November 2, 2021
Calibration is critical for engineers to conduct precise S-parameter measurements using a vector network analyzer (VNA). While a VNA is a highly-linear receiver and has sufficient spectral purity in its sources to make accurate measurements, there are imperfections if calibrations are not performed. As designs continue to extend into higher frequencies, such as 5G Frequency Range 2 (FR2) millimeter wave (mmWave), calibration takes on added importance.
Standard network extraction and de-embedding capabilities are common methods of calibration to characterize and remove test fixture measurement contributions. Such techniques are used for all VNAs, including the ShockLine™ MS46122B series of PC-controlled, compact VNAs that support 1 MHz up to 43.5 GHz.
All three models in the ShockLine MS46122B series are full-reversing, 2-port VNAs aimed at RF and microwave applications within manufacturing, engineering, and education organizations. To accurately conduct necessary S-parameter measurements on RF passive components, some of which are listed in table 1, proper calibration must be conducted.
Type-B Two Tier Calibration
One calibration method engineers can employ to ensure measurement accuracy when using the ShockLine Compact USB VNA MS46122B is type-B, two tier calibration with flex standard network extraction. The method can de-embed a fixture when only partial standards are available. By doing so, the insertion loss will be close to 0 dB.
Type-B, two tier calibration with flex standards is particularly useful when the other end of the fixture is not accessible to standard calibration kits. Using this method, engineers can de-embed the fixture with one (either open or short), two (open and short), or three (open, short, and load) standards. The more standards applied, the more accurate the measurement.
The main purpose of type-B, two tier calibration with flex standards is to extract the .s2p file of the fixture with different or incomplete calibration standards. Algorithmically, it is similar to the Bauer-Penfield technique (full calibration standard: short, open, and load) but an incomplete calibration is performed at the fixture output plane. A general setup for Type-B with a fixture attached to a VNA is depicted in figure 1.
The fixture can be represented by:
- 2 x 2 scattering parameters array denoted as S11, fixture (fixture outer plane reflection on port 1)
- S21, fixture (transmission loss on the fixture from port 1 to port 2)
- S12, fixture (transmission loss on the fixture from port 2 to port 1)
- S22, fixture (fixture inner plane reflection after connecting to reflection standards).
Non-Touching Loop Rule
Engineers can also implement the non-touching loop rule when calibrating. It is a technique that can be applied on any signal flow model. Figure 2 is a network signal flow where the fixture is placed between the VNA and a reflection standard. The system has only one independent variable (a1).
The flow diagram contains “paths” and “loops.” A “path” (P) is defined as a series of lines follow in sequence and in the same direction, so no node is touched more than once. The value of the path is the product of all coefficients encountered in the path. There are also loops in the flow diagram.
The impact of type-B, two tier calibration with flex standards can be seen in a sample measurement conducted using the ShockLine MS46122B. For the measurement, the Compact USB VNA had a starting frequency of 100 MHz and a stop frequency of 10 GHz. One hundred points were acquired. A full 2-port SOLT calibration using TOSLKF50A calibration standard was performed.
In the sample calibration, the short and open are TOSLKF50A calibration standards and its offset is 5.01 mm. The estimated delay for the fixture was set 0 ps and the software automatically measured the fixture delay. The ShockLine VNA GUI will extract a .s2p file using a network extraction tool.
A sample of the fixture .s2p file is shown in figure 3. When using open and short standards, the inner plan, which is S22, is assumed perfect match and also S12 = S21 (assume reciprocal). After de-embedding the fixture, S12 and S21 traces become approximately 0 dB (as depicted in figure 4). S11 is a bit worse before de-embedding because the S11 fixture is not a perfect match.
Using the non-touch loop method with the signal flow diagram, the scattering parameters of the fixture can be mathematically modeled. With Type-B, two tier calibration with open and short standard, the inner plane is assumed to be perfect.
The application for the Type-B two tier calibration with flex standard is a useful tool when the fixture has incomplete calibration standard and the user is only interested in de-embedding the loss of the fixture and measure the device under test (DUT) for the insertion loss.
To learn more, download the application note on type-B two tier calibration with flex standard network extraction.