| Instrumentation to measure the quality of optical connections: measurers of BER |  |  |  |
| Detectores de Señal Optica |
In order to measure the quality of the connections of optical communications instrumentation of high benefits is required habitually. In this article we will concentrate in the procedure of measurement of the error rate (BER) and will present/display some equipment available commercially. Day after day the systems of optical communications they increase to its reach and capacity. At the moment the passage has occurred already to the first 40 commercial systems of Gbit/s, although 160 have been demonstrated in different trials transmissions of Gbit/s by channel and even superiors. In order to be able to measure the quality of signals of high speed all a series is needed instrumentation of high benefits, which supposes an important economic payment. One is laboratory equipment, between which they are: digital, generating optical oscilloscopes of PRBS and measurers of error rate, optical, autocorreladores analyzers of spectrum of pulses, analyzers of plot and protocol, etc. In this article we will concentrate in the measurers To see and the associate foundations custom-made. Measurers of error rate The measurers of error rate are used basically to determine the parameter To see of a connection of optical communications. In order to realise this measurement a pseudorandom sequence becomes congested in the optical transmitter or PRBS and later is moderate in the optical receiver the number of errors that have taken place in the system. To day of today, these equipment reaches rates of bit of 40 Gbit/s and usually has optical interfaces of input/output, because a complete solution is commercialized that includes all a series of devices such as an electronic generator of PRBS, a measurer of electronic error rate, an optical transmitter composed of laser and external modulator, mux/demuxes, an optical receiver, a circuit of recovery of clock, etc. In the market options of manufacturing manifolds can be found.
This factor Q signal noise (SNR) in units of tension or current is known him habitually like relation. In fact, it can even be defined when the noise is not Gaussian, although in the first case a direct relation between Seeing and Q exists. Traditionally BER, it one is moderate based on the SNR. This it is the case of wireless systems or of coaxial cable, where the dominant contributions of noise can not be in the receiver (channel or source noises) and habitually the SNR is proportional to Q. However, in the case of systems Of seeing optical it received average or Prec is moderate based on the optical power. Supposing that the dominant noise is the thermal one of the receiver and that the relation of extinction of the DS is high, then one is fulfilled that factor Q is proportional to Prec and that the SNR in dB is proportional to Prec in dBm. The measures To see begin with the configuration “back-to-back”, by means of which it is able to characterize the quality of the receiver of independent form. For it, the transmitter with the receiver through a variable optical attenuator is connected directly, as it imagines in the figure 4 (a). In this way, and since the quality of the transmitted signal is sufficiently good, the measures To see talk about the benefits of the optical receiver solely. For each value of optical attenuation (that corresponds with a value of Prec), the optimal threshold of decision adjusts and BER it is moderate it. The sensitivity of the receiver is defined as that one value of Prec for which one is obtained BER of 10-9, 10-10 or 10-12. Next the system under test is inserted, figure 4 (b), and the measures are repeated. If the system adds noise or any type of degradation that modifies the signal, then these effects are reflected in the curve To see. It agrees to indicate that any type of attenuation of the system To see will not degrade it by itself, BER because it imagines based on Prec. Nevertheless, a followed attenuation of a process that year noise (for example, an optical amplifier) will bring about a reduction of factor Q and therefore an increase of the error rate. That is to say, a degradation of the curve BER.
Types of degradation Generally, the results Of seeing turn out helpful to determine the type of degradation that happens in the system and to identify their physical origin. Next we will comment some examples. The relation of extinction (extintion ratio, ER) of an optical signal OOK defines as the quotient between the average optical powers of the levels `1 ' and `0 '. In the ideal case, this one is equal to infinite, because the power of the level `0 ' must be null. The degradation of the ER is pronounced in the curve To see like a displacement of the same towards powers majors, as it shows the figure 5 (a). In this case, the penalty of defined power as the degradation in dB of the sensitivity of the receiver can calculate by means of:
where r = P0/P1, being P0 and P1 the average optical powers of the levels `0 ' and `1 ', respectively. For example, for a ER given by r = 0.1 (10 dB), has a penalty of power of 0.87 dB. Equally, the noise ASE of the optical amplifiers of a connection also contributes to degrade the curves To see. On the one hand, the average power of noise ASE produces an increase of the level `0 ', being in a degradation of the ER and a displacement of the curve To see. On the other hand, this noise ASE degrades the diagram of eyes, reason why an increase of the received optical power does not improve the quality of the signal. The result is the appearance of one To see basic below which it is not possible to be lowered. This effect appears reflected in figure 5 (b).
Finally, the interference between symbols (ISI) also causes a egradación Seeing of it (penalty of power and/or basic appearance BER). This one can be due to multiple causes (chromatic dispersion, nonlinearities, etc.), although the analysis of the curves To see is not sufficient to identify the nature of the problem and they require another type of measures such as diagrams of eyes. In this case, the ISI can be identified by means of the dependency of the curve To see with the landlord of the PRBS. For it, comparative measures with patrons of PRBS from 27 - 1 up to 231 - 1, and even with landlord “0101 …”.
Validity of the measures Since BER owns a clear statistical component, it is needed to assure a certain confidence in the measures. As statistical estimator can be used the following expression: where ne is the number of errors and Nb is the total number of bits (sufficiently great). Supposing independent errors, the statistical distribution is binomial. With these data, to assure a level confidence of 95% it is needed to observe a total of 100 errors. This is equivalent to about 3 40 seconds to Gbit/s for 1011 bits (To see = 10-9) and 7 hours for 1015 bits (To see = 10-12). But unfortunately some systems present/display correlados errors, being the most usual case the appearance of bursts of errors brought about by some type of external fluctuation. That is to say, a quite low error rate of time is observed for a long period and it increases suddenly considerably as a result of a burst of errors. In these cases, so that the results of the measures are reasonable, one is due to try to free the cause from this burst and BER to measure it in the period without bursts. Or on the contrary, to take a sufficiently great time interval like so that it includes a significant digit of bursts of errors, from a statistical point of view. For this aim, measuring equipment usually have accountants of intervals of errors, from a statistical point of view. For this aim, measuring equipment usually have accountants of intervals of errors that help in this task. In summary, the measures BER require a methodical process that assures. Luckyly, present measuring equipment offer a great versatility and a great number of options and electrical interfaces of connection as much as optician, facilitating the tasks of design and supervision of the systems and optical communications networks.
|
