Counterfeit F103s

Hello,
Does anyone have any data on counterfeit F103s?

I have a blue pill that reports 128K, which is cool since it is marked F103C8. However, I have an STLINK V2 clone that is marked F103C8 and its flash size register contains 0xFFFF. I also bought some Maple Minis from Ebay (china) and even though they are marked F103CB, they report only 64K flash.

Is there a way to tell if chips we have are real or not? I ordered some F103CBs from Ebay, but now I suspect they will be fake as well. They were $1.90 each and the single piece price at Digikey is more like $6.43.

Thanks!

I would think if you really cared you would buy from an authorized distrubutor. The other alternative is to use the low cost Nucelo / Discovery boards. They would still be low cost (assuming you are in the US) but they will be made with real chips for sure.

If you go the China route, you pays your money and you takes your chances.

Is there a way to tell if chips we have are real or not?

short of asking ST, nothing definitive.

Aside from GD32 chips, those “counterfeits” are likely mis-marked (purposefully maybe?) other STM chips. it could be a STM32FEBK, for example, a lower capacity STM32, a refurbished STM32, or a repackaged STM32 (from die), …

chances of someone other than ST making a brand new STM32 are low.

I have known about fake Atmel chips, but they are quite a bit simpler than the F103. I would not have expected to see so many. I bought several blue pills and some Maple Minis and an STlink clone, all from different sources and they all appear to have fake chips on them. So I am batting 1000.

Guess I will break down and buy a couple from Digi-Key.

Thanks all.

Having chips reporting 128K doesn’t means they are clone. In this forums, many of us discovered that a while ago.
We suspect that ST is using the same die for CB and C8, but, during testing, they mark it C8 if some of the flash cells on the second 64K reports some errors.

When you will get some F103C8 from DigiKey, please report back here what is the flash size is reporting.

There is at least one other thread which discusses this topic

it should be merged, with one of those.

Edit.

See

viewtopic.php?f=39&t=2500&hilit=fake

http://www.stm32duino.com/viewtopic.php?t=110

viewtopic.php?t=1323&start=20

So I am batting 1000.

maybe you should think about buying lottery ticks now.

seriously, we all could use some of your wisdom in identifying fake chips if you don’t mind sharing.

the observation of more ram / eeprom than specified in the datasheet on some stm parts has been around for a long time. I tried that on STM32F030 and then STM8 back in 2014.

the same can be said about usb capabilities as well – you may see some non-USB capable (per datasheet) STM8s used in USB applications natively.

or UID on parts not specified for UID….

obviously, the risk of relying on such things is entirely on you.

Guys,
I understand that a chip sold as 64K that reports 128K is not by itself indication of a fake. My issue were parts clearly marked STM32F103CBT6 that read only 64K. Note, I just read the flash size register I did not try to test the flash for size.

But thinking about it, I am sure the flash size register is programmable by ST, and I would suspect even if they use the same die that they would program it to be the size of the part they mark rather than program it larger. So maybe reading the wrong flash size either more or less is indication of a counterfeit? Just thinking out loud.

I asked ST privately about whether they had heard of fake chips and they did not know of any.

They know that the GD32 are a “clone” but are actually completely different MCU’s, which were designed to be functionally equivalent but not a direct copy, but that is a lot different from fake chips

(Note. The GD32 are not actually functionally equivalent and have a number of small differences which means STM32 binaries don’t always run on GD32 chips)

I’m living in China. and bought more than 100 pieces STM32 chips from over 3 sellers. and Chinese forum, never found faked chips.
But you will get used STM32 chips if you wanna pay lower price. The sellers must told you that the chips are used or refurbished chips.
( Free refund in 7 days on Taobao.com)

BTW, It is only US$0.9 dollar in China. Fake it isn’t good idea on business I think.

and I never noticed flash size of this chip. It is more enough for me.

http://www.stmcu.org/module/forum/threa … 1-1-1.html

he said stm32f103c8 flash size up to 0x0801FFFF, and readable and writeable.

[Fitch – Tue Sep 05, 2017 12:41 am] –
Guys,
I understand that a chip sold as 64K that reports 128K is not by itself indication of a fake. My issue were parts clearly marked STM32F103CBT6 that read only 64K. Note, I just read the flash size register I did not try to test the flash for size.

But thinking about it, I am sure the flash size register is programmable by ST, and I would suspect even if they use the same die that they would program it to be the size of the part they mark rather than program it larger. So maybe reading the wrong flash size either more or less is indication of a counterfeit? Just thinking out loud.

If you say counterfeit with the meaning that the part number may have been scrubbed and a new one printed, then perhaps, still this things are so cheap it’s hard to believe there is profit left on doing that. If you are meaning that someone cloned the MCU, it doesn’t seem like the case so far since STM would have already have heard of them.
Another posibility is that hose are rejects from the factory, perhaps precisely because they were marked with a capacity but programmed to report another, and some way they made it out to the market.
But I do believe they are all STM parts, most likely when the same die, 128KB actual flash, and just the wrong programed size vs label.

