skrf.calibration.calibration.SDDL
- class skrf.calibration.calibration.SDDL(measured, ideals, *args, **kwargs)[source]
Short-Delay-Delay-Load (Oneport Calibration), formulation by Arsenovic et, al.
One-port self-calibration, which contains a known short, a known arbitrary-impedance load, and two partially-unknown delay shorts. The short can be an offset short with known phase, or a zero-length (flush) short. The load can be perfectly matched or an arbitrary impedance standard, as long as it’s known exactly. This is especially useful when an ideal load is impractical to fabricate. The delay shorts have unity reflection magnitudes, but unknown phases (unknown electrical lengths).
See
__init__()for the correct parameters order.This class uses the Arsenovic et, al. formulation [1], which allows either a mismatched or a matched load as the calibration standard. It’s a generalization of both Sigg-Simon’s original SDDM [2] and Liu-Weikle’s modified SDDL methods [3], which are limited to matching or reflective loads only. The latter was originally designed for waveguide calibration with resistance to flange misalignment.
Note
This algorithm is bandwidth-limited due to phase wrapping, conceptually similar to TRL’s bandwidth limit. Some over-determined multi-line variants allow wideband calibration, but this is not implemented in scikit-rf. If a wideband calibration is required (common for coaxial systems), it’s necessary to manually split measurements to multiple subbands and calibrate each band separately.
Band edges may suffer from increased errors. It’s recommended to test line lengths experimentally via simulations first. Choosing an optimal selection of line lengths that covers the widest bandwidth is a non-trivial optimization problem, with several proposed solutions [4].
References
Attributes
Estimate of biased error for overdetermined calibration with multiple connections of each standard. |
|
Return a NetworkSet for each caled_ntwk, grouped by their names. |
|
List of the corrected calibration standards. |
|
Dictionary or error coefficients in form of numpy arrays. |
|
Dictionary of error coefficients for 12-term Model. |
|
Dictionary or error coefficients in form of Network objects. |
|
Dictionary of error coefficients for One-port Error model. |
|
Dictionary of error coefficients in form of Network objects. |
|
Dictionary of error coefficients for 8-term (Error-box) Model. |
|
Dictionary of error coefficients in form of Network objects. |
|
Dictionary of error coefficients in form of Network objects. |
|
The calculated error Network or Network[s]. |
|
|
|
Directivity normalized to the reflection tracking. |
|
Number of ideal/measurement pairs in calibration. |
|
Return any output from the |
|
Returns a NetworkSet for each residual_ntwk, grouped by their names. |
|
Dictionary of residual Networks. |
|
Estimate of total error for overdetermined calibration with multiple connections of each standard.This is the combined effects of both biased and un-biased errors. |
|
Estimate of unbiased error for overdetermined calibration with multiple connections of each standard. |
|
Methods
Short-Delay-Delay-Load initializer. |
|
Apply correction to a Network. |
|
Apply correction to all touchstone files in a given directory. |
|
Apply correction to list or dict of Networks. |
|
Apply correction to a NetworkSet. |
|
Embed an ideal response in the estimated error network[s] |
|
Create a calibration from its error coefficients. |
|
Create a calibration from its error coefficients. |
|
Plot corrected calibration standards. |
|
Plot biased, unbiased and total error in dB scaled. |
|
Plot residual networks. |
|
Remove and return tuple of (ideal, measured) at index. |
|
Remove a cal standard and correct it, returning correct and ideal. |
|
Run the calibration algorithm. |
|
Update current dict of error coefficients. |
|
Write the Calibration to disk using |