Note: This blog uses some names of companies that are everyday names, and I am not implying that they have anything to do with anything illegal. I do not imply and I do not make any claim that any company is doing or has done any of the things I talk about.
Many people get their components from local electronic stores and in the U.S.A. it seems the most popular store is Radio Shack (there seems to be as many RS stores as there are McDonalds). Radio Shack might have sold “surplus” parts long go, but nowadays most of the parts they sell, especially silicon semiconductors such as transistors, are the same parts that are bought by the manufacturers of electronic equipment. You or I can buy a thousand transistors from Arrow, Avnet, Digi-key, Mouser, Newark, Farnell, etc., and pay only 3 or 4 cents, or even less, apiece. Radio Shack buys tens of thousands and packages them in packs of 2 or 3 and charges a couple dollars, and they are what they say they are, a 2N4401 for example. But people like to buy a “bargain pack” of ten or more transistors for not much more than the cost of the 2 or 3 pack. The transistors apparently are often marked with house numbers.
One weasel word I’ve seen used is 2N4401 Type transistor. The word type implies that the transistors may be 2N4401 or a similar type, but not guaranteed to actually be that part number. I say this because many or most times I see a project where the builder says the transistors are 2N2222, but the picture shows that transistor has a black plastic package. The true 2N2222 or 2N2222A came in a metal package, and then later manufacturers put the same chip inside of a plastic package and called it a similar name such as PN2222 or MPS2222. The only true 2N2222 comes in a metal package. But buyers of those bargain packs that say the black plastic transistors are 2N2222 type are being misled.
Where do they get these transistors? From manufacturers that had and order from a company that wanted their house number instead of the usual JEDEC (2N4401), ProElectron (BC337) or Japanese (2SC1815) part number? Often a company may have had a contract or a production run that didn’t use all of the parts, so they try to get what they can by auctioning them off. There are “grey market” parts dealers that acquire these, and somehow they end up in the “bargain packs” that some electronics dealers sell.
One doesn’t know if these house numbered parts actually work like factory parts or if they are “outliers.” For example, an equipment manufacturer has a Quality Control or QC department that receives all shipments of parts, and inspects them for quantity, quality, visually, etc. If the parts are to be used in a piece of important equipment (such as a pacemaker) they may test each individual part, or they may selectively test a few of the parts and use statistics to arrive at a failure rate.
Another example is you look at the parts list for a piece of test equipment, you may see the house part number and in another column they give the nearest equivalent. It may say “2N3055 specially selected”, which means, for example, that the company tested the transistor and found that even though the maximum rating of 60 volts was given in the data sheet, a high percentage of the 2N3055s would be 80 volts or more. So they select these and reject the ones lower than 80 volts and later these rejects get auctioned off to the “grey market” parts dealers. They still say they’re 2N3055, and they still meet the specifications given in the data sheet. But some experimenters may have been using the 2N3055s that will handle 80 volts or more, and when they buy some of these rejected parts, the 2N3055s will not meet their expectations and will not handle any voltage above 80 volts.
Some “bargain packs” also include regular JEDEC etc., numbered parts just to make sure that the buyer doesn’t get a whole package of duds. I have gone to the Radio Shack store and carefully looked at the transistors inside of their bargain packs. The salesdroid has on occasion interrupted his cell phone sales pitch to a customer (gotta make some money with a commission sales) to come over when he saw me shaking the packs to see what was inside. On occasion I found that some packs had several of certain transistors that I wanted. But I think others had the same idea; I found that the packs had been picked over and no longer had what I wanted.
Sometimes the luck would run the other way. I would check the bargain pack to make sure that the “2N4401 type” didn’t include the 2N3904 or BC547 transistors. These may be NPN and work like a 2N4401 in some circuits, but they cannot handle as much current as a 2N4401. I could then avoid bargain packs that had these. The reason I point this out is that the bargain packs have a wide variation on what the buyer is getting, and he might be fooled into getting something that is less than he expected.
The bargain pack of 20 transistors for three dollars amounts to 15 cents apiece, which is 3 or 4 times as much as the cost from a distributor. I just stopped buying the Radio Shack parts and went on eBay to find what I needed. Then I bid on and won “1000 transistors”, and when I received them, I found that the bag was smaller than another bag of 500 similar sized transistors. I counted them by putting them into piles of ten each, and found that there were a total of 400 transistors. I complained to the seller, who claimed he didn’t know (he must not be very observant), but settled by refunding me 60 percent of the cost. I made eBay purchases of other used stuff, such as test equipment; more often than not I got less than I had expected and I came to the conclusion that many of the sellers on eBay were just as bad as some of the grey market sellers – or else they are, or they act like they don’t know anything about it, aka “plausible deniability”. I stopped buying parts from eBay and have since been buying parts from the companies I mentioned earlier, and found that Mouser often had better prices. But I have counted parts and found that they had shipped me parts short anywhere from 1 to a few dozen out of a hundred. Often these were easy to count because they came on tape, so I wonder why they couldn’t make an accurate count. Duh. I still buy parts from them but I’m careful to count some to make sure I’m getting what I paid for.
I have a transistor tester that I can use to find out how well a transistor performs (inexpensive DMMs have a transistor checker, too). The tester shows that there may be a wide difference in the performance of the transistors in the bargain packs. I cut off the lid of a corrugated cardboard box and slit it with a utility knife. This left the corrugations exposed, so that each corrugation could hold a transistor or diode that I had tested and I could write on the cardboard what the test result was. I also built a test jig with a transistor socket and rectified AC with a current limiting resistor, and jacks for my multimeter. I could use this to measure the breakdown voltage of each transistor, up to the peak AC voltage, which is about 160 volts..
I also can see what’s going on in a circuit with an oscilloscope so I know what kind of transistors I’m getting and using. I have plenty of experience with issues caused by low quality and defective parts. I should also add that sometimes it is easier and more productive if I do a repair by replacing most or all of the parts (this is known a shotgunning). This is especially true with older equipment that may have a bad electrolytic capacitor. If one went bad, it is likely that others are getting old and drying out and will soon go bad even if they haven’t already.
Fakes And then there are those that sell counterfeit parts. Some unscrupulous dealer may remove the part number and replace it with the part number of a higher rated part and then sell it for more, making a lot of money. Sometimes this happens because the manufacturer has discontinued making the part, and the grey market dealers have to fake the part. I have read about this happening with the power transistors used in the output stages of high power audio amplifiers.
When I worked for a company that made test equipment, they received bad CMOS chips from a distributor. We assembled the equipment and put the units into burn-in over the weekend. On Monday morning most of the units had failed because of one particular chip. Those chips had to be removed and replaced. But the problem was that CMOS chips were in so high a demand that they had a lead time of the better part of a year, so if the company ordered more chips when they found out they were bad, it would take longer than six months to receive replacements. In other words, the company was having a difficult time getting the orders shipped and hence making the money they would otherwise make. I learned that they had to back date some things, like a month suddenly had more than 31 days! These fakes just show us that when there is a crooked way to make money, someone will find out and take advantage of other unfortunate suckers.
Back to experimenting…