EEVblog Electronics Community Forum

The ST-Link utility can query the chip. Check what it says.

The cheap Blue Pills used to be genuine until someone started remarking STM32s that have half the RAM/flash, and whatever other substitutions they might be performing these days. There's probably overlap in the prices between real and fake.

Thanks for the replies guys. 

A STM32 noob here, infact those are the 1st STM32 in the house 

Noob question, can I use that ST-LINK to upload firmware thru the USB ?

I use Win7 x64 btw.

can I use that ST-LINK to upload firmware thru the USB ?

If the Blue Pill has the STM32duino bootloader on it, you can program it through the USB directly from Windows. Otherwise, you connect the ST-Link to that 90° programming header on the opposite end of the board.

At the ultra-low price point, it's quite likely the chip hasn't had the bootloader flashed on.

I have that exact model, but with RESET spelled correctly.

What has happened in the past is that they use the STMF103C8 instead of the STMF103CB in cheap ST Link clones. The thing is, that the C8 has half the flash compared to the CB. The proper ST Link on Nucleo boards uses a CB chip with 128KB flash. But without much effort, double the whole 128KB can be used, even if the chip is labeled C8. The flash seems to work just fine in many many cases. Background is, that ST probably produces only one chip.

Yours is a F103C8 to begin with. I don't think that there's any false advertising going on. Check what the ST Link utility knows about the chip.

As far as I know, there are no fake ST chips. The Chinese simply use cheaper chips with half the flash. They don't even bother relabeling them.

What has happened in the past is that they use the STMF103C8 instead of the STMF103CB in cheap ST Link clones.

The digit 8 vs the letter B ? That's a horrible choice of part naming...

I have been trying to get some Blue Pills with the CS32 chip on it, but I actually got the real thing. Furthermore they even had the correct value for R10 mounted.

There are some on alix where the pictures show CS32 chip, and the reviewers show the same, where I can see the IC: https://www.aliexpress.com/item/32525208361.html
But who knows what will show up..

When I order STM32F103C8T6 at LCSC.com, in large quantities the price drops to about US$ 1, so when these are mass produced, it sounds possible to use the real thing.

I agree, I haven't seen any evidence that if you get a chip that *has the right chip ID* ie has 128kB of flash, that it is counterfeit. So its easy enough for us here to test, although some other people might be fooled if it came with a GD or CS32 and never used >64kB.

Ok, connected & fired up ST-Link, attached below the result.

So its genuine ? 

Also tested using techman-001's STM32F103C8 Diagnostics  , it has 128KB flash, result -> HERE

It's genuine all right. I replied to your link above with details.

It's a genuine CKS32F103C8T6, which is a Chinese Clone of the STM32F103CB as I understand it.

The clue to it being a CKS32F103C8T6 is that the "lot numbers' are in hexadecimal, unlike genuine STM32F103Cx's.

Quote from: techman-001 on January 14, 2020, 01:22:23 pm

It's genuine all right. I replied to your link above with details.

It's a genuine CKS32F103C8T6, which is a Chinese Clone of the STM32F103CB as I understand it.

The clue to it being a CKS32F103C8T6 is that the "lot numbers' are in hexadecimal, unlike genuine STM32F103Cx's.

LOL , yeah right. 


And you know whats funny ?

The new cheap ST-Link 2 I bought just to program this BluePill, opened it up, lo & behold, a genuine CKS. 

Excellent pictures! I guess this is proof that this "STM32F103C8T6" is in fact fake, being a relabeled CKS32F103C8T6. Your testing this chip with my "STM32F103C8 Diagnostics" binary also proved it has a fully working second 64kB Flash, giving 128kB as indicated by the chips "Flash Size Register"

A genuine STM32F103C8T6 will always indicate it has only 64kB flash in the chips "Flash Size Register", even if it has 128kB of flash. This means programs like your ST-Link will show it as only having 64kB Flash and will refuse to flash the second "hidden" 64kB block unless the ST-Link configuration is modified.

Now the STM32F103CBT6 and the CKS32F103C8T6 are not identical with the CKS throwing up a USB ID error such as :
"Warn : UNEXPECTED idcode: 0x2ba01477 Error: expected 1 of 1: 0x1ba01477" on GDB etc and requiring config file changes before it will work.

see: https://www.eevblog.com/forum/microcontrollers/issue-with-chinese-blue-pill-cks32f103c8t6/msg2767816/#msg2767816

Interesting information about the cheap ST-Link 2, I wonder if just runs the STM ST-Link 2 binary ?

Note that your Windows STM32 ST-LINK Utility picture in one of your previous posts in this thread  reports this chip as a "STM32F10xx Medium-density", meaning it can't tell that the chip is in fact a CKS32F103C8T6.

My free "STM32F103C8 Diagnostics" binary allows a deeper investigation and conclusion of whats in a Blue Pill, especially with feedback like yours 

The Blue Pill story is certainly one of intrigue, of secret deals in low places and the sophistication of the Chinese "fake" industry. I wonder if someone will write a book about it one day ?

