Fake Chips

Hi all,

Is there any reliable method for testing parts for authenticity? I just bought a bunch of LM317T variable voltage regs off Ebay (from a dealer with a solid rep, admittedly - but they could themselves have been fooled) and would like to know if they're genuine or not, as a quick search on the net showed there *are* fake LM317s around.

Would one be safe to assume that any fake part would not be capable of performing at the maximum levels claimed by genuine manufacturers in their datasheets? For example, the 317 is claimed to be good for 1.5A, but if mine burn up at say 1.3A, is that a clear indication that they're fake?

Your thoughts, gentlemen....

Reply to
Cursitor Doom
Loading thread data ...

Are you willing to do something destructive? Remove the epoxy packaging an d look at the chip. For a popular part like an LM317, you can probably fin d a picture of a genuine chip (maybe as they were made 40 years ago). Othe rwise, you could compare it to a known good part.

Some manufacturers have active programs to find counterfeit parts and libra ries of information ueful for identifying thenm (like the way the logos and the numbers got printed). You might contact them with pictures of the pac kaging and the silicon chip.

My buddy at Intel told me that one of the common problems is "NOS" Mil-spec versions of obsolete parts that were really repackaged pulls. They often seem OK, except that they have a higher wear-out failure rate (apparently, catastrophic infant mortality failures are relatively rare, as are the comp letely wrong chip which had been re-marked).

A priori, I would not automatically assume a failure at 1.3A instead of 1.5 A was sufficient evidence, especially in a "new" design with no history of proper cooling, etc. You might have a case if it failed in a known and mat ure design.

Reply to
jfeng

** Authenticity can only ever be proved by tracing the supply path of the device back to the actual manufacturer. Buying from authorised dealers is the only way to be able to do this.

If a device fails to meet specs for the part number, then it is probably "fake". There is no simple test.

** When you buy parts from Ebay, you need to stop worrying about "authenticity" - cos most of them are not.

FYI: The terms "fake" and "counterfeit" refer to parts that are either rejects or relabelled from another much cheaper type that merely used the same package.

.... Phil

Reply to
Phil Allison

No.

High-reliability (aviation, medical, military) parts all have to have traceable sources, and that means paperwork as much as inspections. There have been a variety of counterfeits seen, from currency to rocket struts.

You can only buy an authentic Acme toe-hammer direct from Acme or through a currently-authorized Acme distributor. And, file all of the receipts (and be prepared to ask the distributor for HIS receipts). And invoices. And shipping labels. And purchase orders. Inspect repair/calibration facilities to be sure no one THERE is skimping on the paperwork, or performing procedures that aren't documented formally, or using metric hammer handles for the imperial hammer model.

There are formal procedures for authenticating Bitcoin, but not real physical parts.

Reply to
whit3rd

in my experience with fake parts, they usually have nothing to share with the real one. Usually they bother to fake expensive parts (like out of production mosfets or BJT for example that cost many $$ a piece). I have fake mosfets that really are inexpensive switching high voltage parts, even the pinout is different, so it's easy to spot. BJT are usually much different dies, a quick curve tracer inspection shows them. I guess that nobody would really bother to fake an LM317, it's inexpensive and still in production, so why even think about that? The only gain would come from marking as LM317 some empty TO-220 parts, and that would be easy to spot :-)

read the datasheet, most of the values are "typical".

Best regards Frank IZ8DWF

Reply to
frank

You would think so, wouldn't you? But someone's done it:

formatting link

Reply to
Cursitor Doom

On one chip Acetone took of all the markings in one swipe.

--
Boris
Reply to
Boris Mohar

that's what I said... empty case is the only way to gain from fake LM317. If it even does what's supposed to do, it's not likely to be a fake.

Frank IZ8DWF

Reply to
frank

Wherever possible I try to source parts from a reliable vendor. In the USA, that means (alphabetically) vendors like Allied, Digikey, Mouser or Newark. Companies that buy their parts from the original manufacturer. I don't mind spending a bit more for parts that I believe will be reliable, as I like avoiding reworks, and wondering if the problem was my diagnosis or the new part.

