Dynamic Clock Mode Switch

Copyright .COPYRGT., Texas Instruments Incorporated, 1993. A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owner has no objection to facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the U.S. Patent and Trademark Office, patent file or records, but otherwise reserves all rights in its copyright whatsoever.

This invention relates to dynamic clock mode systems and more particularly to a dynamic clock mode indicator for use with a phase-locked loop circuit.

In the prior art, some processing systems have implemented more than one operating speed. In order to change operating modes, these systems have needed to stop execution by the CPU until a PLL has locked onto a desired clock signal. This has resulted in a loss of operating efficiency. Therefore, there has been a need for systems with an improved ability for transition between a first and second mode of operation; and, there has been a further need for systems with an ability to sense the particular operating mode of a dynamic operating mode system.

The present invention provides a dynamic clock mode switch for switching clock frequencies while allowing continuing operation of a depending system. The switch includes an enable circuit for transmitting an enable signal, a phase-locked loop circuit (PLL) for locking onto an input clock frequency in response to said enable circuit, a PLL lock indicator for receiving a PLL lock signal from said PLL, and a clock multiplexer with a multiplier for multiplying said input clock frequency by a predetermined factor in response to said enable and PLL clock signals.

For a more complete understanding of the present invention and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:

FIG. 1 is a block diagram of the preferred embodiment of a dynamic clock mode switch.

FIG. 2 is a schematic diagram of a PLL with the clock mode switch of the preferred embodiment.

Referring generally to FIG. 1, the preferred embodiment of a dynamic clock mode switch 11 is shown for switching clock frequencies while allowing continuing operation of a depending system shown in simplified from in FIG. 1, and in more detail in FIG. 2. The switch 11 includes a phase-locked loop circuit 15 (PLL) for locking onto 1X clock frequency in response to an enable signal, a PLL lock indicator 29 for receiving a PLL lock signal from the PLL, and a clock multiplexer 19 with multiple input frequencies for multiplying said input clock frequency by two in response to said enable and PLL clock signals.

The switch includes automatic clock reversion circuitry 13 for automatically switching to a 1X clock frequency when the PLL looses lock.

The clock multiplexer delays changing of clock modes with sensing a clock phase boundary in order to avoid a loss-of-state in the CPU.

Switch 11 receives a clock signal CLK2 at node 17 which is input to multiplexers 19, 21. Multiplexers 19, 21 receive as a second input signals 2X.sub.-- CLK2 and 1X.sub.-- CLK2 respectively. Multiplexers 19, 21 are activated by logic circuitry node 27 where the logic circuitry 13 receives as input the output signal from node 23, feedback signal from node 25, CKD enable signal and a PLL lock signal from line 29. Upon enablement multiplexers 19, 21 send output signals to divider circuitries 31, 33, respectively.

The current invention allows clock mode switching to occur dynamically depending on the PLL lock condition and clock-doubling enable bit. There are two signals which control the clock mode: a configuration register bit CCRO<6>, which is an enable bit, or signal , CKD, and a PLL lock indicator. When the enable bit CKD is set, the PLL begins to lock onto the input frequency. The PLL lock indicator will be low until a stable lock has been achieved, in which case it is driven high. When both signals are high, a S-R flip-flop SR1 is toggled upon the next stable clock state. The flip-flop output then switches a clock multiplexer 19 which clocks the core at twice the I/O frequency. The multiplexer 21 is provided to ensure minimum clock jitter between 1X clock generated by circuit 33 and 2X clock generated by circuit 31 by selecting 1X clock output from PLL 15 when in 2X mode.

The dynamic clock switch improves the 1X to 2X transition because the CPU continues to execute, clocked by CLK2, while the PLL is still in 1X mode, and still locking. This prevents extra CPU idle time during the lock time. The invention also prevents loss-of-state if the PLL looses lock, because the CPU will automatically switch back to 1X mode, and be clocked by CLK2. Another feature of the invention is the timing of the clock multiplexer switch. The control waits until the next clock phase boundary to change the clock mode. This prevents glitches and loss-of-state during mode transitions.

The CPU clock mode is selectable by an enable bit, CKD. The PLL lock indicator is also monitored by the control logic. The invention allows the CPU to continue executing instructions in 1X mode while the PLL is locking on to the input clock signal. When the PLL locks, the clock mode will automatically switch to 2X mode if the enable bit CKD is set. The invention prevents any idle CPU time while the PLL is in the locking stage. If the PLL looses lock for some reason, then the control logic will automatically switch to 1X mode while the PLL corrects the phase mismatch. The mode switch occurs on a clock phase boundary so no loss-of-state occurs.