 

Quote from: techman-001 on January 15, 2020, 12:40:05 am

Excellent pictures! I guess this is proof that this "STM32F103C8T6" is in fact fake, being a relabeled CKS32F103C8T6. Your testing this chip with my "STM32F103C8 Diagnostics" binary also proved it has a fully working second 64kB Flash, giving 128kB as indicated by the chips "Flash Size Register"

A genuine STM32F103C8T6 will always indicate it has only 64kB flash in the chips "Flash Size Register", even if it has 128kB of flash. This means programs like your ST-Link will show it as only having 64kB Flash and will refuse to flash the second "hidden" 64kB block unless the ST-Link configuration is modified.

Thanks, I guess its now beyond doubt that my Bluepill boards have the CKS32F103C8T6 isn't it ?

Regarding the laser marking on the chip, it looks pretty convincing, at least to me, also remember the "F" character position, its just weird compared to many that I've seen on the net. 

Quote from: techman-001 on January 15, 2020, 12:40:05 am

Now the STM32F103CBT6 and the CKS32F103C8T6 are not identical with the CKS throwing up a USB ID error such as :
"Warn : UNEXPECTED idcode: 0x2ba01477 Error: expected 1 of 1: 0x1ba01477" on GDB etc and requiring config file changes before it will work.

see: https://www.eevblog.com/forum/microcontrollers/issue-with-chinese-blue-pill-cks32f103c8t6/msg2767816/#msg2767816

Sorry, just tried to install a freshly downloaded STM32CubeIDE 1.2.0 Windows, it gives me error during the installation, and I have no interest to pursue further.

Quote from: techman-001 on January 15, 2020, 12:40:05 am

Interesting information about the cheap ST-Link 2, I wonder if just runs the STM ST-Link 2 binary ?

Note that your Windows STM32 ST-LINK Utility picture in one of your previous posts in this thread  reports this chip as a "STM32F10xx Medium-density", meaning it can't tell that the chip is in fact a CKS32F103C8T6.

What do you mean by that ? See the attached below, screen shots of the ST-Link Utility when upgrading the dongle firmware, its the latest that I downloaded few days ago.

This cheap ST-Link V2 dongle can be upgraded just fine, no problem what so ever. I assume the latest ST-Link firmware was downloaded straight from ST's host by the program.

Quote from: techman-001 on January 15, 2020, 12:40:05 am

My free "STM32F103C8 Diagnostics" binary allows a deeper investigation and conclusion of whats in a Blue Pill, especially with feedback like yours 

Infact, its your fault. 

I have zero experience in STM32, when visiting, I briefly brought up your discovery with my friend, and then he suddenly gave me this two boards, then here I am now. 

Quote from: techman-001 on January 15, 2020, 12:40:05 am

The Blue Pill story is certainly one of intrigue, of secret deals in low places and the sophistication of the Chinese "fake" industry. I wonder if someone will write a book about it one day ?

I guess this will be drowned into the abyss and forgotten by time.

Beyond doubt ? only until some logical poster with contrary proof comes along

Personally I think your chip is a  CKS32F103C8T6 based on the following.
1) STM32F103C8 chips *always* give their flash size as 64kB, NEVER 128kB like yours
2) The LOT numbers are hexadecimal on your chip, STM only use decimal LOT numbers.
3) The Die number is very close to a known CKS32F103C8T6 data in my collection.

We could do with a correctly positioned, clear pic of that chips marking for future reference, if you have the time. The pic you have supplied is pretty good tho and serves as a example of what fake markings look like in this case.

Regarding ST-Link V2, what I meant was that it would seem that the STM ST-Link V2 binary doesn't check for fake chips, perhaps they have no way of telling, don't care or haven't got around to it yet ?

Any self respecting manufacturer would block updates to fake chips with a big fat "FAKE DETECTED" and just stop the update. Note: (just in case STM are reading here), they would NOT try and BRICK the chip like FTTDI stupidly did years ago when they discovered Chinese clones of their chips. They have been universally reviled for that ever since, with a suspected massive loss of market share to their competition.

You have no interest installing STM32CubeIDE 1.2.0 for Windows ? ... well neither do I, being a Forth user

However, those boards are in fact pretty useful and the chips are quite fast, running at 72MHz as they do, have tons of peripherals and even USB.

They aren't recycled STM32's they're brand new fake STM32's on boards costing under $2!

What's not to love apart from the ugly, ugly STM32F10x GPIO register design ?

gpioa.
GPIOA_CRL (read-write) $44444AA8
3|3|2|2|2|2|2|2|2|2|2|2|1|1|1|1|1|1|1|1|1|1|
1|0|9|8|7|6|5|4|3|2|1|0|9|8|7|6|5|4|3|2|1|0|9|8|7|6|5|4|3|2|1|0
0 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0 1 0 1 0 1 0 1 0 1 0 0 0


GPIOA_CRH (read-write) $88854444
3|3|2|2|2|2|2|2|2|2|2|2|1|1|1|1|1|1|1|1|1|1|
1|0|9|8|7|6|5|4|3|2|1|0|9|8|7|6|5|4|3|2|1|0|9|8|7|6|5|4|3|2|1|0
1 0 0 0 1 0 0 0 1 0 0 0 0 1 0 1 0 1 0 0 0 1 0 0 0 1 0 0 0 1 0 0

.... GD is not a clone of STM32 - it is a different chip.

What are the differences ? Mainly.

But its still pretending emulating as STM32 right ? 

For starters, it won't overclock as high: 80 MHz versus 128 MHz for a genuine STM32.

