Anyone familiar with s-parameter measurements knows that there are two steps to the process. First, the measurement instrument needs to be calibrated and then a measurement is performed of the device under test (DUT). This is generally a very inconvenient process that involves making measurements of known standards whereby the instrument utilizes its knowledge of the standards and the measurements it makes of them to produce error terms, which are used to convert the raw measurement of the DUT into its actual s-parameters. There are many types of calibration, such as short-open-load (SOL), used for single-port measurements, short-open-loadthru (SOLT) and short-open-load-reciprocal (SOLR), used for multi-port measurements, and many more. The method of applying the standards to the instrument is usually by making many manual connections of the standards or by using an electronic calibration (ECAL) module. Sometimes a calibration substrate is used with probes, or calibration structures are built into test fixtures.

Calibration has two purposes:

  1. to discover the response characteristics of the transmitter and receiver in either a vector network analyzer (VNA) or time-domain reflectometer (TDR) instrument, and to account for drift of these characteristics over time, temperature, and other environmental changes.
  2. to essentially perform de-embedding of the path between the transmitter and receiver, and the DUT.

These two purposes are quite different in nature and allows for the employment of different methods, as provided in the WavePulser 40iX.

WavePulser 40iX Operation

The WavePulser 40iX takes an innovative approach to calibration in order to make easier and quicker measurements [1]. In explaining this, reference will be made to figure 1, which shows the internal structure of the instrument schematically.

In figure 1, there are two pulser/sampler modules connected through input paths, which are semi-rigid coaxial cables, to the common inputs of the single-pole-six-throw (SP6T) relays. The outputs of the relays have connections to various standards, including short, open, and load standards, as well as a thru standard connection between the ports. The outputs shown unconnected are for other thru standard connections between the other ports of the instrument that are not shown. For each port, one relay output connects to the output path, which is also a semi-rigid cable, leading to the bulkhead connector of the instrument. Finally, user cables connect from the bulkhead connector the DUT.

In the factory, after assembly of the microwave subsection, which is the rigid mechanical structure housing the relays and bulkhead connectors, measurements are made of the s-parameters for various paths. These measurements are performed for the short, open, load and thru standard connections, as well as the path from

the common input of the relay to the bulkhead connector. Similarly, the s-parameters of the user cables are measured and associated with a unique serial number for each cable.

With these s-parameters provided on a memory card internal to the instrument, it is ready to take measurements. Once the temperature has stabilized after power on, the user would select a two-port measurement and press the Go button. In a first step, the instrument calibrates itself by alternately switching the relays between the short, open, load and thru standards and performing acquisitions of impulses applied to the standards. For the two-port measurement, it requires six acquisitions: baseline, short, open, load on both ports simultaneously, the thru driven from port 1 and the thru driven from port 2. All of these measurements have frequency content from true DC to 40 GHz, with the baseline calibration providing improved accuracy at low frequency, which is required for certain applications like automotive Ethernet. These six acquisitions are converted to raw measured s-parameters [2, 3], so called because they are ratios of measurements of incident and reflected waves, but are not calibrated measurements. These raw measured s-parameters and the factory measurements of the standards, are supplied to traditional calibration algorithms like SOLT and SOLR (see [4]) to determine the error terms. Because the standards were measured directly at the common input to the relay, this becomes the internal calibration reference plane, as shown in figure 1. Because of the location of the reference plane, the semi-rigid input path is de-embedded, which allows the input path cables and the pulser/sampler modules to be replaced during repair without any further recalibration of the instrument.

With the internal reference plane established through the internal calibration, the raw measurements of the DUT are made by switching the relays to the output paths and taking two acquisitions, one with pulser 1 pulsing, and the other with pulser 2 pulsing. The raw two-port s-parameter measurements and the error terms are used to calculate a calibrated measurement. But remember the location of the internal reference plane – the calibrated measurement is of the path between the common inputs of each of the relays, meaning it consists of the relays themselves, the output paths, the user cables, and the DUT itself. Using the factory measurements made of path from the common input of the relay to the bulkhead connector of the instrument, these paths are

de-embedded from the measurement to move the reference plane to the instrument reference plane, as shown in figure 1. Finally, the user cables are de-embedded, which brings the reference plane to the DUT, and a calibrated measurement of the DUT is obtained.

