### Counterfeit transistors

[sbrook]

The die is 6mm*6mm which is what was suggested it should be.

Silicone just means a different manufacturing process.

They are CHEAP, but it doesn't mean they're bad ... Sometimes you know they're bad ... like if they had a 2mm * 2mm die! In this case though, the only way to know is to test them! They MAY just be acceptable.

[john_ellis]

Hi Dirk

There is no rule which says a company cannot manufacture a part which was introduced by someone else.

But I agree it is necessary to check whether the second-source manufacturer is making good products or not.

I think your suggestion earlier is a very good one - to check the second breakdown current ratings. This is a good way to see whether a transistor meets its specifications!

there have been many examples about counterfeits that it would be useful to have a second-breakdown test circuit for diy-audio-ers to try. I've sketched a circuit, but so far have not had a chance to explore it.

Perhaps there are other measurements which can be made as well. For example, small die might have a low capacitance. If you can measure Cbc and Cbe this might show (if it is too small) if the chip is possibly too small.

Also, measuring Vbe against Ic (the normal GUmmel plot) might also show if the die size is too small - if the Vbe has to be higher, for example, to give a certain collector current.

And measuring the fT should be possible, to see if a technical match to the spec. has been made.

Unfortunately measurements of Cjc and Vbe(Ic) won't necessarily confirm a problem, especially the Vbe. There is a high temperature sensitivity, and measurements on a "real" and "test" devices need to be made at the same temperature. And manufacturing process variations could change the results, but it may be that a gross difference could indicate a problem.

Measuring Cjc, second breakdown current, breakdown voltages, fT and current gain are tests which should be possible and could be revealing.

cheers
John

[sbrook]

At the same time, one has to be a little careful to determine whether the parameter you're checking REALLY matters for the application depending on whether the critical parameters are acceptable.

For example, most output stages do NOT require high gain, therefore the difference between a transistor with an hfe of 100 and another of 150 may not make much practical difference.

Unless you're very specifically designing with critical tolerances for a particular transistor, then an fT of 100MHz or 150MHz for example is hardly critical in an audio circuit other than ensuring it doesn't take off!

I always take these things in a very practical way ...

You buy some lemons from Spain, but instead you find they're from Morocco. Can you still make lemonade? Yup. Does it taste good? Yup. Does it need a little more sugar? Maybe. OK .. they're usable.

If you buy lemons from Spain and get shipped limes from Costa Rica, then can you make lemonade? Nope - but you can make limeade. Does it taste good? Yup. Will you accept limeade instead of lemonade? Yes - keep em. No - send them back and get a refund.

If you buy lemons from Spain and get shipped dried out lemons from California, can you make lemonade? Nope. Nothing you can do with these ... so send them back and get a refund.

The question is, are these transistors close enough to do the job being asked of them? If they are and it's too much hassle to send them back for a refund, or you might not get a refund, then why not use them! If they aren't good enough to use, then try to get a refund.

[GKU]

Hi Dirk,

I'm a semiconductor device professional.
May I can help you and the others at this case.

First, iscsemi is a well known semicondutcor manufacturer, located in Wuxi China close to the German Infineon plant, specialized on Discrete Power Devices.
The devices in question 2SC2922 and 2SA1216 in the specified device package MT-200 are first designed and manufactured by Sanken Electric in Japan and are on the official product list of iscsemi.
If a semiconductor manufacturer designs a new device it has to be registered with all product parameters and specification as well as the package design at the international semiconductor standardization organization JEDEC.
JEDEC plublishes the registered device on their official publication.
After a certain protection time such devices can be manufactured by other companies to. But - by using the same device designation, the device must be conform with the origin product parameters and specification including the package design.
Iscsemi is registred as origin manufaturer of the mentioned devices at JEDEC.
For sure there are some manufacturing process deviations resulting in slightly different package surface or smooth corners often seen at the metal heatsink. As long as such small deviations don't impact the device parameters and specs, there isn't any reason to be doubtful of the device quality.
But I must mention, due to different semiconductor manufacturing prozesses, which can deviate from the origin process used at Sanken, such device used at Audio application can result in slightly different sound colours, not necessearily important but detectable.

You have also mentioned the 2SC2922 surface has been reworked with sandpaper. The attached photograph support this impression.
Sometimes in production it appears that a wrong designation has been printed, or they aren't in the range of the subcode specs, which require rework.
The manufacturer of these devices has been using laser marking for designation printing which engrave the surface and can't be simply removed with solvents as in the past with ink-printing. Thus correcting the designation require manual rework with sandpaper, a standard procedure with some Chinese manufacturer, but rarerly seen at Japanese companies.