There are presumably other differences. They were discussed on the STM32duino forum, but I'm not sure how much of that was archived.

Thanks, beside they're given free, even I bought them my self, at that price level, I have nothing to complain.

Regarding the STM32 clones, as they had been for a while in many years, I guess they're also keep perfecting it, at least its comparable by STM32 official's specification, aren't they ?

Excellent question, who really knows where the compatibility begins and ends ?

One indicator that the CKS32F103C8T6 must be compatible with the STM32F10x is your Chinese SWD programmer, which runs the official STM ST-Link V2 firmware. That means the USB and GPIO peripherals must be compatible at a minimum.

For a hobbyist however, is the cheap price of $2 for a clone worth the possible months and years they may spend working thru development problems that later resolve to a incompatibility issue ?

My very first STM32F103CB board was a Olimex P103, it cost about $50 but pins are gold plated, the chip is a brand new genuine STM32F103CB with 128kB. The board has a prototype area, rs232 and CAN connector etc. It's a really nice board.

https://mecrisp-stellaris-folkdoc.sourceforge.io/stm32-boards.html?highlight=p103#olimex-p103-board

Ok, It was $48 more than a Blue Pill, but I've never had any issues or doubts about it over the last 7 years.

What is peace of mind worth when developing ? As if the STM32F103 isn't complex enough without worrying about clone issues being the source of your problems ?

And let's not forget that the *genuine* STM32F10x had enough bugs and errata all by itself to cause plenty of sleepless nights for developers ?

You got your Blue Pills for free, and that's definitely the right price. If you need a 'cat feeder', an 'ice sculpture light controller' or anything that's already developed and has a 'library' that works, I say use them.

On the other hand if you want to make something new and original, but it must be cheap, it may be worth checking out Chinese home grown boards such as the 'longan nano' with a RISC-V chip. They're new and exciting, and cheap

https://mecrisp-stellaris-folkdoc.sourceforge.io/gd32vf103.html

St-Link will flash full 128KB no problem, just use something other than ST GUI program, like openocd with custom config

https://web.archive.org/web/20190327110931/http://wiki.stm32duino.com/index.php?title=Blue_Pill#128_KB_flash_on_C8_version

After the genuine BPills the GD32F103 BPills came. GD is not a clone of STM32 - it is a different chip. GD32 BPills came in fall 2015.

Recently the new CKS32 BPills appear. It could be those CKS are re-branded GD32.

You raise some interesting points ...

From https://smdprutser.nl/blog/stm32f103-vs-gd32f103/
----------------------------------------------------------------------------------
The GD32F103C8, manufactured by Gigadevice, has similar specifications and thus many people think the chip is a cloned version of the STM32 part. The chip appears to be the same from a programmers point of view as almost all registers and peripherals are the same. Major differences are the execution speed and flash size. The chip is capable of running a much higher clock frequency (108MHz) and have less wait states for flashrom then the ST counterpart. The size of the flash memory goes up to 3076KB, compared to the 1024KB ST offers.

The differences can be explained by the decapping pictures (done by Zeptobars). The chip appears to consist of two dies: one microcontroller die and a separate flash die. They seem to use just one large single design and a varying size flash chip to serve the whole GD32F103 range. On the controller die there is 128KB SRAM caching the flashrom. This explains the 0 wait states for flash access. This is a similar way the ESP8266 runs (albeit the flash is user replaceable ). Not very surprising as Gigadevice is a big Chinese flash company. Some tweaks to silicon will allow a higher clockspeed (108MHz) and some people report it being 50% faster then the ST part.

Sounds to good to be true! Not very much is found on the internet about this chip, except some people report that not everything is working exactly the same as the STM32
------------------------------------------------------------------------------------

I have the GD32F103C8 datasheet, that's easy to find.

But no SVD anywhere, does anyone have a  GD32F103.SVD or a link ?  (to do  comparisons of peripherals, registers and bitfields with the STM32F103).

Re CKS32F103C8T6:
I can't find any info on the CKS32F103C8T6 anywhere on the net, and even "CKS" doesn't list this chip in their products section.

http://www.cksic.com/en/about/217.html
"China Key System Co., Ltd. (CKS) is a Chinesetechnology company headquartered in Wuxi, Jiangsu province, that designs,develops and manufactures very-large-scale integrated circuits. It was founded in 2008."

Perhaps the "CKS32F103C8T6" is a re-branded "GD32F103C8" as you say ?

GD32F - it has been discussed many times here, afaik. The GD32F is a flashless chip - no flash memory on the cortex chip.
It has got a sram instead of the flash. Upon reset it moves the stored binary from an external 8pin SPI flash (glued on the top of the cortex chip inside the package) to its internal sram and it executes your program off that sram. Therefore it is faster than the stm32 - no flash wait states.
During programming the flash of the GD32 the program gets stored into that external SPI flash.

Is there any way to sort of detect if its GD32F vs genuine STM32 by benchmark code maybe ? Especially on the SRAM vs internal flash ?

So, the "fake" is better than the original?

You may write a short loop in asm toggling a pin, for example, and watch the period with an oscope.

Run STM32f103, GD32f103 and CSK32F103 at for example 128MHz clock, the HAL should set the required number of flash waitstates, something like 4ws with STM32f103 (0ws for 0..24MHz, +1ws for each 24MHz), and 0ws with GD for any clock.

