microchip or atmel ARM's

[Kjelt]

Newer ARM devices from NXP tend to have EEPROM nowadays with pretty good retention time and endurance (100 years and 1million cycles IIRC). One of the advantages of internal EEPROM is that it is slightly harder to hack

If you have sensitive info on the external eeprom you just encrypt the data and perhaps sign it if you want to check the integrity. The key you keep internal to the uC or derive it from that key.
That is so common practice nowadays and costs so little cycles that is not a great disadvantage. I do not know how large the internal eeproms are but now you can choose 1MB flash or whatever you want for instance to store firmware updates. Try that with an internal 8kB eeprom     
BOM cost and PCB surface would count more I think.

[Simon]

Well I got an ST Nucleo-F091RC board and guess what ? yup, nothing supports it! cube MX only goes "down" to a F1 device board and so does the arduino addon, it's like F0 chips don't exist. So basically the message is that rather than totally waste your time listening to the bullshit people come out with is be prepared to go bare metal! because somehow there is always a catch. I mean this is not hard but it is being made hard and I'm wasting more time trying to do it he easy way than just doing it the hard way! Further the cube mx softwarte does not even display properly on my monitor, another piece of junk ware thrown out there to draw people in.

[Kalin]

Well I got an ST Nucleo-F091RC board and guess what ? yup, nothing supports it! cube MX only goes "down" to a F1 device board and so does the arduino addon, it's like F0 chips don't exist. So basically the message is that rather than totally waste your time listening to the bullshit people come out with is be prepared to go bare metal! because somehow there is always a catch. I mean this is not hard but it is being made hard and I'm wasting more time trying to do it he easy way than just doing it the hard way! Further the cube mx softwarte does not even display properly on my monitor, another piece of junk ware thrown out there to draw people in.

You can use the mbed website to program it. Is that an option for your application?

Sent from my SM-G930W8 using Tapatalk

[Simon]

an online program, yea that just gives me the shivers. I'm also not sure how/if it helps me setup the chip or am I doing a lot of low level stuff. Just watching Dave video on doing the same thing and he was about as baffled at the time. This is being made unnecessarily hard.

[Kalin]

an online program, yea that just gives me the shivers. I'm also not sure how/if it helps me setup the chip or am I doing a lot of low level stuff. Just watching Dave video on doing the same thing and he was about as baffled at the time. This is being made unnecessarily hard.

I have found it very easy to start with and I am not an ee. No low-level stuff and very quick to start. After you compile you program it downloads a bin file that you drag to the boards and that's it.

Sent from my SM-G930W8 using Tapatalk

[Simon]

Yea I am just "trying" to download the USB driver, which requires a log in, I am AGAIN waiting a ling time just to log in. ST suck balls big time in terms of anything that is not the chip itself (if I am to believe everything I am being told that I can only verify having jumped through so may hoops I will feel locked in!)

I have written this post in less time than it has taken to log in and fail and now i will have to try again to in to the ShiTware website.

[Kalin]

Yea I am just "trying" to download the USB driver, which requires a log in, I am AGAIN waiting a ling time just to log in. ST suck balls big time in terms of anything that is not the chip itself (if I am to believe everything I am being told that I can only verify having jumped through so may hoops I will feel locked in!)

I have written this post in less time than it has taken to log in and fail and now i will have to try again to in to the ShiTware website.

I guess I was able to avoid that headache using Linux. Drivers weren't an issue.  I think your frustrations probably are a result of windows and the related drivers not because of ST.


Sent from my SM-G930W8 using Tapatalk

[Simon]

Well having to log in to download a driver is not the fault of windows. Hopefully I only have to restart my machine to make it work as the board still does nothing when plugged in. The board that had zero instructions. Yes arm is hard, but this is making it harder.....

Sent from my Moto G (4) using Tapatalk

[Kalin]

Well having to log in to download a driver is not the fault of windows. Hopefully I only have to restart my machine to make it work as the board still does nothing when plugged in. The board that had zero instructions. Yes arm is hard, but this is making it harder.....

Sent from my Moto G (4) using Tapatalk

I didn't? have to install any drivers on Linux. I plugged it in and it showed up immediately as a mass storage device

Sent from my SM-G930W8 using Tapatalk

[julian1]

On linux getting started with stm32 was as easy as apt-get install the arm compilers, apt-get install openocd for the swd comms, apt-get install gdb, apt-get install gmake etc.

Then git clone the libopencm3 libraries. I was up and running and modifying blinky with an arm discovery board in 20 minutes. Getting a completely undocumented/unblogged about Chinese stm32f4 board from ebay - with remote debugging over stlink and with most of the basic peripherals works, writing my own ring-buffer for uart, dac control etc took a few days.

