Page 7 of 24

Re: Where to start STM32H7 support

Posted: Wed Jan 23, 2019 6:12 pm
by maciek
I have tried with either STM32 HAL or FreeRTOS (or both? I don't remember exactly but can check and publish an example) and it was working while connected with loose wires. The card detected, writing and reading OK. Maybe there would be problems with the max. speed, but I haven't been adjusting any parameters and all was fine.

Re: Where to start STM32H7 support

Posted: Wed Jan 23, 2019 7:02 pm
by Giovanni
I will search for a board to attach.

Giovanni

Re: Where to start STM32H7 support

Posted: Wed Jan 23, 2019 11:50 pm
by maciek
Note also that you can connect most of the pins directly to Nucleo's connectors, thus reducing the number of wires. Bend the pins on the card shield and connect only the wires where using another pin can't be avoided.

Re: Where to start STM32H7 support

Posted: Tue Jan 29, 2019 10:36 am
by dwawerek
Hi,
I hava question contected to the topic. I go through older post but I'm not sure, as for now, only Serial Driver module was tested for H7 or maybe You have some UART drivers examples? I tried to start UART drivers( chaneges in mcuconf.h and halconf.h from Blink exmaple for H7 Nuccleo) but I get many errors in hall_uart_lld.h (unkown definitions for example "UART7 DMA streams not defined") .
Best regards
Damian

Re: Where to start STM32H7 support

Posted: Tue Jan 29, 2019 11:12 am
by Giovanni
Hi,

UART is untested on H7, I need to review it. Could you post a note in "bug reports" about this? it is where I look for "things to do".

Giovanni

Re: Where to start STM32H7 support

Posted: Tue Jan 29, 2019 2:54 pm
by dwawerek
I put a note in "bug reports".
Thank You for reply.

Damian

H743 hang in spiExchange()

Posted: Wed Feb 06, 2019 12:36 am
by tridge
I'm bringing up a H743 based board with ArduPilot. I got things building OK, but found that spiExchange() would never return. So I tried the testhal/STM32/multi/SPI test on a Nucleo-H743 board, and got the same issue. The circular buffer test on SPID1 hangs.
Can someone else confirm if that test is working for them? I notice SPI was listed as working by Giovannni back in Jan 2018, so I'm wondering if this is a regression or if it is something in my build environment (I'm using gcc 6.3.1).
The USB_CDC test is working fine on the same Nucleo board.
I also attach 3 patches for build errors I ran across while doing the port
Cheers, Tridge

Re: Where to start STM32H7 support

Posted: Fri Feb 08, 2019 1:25 am
by tridge
does anyone have a download of STM32CubeH7 they can share? The download on the ST site is corrupt.
I'm trying to work out the clock settings to get both the 48MHz OTG clock and an ADC clock in the required range for a H743. I can separately setup either ADC or OTG (and both work), but haven't yet worked out a clock tree that gives me both. I was thinking of trying STM32CubeH7 to see what it generates.
Cheers, Tridge

Re: Where to start STM32H7 support

Posted: Fri Feb 08, 2019 1:30 am
by tridge
tridge wrote:does anyone have a download of STM32CubeH7 they can share?

belay that ... I just realized ST offers downloads of old versions, and those aren't corrupt

fix for PLL DIVP ranges for H7

Posted: Fri Feb 08, 2019 4:18 am
by tridge
I've got a working clock tree now (thanks to the CubeMX clock resolver). It ended up with a DIVP values for PLL2 of 1. That is marked as an error in the hal_lld.h, but as far as I can see from the ref manual, the lower limit of 2 only applies to PLL1.
I attach a patch to fix the limits for PLL2 and PLL3