STM32 Roadmap?

[SiliconWizard]

What is the value of HS USB?

The full speed of USB2 may be overkill for most MCU applications, but the full speed of USB1 can be quite limiting. Especially during debug, where you might want to stream a lot of data to a PC for analysis. I think a lot of people have need for something with room to breath.

FYI, “Full Speed” in the context of USB doesn’t mean “the entire possible speed” as a description, but rather specifically means the 12Mbps mode. Subsequent USB versions don’t redefine the speeds, just add on more, so for example, USB 1.1 Full Speed is 12Mbps, and USB 2.0 Full Speed is also 12Mbps.

Low speed: 1.5Mbps
Full speed: 12Mbps
High speed: 480Mbps

And then I can’t keep track of the official nomenclature in USB 3.0 and later, since it was always confusing, and keeps changing…

I did say "the full speed off USB2", not "full speed USB". Because of funky terminology in the USB specs those are quite different things. They love funky terms in USB specs. Can USB2 send data at anything like 12Mbps (FS) or 480Mbps (HS)? That's also funky terminology.

Terminology is just terminology. Everything has history. USB has a long one already.
The missing link here is the low speed, without which we can't understand why they chose 'Full Speed' in the first place.

USB initially allowed two speeds: low-speed (1.5Mbps) and full-speed (12Mbps). The "full" was obviously to signify the full (max) bandwidth possible on then USB.

Then USB 2.0 introduced high-speed (480Mbps) while still being backwards compatible.

Then USB 3.0... as westfw said.

It's not funky, it's just that it has bagage, like everything else. But if you don't use the standard terminology, people are likely not to understand what you mean.

[harerod]

Speaking of terminology: I still remember a cover page of the German computer magazine "c't" calling that new interface: "Useless Serial Bus". I have used that term ever since.
At that time developers had to upgrade their machines to Win98 to get USB support. We had just started selling our new medical device (dozens of EEG channels), when Intel decided to drop out of the USB device chip game and we had to redesign the hardware and rewrite the firmware around the Cypress EZ-USB. Nobody cared for another USB device implementation in 8051 assembly, which led so the firmware being rewritten in Keil C. Windows driver development for that beast was even worse, blue screens galore. During that development we saw Win98, then ME and finally the rather decent XP.
Since then USB development has matured considerably.

[coppice]

Speaking of terminology: I still remember a cover page of the German computer magazine "c't" calling that new interface: "Useless Serial Bus". I have used that term ever since.
At that time developers had to upgrade their machines to Win98 to get USB support. We had just started selling our new medical device (dozens of EEG channels), when Intel decided to drop out of the USB device chip game and we had to redesign the hardware and rewrite the firmware around the Cypress EZ-USB. Nobody cared for another USB device implementation in 8051 assembly, which led so the firmware being rewritten in Keil C. Windows driver development for that beast was even worse, blue screens galore. During that development we saw Win98, then ME and finally the rather decent XP.
Since then USB development has matured considerably.

I think it got the name Useless Serial Bus because for several years it was entirely useless. Numerous peripheral  products, like mice and scanners, were developed, manufactured and died in the market because there was no way to use them. Peripheral makers must have lost a fortune on this. I used to see so many nice polished products in the computer malls in HK, being test marketed to a public that couldn't make use of them. Millions of PCs had that socket on the back that "was for the future", but definitely not for the present. Intel eventually joked that the US beat the Japanese in WW2 faster than the computer industry got USB into a marketable shape.

The original USB specs were very interesting in what they defined and did not define. Standardised serial ports? Nope. For that legacy reason we still have a bunch of incompatible drivers for those. Standardised ways to take a telecoms E1 or T1 carrying ISDN, and terminate it into a PC, with full plesiochronous support. Yep, 100% defined for the handful of people who might ever use it..

[harerod]

Well, USB-2-Serial became much simpler when FTDI issued their first FT232's. All we had to do was connect a serial EEPROM to host our VID/PID. Brings back memories of FTDI's minuscule stall at Electronica(?), staffed with actual Scottish technical expertise (Andrew?, Steward?).

[coppice]

Well, USB-2-Serial became much simpler when FTDI issued their first FT232's. All we had to do was connect a serial EEPROM to host our VID/PID. Brings back memories of FTDI's minuscule stall at Electronica(?), staffed with actual Scottish technical expertise (Andrew?, Steward?).

My experience was Prolific USB serial ports suddenly became.....er..... prolific, and their driver was already on most machines when you plugged into them.

[SiliconWizard]

I forgot to add USB 4.0.
They have changed the naming convention and that goes: USB4 20 Gbps and USB4 40 Gbps. More explicit.

[SiliconWizard]

Well, USB-2-Serial became much simpler when FTDI issued their first FT232's. All we had to do was connect a serial EEPROM to host our VID/PID. Brings back memories of FTDI's minuscule stall at Electronica(?), staffed with actual Scottish technical expertise (Andrew?, Steward?).

