Author Topic: Cheap Bluepill, very likely it has fake STM32 right ?  (Read 78144 times)

0 Members and 1 Guest are viewing this topic.

Offline BravoVTopic starter

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #100 on: January 30, 2020, 05:02:51 pm »
Send it to noopy  :)

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

Got one!  ;D

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

Now a genuine one would be interesting to see...

Both boards, the Bluepill at 1st post that with the fake STM32F (with "F" at 1st line instead of 2nd), and the ST-Link V2 that has CKS32F103C8T6 are on their way to Noopy.  :P
« Last Edit: January 30, 2020, 05:15:20 pm by BravoV »
 
The following users thanked this post: thm_w, GeorgeOfTheJungle

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #101 on: January 30, 2020, 05:08:14 pm »
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).
Email: OwOwOwOwO123@outlook.com
 
The following users thanked this post: iMo

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #102 on: January 30, 2020, 07:03:25 pm »
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).

I wouldn't worry too much about a cloned Gucci handbag but a genuine STM32F103C8 has 33 peripherals, 466 registers and 3503 bitfields.

Considering just not reading the genuine STM32F10x errata can get one into trouble, that's a curious opinion from an engineer especially as some of the 'clone' differences are rather glaring such as the fact that GD use a serial Flash, glued right on top of the MCU die which loads into Ram at boot up and STM does not.



As they say, "an Elephants  compatibility with a Sherman tank is pretty good, mainly in terms of weight, but they can't breed"
 
The following users thanked this post: newbrain

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 14297
  • Country: fr
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #103 on: January 30, 2020, 07:09:22 pm »
As they say, "an Elephants  compatibility with a Sherman tank is pretty good, mainly in terms of weight, but they can't breed"

 ;D
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4675
  • Country: nr
  • It's important to try new things..
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #104 on: January 31, 2020, 10:52:33 am »
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!!
That is encouraging! Would be great to get some detailed report on the actual compatibility of the GD32F parts. Except never-ending  :blah:  :blah: we only see some naive tests. I do like the GD32F/V concept, but I've been missing concrete information what could be expected in real life..
 

Offline GromBeestje

  • Frequent Contributor
  • **
  • Posts: 276
  • Country: nl
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #105 on: January 31, 2020, 11:38:40 am »
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.
« Last Edit: January 31, 2020, 12:34:38 pm by GromBeestje »
 
The following users thanked this post: thm_w

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #106 on: January 31, 2020, 12:40:58 pm »
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 )

I agree.

Personally, I think a complete and accurate comparison of the STM32F103 with any clones is impossible because there are too many peripherals, registers and bitfields, hence too many permutations.

I can say with confidence however that ONLY a genuine factory STM32F103 is 100% compatible with another genuine factory STM32F103 and because there are no reputable and comprehensive studies available regarding the STM32F103 and the 'clones', any claims of 'compatibility' are only opinions of the author.

I'm using some GD32VF103's at the moment and they appear to be a decent MCU. Many of the peripherals *appear* to be clones of the STM32F103 but every peripheral, register and bitfield has a different name to those used in the STM32F103.

Personally I prefer this as the GD32VF103 will stand on its own laurels, no one will ever claim that a GD32VF103 is in any way compatible with any STM32Fxx even tho some of the peripherals are similar.

In time a new GD32VF103 respin may well be a major seller, and given that the RISCV core is open, OSS community support for the GD32VF103 may dominate other chips.



 
The following users thanked this post: SiliconWizard

Offline GromBeestje

  • Frequent Contributor
  • **
  • Posts: 276
  • Country: nl
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #107 on: January 31, 2020, 01:27:40 pm »
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).
 
The following users thanked this post: iMo

Offline rhodges

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: us
  • Available for embedded projects.
    • My public libraries, code samples, and projects for STM8.
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #108 on: January 31, 2020, 03:46:47 pm »
I am testing the "most recent" blue pills I got from ebay seller "cctv008008" in August 2019.

These are "RESRT" boards. The surface mount components are soldered nicely, but the 6-pin headers for BOOT0 and BOOT1 appear to be hand soldered, two rather poorly. The CPU markings are:

STM32F
103C8T6
9910X RR93
MYS 99 907
(ST LOGO)

When I ran openocd, it complained that the id code 0x2ba01477 was wrong, it expected 0x1ba01477. I think someone has already written about this. At first, I was unhappy that I might not be able to use hardware debugging on these chips. But it looks like changing the code in the openocd config file works:
--- stm32f1x.cfg   2020-01-01 22:34:41.000000000 -0700
+++ stm32fake103.cfg   2020-01-31 08:24:23.754265000 -0700
@@ -31,7 +31,7 @@
       set _CPUTAPID 0x3ba00477
    } {
       # this is the SW-DP tap id not the jtag tap id
-      set _CPUTAPID 0x1ba01477
+      set _CPUTAPID 0x2ba01477
    }
 }

> st-info --probe                             
Found 1 stlink programmers
 serial: 303030303030303030303031
openocd: "\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x30\x31"
  flash: 131072 (pagesize: 1024)
   sram: 20480
 chipid: 0x0410
  descr: F1 Medium-density device

I did get gdb to load an ELF file to flash and it looks like debugging works normally.
Currently developing STM8 and STM32. Past includes 6809, Z80, 8086, PIC, MIPS, PNX1302, and some 8748 and 6805. Check out my public code on github. https://github.com/unfrozen
 
The following users thanked this post: BravoV, iMo

Offline rhodges

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: us
  • Available for embedded projects.
    • My public libraries, code samples, and projects for STM8.
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #109 on: January 31, 2020, 04:15:03 pm »
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 ~~~~~
Currently developing STM8 and STM32. Past includes 6809, Z80, 8086, PIC, MIPS, PNX1302, and some 8748 and 6805. Check out my public code on github. https://github.com/unfrozen
 