PS: you may set manually the number of flash ws in the STM32 lower than required by HAL - called "overclocking"
I did in past with stm32F100 and 103/407.
0ws with F100 (max 48MHz) and 0+1ws/48MHz with F103 worked somehow as I can remember..

Maybe it's time to change the pills from blue to black 
https://www.cnx-software.com/2019/12/24/stm32-black-pill-board-features-stm32f4-cortex-m4-mcu-optional-spi-flash/

Maybe it's time to change the pills from blue to black 
https://www.cnx-software.com/2019/12/24/stm32-black-pill-board-features-stm32f4-cortex-m4-mcu-optional-spi-flash/

In principle an excellent suggestion, the Cortex-m4 is far better in every way than the old STM32F103Cx, but at those prices what's to stop the same things happening all over again, i.e. re-badged poorly documented "compatible" Gigadevice MCU's or worse ?

It's the perfect bait and switch, make a few thousand boards with genuine STM32F4's as loss leaders then when the brandname is well known, quietly switch in the clones ... PROFIT!

I think there is only one way to be sure, if you must have a board, buy a  Nucleo or Olimex and get a guaranteed STM32 MCU or buy the chip only direct from a respected supplier such as arrow.com and put it in your own design ?

If I were a young hobbyist with a limited budget, who wants to start with a 32bit MCUs, I would be pretty happy to get GD/CKS BluePills for $2 incl. postage.

That is the targeted audience for that kind of products. BluePills do not target corporate developers who want to land on Moon or Mars and their budgets are virtually unlimited.

On the other hand I would admit a single BluePill is capable enough to land you on the Moon - as we saw it 50y back. In case you targeted such an exercise I would be pretty confident STM may produce for you a radhard STM32F103C8

Yes, and, on top of that, as most people are going to buy the cheap blue pills you'll have it easier to get help off the inter webs, than with the olimexes. And although I say that, I've bought quite a few Olimexes in my life and never had a problem with them.

BTW, there was already an STM32F103 black pill before the Cortex M4 ones.

People play with that 401/411 BPill already on www.stm32duino.com (the new forum) or mcu.selfip.com (also the new forum) .
Looks promising..
Supported by STM HAL core and stevestrong's github core.

stm32duino.com is back?!?! Cool!

Hi stm32duino community,

after several month down, I'm pleased to announce that stm32duino.com forum is back and alive!

First, I want to thanks Roger Clark for all his wonderful work for the community since this site was created and also for his support and kindness.
The hosting is now managed by ST but the moderation will be done by the community.

The old forum contents can be found here:
https://web.archive.org/web/20190316170 ... duino.com/

due to GPRD it was not possible to restore it. :cry:

Enjoy!

Quote from: langwadt on January 17, 2020, 09:09:34 pm

yeh, price for 401/411 "Black Pill" on Ali $2.79/$3.96, STs budgetary price in qt 10000  $2.13/$2.79

In the "oscilloscope asic" thread I wrote that a "list price" (or budgetary price) "means nothing" in the business. Happy you are showing that prices - it says "somebody" is getting the  chips for a fraction of their budgetary price.. 

yeh the list price is not much more than "is it an expensive part?"

the same stm32f401/411 on digikey is ~$2.50@2500 ...

maybe mr noopy who just did the 399 VREF thread would like to have a crack at it?

*Pill users have been scammed so many times with cheap shoddy Chinese junk and fake chips yet still assume new *Pills will contain genuine parts. It boggles my mind.

Could you show me an *Pill user who ever got a board with fake stm32 chip, plz?
Could you show me a fake stm32 chip, plz?

Quote from: techman-001 on January 18, 2020, 12:30:22 am

*Pill users have been scammed so many times with cheap shoddy Chinese junk and fake chips yet still assume new *Pills will contain genuine parts. It boggles my mind.

Could you show me an *Pill user who ever got a board with fake stm32 chip, plz?
Could you show me a fake stm32 chip, plz?

Of course, proof of both are in this one pic.

The photo shows a fake stm32 chip in a Blue Pill board, you can clearly see the labeling says "STM32F103C8T6" but this chip is not made by STM, therefore it's a fake.

[attachimg=1]

The backstory is here for anyone interested:
https://www.eevblog.com/forum/microcontrollers/stm32f103c8-diagnostic-bootable-binary-freely-available/msg2871198/#msg2871198

What makes it "shoddy"?

Items such as inadequately anchored USB connectors that snap off after a few uses, incorrect value USB resistor (Google “R10 Blue Pill” for more information), header pin coatings that go rusty and FAKE mcu's.

The name "Blue Pill" is the personification of "shoddy".

https://www.yourdictionary.com/shoddy
shoddy definition: The definition of shoddy is something immoral or sordid, or is something badly made. (adjective) Stealing from people who trust you is an example of shoddy behavior.

Quote from: westfw on January 18, 2020, 07:25:38 am

What makes it "shoddy"?

Items such as inadequately anchored USB connectors that snap off after a few uses, incorrect value USB resistor (Google “R10 Blue Pill” for more information), header pin coatings that go rusty and FAKE mcu's.

The name "Blue Pill" is the personification of "shoddy".

https://www.yourdictionary.com/shoddy
shoddy definition: The definition of shoddy is something immoral or sordid, or is something badly made. (adjective) Stealing from people who trust you is an example of shoddy behavior.

