PROCEEDINGS LETTERS critical dependence of the circuit behavior on T is quite interesting to observe.In Fig. 3 (a), ten traces of the output are displayed with T varying from 1.0 to 1.9 in 0.1 step. It is seen that when the input is of sufficient strength, the flip-flop switches at around t = 1.5. But while T = 1.0 does not produce switching, T = 1.1 does, eventually. Fig. 3 (b) shows output waveforms with T varying from 1.00 to 1.09 in 0.01 step.It is seen that the circuit remains in the indefinite region (metastable region) longer and switching eventually OCCUIS if T > 1. 03. Fig. 3 (c) shows the output waveforms when T varies from 1.021 to 1.031. The step is 0.0002 if T is between 1.024 and 1.026, otherwise T varies in step of 0.001. For some values of T , it appears that the circuit would remain in the metastable region for a long time and could not settle until much beyond the normal switching time. By carefully selecting the parameter T , the output waveform can be made to remain in the "metastable" region much longer than what has been shown in these figures. It is interesting to note the similarity of these waveforms with trace Q of Fig. 1 in [ I ] .In conclusion, we have presented here an alternative explanation to the purely probabilistic one advanced by Couranz and Wann concerning the behavior of synchronizers operating in the metastable region. In reality, the detailed operation of the flip-flop should probably take into consideration both the probabilistic and the deterministic behaviors.
REFERESCES[ 1 1 G. R. Couranz and D. F. Wann, "Theoretical and experimental behavior of synchronizers operating in the metastable region,"Absrmct-A simple method is presented for measuring the individual resistances in a loop of resistors without breaking the loop. Such a problem is often encountered in hybrid integrated circuits. Our method is exact and noniterative, in contrast to an alternative method recently suggested in the literature. Also, it requires n number of measurements for an n-resistor loop and hence is canonic.In hybrid integrated circuits, a problem often encountered is that of measuring the individual resistors in a closed loop of n resistors. Swart and Van Wyk [ 11 have recently proposed two methods for solving the problem, both of which involve iterative techniques and the use of a digital computer. In this letter, a simple, yet exact, method is given for such measurements, which does not require any iteration or the use of a computer.Consider a loop of n resistors r l , rz . . r,, as shown in Fig. 1. Let the n nodes be denoted by N , , N , * N,, with the resistance ri between the nodes Ni-, and N i , and let gi = l /~. First consider the case when n is odd, n = 1 cannot form a loop, so let n 2 3. The measurement procedure is as follows. Short-circuit nodes N , and N , and measure the conductance ( G , ) between nodes N , and N , ; obviously, . Similarly measure G,, G 3 , . G,. Where Gi = gi + gi+l is the conductance between the node Ni and the common node obtained by short-circuiting nodes...