Another theory I have is how subcontractors handle duds. Let's say a major optical laser maker subcontracts with "Dumb-Luck" electronics to make 10,000 CD player lasers (optical blocks) for them. Dumb-Luck knows they have to make 13,000 to come up with 10,000 that meet spec. They know they are SUPPOSED to destroy any others, but instead sell them as new "A" stock lasers. They look like real ones, because they are, they just don't meet spec or work at all, and we have no way of knowing. They think they are being good businessmen by getting money for their scrap.

Regards, Tim Bristol Electronics

Reply to
Tim Schwartz

Not reliable, but you can get some clues: Checking for differences in package features and labeling is what I use. It won't determine if the device is authentic, but will detect if there's something obviously wrong or suspicious. An acetone Q-Tip wipe is a good way to test the labeling. I'm not ready to buy an X-ray Fluorescence machine to check for material differences.

Sure. Just build a test fixture for each part and see what it can do. For individual xsistors, a curve tracer might do. For your LM317, a load tester. That's fine for large quantities. However, if you just need a few parts for a repair job, such an approach is impractical. If you can't afford to buy from a reputable vendor, then I suggest you buy from two or more eBay vendors, in the hope that at least one of them might be functional. My luck with this method has been good, but not perfect. I recently needed some 7812CK (TO-3 package) regulators. I bought from two different eBay sources and received obvious counterfeits from both. I then reordered the real part from Mouser, which worked.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

Well, I bought some "Xilinx" 9500-series CPLDs from a Chinese outfit, when I got stuck with a batch of boards and the chip was no longer available.

They mostly worked. I think that they were untested die. A small number, about 5 - 10% would not accept the configuration - they were totally dead. If they did configure, they worked fine.

Well, I don't KNOW these were fake, but the fact that I'd never had more than one Xilinx part ever come up totally dead, lead me to suspect they were. I have no idea whether these were reject wafers, or somebody got hold of the masks and made a batch.

Oh, actually, there was one other detail that pointed to fake. There was no molded-in pin 1 dimple on the chip. All Xilinx parts did have the dimple.

Jon

Reply to
Jon Elson

Ah, yes, the Darkness-Emitting Arsenide Diode, or "DEAD". Seems that was an April 1st story in a respected electronics trade mag, quite some years ago. It was a way to get 100% yield out of the foundry!

But, the above scenario was the genesis of Poly-Paks, about 40 years ago. Floor sweepings, and YOU get to test them!

Jon

Reply to
Jon Elson

I remember seeing the data sheet for the DED in the early 1970s. I think it was from Litronix, but it could have been either Signetics or Monsanto.

Reply to
jfeng

The one from Signetics was the Write-Only memory

formatting link

Reply to
jfeng

Back in the late 90s there was a company called Surburban Electronics that sold fake STR regulators that didn't regulate and supposedly the switching transistors used in Sony tvs. The fake transistors set on fire in under a second. Still don't understand why it was worth counterfeiting such cheap parts.

--
This email has been checked for viruses by Avast antivirus software. 
https://www.avast.com/antivirus
Reply to
Chuck

Suburban still exists IIRC, I just don't buy from them unless it's an obsol ete flyback or something obscure.

I will buy cheap ICs if they're current production but sell in small quanti ties, such as LED controllers and such. Some devices I would hope are just not worth the trouble to counterfeit. Simple three legged devices like tr ansistors, IGBT, mosfets etc are too easy to counterfeit that I won't buy t hem generically unless I simply cannot source them from a reputable supplie r.

Speaking of Sony, how many MCZ3001 smps controller chips have you seen dest roy parts as soon as power is applied? The only reliable source I know of is B&D Electronics. Supposedly, Electronica also has reliable MCZ3001 chip s but since these TVs are pretty much gone, it's no longer a big deal.

Reply to
John-Del

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.