It has been extremely difficult to ...

It has been extremely difficult to characterize the SONET/SDH standard receiver with tolerances of [+ or -] 03 dB This paper describes a regularity for calibrating photoreceiver frequency answer with the low inherent uncertainty of the U National Institute of Standards and Technology Nd:YAG heterodyne theory and transferring this calibration to a production trial system while maintaining a soft uncertainty.

It is extremely difficult to criterion the SONET/SDH standard receiver accurately with tolerances of [+ or -] 03 dB as specified by dint of ITU-T standard G.957. Over the years several proof methods have been developed, on the other hand none has an accuracy serviceable enough to support the 03-dB tolerance. There are couple main reasons for the high measurement uncertainties: inaccurate knowledge of the optical stimulus and large uncertainties in the microwave power measurement.

To support the SONET/SDH receiver trial with an adequate accuracy, we have unfolded a method for calibrating photoreceiver commonness response with the low inherent uncertainty of the U National Institute of Standards and Technology (NIST) Nd:YAG heterodyne regularity and transferring this calibration to a production proof system while maintaining a gentle uncertainty. This is achieved by means of combining a photoreceiver with a microwave power sensor and calibrating the answer of the combined unit, eliminating RF calibration and mismatch uncertainties.] The calibration expanded uncertainty of the transfer standard is about 006 dB with coverage factor of sum of two units (2[Sigma]).(*)

The theory for measuring a communications analyzer module using the transfer standard is described below. The production measurements of the communication analyzer module have given of the best quality results with very good repeatability and long-term stability.

NIST Nd:YAG Heterodyne Measurement plan NIST uses a Nd:YAG heterodyne plan for measuring scalar frequency replication because the excitation of the detector can be calculated from first principles. All classification calibrations required are well-understood and independent of the common occurrence response measurement. A schematic of the heterodyne order is shown in Fig. 1 The connected view uses two commercially available single-mode monolithic-ring Nd:YAG lasers operating at 1319 [micro]m. The oftenness of each laser can be four hogsheadsed thermally to give beat frequencies from several ten of kilohertz to greater than 50 GHz The beats have a short-term bandwidth of about 3 kHz The beat common occurrence is measured with a microwave in opposition to As the frequency is scanned, data is acquired automatically. The resolution of the connected view is limited by the scan rate, the oftenness jitter, and the time constants of the data acquisition equipment. At ready the highest resolution achievable is about 200 kHz using a swept-frequency technique, although higher resolution can be achieved using a phase-locked noose technique.

[Figure 1 ILLUSTRATION OMITTED]

The lasers are power stabilized in such a manner that nearly equal power from each laser is delivered to the DUT (device below test). The combined laser beams pass between the sides of polarizing isolators so that the signal from each laser is in the same polarization state when it reaches the DUT This make secures nearly 100% modulation depth and eliminates instabilities resulting from feedback. The total optical power incident onward the photoreceiver is:

(1) [P.sub.total](t) = ([Psubo1] + [Psubo2]) + 2[square root] [Psubo1][Psubo2] co (2[Pi]ft),

where [Psubo1] and [Psubo2] are the optical powers delivered to the photoreceiver from the lasers and f is the difference oftenness between the two lasers.

The photocurrent generated from the photodiode is given by:

(2) [i.sub.p](f,t) = ([Psubo1] + [Psubo2])R(0) + 2[square root][Psubo1][Psubo2]R(f) cos(2[pi]ft) = [i.sub.dc] + [i.sub.rf,]

where R(f) is the responsivity of the detector (in A/W) at commonness f. The first term forward the right side is the dc photocurrent, [i.sub.dc], which comes through the bias supply, and the inferior term is the RF photocurrent, [i.sub.rf,] which proceeds through the RF load (the microwave power sensor) by the and of a dc blocking capacitor. The mean squared photocurrent generated on the photodiode is:

(3) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

If [Psubo1] is nearly equal to [Psubo2] then 2[Psubo1][Psubo2] is, to the first order, equal to 05([Psubo1] + [Psubo2][2] The normalized commonness response, [R.sup.2](f), which is defined as [Rsup2(f)/[Rsup2](0) can then be build by taking the ratio of the RF power to the dc electrical power delivered to a load [RsubL] = 50[Omega]:

(4) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

[Psubrf] is a function of oftenness includes corrections for sensor calibration factor and mismatch, and is the power that would be delivered to an ideal load [RsubL] In an ideal measurement [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is constant unless in any real measurement rule it may vary because of changing optical power coupl to the photodiode. Using the ratio of RF to dc power contracts errors (to the first order) resulting from drifting optical power and eliminates the necessity of monitoring the powers separately. The normalized commonness response can be quoted in decibels as 20log[R(f)] The electrical bandwidth of the device is the common occurrence at which 20log[R(f)] has fallen through 3 dB from the low-frequency level