Reset Synchronizer


Need for reset synchronizer: The way most of the designs have been modelled needs asynchronous reset assertion and synchronous de-assertion. The requirement of most of the designs these days is:
  1. When reset is asserted, it propagates to all designs; brings them to reset state whether or not clock is toggling; i.e. assertion should be asynchronous
  2. When reset is deasserted, wait for a clock edge, and then, move the system to next state as per the FSM (Finite State Machine); i.e. deassertion should be synchronous
The top level reset sources are mostly asynchronous, both in assertion and during deassertion. The circuit that manipulates the asynchronous reset to have asynchronous assertion and synchronous deassertion is referred as reset synchronizer.

Definition of reset synchronizer: A reset synchronizer synchronizes the deassertion of reset with respect to the clock domain. In other words, a reset synchronizer manipulates the asynchronous reset to have synchronous deassetion.

Figure 1 below shows the schematic representation of how a reset synchronizer is built. It consists of two registers connected in series, the input of first register tied to VDD. The asynchronous reset signal is connected to the Rbar pin of both the register. The output of this circuit has synchronized de-assertion. This synchronised reset fans out to the design.

Figure 1: Reset synchronizer

How reset synchronizer works: When the reset is asserted, it first propagates to reset synchronizer flops. It resets both the flops of reset synchronizer asynchronously (without waiting for clock edge) thereby generating reset assertion for fanout registers. Figure 2 below shows the scenario of reset assertion, and also the timing waveforms associated with assertion of reset.

Figure 2: Reset assertion


Similarly, the de-assertion of reset first reaches the two flops of reset synchronizer. Now, the first flop in chain propagates 1 to intermediate output upon arrival of a clock edge. Upon next clock edge, this signal propagates to the output thereby reaching the fanout registers. The reset de-assertion timing (recovery and removal checks timing) should be met from second stage of reset synchronizer to all the domain registers' reset pins as the deassertion is synchronous.







Some facts about reset synchronizer:

  1. The reset synchronizer manipulates the originally asynchronous reset to have synchronous deassertion.
  2. The reset synchronizer must fanout to all the registers that need to be "OUT OF RESET" in a single cycle. And there must be a single synchronizer for all such flops, otherwise, some flops will be out of reset 2 cycles later, some 3 cycles later; thus, defeating the purpose of reset synchronization.
  3. There can, of course, be multiple reset synchronizers in the design, with the number equal to number of functional clock domains. Each reset synchronizer fans out to all the resettable flops of its own clock domain.
  4. The timing constraints related to a reset synchronizer are discussed here.
Also read:

Quiz: Modeling skew requirements with data-to-data setup and hold checks


Problem: Suppose there are 'N' signals which are to be skew matched within a window of 200 ps with respect to each other. Model this requirement with the help of data setup and hold checks.

As we discussed in data setup and data hold checks, data setup check of 200 ps means that constrained data should come at least 200 ps before the reference data. Similarly, data hold check of 200 ps constrains the constrained data to come at least 200 ps after the reference data. The same is shown pictorially in figure 1(a) and 1(b).


Data check of 200 ps ensures the constrained data to come at least 200 ps before the reference data
Figure 1(a): Data setup check of 200 ps constrains the constrained signal to toggle at-least 200 ps before reference signal toggles.

Data hold check of 200 ps ensures that the constrained data comes at least 200 ps after the reference data
Figure 1(b): Data hold check of 200 ps constrains the constrained signal to toggle at-least 200 ps after the reference signal has toggled.

Now, suppose you apply a data setup check of -200 ps instead of 200 ps. This would mean that the constrained signal can toggle upto 200 ps after the reference signal. Similarly, a data hold check of -200 ps would mean that the constrained signal can toggle from 200 ps before the reference signal. If we apply both the checks together, it would infer that constrained signal can toggle in a window that ranges from 200 ps before the toggling of reference signal to 200 ps after the toggling of reference signal. This is pictorially shown in figures 2(a) and 2(b).

Negative data setup and data hold checks together will ensure a skew check between reference and constrained signals
Figure 2(a): Negative data setup and hold checks of 200 ps
If we combine the two checks, it implies that the constrained data can toggle upto 200 ps after and from 200 ps before the reference signal. In other words, we have constrained the constrained signal to toggle in a window +- 200 ps within the reference signal.

Coming to the given problem, if there are a number of signals required to toggle within a window of 200 ps, we can consider one of these to act as reference signal and other signals as constrained signals. The other signals can then be constrained in both setup and hold with respect to reference signal such that all of these lie within +-100 ps of the reference signal. The same is shown in figure 3 below:

We can constrain the signals to toggle within a window by choosing one of these as reference signal and applying negative data setup and data hold checks for other signals with respect to reference signal
Figure 3: Data checks to maintain skew between N signals