Manual Calibration

While the previous section described the inner workings of how the WavePulser automatically performs calibrated measurements, it is not the entire story. Consider the calibration menu shown in figure 2. To the left of the screen, the calibration type is selected. Usually this is set to Auto, and the DUT measurement proceeds as described previously, with the exception that usually a factory second-tier calibration is applied. The second-tier calibration settings are shown to the right in figure 2 and will be discussed later.

As also shown in figure 2, a manual calibration can be selected, which would invoke the Manual Calibration menu as shown in figure 3. Some discussion of manual calibration is needed to put the second-tier calibration in context.

The Manual Calibration menu allows the selection of all of the information necessary to load the results from a previous manual calibration, or to perform a new one. To perform a manual calibration, a calibration kit is required, consisting of standards. The calibration kit comes with information that describes the models of the standards, which is input through a calibration kit file.

The user connects the standards to the instrument ports, presses the Connect button next to the particular standard/port combination, then presses the Measure button. After the required acquisitions are complete, a check mark is placed in the box for that standard/port combination, and the user continues making measurements until all of the boxes are filled in, after which the combination of the measurements taken and the calibration standard definitions are converted into the error terms for the manual calibration. After a manual calibration, the calibration type shown in figure 2 is left as Manual, and none of the Auto calibration discussed previously is applied. The error terms are applied to raw measurements of the DUT and none of the internal calibration standards are used, nor any of the internal path de-embedding operations. VNA users would be familiar with this process as this is the only way to calibrate it.

User Second-Tier Calibration

  • The user wants to instantaneously improve the accuracy of a measurement. After all, despite the fact that the instrument calibrates to an internal reference plane and calibrated measurements are taken at that reference plane, the remainder of the calculation of the DUT involves de-embedding the output path and user cables. The user cables especially might change over time, mostly from mechanical flexing. Thus, a manual calibration allows a direct, VNA-like calibration that removes this de-embedding step, which relies on a priori knowledge of the path between the pulser/sampler and the DUT.
  • The user is connecting fixtures or probes between the instrument ports and the DUT, and is using calibration as a form of de-embedding.

Item 2 is captured first. De-embedding of fixtures and probes is a common use of calibration. But, it’s a shame that the manual calibration performed is only valid temporarily and must be performed again due to drift of the instrument with time or temperature. This is where second-tier calibration comes in.

Second-tier calibration is selected from the Calibration menu shown in figure 2, which opens the Second-Tier Calibration menu shown in figure 4. This menu is very similar to the Manual Calibration menu shown in figure 3. In fact, the process of making a second-tier calibration is identical to manual calibration up until the final steps.

After performing the steps of applying standards and making measurements as in a manual calibration, the user presses the Perform Internal Cal button, after which the instrument performs an internal calibration to the internal calibration reference plane shown in figure 1. Now, the instrument has calibrations at two reference planes: the internal calibration reference plane and the DUT reference plane.

Once the internal calibration has been performed, ideally very close in time to the completion of the manual calibration, a button to create the second-tier calibration file is activated, as shown at the right of menu in figure 4. The calculation of the second-tier calibration is quite complicated, but is shown stylistically in figure 5. The basic idea of the calculation is to constrain the second-tier calibration error terms such that the internal calibration combined with the second-tier calibration produces the same error terms as the manual calibration.

Because at a given instance in time, the internal calibration combined with the calculated second-tier calibration error terms equals the manual calibration, this means that as the instrument drifts, new internal calibrations can be performed and dovetailed with the second-tier calibration to keep the instrument calibrated. This works because the internal calibration takes care of drift and changes in pulser/sampler performance, while the secondtier calibration is actually performing only a de-embedding operation. Thus, the method works as long as the

de-embedding structure, which consists of everything between the internal reference plane in figure 1 and the DUT, remains unchanged. This is a reasonable expectation.

