What's an STM32GC102CB ?

[tsman]

So IMO, either it was some kind of clone, or some kind of prototyping batch from STM that has leaked from some chinese plant. Would be interesting to have the opinion of STM on this one though....

It would be strange to go to the hassle of cloning a STM32F1 but then still call it a STM32 but change the family to GC. You'd either make your own range or also copy the F family designation. The GD32 devices are clones but they're marketed as a GigaDevices product and completely different inside.

My guess is still that they're a custom STM32F1 made for somebody and old/excess stock has been sold off. The ST-Link cloners were already pennypinching by using STM32F100s which don't officially have USB.

A while ago, somebody asked ST if the incredibly cheap STM32s available from China were counterfeit or real. ST said that there aren't any fake STM32s and that they're all real. They didn't explain why they were so cheap even compared with the bulk order pricing in China. Somebody must be buying a crazy number of these uCs.

[Yansi]

There is a huge market for STM32 and STM8 chips in china, as many of you already noticed them in many products from there already.

I can assure the prices, which these parts sell in large quantities (100ku or 1MU) are tiny compared to what those money hungry distributors often want from you, as an individual.

[amyk]

My guess is still that they're a custom STM32F1 made for somebody and old/excess stock has been sold off. The ST-Link cloners were already pennypinching by using STM32F100s which don't officially have USB.

So the USB controller is present, not officially advertised as working, but often it does? Interesting...

[ataradov]

So the USB controller is present, not officially advertised as working, but often it does? Interesting...

This happens very often. Probably all MCUs on the market have peripherals/features that are not advertised. Often this means that they are not fully compliant. But that's not a big deal for random Chinese manufacturers that never intended to pass compliance test anyway.

[Yansi]

It is way easier than that. It is more cost effective to manufacture less types of dies and market these as multiple devices.  It does not even pay off to split the manufacturing line at the chip level, so peripherals may get disabled/blocked on chip level. The only split in the manufacturing process is at the part labeling.

However, no-one will assure you, that your STM32F102 will have working USB all the time in any way shape or form. It has already happened changes have been made, manufacturing further optimized, so these "hacks" have stopped working suddenly.

The only peripherals that do get really blocked, are those that must conform to stupid export laws, so for example cryptography. You know, some states like Russia and the US of A does not like people to encrypt their properties too freely. 

[tsman]

However, no-one will assure you, that your STM32F102 will have working USB all the time in any way shape or form. It has already happened changes have been made, manufacturing further optimized, so these "hacks" have stopped working suddenly.

Yep. Same deal with the "bonus" flash on some of them. It definitely isn't guaranteed in any way but a lot of people rely on it now which is a bad idea for production.

[Yansi]

It is not a  bad idea, it is a very bad one.  Neither is it guaranteed to work at any time, nor will STMicro give you any support if problems occur. It is solely your responsibility, if you are bending the rules.

It may be fun for homebrewers, but definitely not for production.

[langwadt]

However, no-one will assure you, that your STM32F102 will have working USB all the time in any way shape or form. It has already happened changes have been made, manufacturing further optimized, so these "hacks" have stopped working suddenly.

Yep. Same deal with the "bonus" flash on some of them. It definitely isn't guaranteed in any way but a lot of people rely on it now which is a bad idea for production.

Yeh a place I worked someone accidentally used a section of RAM in DSP that wasn't supposed to be there but it worked because it was the
same die as a part with more ram, one day production stopped because nothing worked.-  The company making the chip had enough volume of the reduced ram version that they had spun a new chip that only had the specified RAM

[SiliconWizard]

It would be strange to go to the hassle of cloning a STM32F1 but then still call it a STM32 but change the family to GC. You'd either make your own range or also copy the F family designation.

Depends on what kind of "cloning" you consider. Redesigning a functionally-equivalent chip would be a hassle, but OTOH you just need to have gotten ahold of genuine bare dies one way or another. A few wafers may have gotten lost...  The packaging itself may be done in any chinese plant that would add the marking they want. That happens.

My guess is still that they're a custom STM32F1 made for somebody and old/excess stock has been sold off. The ST-Link cloners were already pennypinching by using STM32F100s which don't officially have USB.

I also suggested genuine "prototypes" - that could be custom-made chips as well indeed. But I don't know STM's policies when it comes to custom versions of their MCUs. I don't know if they do it. If it's the case here, using/selling custom-made chips for any other use than what was granted by STM is probably illegal anyway.

As to the pricing they have for different markets, it's not surprising. Many companies do that (especially for developing countries markets), and further deals with companies buying in huge volumes is common. Of course they are going to be secretive about the details. This is confidential business information. STM32 MCUs have a significant market share in China, and this is an extremely competitive market, so they have probably done what needed to be done to get there...