*yawn*

Quality is defined as conformance to requirements. A carrier bag (for food shopping) is obviously not the same spec as an executive briefcase, but they both meet their quality targets.

These boards cost $1.86, really WTF do you expect... Maybe you should find some other windmill to tilt at.

I'm not sure if there are F4 clones, but on F303 gigadevice cheaped out by keeping the old F1 peripherals layout rather than copy the STM32F303 peripherals, so GD32F303 isn't drop in compatible with STM32F303, but it does have a Cortex-M4 and FPU (unless you are unlucky and get an early 2017 batch which have no FPU).  I simply developed my code targeting GD32F303 rather than STM32F303 because the BOM cost in that particular application (NanoVNA V2) is critical.

Thanks for the GD information, your NanoVNA V2 looks nicely done

Not that it matters but I  wonder if Gigadevices licensed or purchased the old F1 peripherals from STM as they seem to be using them in a lot of their MCU's?

The Gigadevices  RISC-V GD32VF103  (which also uses most of the F1 peripherals) is quite interesting, I'm playing with one at the moment.
https://mecrisp-stellaris-folkdoc.sourceforge.io/gd32vf103.html#gd32vf103-longan-nano

.... incorrect value USB resistor (Google “R10 Blue Pill” for more information) ...

Hmm ... my pill's R10 is using the right value, isn't that proof its not shoddy ? ... j/k 

Quote from: techman-001 on January 18, 2020, 06:25:55 pm

The Gigadevices  RISC-V GD32VF103  (which also uses most of the F1 peripherals) is quite interesting, I'm playing with one at the moment.
https://mecrisp-stellaris-folkdoc.sourceforge.io/gd32vf103.html#gd32vf103-longan-nano

Just a small correction: you mention that the conditional branches range is +/-2048 bytes, which is not exact. Since instructions can only be aligned to 4 bytes, or 2 bytes for compressed instructions, the 12-bit offset actually encodes an offset that is in multiples of 2 bytes, which is clever, so the range is +/-4096 bytes. (I'm currently actively working on RISC-V so that piqued my attention!)

Otherwise, this little MCU seems nice. One thing I'm wondering about, and haven't really seen figures, is the power consumption. If you have a board, could you measure the current draw while the MCU executes code? I'm just curious.

Thanks for the conditional branch info, I'll update my page

I'll set something up to measure the current drain, all I can really do that be be vaguely informative is a comparison under similar circumstances with a STM32F103, both  running a Forth primes program then post the results here and on my doc site. At least they use the same peripherals with a couple of exceptions.

I took a deeper look at the datasheet: https://www.gigadevice.com/products/microcontrollers/gd32/risc-v/

Comparing with the figures given in the F103 datasheet, it should draw less power/MHz in run mode, but of course there's a number of factors to take into account, such as which peripherals are enabled, etc, and the frequency you ran them at was also not the same obviously. Still, for the STM F103,  they state ~50mA @72MHz with all peripherals enabled (which is not terribly impressive), and for the GD32VF103, 33.7mA @108Mhz (also all peripherals enabled), which looks definitely better. Now that remains to be confirmed with measurements. Thanks for the couple measurements you made, unfortunately the tool you used is not really accurate enough... but we are definitely in the ballpark stated in the datasheets.

Other than power consumption at the typical 3.3V, I noticed two points that are worth considering when using a GD32VF103.

* First, min. VDD is 2.6V (vs. 2V for the STM32F103), which probably also says a thing or two about the process node they must have used. Or maybe it's all due to the flash memory... (which I suspect, see below!)
* Second, the startup time is very significant at 132ms... (and also the wakeup from standby mode which is almost as much...) They are much shorter on a STM32F103. Not sure what explains this as I haven't found clear documentation on the "boot" process. I'm suspecting that the internal flash is first copied to an internal SRAM block (which would allow 1/ cheap flash on a separate die and 2/ easily allow the zero-wait-state that is advertised), but I haven't seen that clearly stated (unless I missed it), and the block diagram makes it look like the core would have direct access to flash memory.

That is the same booting process as the GD32F. I think the GigaDevices has taken the same road with GD32V chips as well (2 chips in the package).

Yeah, several clues are there that indicate it must be this.

It's definitely something to keep in mind. It may not matter in some applications, and may in others.

Noob question, in order to identify STM32F103 vs GD32VF103, say both set @72MHz, run a tight loop, say many-many iterations that will make it significant, between the loop accessing the SRAM at GD32 vs Flash at STM32, and blinks the PC13 led after certain millions of iterations that human eyes can spot it ? 

Sure, but what is the percentage of apps where you have to start your MCU app in XX clocks after a power on?

In significantly less than >100ms, quite a few actually. Can be even more of a problem when getting out of standby mode if you use it, or simply after a reset (from watchdog for instance).
The 2.6V min can also be a concern in many low-power applications (although these days I would probably never use either the STM32F103 or the GD MCU for a low-power design anyway...)
Then, the fact flash is copied at startup can also pose some specific security concerns, so it's not necessarily the best approach for all designs. Sure is cheap though.

All this to emphasize some of the significant differences between the two MCUs (apart from having a different ISA) that could matter depending on specific requirements and that may not be obvious if you don't read the specs carefully.

