Hello Senthil,
Following this thread, I do understand, that the SPI shall be used for DSP boot purposes. From the documentation I gathered, that during SPI boot the PLL is set into Bypass mode by RBL, which means
CoreClk / 6 = SysClk7;
312.5 MHz / 6 = 52.08 MHz;
which also is the SPI Module clock. Latter is further divided down by the Prescaler which is initially set to zero. A zero means div 2, resulting in a SPI Clock rate of 26.04 MHz.
Q1) Is the statement about PLL Bypass Mode correct?
The documentation also says the minimum period for SCK is 3*P2, resulting in a maximum allowed SPI-Frequency of 17.36 MHz. Apparently, this maximum rating is exceeded by simply having the BOOTMODE Pins set to SPI-boot in conjunction with the fastest XCO at CoreClk Input.
Q2) Is 17.36 MHz really the physical limit of the DSP device / SPI subsystem? Many SPI-Flash Memories already allow operation at 80 MHz or 100 MHz to date.
Q3) Why is a physical constraint (here: 1/f(SCK) = 3*P2) denoted in dependency of the XCO Input Frequency (CORECLK P|N), shouldn't it be a fixed indication stated in nanoseconds?
Q4) How can one set the Prescale value to an allowed range when initially booting only with the information provided at the BOOTMODE pins?
When booting from I2C or SPI, the BOOTMODE uses 10 bits (12:3) instead of 7 bits (9:3). The bits (12:10) normally are used to configure the PLL.
Q5) In I2C or SPI Boot are those bits (12:10) overrided / just taken for boot config and not for PLL config?
If the PLL is in Bypass, then it seems that the Cores rather run at the slower Input clock (312.5 MHz) than on the final device clock at 850/1000/1250 MHz.
Q6) Is this correct? Is an additional PLL configuration needed at a later stage?
Thank you for your help,
Andreas