STM32F407 > STM32F427 MIGRATION

[jean_prieur]

Posted on July 08, 2014 at 17:39

Hi everybody!

I want to migrate from STM32F407 to STM32F427, and use the full power of this microprocessor (180MHz instead of 168MHz), because I need more speed ressources.First, I find this new application note:http://www.st.com/web/en/resource/technical/document/application_note/DM00123028.pdfDo you know what is the frequency overdrive mode ?Secondly, I want to set the new maximum frequency on my project (I already use a STM32F427 but at lower frequency). I use an external oscillator of 8MHz.I use this clock parameters for the moment:

SystemCoreClock = 168000000;

PLL_M  = 8PLL_N  = 336PLL_P  = 2PLL_Q  = 7AHBPrescTable[16] = {0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 3, 4, 6, 7, 8, 9};HSE_VALUE = 8000000Can I only change the parameter 

SystemCoreClock

to

180000000

? Because it works but I don't know if the speed of core change.

Thank you very much!

[Tesla DeLorean]

Posted on July 09, 2014 at 15:31

How dangerous is this overclocking ?

Tell me when it fails? The 180 MHz probably wasn't randomly pulled out of a hat, so there is likely some critical path within the design or one of the peripherals that accounts for this number over voltage and temperature, and process variations. ie the Fast or Slowness of the P and N transistors at the boundaries of the process window for the silicon. 192 MHz would have made far more sense, but I'm not sure we've had any of the silicon design team here answering questions.

[jean_prieur]

Posted on July 09, 2014 at 16:01

Ok I'm gonna check if something goes wrong with this clock speed, but for the moment there is no problem with SPI, USART, USB, SDIO and GPIOs.

[Tesla DeLorean]

Posted on July 09, 2014 at 17:35

Looking for gross failures at room temperature, probably not an effective way to find the boundaries.

[jean_prieur]

Posted on July 10, 2014 at 11:21

The STM32F4 is specified to work up to 85°C, I don't know if I will try this temperature at 192Mhz, but maybe why not ? :)

[schauer]

Posted on July 10, 2014 at 16:39

USB isn't going to work natively if you clock the core at 180 MHz.

You might be able to clock it with a HS PHY, but I haven't done that experiment. One could use an ULPI PHY with 180MHz, I do have that running here. In that case you simply do not need the 48MHz (if you do not use the other on board PHY too). The STM32F4 is specified to work up to 85°C, I don't know if I will try this temperature at 192Mhz, but maybe why not ? You may see stability issues in rare situations and/or other strange behaviour sometimes. I'm not sure if I'd use that in production ;)

[jean_prieur]

Posted on July 14, 2014 at 11:40

Hi bs,

What is you method to implement the USB with ULPI PHY?Thanks!

[Tesla DeLorean]

Posted on July 14, 2014 at 14:07

Suggest you review schematics for STM324x9I-EVAL, RedDragon407, or other boards. Typically using

http://www.microchip.com/wwwproducts/Devices.aspx?product=USB3300

series parts.

[jean_prieur]

Posted on July 14, 2014 at 14:17

Ok thanks, so this is a hardware modification. It's too late for me to change the design, so I definitely can't use the 180MHz clock...

[Tesla DeLorean]

Posted on July 14, 2014 at 17:40

Ok thanks, so this is a hardware modification. It's too late for me to change the design, so I definitely can't use the 180MHz clock...

As stated much earlier it's an ''External PHY''

[Tesla DeLorean]

Posted on July 14, 2014 at 17:44

Pushing to moderation AGAIN