Of course, it's likely that most users of "Bluepill" boards wouldn't care about any of the above. (But those, OTOH, will care about the two chips being actually not at all as compatible, at least peripheral-wise, as they could initially assume, which I think was an earlier thread in itself...)

That is the same booting process as with the GD32F. I think the GigaDevices has taken the same road with GD32V chips as well (2 chips in the package).
Btw, a pretty clever decision. No need to cope with slow flash on the chip (like STM does with ART).
An you may design 1GHz mcu with zero ws.

with the down side that RAM uses much more die area than flash, and uses much more power doing nothing like in sleep mode

Quote

the startup time is very significant at 132ms...

Is that in addition to oscillator startup time, or "in parallel" somehow?
(I'll note that the normal Atmel AVR crystal oscillator start-up time is ~65ms.  But the STM32s start on an internal oscillator and you have to enable the crystal oscillator separately.  In theory, that could mean additional time (132ms to copy the flash to RAM, start running, enable the crystal oscillator and wait another few 10s of ms for that to stabilize...)

Good points. One of the things I appreciate about the MSP430 is the 'instant' USART wake up from deep sleep.

And let's not overlook the ESP2866 which also uses the same serial FLASH technique at boot-up as the GD MCU's. That boot-up delay hasn't stopped millions of them being sold and deployed because of their low price, high clock speed and WiFi capability.

GD32F/V includes some sort of ROM (or a state machine) too - the bootloader accessing the external flash (in case of the GD the second die inside the package). Thus ESP and GD are "flashless" mcus.

https://zeptobars.com/en/read/GD32F103CBT6-mcm-serial-flash-Giga-Devices

On topic 

User madias (over there), just got this "It's Fake" letter from STM
https://www.stm32duino.com/viewtopic.php?f=7&t=127

/Bingo

Send it to noopy 

On topic 

User madias (over there), just got this "It's Fake" letter from STM
https://www.stm32duino.com/viewtopic.php?f=7&t=127

/Bingo

Hey Bingo, if you're a member over there, any chance you could tell them about my free diagnostics binary that may help them determine if a STM32F1x is a fake ?

https://mecrisp-stellaris-folkdoc.sourceforge.io/stm32f103c8-diags.html

Cheers:)

Quote from: bingo600 on January 24, 2020, 06:09:06 pm

On topic 

User madias (over there), just got this "It's Fake" letter from STM
https://www.stm32duino.com/viewtopic.php?f=7&t=127

/Bingo

Hey Bingo, if you're a member over there, any chance you could tell them about my free diagnostics binary that may help them determine if a STM32F1x is a fake ?

https://mecrisp-stellaris-folkdoc.sourceforge.io/stm32f103c8-diags.html

Cheers:)

Done

/Bingo

They may take an stm32F103C6 or C4 (32k/16k), sand the markings off the package and print a "fake" marking with ie. the famous 991 KA on it.
Would be interesting to read the IDs of those chips.

Quote from: GeorgeOfTheJungle on January 24, 2020, 07:14:15 pm

Send it to noopy 

or zeptobars https://zeptobars.com/en/ , or electronupdate https://www.youtube.com/user/electronupdate/videos?disable_polymer=1

Holy Moley!

I just spent hours viewing zeptobars, what an excellent decapping site, such clear die pictures, it's like chocolate to a tech such as I

What really shocked me was HOW MANY FAKES there are now !  There are fakes of just about every generic chip these days from 555's to op-amps etc.

Amazing.

https://smdprutser.nl/blog/category/stm-vs-gd32/ some investigation of GD differences

Quote from: techman-001 on January 25, 2020, 09:38:26 am

I does appear that we can use my diagnostic program to tell if a chip marked as STM32F is a fake by reading the LOT NUMBERs.

I have a few reader submitted data that show the LOT Numbers of CS32F103C8T6 chips are in Hexadecimal, ie 0x004D3512 while genuine STM chips are in Decimal only.

I only have a report from one user with a GD32F103 and that chip USB failed to work with the error "Device descriptor request failed".
Does anyone else here have a GD32F103 Blue Pill or board that they could try my diagnostic program on and report the DEVICE ID data ?

I am the user, since I can't get your binary to work with USB enumeration, read GD32 with jlink instead:

UID:
1FFFF7E8 = 2E CB 39 34 38 34 38 0F 33 37 39 38 FF FF FF FF
1FFFF7F8 = FF FF
contains hex similar to cs32 right?

xy = 876202798
wafer number = 56

Flash size register:
1FFFF7E0 = 40 00 = 0x0040 = 64kB

UID [31:0]: X/Y coordinates on Wafer (BCD coded)
UID [39:32]: Wafer number (8-bit unsigned number)
UID [95:40]: LOT number (ASCII encoded)

Thanks for the followup!
I'm assuming that your register 0x1FFFF7E8 contents = 0x2ECB3934    (as the register is 32 bits)?
If you have a USB/3.3 volt dongle and don't mind connecting it to your BluePill unit, I can make you a bootable binary to use instead of the boards non working USB ?

