watch out for counterfeit transistors

i'm compiling a counterfeit transistor identification guide. i've seen a lot of these in the last few years, and the counterfeiters are learning from their mistakes. these transistors are usually knock-offs of popular audio output transistors (such as the 2SA1943 and 2SC5200) or TV horizontal output transistors, or any other semiconductors that are normally in the medium-to-high price range ($5.00 US and up).

generally these devices are cheap devices bought in bulk, then have their original part number removed, and have it replaced by a part number consistent with a higher current/voltage/wattage device. for instance an audio output transistor rated at 200W, 300V, and 15A is counterfeited by taking a similarly packaged transistor (MT100 case style) that is only a 150W, 200V, and 10A device. the original part number is shaved, sanded, wire-wheeled or chemically treated to remove the original ink, or in some cases painted black to cover up the original number. the new number and manufacturer logo, as well as date and beta codes are printed on it. these then are advertised and sold as (for instance) 2SA1943's that are at bargain basement prices.

this does two things. the first thing it does is make lots of money for the counterfeiter. if he bought 1,000,000 pieces for $0.10 of a discontinued or factory reject part, and shaves and re-labels the devices for $0.10 each (total invested=$200.000.00), and sells them for a distributor for $1.00 each, he just made $800,000.00 (and if the faker makes the same deal with 9 others, he made $8,000,000.00).

so the distributor advertises the transistor for $5.00 each, which is $2.00 cheaper than the real one at $7.00 each. here's where the second part of the double-whammy comes into play. the real transistor sells for $7.00, the fake one for $5.00, and 10,000,000 of them just hit the market. well the $7.00 price tag for the real one was "what the market would bear" which influenced the $5.00 price tag for the fake one. now the market is distorted by the sudden glut of 2SA1943s, and just to make sales quotas, the distributors with the real ones have to cut their price to $4.00 each, and then the distributors of the fakes cut their price, and it's a vicious spiral downwards, causing both the distributors and manufacturer of the real ones to take a big financial hit, and also the distributors of the fakes take a big hit from having to drop their prices.

meanwhile service companies are happily scooping up the fake transistors by the truckload, installing them in repairs and charging the end user for the repair. of course the service company has at least a 30-day guarantee on parts and labor.

the happy (at first) customers take home their amps and the first thing they do when they get home is hook it up and turn it up all the way. the amp lasts maybe 10sec, maybe an hour, maybe a week, before zzzzzzzzztpoofffff, the transistors do what comes natural when they're taken beyond their limits and let out the magic blue smoke. now the happy customer becomes an irate one and demands their amp be repaired again for free. so the service company has to buy the transistors again, and complete the repair again. the customer takes it home again and zztsnap, and now there's a vicious cycle until the service company a)gives up and refunds the customer's money, or b) the service company happens to get a pair of real transistors and fixes the amp. during this process the service company may get some of their money back from the distributor, maybe not. or they may figure out that there are counterfeit 2SA1943s out there and find a distributor that carries the real thing. either way the service company has wasted a lot of money in the process.

some distributors may seek out a source of genuine parts, but not before it has become an expensive lesson to learn. other distributors won't care as long as they meet their sales targets.
-----------------------------------------------------------------------------
some things to do when you get a replacement part from a 3rd party source:

1)inspect the original and the replacement carefully side by side. look for slight dimensional differences.... mounting hole diameters, heat sink tab and case dimensions, pin thicknesses and shapes.

2) look for differences in print formatting, letter size, font style, logo, date code format, line thickness, ink density. (sometimes the printing on the fakes looks cleaner and nicer than on the real ones.

3)some fake print washes off with acetone.

4)look for tooling marks, scrapes, scratches, rounded edges, any signs of abrasion or erosion on the surface of the device.

5)electrical testing. check B-E forward voltage (at a constant temp) and compare against a known good one. check C-E capacitance, as this is a fairly linear approximation of the surface area of the silicon inside. again compare against a known good one. the capacitance should be within about +/- 10% of a known good one. curve tracer (if you have access to one),. a curve tracer is an almost instant go-nogo test for comparing transistors (a real transistor curve tracer that shows the characteristic curves of the device, like those shown in a data sheet)

6)destructive testing (taking the case apart to view the actual semiconductor die). this takes a hammer or large bench vise, and a lot of finesse to separate the case from the heatsink tab and semiconductor die. a lot can be seen inside of the case. some fakers just re-label a lower class device, so the die size is usually smaller. while a 3mm square die might not look a whole lot smaller than a 5mm square die, look at it's area. a 3mm die is 9mm², and a 5mm die is 25mm². this matters a lot in current density. another thing that more resourceful counterfeiters (obviously with some kind of limited foundry facilities) is put two 3mm dies on the heat sink tab and parallel them. parallel bipolar transistors have a tendency to "current hog", and eventually one of the dies burns out because it has 90% of the current flowing through it during operation.

i'll get some pictures up as i get them.


one of my "pet" theories is that semiconductor counterfeiting isn't just about the money, but possibly about undermining market competitor's reputations for quality, or worse, as a form of economic warfare of one country against another.