confidentiality". If you like to know more, drop me a mail.

What about those?
They are the original transistors from an old Crest 5000.

/Hugo

I can only verify 2sc2922 and 2sa1216. The ones in the Crest should be real. One of the main difference between a fake and an original is the inability to supply the same power output. The frightening part is this particular fakes look like real.

Regards

 

[JojoD818]

hi all!

here in my country a shop sells both originals and fakes! fakes for about $0.80 and the orig for $4.50, their reason? some people want to buy cheap ones because the origs are very expensive. But if your amp blows up while using the fakes, the store will not be liable she said.


jojo

 

[mcp]

Inside view

I finally cracked open the fake Sanken and compare it to an original. The differences are quite astounding. You can see them at:

http://www.ampslab.com/projects.htm

Regards

 

[grataku]

What about those?
They are the original transistors from an old Crest 5000.

/Hugo

Yes, those are the real ones. It's not really hard to tell, come on.
The forgerer had to give it away by using NY as the selection code.

 

[Elso Kwak]

Re: Inside view

I finally cracked open the fake Sanken and compare it to an original. The differences are quite astounding. You can see them at:

http://www.ampslab.com/projects.htm

Regards

Hi Michael,
I am sorry but my gutsfeeling tells me it is the other way around.
I feel the transistor with the die directly attached to the collector case is the real one and the one with the TO-220 case attached to the larger plate is fake.
Maybe others know more about the inwards of transistors.
Please be careful, wash your hands!
You might be exposed to Berylium that is highly toxic.

 

[grataku]

Elso,
In fact, they are probably both fakes! The selection code is not convincing (i.e. the 1 number and 2 letters) but most of all the weight doesn't match the weight indicated on the datasheet that is ~18 grams.

 

[Jocko Homo]

Draw your own conclusions.....

These are samples that came directly from the Sanken rep, so I would assume that they are genuine.

I am not cutting them open, though..........

 

[mcp]

Hi Elso,

Thanks for the warning.

I can assure everyone that the one on the right is the genuine Sanken. I have a few hundred pairs of the originals. I've tested them in prototypes, I've used them in finished designs, I've used them in 1pair, 2pairs, 3pairs, 4pairs... up to 9pairs per channel. Most were in Class AB, some in full Class A.

There were some points I did not put in my site for I wanted to focus on just the main issues. Perhaps this is the right time to share them with everyone.

First, I tried sawing through the plastic with a normal hacksaw. In the original, it was very tough. When that didn't work, I pounded it with a hammer. It took quite a fair bit of pounding before the original started to crack, but it was much easier with the fake. The plastic on the fake is brittle.

Remember the shiny coating at the back of the fake.

When I filed the base, the shiny coating started to flake off easily, in fairly small pieces. It is actually low quality chrome plating.

If you look at the inside, the original die is much bigger. It is embeded in a TO3P back plate which is then bonded to a bigger plate. This is a costlier process and I can only guess that it provides better heat transfer. Fake manufacturers do not incur extra cost in this manner, especially when the features are hidden. Don't forget, they are producing this in huge numbers, so every cent counts.

Perhaps Jocko can verifiy this with the rep, or better still, get them to open one for sample.

Regards

 

[Netlist]

Yes, those are the real ones. It's not really hard to tell, come on.
The forgerer had to give it away by using NY as the selection code.

Maybe I confused people here with my post.
I actually meant that original transistors that came out of a Crest can’t be fake.

/Hugo

 

[grataku]

Hi Elso,

Thanks for the warning.

I can assure everyone that the one on the right is

Since the weight you reported was 10g, your balance must be a little off.

I wonder why they didn't stamp the selection code in the middle. The brand stamp was almost perfect, why get sloppy on the code? That way it would be almost impossible to notice the fake by browsing through a pile semiconductors. Maybe that's why they don't do it...

 

[mcp]

Originally posted by grataku
Since the weight you reported was 10g, your balance must be a little off

That ain't bad, coming off a $10/- weighing scale