I'm doing a improved V2 atm, any chance you could followup with the 32 bit Hex contents of these registers as you're the only one with a GD32F103 Blue Pill who has responded ? 
0x1FFFF7E0  =
0x1FFFF7E8  =   
0x1FFFF7EC  =
0x1FFFF7F0  =
Here are the contents of one of my STM32F103C8 MCU's for comparison but as we know your numbers will be different.
id (STM32F103C8)
0x1FFFF7E0  = 0xFFFF0040
0x1FFFF7E8  = 0x066FFF53
0x1FFFF7EC  = ox55507150
0x1FFFF7F0  = 0x87042957


Sure, Ali assumes the STM document might be a fake 

Quote from: bingo600 on January 24, 2020, 06:09:06 pm

On topic 

User madias (over there), just got this "It's Fake" letter from STM
https://www.stm32duino.com/viewtopic.php?f=7&t=127

/Bingo

Wuutt ...

Aliexpress rejected the Official STM document , about fake chips 

/Bingo

They must still be in shock over at Alibaba at this unbelievable claim from Madias?

Hmm, that reminds me of a story my mother read to me as a small boy ... but what was the name of it ?

Ahh, I remember now, it was "Alibaba and the 40 honest people" !

What if the fab in China or Taiwan that makes the fakes makes the original STs too? Then maybe ST themselves are selling fakes too... What a mess!

What if the fab in China or Taiwan that makes the fakes makes the original STs too? Then maybe ST themselves are selling fakes too... What a mess!

Don't buy from Taiwan or China perhaps ?

I'm still using my hoard of 480 STM32F051's (64kB flash. 8kB ram, QFN32) I bought from Avenet (Dallas) in 2014 for $0.56 USD each. They're sealed in chip carriers in silver packs with nitrogen gas and special moisture monitors, plus each pack includes a traceable printout for the product, utterly professional.

This is a far cry from the usual Chinese chip purchase which usually come in a envelope.

The shipping was $65 at the time which was a lot back then but I amortized it over all the chips in that shipment. The other amazing thing is that the order was 3 days from payment to my door! I can't even get stuff in that time using the Australian Postal service from Melbourne, which is only 1000km away from me.

I've also bought from Arrow.com which have really low prices from time to time on old/excess stock etc. They are FAST delivery with similar packaging to Avenet.

No way either of those two business above supply fake chips.

If you want a "genuine" chip with 128/256KB of flash, search GD32F103C8T6 on aliexpress. I have not seen a fake GD32 yet 

I'm done with F103's - Chance of fake is to big.

I'll be buying F401/f411 boards ... Until they fake those too. 

/Bingo

Not sure about the CKS32, but the GD32 peripherals compatibility with STM32 is pretty good. I've used the GD32 in a real design that uses DMA ADC, DMA SPI, USB, flash writing, and all the timers, all using libopencm3 and based on the STM32 documentation. I've yet to notice any difference in any of the peripherals. All existing STM32 code runs unmodified on the GD32. Unmodified STM32F103 binaries run on the GD32F303 too!!

I would say don't worry too much about clones. You most likely won't ever notice the difference unless you check the idcode or use ST's tools which check idcode (which btw I don't recommend using because they force you to write spaghetti code that is unmaintainable).

As they say, "an Elephants  compatibility with a Sherman tank is pretty good, mainly in terms of weight, but they can't breed"

 

I'm working on comparing some *32F103 parts. ( https://www.eevblog.com/forum/other-blog-specific/*32f103-comparison ) 
Indeed, any test is likely to only cover a small portion of the chip. Each peripheral has a dozen of options, and there are a lot of corner cases. Like a communication peripheral that works fine when everything is connected right, but the error handling is implemented differently. So to run a complete test, it's hard to cover all possible corner cases.

Also from hardware perspective it is hard to test many combinations of crystals and capacitors. I've already run into some issues when exchanging STM32 and GD32 ( https://www.eevblog.com/forum/microcontrollers/crystals-and-caps )

edit: it seems the * broke the auto url tags, fixed it.

addition: also one thing to keep in mind regarding STM32 vs GD32, the voltage range differs.

I'll only be able to tell it's compatible with my code. I'll be doing some DMA transfers to timers to control WS2812 LEDS for example. I'll have to think about some more tests to do, but it will never test every possible configuration. Also, thinking about some tests to identify the part. In the end, all it will say whether it's good enough for my purposes. So far, in my opinion, only STM and GD are worth considering. That is based on the presence of documentation and availability of the chip itself. 

The parts I've under test are sold as APM32F103 and CS32F103. For this test I don't have any clones labelled as ST parts that I am aware of.

So far I've only seen the GD32VF103 on Longan Nano boards, but I've not yet seen the chip itself for sale. My thought about the chip is it to be a GD32F103 with the ARM core replaced with a RISC-V code, and the peripherals and pinout remain unchanged. So you could replace the chip in an existing design (if it has a full JTAG header available, as no ARM means no SWD).