The circuitry includes a PLL 15 for doubling the input clock frequency, a PLL lock indicator signal, LOCK delivered on line 29, a PH.sub.-- bound signal to indicate clock switch condition, an STBL signal to switch the clock multiplexers 19 and 21, clock dividers C-DIV1 and C-DIV2, input clock CLK2, clock-double enable bit CKD, select indicator signal SEL.sub.-- 2X and !SEL.sub.-- 2X where "!" herein denotes inverse, i.e., the inverted form of the designated signal it precedes, clock divider C.sub.-- DIV3 for dividing PLL output, 2X mode clocks 1X.sub.-- CLK2 and 2X.sub.-- CLK2, CPU clocks PH1, PH2, PH1.sub.-- IO, PH2.sub.-- IO, SR flip-flop SR1, logic gates A, B, C, D, E, F, SR1 controls !SET, !RESET, clock delay DEL1, PLL output PLL.sub.-- OUT. Note that in the drawings, the inversion of a signal is indicated by a bar over the signal name.

In operation C.sub.-- DIV1 accepts the output of multiplexer 19 and generates non-overlapping core clocks PH1 and PH2. C.sub.-- DIV2 accepts the output of multiplexer 21 and generates non-overlapping I/O clocks PHI.sub.-- IO and PH2.sub.-- IO. Multiplexer 19 accepts STBL as a select signal, multiplexer 21 accepts STBL as a select signal, multiplexer 19 accepts 2X.sub.-- CLK2 as its STBL=1 input, multiplexer 19 accepts CLK2 as its STBL=0 input, multiplexer 21 accepts 1X.sub.-- CLK2 as its STBL=1 input, multiplexer 21 accepts CLK2 as its STBL=0 input. NOR gate B accepts !PH2.sub.-- IO, 1X.sub.-- CLK2, 2X.sub.-- CLK2 and generates PH.sub.-- BOUND. NAND gate A accepts LOCK and CKD signals and generates !SEL.sub.-- 2X. Inverter gate E accepts !SEL.sub.-- 2X and generates SEL.sub.-- 2X. NAND gate C accepts PH.sub.-- BOUND, !SEL .sub.-- 2X and generates !RESET. NAND gate D accepts PH.sub.-- BOUND, SEL.sub.-- 2X, and generates !SET. Inverter gate F accepts PH2.sub.-- IO and generates !PH2.sub.-- IO. SR1 accepts !SET and !Reset and generates STBL. PLL 15 accepts CLK2 and 1X.sub.-- CLK2 and generates PLL.sub.-- OUT. C.sub.-- DIV3 accepts PLL OUT and generates 1X.sub.-- CLK2. DELl accepts PLL.sub.-- OUT and generates 2X.sub.-- CLK2.

The PLL feedback is provided by a 50% duty cycle signal 1X.sub.-- CLK2. PLL 15 is completely bypassed in 1X mode and CLK2 is used. In 2X mode, PLL provides 1X.sub.-- CLK2 and 2X.sub.-- CLK2 for dividers C.sub.-- DIV1 and C.sub.-- DIV2 inputs. The CPU and I/O clocks are outputs from the dividers, respectively.

In operation the following sequence occurs when the system requests 2X mode: the CPU executes in 1X CLK mode, software writes to register and sets CKD, PLL is activated and begins locking-on to CLK2, the CPU continues to execute in 1X mode, PLL locks and sets LOCK, SEL.sub.-- 2X goes high and !SEL.sub.-- 2X goes low, PH.sub.-- BOUND goes high when !PH2.sub.-- IO, 1X.sub.-- CLK2, and 2X.sub.-- CLK2 are low, !SET goes low, STBL goes high, multiplexer 19 selects 2X.sub.-- CLK2, multiplexer 21 selects 1X.sub.-- CLK2, C.sub.-- DIV1 generates PH1, PH2 at the CLK2 frequency, and C.sub.-- DIV2 generates PH1.sub.-- IO, PH2.sub.-- IO at half the CLK2 frequency.

In operation the following sequence occurs when the system requests 1X mode or the PLL looses LOCK: the CPU executes in 2X mode, software writes to register and resets CKD, PLL is deactivated and resets LOCK, SEL.sub.-- 2X goes low and !SEL-2X goes high, PH.sub.-- BOUND goes high when !PH2.sub.-- IO, 1X.sub.-- CLK2, and 2X.sub.-- CLK2 are low, !RESET goes low, STBL goes low, multiplexer 19 selects CLK2, multiplexer 21 selects CLK2, C.sub.-- DIV1 generates PH1, PH2 at half the CLK2 frequency, and C.sub.-- DIV2 generates PH1.sub.-- IO, PH2.sub.-- IO at half the CLK2 frequency.

While this invention has been described with reference to illustrative embodiments, this description is not intended to be construed in a limiting sense. Various modifications of the illustrative embodiments, as well as other embodiments of the invention, will be apparent to persons skilled in the art upon reference to this description. It is therefore contemplated that the appended claims will cover any such modifications or embodiments as fall within the true scope of the invention.