70592C Datasheet, PDF (135/314 Pages) Microchip Technology – High-Performance, 16-bit Microcontrollers

PIC24HJXXXGPX06A/X08A/X10A
9.2 Clock Switching Operation
Applications are free to switch between any of the four
clock sources (Primary, LP, FRC and LPRC) under
software control at any time. To limit the possible side
effects that could result from this flexibility,
PIC24HJXXXGPX06A/X08A/X10A devices have a
safeguard lock built into the switch process.
Note:
Primary Oscillator mode has three different
submodes (XT, HS and EC) which are
determined by the POSCMD<1:0> Config-
uration bits. While an application can
switch to and from Primary Oscillator
mode in software, it cannot switch
between the different primary submodes
without reprogramming the device.
9.2.1 ENABLING CLOCK SWITCHING
To enable clock switching, the FCKSM1 Configuration
bit in the Configuration register must be programmed to
‘0’. (Refer to Section 21.1 “Configuration Bits” for
further details.) If the FCKSM1 Configuration bit is
unprogrammed (‘1’), the clock switching function and
Fail-Safe Clock Monitor function are disabled. This is
the default setting.
The NOSC control bits (OSCCON<10:8>) do not
control the clock selection when clock switching is
disabled. However, the COSC bits (OSCCON<14:12>)
reflect the clock source selected by the FNOSC
Configuration bits.
The OSWEN control bit (OSCCON<0>) has no effect
when clock switching is disabled. It is held at ‘0’ at all
times.
9.2.2 OSCILLATOR SWITCHING SEQUENCE
At a minimum, performing a clock switch requires this
basic sequence:
1. If desired, read the COSC bits
(OSCCON<14:12>) to determine the current
oscillator source.
2. Perform the unlock sequence to allow a write to
the OSCCON register high byte.
3. Write the appropriate value to the NOSC control
bits (OSCCON<10:8>) for the new oscillator
source.
4. Perform the unlock sequence to allow a write to
the OSCCON register low byte.
5. Set the OSWEN bit to initiate the oscillator
switch.
Once the basic sequence is completed, the system
clock hardware responds automatically as follows:
1. The clock switching hardware compares the
COSC status bits with the new value of the
NOSC control bits. If they are the same, the
clock switch is a redundant operation. In this
case, the OSWEN bit is cleared automatically
and the clock switch is aborted.
2. If a valid clock switch has been initiated, the
LOCK (OSCCON<5>) and the CF
(OSCCON<3>) status bits are cleared.
3. The new oscillator is turned on by the hardware
if it is not currently running. If a crystal oscillator
must be turned on, the hardware waits until the
Oscillator Start-up Timer (OST) expires. If the
new source is using the PLL, the hardware waits
until a PLL lock is detected (LOCK = 1).
4. The hardware waits for 10 clock cycles from the
new clock source and then performs the clock
switch.
5. The hardware clears the OSWEN bit to indicate a
successful clock transition. In addition, the NOSC
bit values are transferred to the COSC status bits.
6. The old clock source is turned off at this time,
with the exception of LPRC (if WDT or FSCM
are enabled) or LP (if LPOSCEN remains set).
Note 1: The processor continues to execute code
throughout the clock switching sequence.
Timing sensitive code should not be
executed during this time.
2: Direct clock switches between any pri-
mary oscillator mode with PLL and
FRCPLL mode are not permitted. This
applies to clock switches in either direc-
tion. In these instances, the application
must switch to FRC mode as a transition
clock source between the two PLL modes.
3: Refer to Section 7. “Oscillator”
(DS70227) in the “dsPIC33F/PIC24H
Family Reference Manual” for details.
9.3 Fail-Safe Clock Monitor (FSCM)
The Fail-Safe Clock Monitor (FSCM) allows the device
to continue to operate even in the event of an oscillator
failure. The FSCM function is enabled by programming.
If the FSCM function is enabled, the LPRC internal
oscillator runs at all times (except during Sleep mode)
and is not subject to control by the Watchdog Timer.
If an oscillator failure occurs, the FSCM generates a
clock failure trap event and switches the system clock
over to the FRC oscillator. Then the application
program can either attempt to restart the oscillator or
execute a controlled shutdown. The trap can be treated
as a warm Reset by simply loading the Reset address
into the oscillator fail trap vector.
If the PLL multiplier is used to scale the system clock,
the internal FRC is also multiplied by the same factor
on clock failure. Essentially, the device switches to
FRC with PLL on a clock failure.
© 2011 Microchip Technology Inc.
DS70592C-page 135