Continuing with the same board, I have the USB Forth diagnostics loaded.

stm32id
Die xy coords: 81699
Wafer Number: 45
Lot_num ascii encoded  [23:0]: 0x004D3812  | M 8 .
Lot_num ascii encoded [55:24]: 0x004E4B43  | . N K C

Flash Size Register = 131072 Bytes

Testing 64kB Flash block: 0x10000 - 0x1FFFF
Erasing
Flash with 1010101010101010 (0xAA)
Testing for 0xAA - OK
Erasing
Flash with 0101010101010101 (0x55)
Testing for 0x55 - OK
Erasing
~~~~~ ALL TESTS PASSED ~~~~~

I'll only be able to tell it's compatible with my code. I'll be doing some DMA transfers to timers to control WS2812 LEDS for example. I'll have to think about some more tests to do, but it will never test every possible configuration. Also, thinking about some tests to identify the part. In the end, all it will say whether it's good enough for my purposes. So far, in my opinion, only STM and GD are worth considering. That is based on the presence of documentation and availability of the chip itself. 

The parts I've under test are sold as APM32F103 and CS32F103. For this test I don't have any clones labelled as ST parts that I am aware of.

So far I've only seen the GD32VF103 on Longan Nano boards, but I've not yet seen the chip itself for sale. My thought about the chip is it to be a GD32F103 with the ARM core replaced with a RISC-V code, and the peripherals and pinout remain unchanged. So you could replace the chip in an existing design (if it has a full JTAG header available, as no ARM means no SWD).

Agree again, if ones project tests ok on any chip then all is good

I haven't seen any pricing on the GD32VF103 chips either, and frankly STM32 will do me perfectly for the next few years.

I think GD have quite a battle on their hands as ARM own the embedded market, are cheap (I believe that individual chip prices are a very minor part of any project cost), reliable and very well documented.

It will be a interesting new decade for embedded indeed

Has anyone tried using the flash SRAM mirror on the GD32 as actual ram? I haven't yet figured out how to write to it. All writes seem to be ignored as I'm reading back the old value (this is with a volatile pointer pointing to a flash address). If I unlock the flash for writing it still ignores the write unless I erase the page first. It seems to track which pages are erased because only the first write after each erase sticks (rather than some AND/OR of the old and new value). Writes to the aliased address (starting at 0x0) instead of the flash address (0x08000000) seem to be always ignored even with a prior erase and with flash writing enabled. I'm looking for a way to make it think the flash is smaller and use the remaining space as ram.

There are 2 different physical blocks of sram inside - sram_sram and flash_sram. Most probably the access to flash_sram is limited. Otherwise they would have created a single contiguous block of sram for both flash and sram with uniform access..
PS: the picture of CBT6 on zeptobar site shows 4x32kB flash_sram, 2x10kB sram_sram and 4x512B reg_sram(?) blocks..

I'm sure someone at CKS knows what's going on in their chips ?

Is it too much to ask for a detailed datasheet explaining what they are doing so designers can use that knowledge in their designs, or is the clone world all about secrets due to counterfeiting etc ?

I'll take a genuine STM32  any day over one of these poorly documented Chinese 'clone' chips.

Quote from: techman-001 on February 01, 2020, 09:37:17 am

Quote from: imo on February 01, 2020, 08:48:15 am

There are 2 different physical blocks of sram inside - sram_sram and flash_sram. Most probably the access to flash_sram is limited. Otherwise they would have created a single contiguous block of sram for both flash and sram with uniform access..
PS: the picture of CBT6 on zeptobar site shows 4x32kB flash_sram, 2x10kB sram_sram and 4x512B reg_sram(?) blocks..

I'm sure someone at CKS knows what's going on in their chips ?

Is it too much to ask for a detailed datasheet explaining what they are doing so designers can use that knowledge in their designs, or is the clone world all about secrets due to counterfeiting etc ?

I'll take a genuine STM32  any day over one of these poorly documented Chinese 'clone' chips.

The flash ram thing is about GigaDevice parts (GD32), not about the China Key Systems part (CKS32)? GigaDevice has datasheets, user manuals, errata, etc. on their website. I'd say GigaDevice is good. Using the flash shadow RAM as generic RAM, it's not the intended use, but is more like a hack to get more RAM then advertised for the part. This idea has been discussed since the architecture of the GD32, having an external flash with a shadow RAM had been pointed out. So far, I've only read about people wondering whether it is possible, but I haven't seen anyone do it yet.

However, the CKS32 or CK32 (are they the same part of different parts?) I haven't even found evidence it is even China Key Systems manufacturing them. There is no mention of the part of their site. So, manufacturer unknown, no datasheet or any other documentation. The CKS32 part is a no go.

I only have GD32VF103's (risc-v) and the docs are ok but there are some major errors, such as pasting the reset method from the GD32F103 into the GD32VF103 doc when it's totally different as far as I can tell.

GD do list their chips on their website with doc and I quite like the GD32VF103 myself, I think it may have a bright future, time will tell. They certainly seem like a bonafide and honest chip maker to me.

As you said in a earlier post, be nice to see the GD chips for sale standalone. Perhaps a respin to fix the USB loader issue first ?

Totally agree re the 'CKS' CS32, as you say it's not listed on the CKS website, but it's for sale everywhere. I wonder why the CKS website doesn't have a notice about this chip, some kind of advisory 'be advised this is not a CKS chip'  etc ?

Definitely seems a bit dodgy and the 'CKS' CS32 chips are now in Blue Pills also.

Hi all,

have done some more decapping:

https://www.richis-lab.de/STM32.htm

First one is one of the confirmed fake STM32.

Second one is from an other blue-pill-board.

Third one is from Aliexpress.

Fourth one is from an ST-Link V2.

Except the first one are all original STM32.
In the upper left corner of the fake-die there is a really small thing. Possibly some manufacturer marking. Possibly only dirt…
You clearly can see the difference between the dies.




Does anybody know the figures in the silicon art?