An RF cable test assembly.
This requirement for high measurement stability is particularly relevant for the design of cable test assemblies, which are in daily use in component and assembly shops and in test labs. Thousands of repeated bending, mating and demating cycles as well as permanent wear brought about by manual handling can quickly degrade such measurement equipment.
Therefore, the SUCOTEST 18 cable assembly has been developed to meet the demands of high volume return and insertion loss measurements, and fulfill the requirement for effective and efficient operation. In particular, it withstands repeated mating/demating cycles and cable wear during testing. The assembly is completely designed and manufactured in-house, which means 100 percent process control of the individual components as well as an ensured and controlled compatibility. Its precision facilitates constant measurement results, thus reducing the overall measurement costs.
Furthermore, the newly designed N nut, a short taper sleeve and a cable that can be positioned without spring-back, gives the assembly unique handling characteristics. A comprehensive range of assemblies in three lengths with three different male connector configurations (N to N, SMA, N to SMA) as well as an N and SMA precision female-female adaptor series completes the range.
This article emphasizes the significance of the cable test assemblies' return and insertion losses. To illustrate the point the specifications for the SUCOTEST 18 are shown in Table 1.
[FIGURE 1 OMITTED]
SUCOTEST 18 UNDER TEST
The cable assembly has been specifically designed for return loss (RL) and insertion loss (IL) measurements during test procedures in component and assembly shops, test labs and with automatic test equipment. In all of these applications, in order to perform accurate and reproducible measurements with a network analyzer, it is essential to use measurement assemblies offering absolute electrical stability. This is particularly important for both manual and automated measurement applications, since the assemblies are moved during initial calibration as well as after each measurement.
When instabilities do occur they can generally be attributed to two basic phenomena. On the one hand, the insertion loss of the assembly may change in a predictable and continuous manner when the assembly is bent or subjected to torsion. Conversely, undefined insertion losses may occur when the assembly is moved or exposed to vibration.
To quantify and, therefore, control these factors, Huber + Suhner performs three tests on each cable assembly lot produced. First, to establish insertion loss stability versus bending, the device under test (DUT) is connected to the network analyzer. The insertion loss is measured and recorded in order to show the changes relative to the original position during the remainder of the procedure. This is followed by clockwise bending by 180[degrees], bending back into the original position, anticlockwise bending by 180[degrees] and bending back into the original position. These various positions are illustrated in Figure 1.
In each position, the changes relative to the original position are recorded, and the maximum values are compared with the assembly specifications. The results for a highly stable cable are shown in Figure 2. This test is based on the Standard IEC 60966-1, Procedure 126.96.36.199 Bending, but, as mentioned, it is the insertion loss that is measured and assessed instead of the phase.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
The test to evaluate the insertion loss stability versus twisting is based on the Standard IEC 60966-1 Procedure 188.8.131.52 Twisting, with measurement of the insertion loss instead of the phase. The procedure is the same as that for the bending test, but the DUT is twisted by 180[degrees] in both directions, as shown in Figure 3, with the results for a highly stable cable shown in Figure 4.
Finally, the test to establish the insertion loss stability versus fast movement (shaking) is carried out using the device shown in Figure 5, and reveals stability problems of the 'undefined' type. The assembly is connected in the original position to the network analyzer as described for the two previous measurements and clamped in the center of the resulting arc in a purpose-designed machine, which shakes the assembly at a precisely defined frequency and amplitude. During and after shaking, the changes in the insertion loss are not allowed to exceed the specified value. Again, the results for a highly stable cable are shown in Figure 6.
[FIGURE 5 OMITTED]
To achieve the maximum insertion loss stability, a very precisely matched cable construction is needed. The individual components and materials must be consistent and of high quality, and the assembly operation requires specialized tools and machines.
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
When test cables are used together with a network analyzer, it is often said that even if the cable has a mismatch, this will be corrected mathematically during the calibration procedure. This is partly true so long as the cable is not moved after calibration. However, there can still be problems with local discontinuities caused due to either the inductive interface, the cable entry or, for instance, an excessive value of the insulator bead dielectric constant.
Figure 7 shows two mated connectors with local discontinuities. If the curve is above 50 [ohm], the discontinuity is inductive, otherwise it is capacitive. Point A represents a typical inductive discontinuity due to an air gap at the interface plane between both connectors. If the discontinuities are too close together, multiple reflections may occur as within an interferometer.
Figure 8 illustrates the situation with two discontinuities,
[GAMMA]1 = coefficient of reflection due to a discontinuity at the interface
[GAMMA]2 = coefficient of reflection due to a residual discontinuity in the test cable assembly
z = distance between [GAMMA]1 and [GAMMA]2
The overall [GAMMA] seen at the left disontinuity is given by
[FIGURE 8 OMITTED]
[GAMMA] = [[GAMMA]1 - [GAMMA]2[e.sup.-2j[PHI]]]/[1-[GAMMA]1[GAMMA]2[e.sup.-2j[PHI]]] (1)
After calibration, the reflection, [GAMMA], is arbitrarily set to 0 with the use of correction coefficients. The calibration standards applied are high end products, and the mating dimension is usually very small. Unfortunately, Equation 1 shows that the overall reflection depends on z and the value of [GAMMA]1. ([GAMMA]2 is assumed to be fixed because it is incorporated in the test assembly.) In practice, z and [GAMMA]1 are larger than during the mating with a calibration kit connector.
[FIGURE 9 OMITTED]
Figure 9 shows a numerical application of Equation 1
[GAMMA]1 = -10 dB
[GAMMA]2 = -25 dB, -17 dB and -12 dB
z = varies from 0 to 0.1 mm
f = 20 GHz (2)
As a result, it can be seen that what might be interpreted as a failure of the DUT may in fact be the result of a multiple reflection between the test assembly and the interface or between the test assembly and one discontinuity of the DUT. Moreover, the discontinuities are frequency-dependent, which might lead to an excessive measured ripple in the transmission of an amplifier.
It has been shown that to obtain a good cable design that keeps all the discontinuities of the cable assembly within an acceptable level it is very important for the test assembly supplier to have tight control over the entire assembly design (connector and cable). As has been demonstrated through the SUCOTEST 18 cable test assemblies, Huber + Suhner is in a position to measure in the time domain and then to adjust the compensations between the different 50 [ohm] subsections. Another and complementary way is to perform simulations (mode matching network and inverse Fourier transformation) or to proceed directly in the time domain with a 3D transient solver.
With the aid of these tools, all the discontinuities of cable test assemblies are minimized and not merely compensated for, therefore increasing the quality of measurements by network analyzers.
TABLE 1 SUCOTEST 18 SPECIFICATIONS Up to 2.01 to 4.01 to 6.01 to 12.01 to 2 GHz 4.0 GHz 6.0 GHz 12.0 GHz 18.0 GHz (dB) (dB) (dB) (dB) (dB) Return Loss N >30 >28 >25 >21 >19 SMA >30 >28 >25 >21 >19 Insertion Loss Stability Shaking <0.03 <0.03 <0.03 <0.03 <0.03 Bending <0.03 <0.04 <0.04 <0.05 <0.05 Torsion <0.03 <0.04 <0.04 <0.05 <0.05
Huber + Suhner, Herisau, Switzerland, +41 71 353 4353 or e-mail: firstname.lastname@example.org.
Circle No. 306
HUBER + SUHNER
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Cables & Connectors Supplement|
|Date:||Mar 1, 2004|
|Previous Article:||2004 connector, cable and cable assembly survey: this is the second survey of connectors, cables and cable assemblies sponsored by Microwave Journal....|
|Next Article:||New generation right angle connectors and adapters.|