So maybe reading the wrong flash size either more or less is indication of a counterfeit?

looks like “you suspect that they are fakes”. that’s quite different from “they are fakes”.

most of us simply don’t have the know-how or the hardware to identify fake chips, especially mcus. the lucky few so far have not been able to say how they identify fake mcus with such a high degree of confidence, unfortunately.

BTW, It is only US$0.9 dollar in China. Fake it isn’t good idea on business I think.

here you have the answer as to how / why those stm32f103 boards are so much cheaper than the bare chips you can buy off digikey or arrow,

Thank you all for your replies. I originally had hoped that someone knew how to identify a fake chip, maybe a font or logo wasn’t right, maybe the package is a bit off. Anyway, perhaps all these chips a real, but they are retested from rejected parts? It would explain why you can buy a board with a chip on it shipped from China for less than the 1000 piece price from a US distributor. Might also explain flash size registers with incorrect or bogus data in them.

Anyone find anything else odd about their blue pills or Maple Minis aside from flash size?

I dont think low price from companies in China equates to fake parts.

These companies often buy their parts in 1000+ quantities and you are not paying the shipping costs, plus warehousing , plus operating cost of a company in America

but they are retested from rejected parts?

Or those parts are real but just someone has figured out a more efficient way to compete, or is willing to accept a lower margin? Or …..

Why do fake parts have to be the only plausible explanation?

I’m moving this to Off Topic, as really its a discussion for an official STM site, as it has no appreciable impact on using Arduino on the STM32 series of MCU’s

Just in case anyone is interested, I purchased some STM32F103CBT6 parts from Arrow Electronics and from EBay seller polida2008. The Arrow price was almost $6, the EBay price was $1.90. I built a test fixture with a socket so I could verify them before soldering them in to anything, and all appear to be the genuine article. My hat goes off to the EBay seller, $1.90 is better than buying a whole reel in the US.

[Fitch – Sat Sep 23, 2017 11:21 pm] –
Just in case anyone is interested, I purchased some STM32F103CBT6 parts from Arrow Electronics and from EBay seller polida2008. The Arrow price was almost $6, the EBay price was $1.90. I built a test fixture with a socket so I could verify them before soldering them in to anything, and all appear to be the genuine article. My hat goes off to the EBay seller, $1.90 is better than buying a whole reel in the US.

@fitch

This is the consensus…

There do not appear to be any cloned or counterfeit STM32s..

Chinese companies can currently source and supply components at costs far below vendors in the USA, Europe etc etc

Whether of not this supply dries up over time, still remains to be seen. But you may as well take advantage of it until that happens

counterfeit chips are like compiler bugs: they are incredibly easy for newbies to spot.

[dannyf – Sun Sep 24, 2017 12:24 am] –
counterfeit chips are like compiler bugs: they are incredibly easy for newbies to spot.

Oh that explains why I have found so many compiler bugs!

that explains why I have found so many compiler bugs

That means you are exceptionally gifted.

my guess is, it is probably more difficult and more expensive to make even clone , compatible or for that matter ‘fake’ chips compared to simply buying the real ones, unless of course if the fake chips totally doesn’t work and are simply duds

I would think it unlikely that the STM32F103XXX that we see are clones or copies, since that would imply some pretty sophisticated entity had cloned them.

However they might conceivably be “recovered” in some “hot wok” chip recovery enterprise in the back streets, from some common consumer grade product(s).

Given that they have a relatively low value, I would suggest that there are much better things to clone, with much broader appeal (Nike trainers for example).

Or they are surplus parts from a project, bankrupt sales, recovered parts, or simply from someone who is more efficient or wants to use this for marketing purposes, ….. The list goes on and on.

The issue with the “counterfeit!” Clamants is that once they rule every other possible explanations, they are left with 120 percent sure that anything priced lower than what digikey or arrow charges for the same part is a counterfeit.

Only if they understood what it took to counterfeit something like a 32bit chip, ….

tests will be done via decap to try to make a definitive answer to the question stm32f103
https://www.eevblog.com/forum/projects/ … -your-help!/

to make a definitive answer to the question stm32f103

sending those chips to ST would be a definitive answer.

ST will never answer

p.s. this is the correct link
https://www.eevblog.com/forum/projects/ … our-help!/

[testato – Mon Jul 16, 2018 6:34 am] –
ST will never answer

maybe yes maybe not

[fpiSTM – Mon Jul 16, 2018 8:36 am] –

[testato – Mon Jul 16, 2018 6:34 am] –
ST will never answer

maybe yes maybe not

When is an answer not an answer….