STM32C0 series?

[ataradov]

Everything in a car's trunk is still either 12V power or 5V signal

But you never connect MCU directly to the car's system. In most cases you need a transceiver (CAN/LIN)  anyway. The number of typical I/O is typically very low, and even if you have to put level shifters there, it always offsets the cost of finding other 5V components. MCU is not the only things in the system. So, one way or the other you would have to level shift something.

5V logic was often used to improve noise immunity, but with improvement in design techniques this is not a factor anymore.

I review a lot of automotive designs, absolute majority is based on the 3.3V logic.

[JPortici]

I interface directly with sensors and actuators. 3V3 designs cost me around 20% more than a 5V design, if the 5V design is possible EG the combination of peripherals and comms are available in a 5V microcontroller, usually it is (dsPIC33EV series). The savings (if any) from the 3V3 microcontroller are canceled out by increased costs due to more components - as in more BOM lines-, more silicon (such as the additional regulator, the additional mosfet drivers if i can't directly drive low side mosfets anymore) and the increased board space.

[PatocheHAM]

Delivering both, automotive and non automotive MCUs, I can say that Automotive is still using by far 5V MCUs.  It may eventually depends on the usage.
Car body are 5V definitly, mainly for legacy and BOM cost reasons. In some case when displays are present, 3.3V supply are existing and therfore a 3.3V device with 5V compatible I/Os will be acceptable.
For after market devices that won't require AEC-Q100 and the car makers guarantees, we see 3.3V devices with aven non-auto grade components.
 
 

[PatocheHAM]

"The DIP28 pinout is designed to be as compatible as possible with the ATMEGA328 8‐bit microcontroller." This is like board vendors putting shield connectors on all kits. I have never seen anyone actually use them for anything. The attractive part of the Arduino is not connectors or the MCU, it is the firmware ecosystem. And I'm thinking ST is not going to do much work there, and tell the story how Cube is better.

It looks like a 5V I/O device. Pretty pointless if you ask me. If it is extremely cheap, there would be uses, but if it costs as much as comparable devices, seems strange to me to bring such a device to market.

They probably thought Atmega's would go completely out of stock during the shortage, so they spun this up as an idea to capture the marketshare.
Shields I've never personally used, but, there is a lot of value in having a standardized pinout. Also having lots of pin headers is never bad on a dev board.
https://www.google.com/search?q=nucleo+shield

32-bit MCUs are already massively cheaper than 8-bit MCUs. The only way you can move 8-bit people to 32 bits is if you dumb down all the peripherals to the 8-bit level. But that does not seem like a good idea.

Some are some aren't, the cheapest MCUs are 8-bit Padauks right.

I do like the fact that i can move up/down the line in microchip parts with no redesigns, the same pinout (or with minor differences like a VDD pin becoming a GPIO with zero alternate functions) is used across most 8bit/16it and some 32bit parts. I welcome standardised pinouts (i was thinking of replacing a MHCP part with an atmel part that had almost the same pinout but swapped USB pins. SWAPPED. ugh

This is the idea with the STM32C0.  The pinout compatibility with the STM32G0 ( going form 16K Flash to 512KB) is granted. therefore for basic applications the STM32C0 will stretch the ST cortex M0 offer down to the level of the cheapest 8-bit devices (e.g. the STM8).  I cannot tell much as the device is not officially released. But the STM32 users will be able to keep their developments environment and software libraries for small applications.

Any questions?
   

[uer166]

Delivering both, automotive and non automotive MCUs, I can say that Automotive is still using by far 5V MCUs.

That is indeed still the case, surprisingly. Most body controllers I've seen in the latest and greatest EVs are SPC5 series, 3-5V range. ST just released a new series, Stellar, with Cortex-R and a Cortex-M7, seems to be direct competition to some TMS320 jobs, not sure if they finally dropped the 5V on those but the chips sure do look cool.

[AndyC_772]

Oh, please tell me these parts can replace dsPIC33EV, you'll make my day.

150C temperature rating and 5V I/O? Pretty please?

[Scrts]

Oh, please tell me these parts can replace dsPIC33EV, you'll make my day.

150C temperature rating and 5V I/O? Pretty please?

Look up Renesas RH850 and FL78, but not sure about their stock either...

[PatocheHAM]

Oh, please tell me these parts can replace dsPIC33EV, you'll make my day.

150C temperature rating and 5V I/O? Pretty please?

Look up Renesas RH850 and FL78, but not sure about their stock either...

The mission profile of the STM32C0 is Consumer / Industrial. not automotive. the spec max will reach 125°C TA . FOr 150dgC I suggest to check either the STM8A (but It may not repalce the dsPIC33.) you may wan to check the MCU auto portfolio of ST. STPC56.

[alpha-zero]

From STM32C031.svd:
  <name>STM32C031</name>
  <version>0.2</version>   
  <description>STM32C031</description> 
  <cpu>
    <name>CM0</name>
    <revision>r0p0</revision>
    <endian>little</endian>
    <mpuPresent>true</mpuPresent>
    <fpuPresent>true</fpuPresent>
    <nvicPrioBits>4</nvicPrioBits>
    <vendorSystickConfig>false</vendorSystickConfig>
  </cpu>

[ataradov]

This is just an invalid SVD file, not that anyone really cares, half the stuff in SVD files is junk that is ignored by most implementations. They may have some FPU, may be on the IOBUS or just as a peripheral. But it is not correct to mark it using this field.

Here is how it is described in the SWD doceumentation: "Indicate whether the processor is equipped with a hardware floating point unit (FPU). Cortex-M4, Cortex-M7, Cortex-M33 and Cortex-M35P are the only available Cortex-M processor with an optional FPU."

This list is missing Cortex-M55, but Cortex-M0+ is definitely not there.

[thm_w]

Pages are all up now:
https://www.st.com/en/microcontrollers-microprocessors/stm32c0-series.html
https://www.cnx-software.com/2023/01/13/low-cost-stm32c0-32-bit-microcontroller-aims-to-displace-8-bit-mcus/

"consistent pinout with G0, same IP platform"
3.3V, 2.5MSPS ADC.
Seems like a near pointless marketing extension of G0.

In terms of st estore pricing:
- STM32C011 is $1.01
- STM32G050 $1.24
- STM32F030 $1.11

Although these prices are inflated as you can get F0/G0 from LCSC for <50c each.
So should indicate a ~10% price reduction in bulk volumes.

[ataradov]

Just 5V tolerant and not even for all I/Os. It is good to have more options, but there is nothing that really stands out here. Not sure why they needed a separate series for that.

The datasheet from Mouser  has "ST RESTRICTED – SUBJECT TO NON-DISCLOSURE AGREEMENT – DO NOT COPY"  all over it. LOL.

[uer166]

Just 5V tolerant and not even for all I/Os. It is good to have more options, but there is nothing that really stands out here. Not sure why they needed a separate series for that.

The datasheet from Mouser  has "ST RESTRICTED – SUBJECT TO NON-DISCLOSURE AGREEMENT – DO NOT COPY"  all over it. LOL.

This series is explicitly to replace 8-bitters, these parts are supposedly noticeably lower cost than any other STM32, but you'll probably only care if you're buying in the millions.

[ataradov]

But G0 series did the same exact thing. There may be some further price reduction, but I really don't see why this could not have been done within G0 series. This C0 just creates more confusion.

[uer166]

But G0 series did the same exact thing. There may be some further price reduction, but I really don't see why this could not have been done within G0 series. This C0 just creates more confusion.

G0 has some decent analog features, maybe they want to differentiate their analogey G-series (I know I personally associate it with true mixed signal), from a bottom-of-the-barrel MCU.

[PCB.Wiz]

....Seems like a near pointless marketing extension of G0.

In terms of st estore pricing:
- STM32C011 is $1.01
- STM32G050 $1.24
- STM32F030 $1.11

Although these prices are inflated as you can get F0/G0 from LCSC for <50c each.
So should indicate a ~10% price reduction in bulk volumes.