What do you again mean by "USB 2". FTDI's first chips were compatible with USB 1.1 only AFAIR.
And yes, that made adding USB to a device much easier. They had also released  a parallel FIFO version at about the same time (FT245), which allowed nearly the full bandwidth in USB FS, about 1 MBytes/s.
You could do the same with Cypress EZ-USB MCUs at the time which were pretty much the only other "simple" solution available, although it was more involved and didn't come with usable drivers.

FTDI's first USB 2.0 HS (high speed) came much later, around 2007 IIRC. Of course to get the benefit of HS you would need to use a parallel FIFO mode. Even on their latest chips, the UART <-> USB function tops at 3 Mbits/s, for which the bandwidth of FS is enough.

And then they released USB 3.0 SS chips (FT6xx). At this point, alternatives (such as Cypress FX3) are much more expensive actually. I think there are some chinese interface chips supporting USB SS, but I don't have references in mind.

I don't know if they'll ever release chips allowing the new speeds of USB 4. That's freaking fast.

[tooki]

Well, USB-2-Serial became much simpler when FTDI issued their first FT232's. All we had to do was connect a serial EEPROM to host our VID/PID. Brings back memories of FTDI's minuscule stall at Electronica(?), staffed with actual Scottish technical expertise (Andrew?, Steward?).

What do you again mean by "USB 2". FTDI's first chips were compatible with USB 1.1 only AFAIR.

I suspect that by “USB-2-Serial”, they meant “USB-to-serial”. Just a very unwise choice of spelling, methinks.

[SiliconWizard]

Well, USB-2-Serial became much simpler when FTDI issued their first FT232's. All we had to do was connect a serial EEPROM to host our VID/PID. Brings back memories of FTDI's minuscule stall at Electronica(?), staffed with actual Scottish technical expertise (Andrew?, Steward?).

What do you again mean by "USB 2". FTDI's first chips were compatible with USB 1.1 only AFAIR.

I suspect that by “USB-2-Serial”, they meant “USB-to-serial”. Just a very unwise choice of spelling, methinks.

Ouch
But at least maybe the history I gave helped some remember/or be aware of the offering.

[harerod]

... they ...

Don't do that to me, please. 

[alm]

... they ...

Don't do that to me, please. 

You can list your preferred pronoun in your profile / signature if you care that much

[tooki]

... they ...

Don't do that to me, please. 

So… don’t use English to you?!?
In English, the singular ”they” is a standard way of referring to someone whose gender you do not know. That usage is well-documented, going back to at least the 14th century.

(This is, incidentally, precisely the reason “they” has also become the dominant gender-neutral pronoun, beating all the invented ones like “ze”: it’s only a minor extension to a familiar usage. But that doesn’t mean it’s lost its original meaning!)

[harerod]

Tooki, I simply love our occasional exchanges about linguistics.
Sure, a lazy "they" is one viable device, in case you couldn't possibly know the referee's sex and wanted to churn out another quick post. However, one might venture that in a forum like this, the oldfangled "generic he" might be an even better bet.

Since I signed our private exchanges with my given name and having set my  gender in my forum profile, I permit myself to politely ask you to do this old grumpy engineer the courtesy of using that information.

https://en.wikipedia.org/wiki/Gender_in_English
https://en.wikipedia.org/wiki/Singular_they
https://en.wikipedia.org/wiki/Gender_neutrality_in_languages_with_gendered_third-person_pronouns#Generic_he

[tooki]

Tooki, I simply love our occasional exchanges about linguistics.
Sure, a lazy "they" is one viable device, in case you couldn't possibly know the referee's sex and wanted to churn out another quick post. However, one might venture that in a forum like this, the oldfangled "generic he" might be an even better bet.

Since I signed our private exchanges with my given name and having set my  gender in my forum profile, I permit myself to politely ask you to do this old grumpy engineer the courtesy of using that information.

As a native English speaker, I respectfully decline to strike a correct, valid usage of generic English terminology from my vocabulary (even if just for one person). I’m happy to use “he” referring to you in the future, but can’t guarantee I will remember the username and make the association.

Honestly, I didn’t remember your username or name, and I’m not going to check someone’s profile to see if they happen to have set their gender. It simply doesn’t matter that much when a perfectly serviceable solution has been in continuous use for 700 years.

[paulca]

I had a meeting with a local manager today and relevant SMEs, they were epic.
I had a meeting with a local manager today and relevant SMEs, he/she was epic.
It matters.  Plurality.

[tooki]

I had a meeting with a local manager today and relevant SMEs, they were epic.
I had a meeting with a local manager today and relevant SMEs, he/she was epic.
It matters.  Plurality.

Irrelevant to the issue at hand, it’s the well established singular they.

[gmb42]

Can all the pronoun discussion be moved to the appropriate place (or even better deleted) and then we can get back to discussing the STM32 roadmap?

[Sacodepatatas]

It's not about freelancer. Tried to access with business account.

Not sure they use some kind of  process to only allow  preference based methods based on location, company name etc..
They probably ran out of free sample C0 stock.

