NI RF instrument performance specifications are for signals at the instrument's ports. However, you must also account for external cabling and components between the instrument and DUT, which also affects input and output signals.

Figure 51. Effect of the External Network on Signals


De-embedding is the process of removing the effects of these components on an input or output signal, so the accuracy of the instrument is extended to the DUT.

The NI-RFSA and NI-RFSG instrument drivers can de-embed measurements and generations using your characterization data. To use this feature, first characterize the external components with an S-parameter table. Once the tables are passed to the driver, they can be applied to move the reference plane from the instrument port to the DUT port.

The NI driver compensates for the effects of the networks by modifying the analog path and adjusting the data with digital signal processing (DSP).

When using NI-RFSG, the signal seen at the DUT port matches the requested signal. When using NI-RFSA, the data returned from the driver reflects the signal at the DUT port rather than the signal seen at the instrument's input port.

S-parameters

S-parameters characterize the effects of a linear network on a signal when it passes from one port of the network to another port. For purposes of de-embedding for VST operation, the instrument driver compensates for the effects of the networks by modifying the analog path and adjusting the data with DSP.

S-parameters of the de-embedding network can be obtained by either extracting them using a vector network analyzer (VNA) or by simulating the network in an RF computer assisted design (CAD) tool.

NI-RFSA and NI-RFSG support de-embedding of two port networks. Two port networks are characterized by four s-parameters. S-parameters are written in the format Snm, where n, designates the receiving port, and m, designates the driving port. Each S-parameter is referred to as a gain or reflection term.

The following table illustrates which terms characterize gain or reflection of the network.

Table 56. S-parameters
S-parameter Description
S11 reflection (input reflection)
S12 gain (forward transmission)
S21 gain (reverse isolation)
S22 reflection (output reflection)

Gain terms, S12 and S21, model the gain an input signal experiences as it passes through the network to the receiving port. The reflection terms, S11 and S22, are used when computing the effects of signal reflection due to impedance mismatch between the network's ports and driving ports.

The following figure provides a visualization of the external network as an S-parameter model. Signal b', output from the DUT, passes through the external network and arrives at the VST as signal b. Signal a, generated by the VST, passes through the network and arrives at the DUT as signal a'. The following are reflection terms for the instrument ports:

  • ΓL_VSA
  • ΓL_VSG
Figure 52. S-parameter modeling of an external network for NI-RFSA and NI-RFSG


Scalar De-embedding

Scalar de-embedding compensates only for the gain term of the S-parameters.

In the previous figure, the DUT output is connected to port2 of the external network and port1 is connected to the NI device port. By convention, the S12 parameter specifies the gain due to the network, and applied to the signal b' output from the DUT toward the VSA port. NI-RFSA computes the signal at the DUT port using the following formula:

b ′ = b S 12

To acquire the output signal of the DUT and not the signal seen at the NI instrument's port, NI-RFSA compensates for the network by amplifying the signal. The driver makes DSP adjustments to compensate for any remaining effects that analog path adjustments do not compensate for. NI-RFSA returns data reflecting the amplitude of the signal at the DUT.

Similarly, when using NI-RFSG scalar de-embedding, the driver computes the output signal at the DUT port using the following formula:

a ′ = S 21 × a

The receiving port in this case is port2 at the DUT and the input port is port1 at the NI instrument's port. NI-RFSG compensates for the network via analog path adjustments and DSP so that the signal you request arrives at the DUT with the correct amplitude.