Likewise, for arduino, I never installed any external party's application/dev tools. Likewise for basic fpga and verilog development on the ice40.

The more languages and software that you need to code for and support - the simpler you want the development toolkit to be, and the smaller the dependency footprint. My first instinct would be to run a mile from any vendor created gui to manage auto-magically handle the configuration and make things 'simple'.

Honestly how does anyone put up with rebooting to install some stupid proprietary comms driver?

Disclaimer - I am not a professional embedded developer.

[Throy]

Well I got an ST Nucleo-F091RC board and guess what ? yup, nothing supports it! cube MX only goes "down" to a F1 device board and so does the arduino addon, it's like F0 chips don't exist. So basically the message is that rather than totally waste your time listening to the bullshit people come out with is be prepared to go bare metal! because somehow there is always a catch. I mean this is not hard but it is being made hard and I'm wasting more time trying to do it he easy way than just doing it the hard way! Further the cube mx softwarte does not even display properly on my monitor, another piece of junk ware thrown out there to draw people in.

After taking about 10 seconds to start a new project and use the board filter built into Cube, I found your ST Nucleo-F091RC that isn't supported...

[Simon]

ok I have found it also, although an updates was downloaded as well before it let me open it. This software was never made for 4K screens and reading text in it is not easy. In any case it's the same result as reading what chip you have on the board and just choosing it in the chip setup menu instead of board setup menu, using the board selection does not relate functionality to the headers or anything like that if for example you are planing to design a PCB for the nucleo to plug into. The software I doubt was written by ST themselves as there is a select your chip brand option.

I have downloaded SW4STM32 and not got very far with it yet, I'm still trying to find some example code just to see what a project looks like, naturally the one and only document out of several on the page for this nucleo board that is any use does not match reality, Not that I am at all surprised.....

[Simon]

oh of course if I select my nucleo board instead of just the chip it tells me what pin the button and LED's are on, because the non existent documentation for this board does not tell you this sort of stuff....

[Simon]

oh and suddenly SW4STM32 wants to install a ton of "updates" I suspect it's the STM32 specific bits of what is otherwise a generic IDE called eclipse.

[ebclr]

Wanna play easy and fast with ST-Nucleo-F091RC

http://developer.mbed.org/platforms/ST-Nucleo-F091RC/

[Kjelt]

Simon really 
St has board schematics bom and even pcb layout for all its boards, no other manufacturer does that.
I am going to stop reading this topic because it looks you are prejudiced to the whole thing and have your mind already made up that it s*cks so everything you do will only confirm this.
Please go use At el or NXP to find out they also suck and dont have a 1,23 ready to go solution for you.

Just want to add one thing if you go baremetal as you call it most programmers only do the sunny day scenario. That works till you meet rain. The ST HAL is bloated I agree but it does handle sunny and rainy day scenarios and all possible rainy day scenarios cost five times more code than the sunny day. So if you want realibility 24/7 operation you better add all recovery options or you have to add the watchdog and accept hard resets on occasions.

[Simon]

Simon really 
St has board schematics bom and even pcb layout for all its boards, no other manufacturer does that.

I am yet to find a scematic, I have been through all of the PDF's on the Nucleo-F091RC page, none have a schematic and infact none have any information that is very useful. The headline document has one page of specs, half a page of waffle, one and a half pages of lists of other Nucleo boards available and a page of disclaimers, nothing really useful about the board.

I am going to stop reading this topic because it looks you are prejudiced to the whole thing and have your mind already made up that it s*cks so everything you do will only confirm this.
Please go use At el or NXP to find out they also suck and dont have a 1,23 ready to go solution for you.

Indeed ST does seem to offer the best "entry" solution as badly formed as it is, I am sorry if the cube software is just about legible on my screen. It is indeed a great idea but after that I need to work in an IDE, after creating 2 accounts I have downloaded the "link" driver that has done nothing for me, my nucleo board still sits here unrecognized by the computer. I have down loaded the SW4STM32 or A6 or whatever it is called and will be looking at that when i have more time.

Just want to add one thing if you go baremetal as you call it most programmers only do the sunny day scenario. That works till you meet rain. The ST HAL is bloated I agree but it does handle sunny and rainy day scenarios and all possible rainy day scenarios cost five times more code than the sunny day. So if you want realibility 24/7 operation you better add all recovery options or you have to add the watchdog and accept hard resets on occasions.

