Hello! I've tried some times ago fixing a Marantz 1060. It blew all the new MJ15003 I put in. After changing them to 2N3442, it worked fine.
So I thought about counterfeit transistors, and read that MJ15003 are probably the most commonly faked.
I decided to open up a shorted MJ15003 from the amp and a new 2N3442 that left over because I broked a leg of it.
The cover of the MJ went off easily in a vise, the 2N required the use of a saw. The die in the MJ was covered in white silicone, and it measures 4x4mm, under it there was a little copper disc (I read that the disc must be large about the size of the entire transistor, this one was very small). In the 2N, a device rated for less than the powerful MJ, the die was identical in dimensions, no silicone, and a big copper disc under it. I also become surprised when I saw the little wires connecting the MJ collector to the die. And these can carry 16A? I'm not sure about that... 2N had bigger wires, still giving a reduced max Ic.
Chances that I get counterfeits? They gave me much trouble, as Anatech knows...
(sorry, no pictures of them)
So I thought about counterfeit transistors, and read that MJ15003 are probably the most commonly faked.
I decided to open up a shorted MJ15003 from the amp and a new 2N3442 that left over because I broked a leg of it.
The cover of the MJ went off easily in a vise, the 2N required the use of a saw. The die in the MJ was covered in white silicone, and it measures 4x4mm, under it there was a little copper disc (I read that the disc must be large about the size of the entire transistor, this one was very small). In the 2N, a device rated for less than the powerful MJ, the die was identical in dimensions, no silicone, and a big copper disc under it. I also become surprised when I saw the little wires connecting the MJ collector to the die. And these can carry 16A? I'm not sure about that... 2N had bigger wires, still giving a reduced max Ic.
Chances that I get counterfeits? They gave me much trouble, as Anatech knows...
(sorry, no pictures of them)
Check em against the pictures on http://sound.westhost.com/counterfeit.htm (at the bottom) - it is highly likely you got fakes though.
This counterfeit transistor problem is becoming a real epidemic.
I propose testing any newly obtained transistors with a jig that would separate a real from a fake.
For example, for an MJ15003 I would test the one thing that sets this transistor apart from others and makes it expensive: the SOAR curve.
Subject the part to simultaneous voltage/current combinations which are just inside the DC SOAR curve of the geniune part. The good will survive, the fake shouldn't.
You need to choose V/A combinations which lie in the moderate current/moderate voltage region (rather than hi V lo A or vice versa.) The near-axis extremes are often well handled by many devices, but it's in the middle of the graph that cheap 3055's will have difficulty. Because the MJ15003 has a more "bulging" SOAR curve in the middle, and cheaper parts are more shallow in that area.
Let me know what you guys think, maybe we can come to a concensus as to a "standard" screening test for fakes.
Adrian
I propose testing any newly obtained transistors with a jig that would separate a real from a fake.
For example, for an MJ15003 I would test the one thing that sets this transistor apart from others and makes it expensive: the SOAR curve.
Subject the part to simultaneous voltage/current combinations which are just inside the DC SOAR curve of the geniune part. The good will survive, the fake shouldn't.
You need to choose V/A combinations which lie in the moderate current/moderate voltage region (rather than hi V lo A or vice versa.) The near-axis extremes are often well handled by many devices, but it's in the middle of the graph that cheap 3055's will have difficulty. Because the MJ15003 has a more "bulging" SOAR curve in the middle, and cheaper parts are more shallow in that area.
Let me know what you guys think, maybe we can come to a concensus as to a "standard" screening test for fakes.
Adrian
In my experience, fake (OnSemi or Motorola) MJ15003s or MJ15003 from unknown manufacturers measure totally differently to genuine parts- The latter has a much more linear Ib vs Ic curve.
You would need a high current power supply, function generator, oscilloscope, base and emitter resistor (say 100 ohm and 1 ohm) and a buffer (e.g. BUF634) to perform the test.
The best way to ensure you get genuine parts is to buy direct -- from OnSemi, Fairchild, Linear, ......
Most difficult to locate reliable souce are Toshiba FETs and BJTs. I have also had fake Toshiba's; but they did not escape the linearity test.
Patrick
You would need a high current power supply, function generator, oscilloscope, base and emitter resistor (say 100 ohm and 1 ohm) and a buffer (e.g. BUF634) to perform the test.
The best way to ensure you get genuine parts is to buy direct -- from OnSemi, Fairchild, Linear, ......
Most difficult to locate reliable souce are Toshiba FETs and BJTs. I have also had fake Toshiba's; but they did not escape the linearity test.
Patrick
I'm trying to locate some of these MJ 15003/4 or
MJ 15015/16 for a customer and have located
some MEV as the manufacturer, just followed the
link to Rod Elliot's page and saw there were fakes
of the 2N 3773 with the MEV name on it. Before buying
these I'd like to know about the company MEV and if
these could be fake. Or experiences with usage of MEV items.
Steve @ Apex Jr.
