The Glitchy Clock Multiplexer
Switching between a fast PLL clock and a slow oscillator with a plain mux can produce a pulse shorter than either clock. That runt pulse can corrupt every flip-flop on the output clock.
The solution module must safely switch the output clock between two free-running asynchronous clocks based on a select signal. When the select signal changes, the module must ensure a clean, break-before-make handoff. The current clock's enable must be fully de-asserted and synchronized into the new clock's domain before the new clock's enable can be asserted. This prevents any overlapping high periods that would cause a glitch.
To achieve this, you must implement a classic two-stage synchronizer feedback loop. Each clock domain requires its own two-stage synchronizer to process the enable signals. To prevent the enable signal from changing while the clock is high, all synchronization flip-flops must be triggered on the negative edge of their respective clocks.
Port Table
| Signal | Direction | Width | Description |
|---|---|---|---|
clk_a |
input | 1 | First asynchronous free-running clock |
clk_b |
input | 1 | Second asynchronous free-running clock |
rst_n |
input | 1 | Asynchronous active-low reset; drives output to 0 and clears all internal registers |
sel |
input | 1 | Clock select signal; 0 selects clk_a, 1 selects clk_b |
clk_out |
output | 1 | The multiplexed clock output |
Worked Trace
- Event 1:
rst_n=0,sel=0,clk_a=0,clk_b=0 →clk_out=0 - Event 2:
rst_n=1,clk_anegedge 1 → A-domain sync stage 1 loaded - Event 3:
clk_anegedge 2 → A-domain sync stage 2 loaded, domain A enabled - Event 4:
clk_agoes high →clk_out=1 (followingclk_a) - Event 5:
sel=1 → select changes to B, but domain A remains enabled - Event 6:
clk_anegedge 1 → A-domain sync stage 1 cleared - Event 7:
clk_anegedge 2 → A-domain sync stage 2 cleared, domain A disabled - Event 8:
clk_bnegedge 1 → B-domain sync stage 1 loaded (since domain A is now disabled) - Event 9:
clk_bnegedge 2 → B-domain sync stage 2 loaded, domain B enabled - Event 10:
clk_bgoes high →clk_out=1 (followingclk_b)
Constraints
- All synchronization flip-flops must be negative-edge triggered.
- You must use exactly two flip-flops in sequence for each clock domain's synchronizer.
- The design must not produce any pulses shorter than the high-time of either clock.
rst_nmust asynchronously clear all internal registers to 0.- The final output must be a combinational logic function of the clocks and the synchronized enable signals.
Topics
Solve this problem
Write the module in Verilog, SystemVerilog or VHDL. Your submission is compiled and simulated against the problem's testbench. You get the waveform back, not a stored answer.
This problem is part of Codiode Pro. The statement above is free to read.
The circuit builder and code editor need a desktop screen. On a phone, read the problem here and open it on a laptop to solve.
Part of Clock Domain Crossing and Synchronization
Problem 10 of 36 in the Clock Domain Crossing and Synchronization track.
Related problems
Browse all problems · Verilog practice problems · Learning tracks