Recommended actions:
1. To check if your devices corresponds to the origin product specifiaction and if you don't have the ability to measure the important product parameters, you should send some examples to an institute of an university having an electronic faculty, which can easy check the parameters.
2. If you board such devices from an official German distributor, who sold the devices under the origin of Sanken, he is obliged to retake them, due to incorrect sales conditions.
3. General purchasing recommendations:
3.1 Buy only from serious sources as from official distributors
3.2 If you buy over ebay, I would not recommend any purchase from private institutions.
3.3 If a device has been aquired from an official distributor under a specific product and manufacturer designation (Sanken), the delivered devices must correspond to the origin device. In case the original device isn't on stock, the supplier must ask you prior to shipment, if an equivalent type can be send.
The German disti Reichelt offer such devices in their "Japan Transitor" list. In this case it "must" be Japanes origin. The product designation only (2SC2922 f.i.), first registred from Sanken, isn't automatically a Japanese origin product, but if mentioned Japanese product, Japanese must be in.

Cheers

Gerd

[Onvinyl]

Thanks for these splendid informations!
Rüdiger

[Eva]

Since these are with no doubt the fake inferior parts that are hurting us so much, I wonder if we have some fake seller friends here trying to avoid the embarrasement...

Genuine MT-200 parts must look like this:


And I also have plenty of experience with these parts...


All these are fakes, fortunately I managed to return the fake lot and get a refund:


All the transistors manufactured with the withe-goo process are unreliable fakes. Genuine parts sourced from a quality process have either a thin transparent resin coating over the die, or no coating at all.

This page contains plenty of fake vs genuine pictures demonstrating the previous fact:
http://transfal.tripod.com/indexold.html

[bob_v5]

A posible reason for relabeling- operator boredom. When I used to work in a plastics factory, one of the parts made there was the front panel of set top boxes. The machine for printing the labels got mis-aligned over time. The people running the machine were low paid, bored, and often on drugs or just plain stupid. Hundreds of incorectly printed front panels could be packed during the night shift, leaving someone with the job of removing the ink from the panels to be reprinted. We used solvents and rubbed them with a cloth. But if asthetics werent important the writing could have been removed much quicker with sand paper!

[small_boy]

Hi Gerd,

first, thx for the very good informations.

And its right that the label is printed with laser.

I then can hope thats the 2SC2922 is from ISC and the transistors hold the spec´s from the Datasheet (that i can´t download on homepage from ISC, there is no 2SC2922 on it

And when the sound is a little bit different to Sanken, with that i can living.

important for me is that this 2SC2922 are no fake or Counterfeit and i can use it in a amp and the Amp don´t burn out after one hour hard sound of AC/DC ))

greetings from germany

Dirk

[small_boy]

Hi Eva,

you have many fakes of sanken. OK, fake of sanken Transis, i can understand. Because sanken are good and expensive. But ISC isn't so known as sanken, why fake this.

but when i look on the side from your link, in the future I must have afraidly to buy a Transistor, because there counterfeit all what it give on the market.

Nice photo of the sanken transis, pls send me 10 of 2Sc2922 and 10 of 2SA1216 ! )))

Dirk

[Eva]

They don't have a datasheet for a good reason... lol...

I think they are packing a "generic" power die in a fake MT200... They use to produce several different transistor models with the same generic die. Fakes of one model are frequently relabeled to supply demand of another model. They don't care about the performance differences between models since no one matches the specs of the genuine devices anyway. Forget about Sanken performance (high and linear hFE, 50Mhz Ft) which is exclusive of Sanken LAPT process, what you have is MJ15015/MJ15016 1970's performance (then, why not using these other devices?)

I tried a couple of the Sanken fakes shown in the last picture in one of my amplifier prototypes and it was not even stable. It oscillated due to the poor bandwidth of the fakes while it was unconditionally stable with genuine parts.

Why manufacturing 2SC2922 and 2SA1216 under another name? Because Sanken parts are of great quality and widely used, and there will be always people willing to buy the same part number with another brand for less money and expecting them to be as good the originals... It's like buying a Renault with a BMW sticker and believing you have a BMW.

If you ask them for a lot of 2SC3264 and 2SA1295, or any other MT200, they will sand the cases again and sell you the same parts relabeled to read what you asked for...