Hfe of of 8. Sorry, defective parts. 2N5550's were below spec, but just barely, so I would say just crappy parts. Replacements were all within spec. Low end, but within.
Maybe this counterfeit subthread can be spun off to a separate thread, like the cars thread was? On to the topic:
If the "source" of a part is not an authorized distributor for the manufacturer of the part you're buying, don't buy the parts from that dealer. Buying from them may be "convenient" like the local mini-mart convenience store, but it's clearly NOT convenient if you get a fake. It appears the chance of getting a fake this way is high and increasing.
If you get a fake from an authorized distributor, immediately tell THE MANUFACTURER all the particulars, what part and what distributor, send the manufacturer one of the fakes, and let them get their attorneys and what-not on it to clear up its relationship with the distributor. Doing these would give manufacturers and distributors plenty of motivation for making sure fakes don't get into the distribution system, as if they didn't already have enough motivation.
Yes, there's huge amounts of money involved (both to be made in the crime, AND money and lives to be lost on the end-product side), but it seems if people went through proper procedures (even taking a distributor to court when the wrong part arrives, as it can cost substantial money in delaying a pro this would be a lot less of a problem.
The Raspberry Pi site shows two problems related to this - the manufacturer (board stuffer) "helpfully" substituted a cheaper Ethernet connector without a built-in transformer (what the article calls "magnetics"), resulting in a non-isolated Ethernet connection
Manufacturing hiccup | Raspberry Pi
And there's this so-called "China Export" symbol that looks remarkably, confusingly similar to the European CE mark:
Easter picture post | Raspberry Pi
Wikipedia has this covered (it appears the diagram in the article is taken from here):
CE mark - Wikipedia, the free encyclopedia
There's obviously big problems when parts are shipped to subcontractors and other countries to be stuffed, as while the parts are in "someone else's hands" cheaper ones can be substituted, but what comes from the distributors should be good.
If the "source" of a part is not an authorized distributor for the manufacturer of the part you're buying, don't buy the parts from that dealer. Buying from them may be "convenient" like the local mini-mart convenience store, but it's clearly NOT convenient if you get a fake. It appears the chance of getting a fake this way is high and increasing.
If you get a fake from an authorized distributor, immediately tell THE MANUFACTURER all the particulars, what part and what distributor, send the manufacturer one of the fakes, and let them get their attorneys and what-not on it to clear up its relationship with the distributor. Doing these would give manufacturers and distributors plenty of motivation for making sure fakes don't get into the distribution system, as if they didn't already have enough motivation.
Yes, there's huge amounts of money involved (both to be made in the crime, AND money and lives to be lost on the end-product side), but it seems if people went through proper procedures (even taking a distributor to court when the wrong part arrives, as it can cost substantial money in delaying a pro this would be a lot less of a problem.
The Raspberry Pi site shows two problems related to this - the manufacturer (board stuffer) "helpfully" substituted a cheaper Ethernet connector without a built-in transformer (what the article calls "magnetics"), resulting in a non-isolated Ethernet connection
Manufacturing hiccup | Raspberry Pi
And there's this so-called "China Export" symbol that looks remarkably, confusingly similar to the European CE mark:
Easter picture post | Raspberry Pi
Wikipedia has this covered (it appears the diagram in the article is taken from here):
CE mark - Wikipedia, the free encyclopedia
There's obviously big problems when parts are shipped to subcontractors and other countries to be stuffed, as while the parts are in "someone else's hands" cheaper ones can be substituted, but what comes from the distributors should be good.
When I was a QE back in industry we tested samples of everything. We almost never found a defective part. (Well Unitrode z5u's) I ran the corrective action for a FA lab. Just about every in-board failure could be traced to an application error or handling error. Parts were good. We did happen to use a lot of 2907's and 2222's. That was the 70's and 80's BC (Before China). To be clear, the very same foundries are making the top quality parts.
I remember Soviet transistors 2T-208 designed especially for some kind of switching purposes. They were almost reversable. However, it was not reflected in "civil" datasheets, but break-down voltages and betas were almost the same.
Toshiba made a part for shunt clamping that had a high reverse beta. I'm not sure if it is still around.
EDIT: the 2SC2878. A base-emitter breakdown voltage of -25V, a 1 ohm on resistance at Ib = 5mA. Reverse beta of around 200.
EDIT: the 2SC2878. A base-emitter breakdown voltage of -25V, a 1 ohm on resistance at Ib = 5mA. Reverse beta of around 200.
Last edited:
About two years ago, I purchased some 7815/7915 v. regs, I needed them for some low tech power supplies. I don't remember whose they were supposed to be. Anyway, I have been using them for so long, I took them for granted. I installed them and - nothing. Dead as a door nail.
After some juggling around, I discovered that the pins had changed places. But they kept the original markings. God knows what that was and from where. Testing them with their pins changed, they worked perfectly. Still do. Fortunately, I bought only two pairs.
Call me conservative, but I don't like that type of surprise.
After some juggling around, I discovered that the pins had changed places. But they kept the original markings. God knows what that was and from where. Testing them with their pins changed, they worked perfectly. Still do. Fortunately, I bought only two pairs.
Call me conservative, but I don't like that type of surprise.
Last edited:
Actually that is not so unusual. Different manufacturers have produced these 'standard' devices with different pinouts. That's the trap if you check the pinout on a data sheet: use the wrong manufacturers sheet and you'r ehh, done in.
