Counterfeit parts..

[hamster_nz]

I have been reading an interesting PDF on counterfeit parts.

https://www.cti-us.com/pdf/CCAP-101InspectExamplesA6.pdf

Lots of pictures and examples.

As an example of how little I know, I had never heard of lead extensions before (I am sure they are old hat to some here!)

[ataradov]

Power transistors are commonly recycled this way. But on that case they are actual genuine devices. They are just being sold as new instead of used.

[wraper]

Power transistors are commonly recycled this way. But on that case they are actual genuine devices. They are just being sold as new instead of used.

This might be genuine as well. Chinese replace marking not only to necessarily put fake part number but also to make bunch of parts with proper part number but fake batch number and fresh date code. So parts appear new and from single batch.

[MyHeadHz]

I've wondered about counterfeits for a while.  Silly questions like whether they functioned at all and whatnot...   That PDF answers those questions and a lot more.  Thanks!

[dzseki]

I've wondered about counterfeits for a while.  Silly questions like whether they functioned at all and whatnot...   That PDF answers those questions and a lot more.  Thanks!

It completely depends on luck. It might be an original part but just relabelled to appear to be new (best case), the sometimes they upspec the part by relabelling (at least still functional), and of course it can be a totally random part relabelled.
The best I've seen so far was an IC that should have been in a 16 lead CDIP package, instead it was in a 18 pin CDIP package 

[wraper]

The best I've seen so far was an IC that should have been in a 16 lead CDIP package, instead it was in a 18 pin CDIP package 

I've seen one stepper driver from Toshiba relabeled as another model from Toshiba too, pure sabotage. Somehow gut feeling (was nothing wrong with appearance) made me check terminals with multimeter. And part turned out to have GND pins in different positions, etc. Saved me burned PCB traces. It was not even sanded, apparently their removed ink marking with solvent from older part made before they started to use laser, and applied new marking with laser.

[SiliconWizard]

 

[Gyro]

What's the betting on the Rubycon one inside being a fake too! 

[james_s]

What's the betting on the Rubycon one inside being a fake too! 

In my mind I'm picturing an even smaller capacitor stuck inside that empty can 

[wraper]

I doubt that capacitor picture is real. It's like two decades old and with unknown source, and the only picture I know showing smaller cap within snap in capacitor. Usually counterfeiters just put fake sleeve on some crappy or salvaged capacitor. Though there are known cases when radial capacitor is placed into barrel with rubber plug to make it an axial capacitor which are barely produced nowadays

[wraper]

What's the betting on the Rubycon one inside being a fake too! 

Have you ever even seen a real Rubycon in blue color?

Dunno as of currently produced but I'm certain there where blue Rubycon in the past.

[ANTALIFE]

I recently ran into a couple of counterfeit FRAM IC's. They acknowledged all I2C commands but when it came to actually reading the data from any address it would always return 0x00

https://www.eevblog.com/forum/projects/mb85rc64ta-fram-trouble-writingreading/

[peter-h]

I've had fake Hitachi H8 processors.

The lead frame was maybe 0.5mm bigger so they (OTP H8/332) didn't fit into the programmer very well.

They were just an empty package with leads.

We got about 10k of them, from one of the used stock cowboys in the US. They obviously didn't know they got this stock. The packages were labelled fairly well but not identically to Hitachi.

The funny thing was that there was a type on the P/N on the package - something like H8/342 - which didn't exist. It's like the old joke about currency counterfeiters printing some £7 notes

Luckily we had a credit account with the supplier, so we didn't get conned out of some $50k. They cancelled the credit account immediately... and we never bought from any other US used stock sellers afterwards unless they offered credit.

[peter-h]

Yes, USD 5 was a good price, when the UK disti price was GBP 9.

[jmsc_02]

I also like this presentation. it is very well documented and very easy to follow: https://mttc.jpl.nasa.gov/files/NASA%20Counterfeit%20Training_unlimited%20distribution%20handout.pdf

[peter-h]

When it comes to microprocessors, the counterfeit ones are usually non functional - just an empty package.