Indeed which is why I have been persevering as i don't really want to learn a 1000+ page document of registers if I can help it and then spend hours writing setup code as I have had to do just for a "simple" AVR.

However, I am encountering the same sorts of problems Dave did and documented in a video about how broken the links are between various aspects. I am not a toolchain expert and thought I was going to start learning to program ARM micros rather than learn how to make different parts of a tool-chain talk to each other. I have read countless pages that just send me to other sites that just send me to other sites. As Dave said there is no option to just download something that works to get you going. Sure the reward might be better than Atmel at the end but it's one hell of a ride until I figure out how to make all of the components of the tool chain work together.

Considering they taunt you with the idea that it is arduino compatible in layout, it is quite deflating to find that this statement is purely for the marketing spin as setting up with ST is more complex than bare metalling another manufacturer never mind with an arduino!

[mrm2007]

Hi,

This is Nucleo-64 series user manual (Nucleo-F091RC is one of them)-UM1724 :

www.st.com/resource/en/user_manual/dm00105823.pdf

At the end of the document  (Appendix A) you can see  the schematics.

Sent from my E2333 using Tapatalk

[Simon]

Thank you, that looks like a useful document, looks like they didn't include it on the specific board page as it covers the whole range.

[Kjelt]

Just checked that indeed is an omission of st, normally there are three zip files one with the gerbers one with the bom, both are there for your board and one with the schematic which is missing from your page for your board.

Perhaps I am just lucky getting the chain to work for my two L4 boards in two evenings or the many years experience using toolchains, I don't know.
Fact is if you look at the many topics on the st forum and here on this forum using these chips is pretty serious stuff with a lot of pitfalls and challenges which take many hours to investigate experience to solve. The same for Atmel or NXP or other vendors.

[hans]

Yes, the PDF linked by mrm2007 is on the Nucleo product page.
It's called the user manual , last PDF on product page.

I have to agree with Kjelt. This topic has tendency to turn into "where can I find this pdf, its not on the page, much complicated, so sucks". Try to be more resourceful. We all went through the same hoops, it's not going to change a thing between vendors because it's all the same but with a different flavour of sauce. It's not ideal, but finding a way through the system helps.

And yes Microchip/AVR will have the same thing. If you look at a recent PIC24 or PIC32 datasheet, it wont tell much specific about the peripherals anymore. It is in a separate document. Maybe if you're lucky a brief overview of the registers are in there, but it does not explain in an animated way what each bit does.
Atmel used to publish summary datasheets for their parts. That can be confusing as well.

Overcoming 4K difficulties: maybe try Windows magnifier?
I know it sounds stupid, but in XFCE on Linux it has the magnifier hotkeyed to ALT + Scroll. If my eyes are fatigued, I'm sitting far from my screen (like chair back is all the way down), or just plain lazy, it's so helpful.
And no I don't even have 4K or 1440p that is.

[Simon]

Yes it is a mine field, which is why you want the bits that don't need to be made hard, made hard. my understanding is that I need and IDE to hold it all together, a compiler and something from the chip vendor (header files)to give sensible names to pins and registers rather than address numbers.

I have several choices of IDE non of which are probably setup for any particular vendor as they are generic. I believe they come with compilers so it's just a case of getting them to talk the language of the standard definitions for that chip. Of course if I could get my programmer or whatever they care to call the bit on the board that connectors to the computer that you can snap off to even be recognized that would be a bonus.

MBED is not an attractive option for me, I need something to use professionally, I'd not use MBED any more than I would use circuit maker

[Simon]

Yes, the PDF linked by mrm2007 is on the Nucleo product page.
It's called the user manual , last PDF on product page.

Well maybe I just dismissed it mistakenly when it looked like just another generic bit of stuff like every other PDF on that page as it starts off looking more like a catalogue than anything else.

[Simon]

Overcoming 4K difficulties: maybe try Windows magnifier?
I know it sounds stupid, but in XFCE on Linux it has the magnifier hotkeyed to ALT + Scroll. If my eyes are fatigued, I'm sitting far from my screen (like chair back is all the way down), or just plain lazy, it's so helpful.
And no I don't even have 4K or 1440p that is.

It's not as simple as that, initially, although it seems to have fixed itself a bit since the update stuff did not fit properly on the side bare as it failed to layout properly. After the update it looks a bit better although the search box is too small. It's not size of text it's actual graphical problems where it can't cope with unexpected resolutions and some bit's adapt whereas others don't.

[Simon]

ahem, just plugged a second nucleo board in and it's all lights a blazing. I'm thinking the first one might be a dud, or something is up with the computer.