Same happens with small signal transistors, like the BC5xx types popular in Europe. There are at least two different pinouts for the nominally same device.
jan
Well, frankly, that was the first and only time I ran into that.
I've never ran into this with transistors, but on the other hand, I always buy transistors from Germany and in some volume (no biggie given their almost ridiculous price), because I need larger populations to pair them up. I make it a point that my BC transistors must be from either Siemens or Philips, period.
That bit me once too. It appeared that a device was knocked off, possibly incorrectly, designed into some product that became quite successful, prompting the original manufacturer to issue a new version with a slightly different pinout matching the knockoff, so that they could participate in the business!
Does anyone remember the good old days, when Motorola would have the pinout of bipolars in debossed letters on the TO-92 package? This led someone in an engineering stockroom at UCLA to suppose EBC was the name of the manufacturer, and had labeled a bin of them thus 😀
Now, that I do know.
When Medium Density Fibreboard first appeared in the mid-seventies, some illiterate customs official received a package from Japan full of MDF. He had no clue what it really was, and "Medium Density Fibreboard" was too big for him to swallow. But he saw the "Made in Japan" logo, so he called it "Mediapan". To this day, it is locally known as "mediapan"; say MDF and only the loudspeaker freaks will know what you're talking about, say "mediapan" and everybody knows what you are talking about.
I remember the part now --- one of the SM versions of the TL431, the bog-standard shunt regulator.
Remember "MADE IN USA"? Usa being a place in Japan, back when Made in Japan was more-or-less a pejorative.
Remember "MADE IN USA"? Usa being a place in Japan, back when Made in Japan was more-or-less a pejorative.
I once fried some VN2222 FETs before discovering that there were 2 of different devices from different manufacturers, including different breakdown voltages.
In which case I have to ask - how come two different devices have the same name?
As I understood you, they were only roughly the same, i.e. similar.
One more example:
LM4562 - Dual High Performance, High Fidelity Audio Operational Amplifier
http://semicon.njr.co.jp/njr/hp/fileDownloadMedia.do?_mediaId=9720
Both are opamps, and both were targeted on audio market.
I think this is a dfferent matter. NJR is far too big and serious a company to rip anybody off, and/or stoop to making low cost copies. I think this is a matter of licencing, quite common practice in the semiconductor industry.
Just as Motorola/ON Semi MJL 3281 is a licenced version of Toshiba's 2SC3281 (also somewhat improved, instead of 150W as the original, the MJL is rated at 200W, a welcome improvement).
As a matter of fact, and I think Scott and John can confirm this, quite a number of op amps we normally associate with for example AD were in fact designed by PMI (Precision Monolithics Inc.), but since they just design and have no manufacturing facilities, others actually manufacture them. The venerable OP27 and OP37 were in fact designed by PMI, if they even exist still.
LF 411, for example, is manufactured by National Semiconductor, ITT, Texas Instruments and I think ON Semi as well (but not sure). Cross licencing is an everyday occurrence in the industry.
No, it is not a matter of licensing. It is a matter of using of the same number for different opamp. Like in case of VN2222, use of the same name for different MOSFET.
There's licensing of parts, but I think very few part numbers are registered trademarks. I recall Intel trademarked 486 because AMD made earlier compatible processors, but then Intel went to Pentium, allegedly because a word is easier to protect than a number.
Cross licensing does happen a lot (most likely due to the proliferation of patents more than anything else), but I agree with Wavebourn, that's not necessarily what has happened when you see a different manufacturer of a pre-existing part number. I'm guessing there are too many part numbers to economically register and protect as trademarks. Even so, a 741 or 555 with different letters in front of it is arguably not the same trademark. But if they had had any idea how popular they would become, they would have had attorneys way back when trying to trademark the numbers 741 and 555, or else rename them something like"Ubiquitous OpAmp" or "Universal Timer" and trademark the names.
I recall the name "Transputer" was chosen for a microprocessor with the idea that that company's particular architecture would become as ubiquitous as transistors. It looks like that's happening with ARM.
There were the National Semiconductor manuals from about 30 years ago with about a hundred trademarks in the front cover, covering many features of their parts. Probably the only one anyone remembers is "Tri-State."
ETA: And there's two (at least) distinct points here: There's licensing of the name, and there's licensing of functionality. A "new" part with a similar name/number may be licensed for either, both, or neither from the original rights holder(s).
Cross licensing does happen a lot (most likely due to the proliferation of patents more than anything else), but I agree with Wavebourn, that's not necessarily what has happened when you see a different manufacturer of a pre-existing part number. I'm guessing there are too many part numbers to economically register and protect as trademarks. Even so, a 741 or 555 with different letters in front of it is arguably not the same trademark. But if they had had any idea how popular they would become, they would have had attorneys way back when trying to trademark the numbers 741 and 555, or else rename them something like"Ubiquitous OpAmp" or "Universal Timer" and trademark the names.
I recall the name "Transputer" was chosen for a microprocessor with the idea that that company's particular architecture would become as ubiquitous as transistors. It looks like that's happening with ARM.
There were the National Semiconductor manuals from about 30 years ago with about a hundred trademarks in the front cover, covering many features of their parts. Probably the only one anyone remembers is "Tri-State."
ETA: And there's two (at least) distinct points here: There's licensing of the name, and there's licensing of functionality. A "new" part with a similar name/number may be licensed for either, both, or neither from the original rights holder(s).
Last edited:
- Status
- Not open for further replies.