Making a functional micro is a huge amount of work. The chinese almost never do anything difficult if they can make a fast buck easily.

The way you do it is you buy say 10k of some part from a disti. You then return it for a refund. But you return to them the duff parts. They put the duff ones into stock and somebody else ends up with them. By the time it is discovered the source of the duff ones is long gone, having made millions simply by running a chip overmoulding and marking machine.

It is particularly nasty because the duff parts may be undiscovered for a long time, if your stock turnover is slow. So you can't claim under warranty because it may have expired. We used to always open the packages and program one device from each one.

Unfortunately returning stock is not difficult, if you pick the right channel.

[SiliconWizard]

I also like this presentation. it is very well documented and very easy to follow: https://mttc.jpl.nasa.gov/files/NASA%20Counterfeit%20Training_unlimited%20distribution%20handout.pdf

Fun to see that USA is in the top 10...

[DFC]

OK, I've read through this thread as well as both Pdf's.I don't know, maybe I'm nieve but I find it mind boggling the amount of effort put in to fake items where amount of money made doesn't justifiy the amount of work put in.Just think about it. How many pcbs has to be stripped to build up a reasonable amount of ic' s to sand, sandblast, then to reprint whatever the hell it is they want to put on there. Not to mention the lead extensions. Just think about what needs to be done in order to achieve this at a scale where it becomes worthwhile. I don't think anyone in their right mind would dispute the major source of these faked parts which is kinda sad especially after reading all 32 pages of the "Trade war" thread.

[ataradov]

Well, everything is possible if you have unlimited amounts of virtually free labor.

[2N3055]

What's the betting on the Rubycon one inside being a fake too! 

Have you ever even seen a real Rubycon in blue color?

I did, but some years ago. They have current line HBX in blue but different shade..

[james_s]

When it comes to microprocessors, the counterfeit ones are usually non functional - just an empty package.

I think more often they take whatever useless random IC they can find cheaply in the right package and re-mark it as a microcontroller. I recall several years ago one of the maker oriented companies got screwed with a batch of AVRs that were actually motherboard buck regulator ICs.

[all_repair]

Yes, USD 5 was a good price, when the UK disti price was GBP 9.

Likely your supplier sourced around.  And at that qty and price, it was a good invitation to laser marked something to fulfill your order.  Your supplier also got hit, likely their offered price was less than USD2.50, or maybe even less than USD1.00.  The one who actual laser marked is usually not the one making the bulk of the profit.  They are paid about the labour rate.  The last 2 contacts before the actual buyer normally take the largest cut.

I got hit for mostly my MOSFET.  I suspect some I bought from E14, DigiKey and Mousers are also fake.  Ended up those 2nd hand MOSFETs from China are more reliable than the brand new.  I did not keep good record for them, mostly used for repair.  New one are failing after 1 to 3 hours.

[JackJones]

When it comes to microprocessors, the counterfeit ones are usually non functional - just an empty package.

I think more often they take whatever useless random IC they can find cheaply in the right package and re-mark it as a microcontroller. I recall several years ago one of the maker oriented companies got screwed with a batch of AVRs that were actually motherboard buck regulator ICs.

It was Sparkfun, ATmega328s turned out to be NCP5318 buck regulators.

Story in four parts:
https://www.sparkfun.com/news/350
https://www.sparkfun.com/news/364
https://www.sparkfun.com/news/384
https://www.sparkfun.com/news/395

Interestingly the die inside wasn't a released version but "pre-release engineering material". Makes me wonder how someone got their hands on them?

[peter-h]

AFAICT the main supplier we had never got hit. The duff parts came from a new source.

The reason so many H8s were being sold off is most likely one of

- a large user sold off excess stock

- Hitachi's last time buy announcement forced stockists with stock to dispose of it at any price

It was a 1980s-era processor after all

The stock was about 10 years old. Perfectly packed. There is a lot of good stuff to be found, mostly in the US. But you have to be careful and never pay in advance for any large purchase.

[Deodand2014]

A video discussing how to determine if a particular chip is rebadged or not.



Basically, acetone is your friend...