Rian Malcovich

LPC824, PLL post divider is not working?

Discussion created by Rian Malcovich on Jul 9, 2016
Latest reply on Jul 11, 2016 by Rolf Meeser

Hello,

 

After several hours of struggle and confusing manual, here is the problem on LPC824:

 

According to LPC82x User Manual, pag. 56, the PLL block have some high frequency oscillator (called CCO) that can work (can be locked) in FCCO = 156...320 MHz range.

 

After CCO, exists a post divider that can be set to 4 divider values (1,2,4 and 8) automatically further divided by 2 for 50 % duty-cycle, resulting in real P = 2,4,8 or 16. This is for obtaining FCLKOUT (MAIN CLOCK) that is limited to 100MHz.

 

To close the PLL loop, exists also a M divider (called feedback in user manual) that can be set such as real divider values = 1...32

 

Now when injecting 12 MHz from IRC into PLL and set M = 5, MAIN CLOCK will be 60 MHz with no effect of changing P value. That is, any value set to P the clock will be 60 MHz (IRC * M). Since FCCO can be locked only in 156...320 MHz range, we deduce that lock is produced at 240 MHz and real P is 4 (P = 2, 2*P = 4).

 

We enabled CLOCKOUT to monitor MAIN SYSTEM CLOCK and this indeed outputs 60 MHz, 72 MHz or whatever M is set to multiply the 12 MHz IRC.

 

Any value programmed in P bits (pag. 35, SYSPLLCTRL register) have no effect.

 

Taking example from table 57 (pag. 58) if keep M=5 and set P = 4 (real P = 8) the CCO should (still) lock on 240 MHz but MAIN CLOCK should be 30 MHz (well under 100 MHz limit). Instead, example keep the MAIN CLOCK = 60 MHz and modify the SYSTEM CLOCK from SYSAHBCLKDIV register.

 

Appears that P divider is just not working, and set at fixed value P = 2 (real P for PLL formula = 4) regarding PSEL bits setting in SYSPLLCTRL register.

 

Can someone confirm this?

Tnx,

Outcomes