Offline rhodges

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: us
  • Available for embedded projects.
    • My public libraries, code samples, and projects for STM8.
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #110 on: January 31, 2020, 04:29:21 pm »
Here is another board from the same order:

stm32id
Die xy coords: 594193
Wafer Number: 45
Lot_num ascii encoded  [23:0]: 0x004D3812  | M 8 .
Lot_num ascii encoded [55:24]: 0x004E4B43  | . N K C

Only the XY coordinates changed.
Currently developing STM8 and STM32. Past includes 6809, Z80, 8086, PIC, MIPS, PNX1302, and some 8748 and 6805. Check out my public code on github. https://github.com/unfrozen
 

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #111 on: February 01, 2020, 06:16:53 am »
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 :)
 

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #112 on: February 01, 2020, 06:40:38 am »
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 ~~~~~

Thanks for the feedback  :-+

The only chips I've seen that:
1. connect to the diagnostic image with USB
2. advertise 128kB Flash
3. contain "M" in the "Lot_num ascii encoded  [23:0]:"

are CKS32F103C8T6, so I think that's what you have.

Given the USB Id's are also wrong, it's looking like FAKES once again as the chips are marked STM32F103.
« Last Edit: February 01, 2020, 06:42:56 am by techman-001 »
 

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #113 on: February 01, 2020, 07:49:04 am »
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.
Email: OwOwOwOwO123@outlook.com
 

Offline iMo

  • Super Contributor
  • ***
  • Posts: 4675
  • Country: nr
  • It's important to try new things..
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #114 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..
« Last Edit: February 01, 2020, 09:01:51 am by imo »
 

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #115 on: February 01, 2020, 09:37:17 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.
 

Offline GromBeestje

  • Frequent Contributor
  • **
  • Posts: 276
  • Country: nl
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #116 on: February 01, 2020, 10:11:09 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.
 

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #117 on: February 01, 2020, 10:31:46 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.
 

Offline GeorgeOfTheJungle

  • Super Contributor
  • ***
  • !
  • Posts: 2699
  • Country: tr
« Last Edit: February 01, 2020, 07:47:21 pm by GeorgeOfTheJungle »
The further a society drifts from truth, the more it will hate those who speak it.
 

Offline Noopy

  • Super Contributor
  • ***
  • Posts: 1707
  • Country: de
    • Richis-Lab
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #119 on: February 02, 2020, 02:35:03 pm »
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?

 :popcorn:
 
The following users thanked this post: 3roomlab, thm_w, GeorgeOfTheJungle

Offline martinayotte

  • Regular Contributor
  • *
  • Posts: 64
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #120 on: February 02, 2020, 04:15:25 pm »
It looks like rotated fishes ...  :-DD
 

Offline nigelwright7557

  • Frequent Contributor
  • **
  • Posts: 689
  • Country: gb
    • Electronic controls
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #121 on: February 02, 2020, 06:05:46 pm »
I'm done with F103's - Chance of fake is to big.

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

/Bingo

I bought in some Chinese irfp240/9240 transistors because they were slightly cheaper.
They lasted 10 minutes until a spike on the mains blew them up.
I replaced them with parts from a "reputable" uk dealer and still going over a year on.
While these copy parts are often functional you dont know what corners have been cut.
I opened one up and the die is about half the size of the genuine part.

The moral of the story is pay a little more and get genuine parts and their will be much less heartache.
I now just use Farnell and RS Components in the UK and not had any problems.








 

Offline GeorgeOfTheJungle

  • Super Contributor
  • ***
  • !
  • Posts: 2699
  • Country: tr
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #122 on: February 02, 2020, 08:00:42 pm »
Does anybody know the figures in the silicon art?

That's what engineerds do!
The further a society drifts from truth, the more it will hate those who speak it.
 

Offline techman-001

  • Frequent Contributor
  • **
  • !
  • Posts: 748
  • Country: au
  • Electronics technician for the last 50 years
    • Mecrisp Stellaris Unofficial UserDoc
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #123 on: February 02, 2020, 09:20:22 pm »
I'm done with F103's - Chance of fake is to big.

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

/Bingo

I bought in some Chinese irfp240/9240 transistors because they were slightly cheaper.
They lasted 10 minutes until a spike on the mains blew them up.
I replaced them with parts from a "reputable" uk dealer and still going over a year on.
While these copy parts are often functional you dont know what corners have been cut.
I opened one up and the die is about half the size of the genuine part.

The moral of the story is pay a little more and get genuine parts and their will be much less heartache.
I now just use Farnell and RS Components in the UK and not had any problems.


While viewing the outstandingly clear decapping PR0N pics at https://zeptobars.com/en/ the author comments noted that many of the clones (and there are a LOT) whilst quite nicely designed, seemed to view silicon die area reduction as a priority.

In other words, many of the 'clone' dies are a lot smaller than the genuine articles so in the case of your irfp240 clones this reduction in die size probably didn't bode well when dealing with hundreds of volts ?

 

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Cheap Bluepill, very likely it has fake STM32 right ?
« Reply #124 on: February 03, 2020, 05:08:53 am »
Die size reduction on the GD32 is because of using a newer process. The fake transistors you got are most likely relabeled crap and not a proper "clone". If you go the effort of actually designing a clone ASIC there is no reason to skimp on reliability because it doesn't cost more to design it properly.
Email: OwOwOwOwO123@outlook.com
 
The following users thanked this post: thm_w


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf