ST beat NXP again in the MCU battle

[tszaboo]

NXP proudly announced the LPC800 series, not so long ago, with a DIP8 package 32 bit microcontroller, prices are low.
Yesterday, ST announced the STM32F030 series.
It is lower price, has a 12 bit ADC, up to 40 something GPIO in contrast to 18, up to four times the flash, and higher speed.
And in the meantime, the Zero geckos, and the Kinetis L0 has arrived.
So what do you think, who will win the Cortex M0/M0+ battle?

[BravoV]

So what do you think, who will win the Cortex M0/M0+ battle?

How about you answer that question 1st, which one do you preferred ? 

[ElectroIrradiator]

Not sure I can follow your reasoning?

NXP have had numerous M0/M3/M4 families out there for a while, including a variety of M0 MCUs in various packages, speed grades and other capabilities. They even have the kit friendly LPC1114FN28 in a DIP-28, running at up to 50 MHz clock speed.

12 bit ADCs on the low end are nice but not particularly unusual.

[tszaboo]

I think if ST starts shipping, we have a clear winner for the moment, judging by Digikey prices, and reel/tray quantities. But maybe it is worth doing a sub-dollar MCU selection table, and ask vendors for real pricing.

Not sure I can follow your reasoning?

The LPC1114 is about 50% more expensive for almost the same setup.

[dannyf]

So what do you think, who will win the Cortex M0/M0+ battle?

Given the title yo have given to the thread,

ST beat NXP again in the MCU battle

it doesn't seem to me that you have given us the chance to think at all.

[free_electron]

Who will win ? ARM ! As all these processors make them money !

[dannyf]

But maybe it is worth doing a sub-dollar MCU selection table,

there are many such chips, very few are known outside of China.

[BravoV]

Who will win ? ARM ! As all these processors make them money !

Damn, you're so right, and pssstt ... I'm also a happy camper just to be tiny-weeney part of it, hint -> check the date.    

[nctnico]

I think if ST starts shipping, we have a clear winner for the moment, judging by Digikey prices, and reel/tray quantities. But maybe it is worth doing a sub-dollar MCU selection table, and ask vendors for real pricing.

Not sure I can follow your reasoning?

The LPC1114 is about 50% more expensive for almost the same setup.

I'd rather pay more for something which is actually working. The details are in the fine print.

[BravoV]

I think if ST starts shipping, we have a clear winner for the moment, judging by Digikey prices, and reel/tray quantities. But maybe it is worth doing a sub-dollar MCU selection table, and ask vendors for real pricing.

Not sure I can follow your reasoning?

The LPC1114 is about 50% more expensive for almost the same setup.

I'd rather pay more for something which is actually working. The details are in the fine print.

Maybe he means it for enthusiast or hobbyist perpective ? 

[baoshi]

Curiously STM32 price in China is already (1) order of magnitude lower than DigiKey

But maybe it is worth doing a sub-dollar MCU selection table,

there are many such chips, very few are known outside of China.

[westfw]

Yesterday, ST announced the STM32F030 series.

Um.  You mean like back in July?  "32 bits for 32 cents" ?
The LPC800 was interesting for reasons other than price that "yet another ARM in a 64pin lqfp" isn't...
(Heh.  ARM is the best performing stock in my personal portfolio.  Up over 600% since I bought in 06.  Microchip up 76% since 09 is second best; I actually made more on the MCHP; I should've bought more ARM.  They've both been paying dividends, too...)

[tszaboo]

But maybe it is worth doing a sub-dollar MCU selection table,

there are many such chips, very few are known outside of China.

Yes, but as far as I'm interested, only Cortex M3/0/0+ would make it to this table. I know, that there is PIC10 or MSP430 which is cheaper than the cheapest M0, but I also have to keep in mind, that they offer so much more, that I never would consider a PIC10. And I'm well avare, that there are a lot 8051 based, or even 4 bit MCU, which are cheaper than a not gate from Texas instruments, but for some magical reason, I would like to buy the chips. And download the datasheets. And have programming cables, and ICPs, with proper documentation. ARM does that, and to top this, you can give the board to an "out of university" programmer to make the code.
Not to mention, NXP and ST are EU companies.

Yesterday, ST announced the STM32F030 series.

Um.  You mean like back in July?  "32 bits for 32 cents" ?

I dont know, I got the mail from them yesterday, "Microcontroller news from STMicroelectronics November 2013". So far no stocks.

I made the question, because the battle is far from over. We are still waiting for Atmel, the D20 is positioned above these, and the name also suggests, that something is coming below that.

[dannyf]

but as far as I'm interested, only Cortex M3/0/0+ would make it to this table.

So you are really asking us to compile a list of sub-dollar mcus that YOU are interested in.

I don't think you will find a lot of interests in an exercise like that.

[tszaboo]

but as far as I'm interested, only Cortex M3/0/0+ would make it to this table.

So you are really asking us to compile a list of sub-dollar mcus that YOU are interested in.

I don't think you will find a lot of interests in an exercise like that.

No, thank you I am perfectly capable of doing that. I was asking for opinions. Like the sentence: "what do you think"

[westfw]

So far no stocks.

Mouser claims to have some STM32F030 chips as well as the LPC800 chips.
I gotta say that I don't find the price difference between a $1.15 chip (LPC800 in SOP20, q100) and a $0.89 chip (STMF030 in SSOP20, q100) to be very limiting.

[mikeselectricstuff]

At the lower end, factors other than chip cost an be signifcant, like does it have a sufficiently accurate internal osc to not need an external one, and how easy/quick is production programming.
I've sometimes designed in a PIC primarily because I can get them preprogrammed and marked from Microchip for a few pennies. If you can build in most of the test functionality, your  entire post-production program/test cycle can become a simple as just turning it on. And of course this could even mean just applying power turning on a whole panel of boards to test multiple units at once.
On a current project, the PIC also programs the onboard CPLD when first powered up. 
Other manufacturers should learn from this, though they seem to be stuck in the old "buy only through distributors" mindset.
 

[dannyf]

thank you I am perfectly capable of doing that.

Don't thank me: I was merely helping you clarify exactly where you were asking.

I was asking for opinions. Like the sentence: "what do you think"

More precisely, you were asking what sub-dollar mcus we think you would be interested in. I would say you are the only one in the world qualified to answer that q question.

[tszaboo]

At the lower end, factors other than chip cost an be signifcant, like does it have a sufficiently accurate internal osc to not need an external one, and how easy/quick is production programming.

That is a very interesting insight. The LPC800 has a 12 MHz +/- 1.5 % accurate internal oscilaltor, while the STM32F030 is 8Mhz +/- 5% which is user trimable to 1%. I prefer the LPC. With the STM it could easily be that if you have a serial port, the timing will be so incorrect, that your communication fails.

As for programming, I prefer serial port bootloaders and dedicated in-circuit-serial-debug interfaces for development. But indeed, a preprogrammed MCU can be an advantage. Silica and other vendors offers programming services. But I still prefer doing programming during the ICT testing.

[timb]

I love the FRAM series MSP430, but one issue is you can't pre-program them due to the possibility of bits flipping during reflow.

[Kjelt]

I gotta say that I don't find the price difference between a $1.15 chip (LPC800 in SOP20, q100) and a $0.89 chip (STMF030 in SSOP20, q100) to be very limiting.

But that is not their main target market. The target market is single companies that buy millions a year.
If you have those numbers the prices are totally different and if ST sells (as an example) for $0.10 less then NXP that is a difference of $100000 per million pieces.

There are very strange decisions that are being made at that level, as an example for our last product we wanted a 32 bits cortex M3 with 128KiB ROM and 16KiB RAM but were given a uC that had more then double that because it already was used in three other products. The bigger chip was (combined in the total bulk buyin) actually much cheaper then the smaller chip we first requested. That is fortunate, the other way round also sometimes happens and then we need a lot of time for sw engineering or limit the product requirements to make ends meet.

[poorchava]

So far no stocks.

Mouser claims to have some STM32F030 chips as well as the LPC800 chips.
I gotta say that I don't find the price difference between a $1.15 chip (LPC800 in SOP20, q100) and a $0.89 chip (STMF030 in SSOP20, q100) to be very limiting.

kamami.com.pl also claims to have some STM32F030's (which for some reason doesn't show up on English version of their web shop)

[tszaboo]

If you have those numbers the prices are totally different and if ST sells (as an example) for $0.10 less then NXP that is a difference of $100000 per million pieces.

Not to mention, that they are making a lot of ASICs obsolete, or at least not recommended anymore. The cheapest 12 bit ADCs at digikey are 0.89 USD, this MCU is cheaper. The cheapest 16 bit GPIO extender costs 0.62 USD @10000, the MCU becomes cheaper. And the list goes on. It can really turn the industry upside down.

[Kjelt]

It can really turn the industry upside down.

I hadn't looked at it from that point of view, its kinda rediculous isn't it? A 32 bits micro cheaper then a simple I/O expander.

[nctnico]

And more versatile as well. A couple of years ago I designed a keypad (based on touch thru metal sensors) which needed to be controlled/read by I2C. When looking for parts an ARM controller from NXP was much cheaper than an I2C ADC. And I would need some opamps as well. The end result is a PCB with about 10 passive components, a microcontroller and a connector. It can't be any simpler.