Perhaps they could consider having separate sessions for different regions in the world... 
Would not mind the loss of opportunity to get a C0 board that we probably never would be able to get anyhow, but the course would be interesting.

I've followed the course in English the past 20th of this month although i couldn't pay so much attention because of personal matters, so i wanted to watch the course again once being more relaxed). During the course we were told that the 3h recording would be available in a brochure sent to our emails besides the link for ordering the Nucleo Board samples. That was 4 days back. Still no news from ST Micro besides some email advertisements regarding more courses. I'm starting to think that i've been mocked. But, let's wait some more days...

[peter-h]

Well, USB-2-Serial became much simpler when FTDI issued their first FT232's. All we had to do was connect a serial EEPROM to host our VID/PID. Brings back memories of FTDI's minuscule stall at Electronica(?), staffed with actual Scottish technical expertise (Andrew?, Steward?).

At least one of these is still sitting somewhere in Scotland, and replying to emails.

The commercial side of FTDI is long gone, to Taiwan I believe.

I still use these chips but the FT232 was replaced by the much cheaper FT231. FTDI are today a really opportunistic outfit and are gauging people with FT232 pricing. Mouser have got stuck with vast stocks of FT232 when the chip shortage bubble blew up; they will probably flog them off to the 2nd tier disti cowboys in the US and elsewhere. Off topic anyway.

32F4 USB works OK, with the ST Cube IDE supplied code and with lots of patches (found on google) added. I have a pretty solid CDC and MSC running. I would have never even dreamt of doing USB without that code. It's horrible.

It's almost never wroth the hassle to use CDC. When i have lots of data i use raw Bulk endpoints, as I've finally managed how to make windows recognize my device so it gets enumared as a WinUSB device, no driver needed (the application uses libusb, but that's another story)
Raw bulk endpoints: Copy raw memory to endpoint, arm endpoint. let the controller and pc figure out the rest. couldn't be simpler for the MCU. CDC devices instead, at least on the OS side, have to behave as real COM ports with all the delays and buffering

What if you actually want a VCP under windoze? It's a very simple API for apps to use. The delays probably don't matter in most real applications.

[SiliconWizard]

Definitely use a FT230X or FT231X these days. They're cheaper, support more, draw less power and have not been cloned whatsoever (yet at least) that I know of, so you're relatively safe.

[westfw]

USB-2-Serial became much simpler when FTDI issued their first FT232's

Well, aside from the fact that they were pretty hard to get; not carried by the usual distributors...

[harerod]

Quote

USB-2-Serial became much simpler when FTDI issued their first FT232's

Well, aside from the fact that they were pretty hard to get; not carried by the usual distributors...

Well, I guess we (a medical devices outfit with a couple of hundred employees) got them either directly from FTDI or via our standard distributor Spoerle. Fascinating how stories a quarter century old stick in the memory.

STM32 availability seems to be improving. Only yesterday ST's website told me that 2000 units of a certain component were available from Future. After clicking on the provided link, I was told that Future is sold out and awaits new stock by end of this year. Really smooth database links...

[Sacodepatatas]

I've followed the course in English the past 20th of this month although i couldn't pay so much attention because of personal matters, so i wanted to watch the course again once being more relaxed). During the course we were told that the 3h recording would be available in a brochure sent to our emails besides the link for ordering the Nucleo Board samples. That was 4 days back. Still no news from ST Micro besides some email advertisements regarding more courses. I'm starting to think that i've been mocked. But, let's wait some more days...

After 8 days, the e-mail message with the aforementioned  links eventually arrived to my e-mail box. The recording of the workshop has been uploaded to Youtube by ST, but the link has not been published for listing (although it's publicly accesible). I'll respect the ST's will and won't put a link here in this post, but I might reply to any PM asking for it, with a reply containing the link.

[grantb5]

IYDM, Taking this thread a bit back on track.

2. H5 series fully pin to pin compatible with F4 line. Secure boot/TEE, Azure & AWS IoT previsioning with a key and certificate at STM Factory level
3. U5 now scalable to 4MB flash and 2.5MB SRAM for graphics framebuffer and NeoChrome GPU accelerator (upgrade from ChromArt)

Thank you! 

I'm in the opposite camp as the OP. I'm just starting with STM32 but did one learning project with F103 because of the plethora of resources for it. I'm actually looking for a fairly cheap *new* STM32, like $5-ish in quantity. The U575 looks good price-wise compared to the older parts, BUT I need a pretty large TFT LCD. I don't see any STM32's that I can afford that would have enough memory for an 800x480 TFT compared to say, ESP32, but I don't want to go the ESP32 route. The H563 also looks good and is a bit cheaper. Most of the gee whiz stuff on those micros I don't need (I do need USB Device, full speed).

[langwadt]

FTDI's first USB 2.0 HS (high speed) came much later, around 2007 IIRC. Of course to get the benefit of HS you would need to use a parallel FIFO mode. Even on their latest chips, the UART <-> USB function tops at 3 Mbits/s, for which the bandwidth of FS is enough.

the HS chips can do 12MBaud