All of the mathematical complication notwithstanding, it’s worth mentioning that the second-tier calibration is actually calculated and applied in a slightly more intricate manner, in that the path from internal reference plane to the instrument reference plane is de-embedded. This does not change the accuracy because any errors in this de-embedding are taken up by the second-tier calibration. In other words, it is constrained such that everything adds up to the manual calibration.

Factory Second-Tier Calibration

When the WavePulser 40iX is calibrated in the factory, this same process of performing a manual calibration and converting it to a secondtier calibration is performed with the user cables that are measured and provided with the instrument. Here, one more twist on the algorithm is made such that the cables are added to the standards applied, setting the second-tier calibration reference plane at the instrument reference plane. Thus, the DUT calculation steps become:

  1. Perform the internal calibration.
  2. Perform the raw DUT measurement.
  3. Apply the internal calibration to the raw DUT calculation to produce first-tier calibrated measurement of the path from the internal calibration reference plane to the DUT.
  4. De-embed the path from the internal calibration reference plane to the instrument reference plane.
  5. Apply the factory second-tier calibration error terms to provide a second-tier calibrated measurement at the instrument reference plane.
  6. De-embed the user cables to provide a calibrated measurement of the DUT.

This is obviously very complicated, but if one is following these steps, one would observe that if everything were measured perfectly without the factory second-tier calibration, the error terms used in step 5 would do absolutely nothing. In practice, the factory second-tier calibration makes slight improvements to the calibration accuracy.

The bundled configuration of the WavePulser 40iX ships with a compact Rosenberger RPC-2.92 female calibration kit. This calibration kit can be used to perform user and factory second-tier calibration of the instrument on a periodic basis. To do this, the following steps are followed:

  1. Ensure that the user cables with the correct serial numbers and s-parameter files are attached to the unit with the matching color band at the instrument port. The serial numbers can be found on the Instrument Setup tab.
  2. Select the Second-Tier as the calibration type in the Calibration menu in figure 2.
  3. Set up the number of ports, number of averages and the acquisition length. For a factory calibration, the ports must be 4.
  4. Under Load/Save Calibration Files Path,2 select a folder to contain the calibration files. Usually these are saved in a new directory under c:\LeCroy\WavePulser\UserCalibration.
  5. Under Use Cal Kit File, browse to the calibration kit file for the Rosenberger calibration kit. It is installed automatically with the software in the directory: c:\LeCroy\WavePulser\UserCalibration and has the file extension .cstd.
  6. Attach the calibration kit to the ports at the ends of the cables, press the Connect button, and take acquisitions by pressing the Measure button in whatever order desired, with the goal of filling in all of the boxes with a check mark. Usually, by using the thru standard between two ports, in conjunction with two reflect standards on the other ports, the calibration can be accomplished in exactly six measurements.
  7. Perform the internal calibration. This step also performs the baseline calibration by connecting the internal load with no pulsers pulsing.
  8. Create the second-tier calibration file.
  9. Press the Convert To Factory Cal button.

At the end of this procedure, the user second-tier calibration file is created and written to the memory card internal to the instrument with the name X_FactorySecondTierCalibration.l12t, where X stands for the instrument serial number. If the user performs the factory second-tier calibration for the first time, the factory calibration data provided by Teledyne LeCroy will be renamed with the suffix _Original. Otherwise it will be renamed with a suffix indicating its creation date. This allows older calibrations to be recovered in the event that problems are found with the new calibration.

By following this procedure, WavePulser users have the unique advantage to periodically maintain the unit calibration themselves without having to send the instrument back to the factory.


The WavePulser 40iX combines factory measurements of internal standards with automatic internal calibration to set an internal instrument reference plane. This internal calibration is combined with de-embedding and second-tier calibration to remove the effects of cables and fixtures in the path in the measurement of the DUT. This combination of calculations gives the user the advantage of taking measurements from DC to 40 GHz in minutes, and frees the user from cumbersome and time consuming calibration of the instrument. Methods are provided within the instrument to perform manual calibrations when these are required and to apply these as second-tier calibrations so that the de-embedding effects of the manual calibration can always be used.

Finally, manual calibrations can be converted to factory second-tier calibration equivalents to avoid the need to return the WavePulser to the factory for periodic recalibration.