MJ 15015/16 for a customer and have located
some MEV as the manufacturer, just followed the
link to Rod Elliot's page and saw there were fakes
of the 2N 3773 with the MEV name on it. Before buying
these I'd like to know about the company MEV and if
these could be fake. Or experiences with usage of MEV items.
Steve @ Apex Jr.
In my part of the world, I am unable to buy direct from the mfr, (unless I buy 1000).
They will send me to a distributor.
One of their most trusted distributors, however, Digi-Key, has been reported to ship counterfeit MJ15003/4 to a reader who was writing about this in a counterfeit semi forum.
So I don't know what to think any more.
Digi-Key will accept returns on semiconductors, if you claim they are fake, so there may be no money loss there, but it still puts the burden of testing on the buyer.
Adrian
They will send me to a distributor.
One of their most trusted distributors, however, Digi-Key, has been reported to ship counterfeit MJ15003/4 to a reader who was writing about this in a counterfeit semi forum.
So I don't know what to think any more.
Digi-Key will accept returns on semiconductors, if you claim they are fake, so there may be no money loss there, but it still puts the burden of testing on the buyer.
Adrian
I've always used Allied Electronics
http://www.alliedelec.com
...don't know of any connterfeits from them, but I guess anything's possible.
http://www.alliedelec.com
...don't know of any connterfeits from them, but I guess anything's possible.
I got burned with fake MJ15003/4 too!
I stupidly bought some cheap MJ15003 and MJ15004 from China on eBay, should have known better. The supplier is still being helpful and replying to emails at the moment but I haven't got my replacement or refund yet, I'll keep you posted.
I have checked all the transistors they sent me, the 200 transistors labelled ON Semiconductor and the 20 labelled Motorola. I am afraid none of them are genuine. They are all a cheaper, lower specification TO3 transistor labelled as the high spec MJ15003 and MJ15004. I'll bet they have been taken from the rejects bin of a much lower spec transistor product run. As a quick check I found it is possible to tell a genuine one from a fake one by weight. On my cheap kitchen scales, the genuine one weighs about 12 grammes, a fake one weighs only about 10 grammes due to the lack of big heat spreader. This is no replacement for a proper voltage and power test though. I tested the power handling of the transistors using the rig pictured below. This set up dissipates about 30 Watts max and carries a max current of 8.5 Amps through the transistor. it runs off a 12 volt lead acid to get the Amps. An MJ15003/4 should stand 5 Amps at 50 Volts = 250 Watts when bolted to a good heatsink. The internal connection wires on the fakes are not big enough to carry the current let alone the silicon die. The transistors they supplied burnt out at about 18 Watts. The reverse breakdown voltage I tested was right on the rated limit of 140 Volts, a genuine MJ15003/4 would stand much more than that for a good safety margin.
The fakers rely on most small electronics firms not having a high current and high Voltage transistor tester. I made myself a tester that checked transistors at 3 Amps, I thought this would sort out the fakes, how wrong I was. Ordered 20, they passed with my checker so I ordered 200 at 75p each. Put some in a 250 Watt amp and they popped like a bowl of giant rice crispies! My tester worked at about 3 Volts at 3 Amps = 9 Watts - This is not enough. You need to bolt them to a big heatsink and put a few more Watts through them.
I stupidly bought some cheap MJ15003 and MJ15004 from China on eBay, should have known better. The supplier is still being helpful and replying to emails at the moment but I haven't got my replacement or refund yet, I'll keep you posted.
I have checked all the transistors they sent me, the 200 transistors labelled ON Semiconductor and the 20 labelled Motorola. I am afraid none of them are genuine. They are all a cheaper, lower specification TO3 transistor labelled as the high spec MJ15003 and MJ15004. I'll bet they have been taken from the rejects bin of a much lower spec transistor product run. As a quick check I found it is possible to tell a genuine one from a fake one by weight. On my cheap kitchen scales, the genuine one weighs about 12 grammes, a fake one weighs only about 10 grammes due to the lack of big heat spreader. This is no replacement for a proper voltage and power test though. I tested the power handling of the transistors using the rig pictured below. This set up dissipates about 30 Watts max and carries a max current of 8.5 Amps through the transistor. it runs off a 12 volt lead acid to get the Amps. An MJ15003/4 should stand 5 Amps at 50 Volts = 250 Watts when bolted to a good heatsink. The internal connection wires on the fakes are not big enough to carry the current let alone the silicon die. The transistors they supplied burnt out at about 18 Watts. The reverse breakdown voltage I tested was right on the rated limit of 140 Volts, a genuine MJ15003/4 would stand much more than that for a good safety margin.
The fakers rely on most small electronics firms not having a high current and high Voltage transistor tester. I made myself a tester that checked transistors at 3 Amps, I thought this would sort out the fakes, how wrong I was. Ordered 20, they passed with my checker so I ordered 200 at 75p each. Put some in a 250 Watt amp and they popped like a bowl of giant rice crispies! My tester worked at about 3 Volts at 3 Amps = 9 Watts - This is not enough. You need to bolt them to a big heatsink and put a few more Watts through them.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.