But G0 series did the same exact thing. There may be some further price reduction, but I really don't see why this could not have been done within G0 series. This C0 just creates more confusion.

Yes, the visible prices are not low cost at all, and their decision to release a new low pin part, without wide Vcc is very strange, given almost everyone else releasing 8~20 pin parts chasing 8-bit displacement, is doing 1.8~5.5 range parts.

It's not even especially low power, it says
Stop mode In Stop mode, the device achieves the lowest power consumption while retaining the SRAM and register contents
Most Stop Mode values are <110uA at 25C and <1.1mA (!) at 125'C

The two lower power modes than that, lose information.
The lowest power Shutdown mode works like a CMOS switch in the power rail, so MCU restarts.

[I wanted a rude username]

Seems like a near pointless marketing extension of G0.

It's different enough to warrant its own letter. For example the clock is HSI48, which means you don't get much benefit running it at 16 MHz like you do with the L0 and G0 with their HSI16. It's a cost-optimised 8-bit-killer, just as ST say it is.

[tim_]

Just 5V tolerant and not even for all I/Os. It is good to have more options, but there is nothing that really stands out here. Not sure why they needed a separate series for that.

If you look at the size of the WLCSP products to the STM32G you can see that the die size of the comparable STM32C is significantly smaller (4mm² vs 2.4mm²), while using the same 90nm technology. So removing periphery and possibly optimizing layout more is basically a huge cost reduction. The purpose of the device is as stated - compete with 8bit in price. It's a STM8S killer.

[westfw]

Just 5V tolerant and not even for all I/Os.

It looks to me like all the I/Os are 5V tolerant.  They've got designations for non-5vT pins, but all the actual pins are marked as "FT" (except RESET?)
Trying for name similarity to Microchip's SAMC and PIC32C product lines?  (although IIRC both of those have actual 5V operation.)

[ataradov]

PIC32C has nothing to do with voltage tolerance. "C" stands for "Cortex-M" here. There are a ton of 3.3V-only devices in PIC32C series and all Cortex-M devices would be PIC32C. PIC32CM MC are 5V capable because they are essentially SAM C devices with minimal improvements.

Their wording implies that not all pins are tolerant, so may be they expect higher pin count devices. But this is not all that important. In a lot of cases where you need 5V operation simply being tolerant is not enough. It might work as a transition device, I guess.

[SiliconWizard]

Note that the C0 series has pretty small flash size, and given the bloat of the HAL, using it with the C0 is probably going to be funny. For those who can't live without the HAL.

[DavidAlfa]

 The HAL problem is really withthe RAM, with all those handlers.
It's pretty obvious you'll have to use LL with small devices.
Not that hard since their peripherals are usually way simpler than say H7 series.

[ataradov]

And people transitioning from STM8 have to be used to that anyway.

[cv007]

G0 has some decent analog features, maybe they want to differentiate their analogey G-series (I know I personally associate it with true mixed signal), from a bottom-of-the-barrel MCU.

A number of peripherals already go missing depending on which G0 you are talking about, so you do not get DAC/COMP/RNG/AES/etc for a G031, and a G030 will be even more bare bones. Pretty hard to use a generic reference to a G series when the whole line of G may differ quite a bit, and seems to be the case with many of their series.

[thm_w]

Seems like a near pointless marketing extension of G0.

It's different enough to warrant its own letter. For example the clock is HSI48, which means you don't get much benefit running it at 16 MHz like you do with the L0 and G0 with their HSI16. It's a cost-optimised 8-bit-killer, just as ST say it is.

I'll believe it when I see it, right now no one is going to be replacing a <15c STM8 with a $1 C0.

Looks like its being discussed in the 10c WCH RISCV thread: https://www.eevblog.com/forum/microcontrollers/wch-$0-10-risc-v-mcu/125/
There is also the PUYA <10c ARM ICs as well: https://www.lcsc.com/product-detail/Microcontroller-Units-MCUs-MPUs-SOCs_PUYA-PY32F002AA15M6TU_C5292058.html