Abysmally Poor L7805CV Performance.. Counterfeit or "CV" variant limitation?

[wraper]

I suspect that the ebay 7805 what you tried might be a "reject" from the factory production and not an outright clone.
The fact that it "sort of works" is proof.

Yeah, dream about it. They are just some cheapest 7805 from some Wung Hung Low factory marked as ST. Most of the fakes sort of works, otherwise it would be impossible to run such business. And actually this it the worst part about counterfeits, if they just wouldn't work, the damage would be tiny, then they wouldn't find a way into the final product which is sold and then start to fail after some time.

I also suspect that the lettering at the end like A,C, CV, CT, ACT etc might have to do with their performance in a circuit.

Open the datasheet to find this is not a case.

[krish2487]

Quote from: wraper on Today at 07:58:32 AM

Why not??
They differ in their tolerances, temperature ranges, output current capability to quote a few.
If it is the same silicon die then

why not a "one ring to rule them all " kind of solution. 
Why different variants??
More importantly how are they meeting different specs??

(Before anyone tries to launch a seperate tirade, I do know what is component binning and how it works)

[wraper]

Why not??
They differ in their tolerances, temperature ranges, output current capability to quote a few.
If it is the same silicon die then why "one ring to rule them all " kind of solution. 

Yes it is the same silicon (if you are talking about parts from particular manufacturer).

[sleemanj]

If you spend any time on Taobao, you will see things like two listings for regulators from the same seller, one far far cheaper than the other. 

The difference is the cheap one is sold as "domestic quality" and the expensive one as "imported" or "100% genuine" or some such moniker.

Note the price of the 7805 on the left, and the price of the 7805 on the right, from the same seller...


The left one is an actual ST part...



The right one is a "domestic" part (read: fake)...

[Alex Eisenhut]

Here's a known good part from my Private Reserve. Note the accurate etching of the logo.

[wraper]

Here's a known good part from my Private Reserve. Note the accurate etching of the logo.

Looks genuine, but rather old part made in 1997, therefore font and marking format is different from newer parts.
http://application-notes.digchip.com/005/5-9794.pdf

[amyk]

Open them up and take some die photos if you can...

http://www.righto.com/2014/09/reverse-engineering-counterfeit-7805.html

[Mechatrommer]

Open them up and take some die photos if you can...
http://www.righto.com/2014/09/reverse-engineering-counterfeit-7805.html

thats the link i was refering to in the OP number (1)... the problem is i dont have the luxury he has. i hope with my report here, someone with the right tools will bother to get one or two of this fake and make the die teardown report...

latest test is i tried to feed the fake L7805CV with lower Vin through smps/buck convertor. i need to feed minimum around 6.5V to get stable 5Vout (under minimal or stable load), that is ~1.5Vdrop. but the transient servo respond still abysmal like before, so this doesnt help... but when feeding the servo direct from the 5Vout buck, with a little help from an output capacitor, the transient respond is super solid. +1 to china made cheapo smps/buck converter module...

[BravoV]

Too bad you're not living within neighborhood, other wise you just can grab few of these from my private reserve vintage collection. Old stuffs with shiny gold plated pins. 

[Fungus]

China has many legit component sources. Sometimes Alibaba/Aliexpress is just fine.

But China Ebay is full of counterfeit components, especially the types I listed: mosfets, simple ICs such as regulators, LEDs, and electrolytic capacitors. It's purely stupid to buy those, but you are free to do so if you want to.

There's good sellers on there, too. Maybe the problem is that you select "cheapest first" in the sorting options and pick the top of the list. I've bought hundred of lots of components on eBay and could count the times I've thought "junk!" on one hand (and even then you can usually get a refund, no questions asked).

[Fungus]

I suspect that the ebay 7805 what you tried might be a "reject" from the factory production and not an outright clone.
The fact that it "sort of works" is proof.

I keep on hearing this but is it really true? Are ST Micro (for example) producing so many bad parts that there's a thriving market based around raiding their dumpsters?

I'd expect ST to take a long hard look at their production lines if that was happening.

[Ian.M]

More likely its the usual far east scam of remarking something vaguely similar, as the premium product.   Do the clones meet spec if tested as a 78M05?

[krish2487]

It is a common practice.
Some of the biggest players in the silicon fab business are prone to falling prey to this.
Usually, it is more likely that unscrupulous employees operate and manufacture in the "off" hours at the plant, or a rejected batch is "missing" and surfaces as a legit component.


There have been many such reported incidents from TI and ST, IIRC there was one such similar case on this forum as well, though long back...




Quote from: Fungus on Yesterday at 07:18:02 PM

>Quote from: krish2487 on Yesterday at 07:50:54 AM

I suspect that the ebay 7805 what you tried might be a "reject" from the factory production and not an outright clone.
The fact that it "sort of works" is proof.

I keep on hearing this but is it really true? Are ST Micro (for example) producing so many bad parts that there's a thriving market based around raiding their dumpsters?

I'd expect ST to take a long hard look at their production lines if that was happening.

[Mechatrommer]

More likely its the usual far east scam of remarking something vaguely similar, as the premium product.   Do the clones meet spec if tested as a 78M05?

i dont have original 78M05 to compare with. but i also bought the immitation 78M05 from the same seller, the result is double abysmall as in the picture.... the logo instead of ST, looks like SI or 51. i guess there is area 51 factory in china doing this dirty business...

i guess someone should make a sticky thread on "the counterfeit ICs" compilation in this forum. there are discussion before about the same matter, namely FT2232 usb->serial, a CC1101 radio transceiver chip iirc...

[Tomorokoshi]

Part of the process of counterfeiting electronic components:

http://www.asq.org/asd/2009/03/compliance/counterfeit-parts.pdf

A lot of it comes from "recycled" electronic scrap.

[jitter]

yup i also noticed the weirdness of the logo, so i take photo to compare with my other ST's LM317 chip logo (see attached photo). someone may slam if my LM317 is also a counterfeit?

The "MAR" on your 7805 and the fact that they were sourced from China instantly made me suspicious.
On Chinese made ST components, I would expect to see "CHN" or "CHINA". "MAR" probably means they were made in Morocco (in some languages the first "o" is an "a" instead, in e.g. mine we would write "Marokko"). The "MRC" pressed in the little circle of your 317 likely also indicates Morocco as the country of origin.

The counterfeiters probably copied the print from a 7805 of Moroccon origin and put on a more recent datecode. But this datecode yields another clue: four digits instead of the three on the genuine item. Another thing: I haven't seen ST voltage regulators from Morocco since last decade. Not strange, the Moroccon fabs were supposed to be shut down in 2010. So "1418" on a Moroccon IC???

Your LM317 might be genuine, but I'd say the 7805 has "fake" written all over it.

For comparison some pics of ST voltage regulators from several vintages and origins. Except the L4805CV, all are from old surplus stock at our company and 99.999% sure to be genuine.
Note that the datecoding system seems to have changed around 2000. It looks like it was four digits buried inside the lot trace code before and later separated from it in three digit format.
Also note that despite origin or age, the ST logo doesn't seem to vary much between them.

Pics:
- A 7805 of unknown origin, presumably from the 18th week of 1992.
- A 4805 from Malaysia, presumably from the 15th week of 2002.
- A 7809 from Morocco, presumably 18th week of 2005, note: now three digits. I've also seen this format on other more recent ST stuff.
- And last but not least a Chinese 7812 of presumably the 9th week of 2000.