70F6, can't configure RCC for use with USB

[mike louaillier]

Posted on July 18, 2017 at 23:12

Hello i'm using CUBE 4.22 and i keep getting a conflict (see jpg bellow). when i click the checkbox USB->Device (FS), the TIM1 says ''!'' and shows several conflicts. and when i click on the Clock Configuration tab, the clock screen appears and i see a dialog box that says ''do you want to run the automatic clock issues solver?''. i click ''yes'' and then it says ''cannot find solution for some constraints''.

is there a way to resolve this problem? thanx

Note: this post was migrated and contained many threaded conversations, some content may be missing.

[Kraal]

Posted on July 19, 2017 at 10:15

Hi !

You didn't choose the right sub-family to have the crystal-less USB. Take a look there 

http://www.st.com/en/microcontrollers/stm32f0-series.html?querycriteria=productId=SS1574

Only F0x2 and F0x8 have this mode.

Also on the nucleo you can have the clock coming from the ST-Link part, so you might have not noticed that you were using an external clock.

[mike louaillier]

Posted on July 19, 2017 at 22:19

well thanx for that but this is seriously messed up.

http://www.st.com/en/microcontrollers/stm32f0-series.html?querycriteria=productId=SS1574

page does say that the 070 chips do NOT have a ''crystal-less'' USB. but 

http://www.st.com/content/st_com/en/products/microcontrollers/stm32-32-bit-arm-cortex-mcus/stm32-mainstream-mcus/stm32f0-series/stm32f0x0-value-line.html?querycriteria=productId=LN1826

page says that they DO. the 

http://www.st.com/content/ccc/resource/technical/document/datasheet/31/4f/10/5f/a5/32/46/42/DM00141386.pdf/files/DM00141386.pdf/jcr:content/translations/en.DM00141386.pdf

says that ''the clock source must use an HSE crystal oscillator'' but no the NUCLEO-F070RB does NOT have a crystal. here (see below) are the parts from the 

http://www.st.com/content/ccc/resource/technical/document/user_manual/98/2e/fa/4b/e0/82/43/b7/DM00105823.pdf/files/DM00105823.pdf/jcr:content/translations/en.DM00105823.pdf

that illustrate the hardware:

in the image you can see crystal X1 connected only to the STLINK interface, you can see crystal X2 which is 32kHz, and you can see crystal X3 which is SUPPOSED to be connected to the 70RB chip according to the schematic, but it is clearly absent on the board itself. and yes i checked and that's what my board looks like. there is no crystal X3. the board is MADE by ST, and i did have a working USB connection on it, so i still dunno what to think. either the ST people need to proofread their docs, or there just isn't a definitive answer to this question. the circuit i'm making uses a 500kHz PWM and i have to be certain that the USB signals will have enough fidelity to work.

stm32f070f6 no crystal ambiguity contradictory docs documentation timing

[Chris1]

Posted on July 19, 2017 at 22:44

As Kraal suggested, your Nucleo board is likely using the MCO (Master Clock Output) from the ST-Link.  Although the F07 is not using the crystal directly, it's essentially the same as the 8 MHz crystal clock. 

You should be able to put an oscilloscope probe on 'U5B pin 5' or the SB50 bridge and see an 8 MHz clock.

[mike louaillier]

Posted on July 19, 2017 at 23:26

OK thanx i found what you're talking about. yes there is an MCO signal going to the 70RB chip. so it does have an external clock pulse. i been misled because of implications that i never tried to verify. the X3 is missing because it's an option in case i separate the STLINK from the uC. still,

http://www.st.com/content/st_com/en/products/microcontrollers/stm32-32-bit-arm-cortex-mcus/stm32-mainstream-mcus/stm32f0-series/stm32f0x0-value-line.html?querycriteria=productId=LN1826

page says that they 'recently added crystal-less USB' to the STM32F0x0 Value line.

that sucks because my PCB is really small, n now i gotta add three more components. and the crystal is a big-ass part. damn....

[Chris1]

Posted on July 19, 2017 at 23:44

I can appreciate the negative surprise.  Did you check to see if the similar parts with 'crystal-less USB' Kraal suggested may work, for example, STM32F072 [48 pin package, however] or STM32F042?

[mike louaillier]

Posted on July 20, 2017 at 01:26

yeah that's a thought. it ***** having to figure this **** when i'm almost done with the whole development. the 42 chips are considerably more expensive. but if i have to pay for a crystal then maybe i should go with that. now i gotta go back and refigure a whole lotta stuff.....

[mike louaillier]

Posted on July 20, 2017 at 07:24

well i got the LED to blink so it is running properly. but it ***** because i am gonna have to get a crystal. or change to a 042 chip. OH THE HUMANITY.....

[Willem La Grange]

Posted on July 20, 2017 at 12:20

I cant imagine that USB HS (480Mbit/s) can ever work without a crystal. You may get USB FS (12Mbit/s) working without a crystal but I never even tried because it really sounds like an invite for production problems to me.

Try something like this crystal if you have PCB problems: ABM3B-8.000MHZ-10-1-U-T  (Abracon LCC)

[Kraal]

Posted on July 20, 2017 at 12:29

Since we are talking about an F0, USB HS was out of question from the start. However I have used almost successfully(*) the crystal-less mode with an F042 on a custom board (HID stuff, hobby grade). I agree with you that for a real product which will see a different temperature range and so on, I would go with a real crystal. The crystal ST choose for the Nucleo-32 board is quite small.

(*) Almost, because the enumeration process is taking 2 seconds where it should be negligible. After enumeration everything works as expected...

[mike louaillier]

Posted on July 20, 2017 at 20:01

thanx i'll get one it looks like the smallest i can get. but i'm also looking at one of 

http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail&itemSeq=233671609&uq=636361512982254510

because it's a through hole and i can position it over the uC chip and save board space that way.