Purpose : Provide examples of Scalar Calibration Techniques as alternatives or supplements to the Vector Cal Techniques.
For certain fixture configurations, conventional 1 port reflective s-parameter calibrations are not possible. This occurs due to several possible factors.
1.) The RF resource attached to the fixture path is Scalar: RF4, 5 and 8 are source or receive only ports. There are no couplers and therefore cannot have an s-parameter calibration.
2.) The fixture has a large attenuator in the path: A large (typically >6dB) attenuator makes it difficult to do a reflective 1-port calibration.
3.) The path is not reciprocal: The presence of an amplifier or circulator cannot be calibrated with a reflective 1-port calibration.
To address these fixture paths we have two types of fixture calibrations; AC Calibrations and Scalar Calibrations. This note will focus on Scalar Calibrations.
All of the typical conventions of Fixture calibrations apply. The path name is created by the fixture editor, It needs to be reset during the cal plan compilation, It must be saved after etc.
In writing these calibrations, more care must be paid to the power levels involved. The standard s-parameter cal plans are already optimized for most systems. In this case, the use of large pads, or amplifiers forces the developer to pay attention to IF Gain and Receive attenuation. These example test plans are not optimized for all systems, amps and attenuators. Optimize the measurements, just as you would for a conventional testplan.
The key to being able to implement any calibration is the ability to create a "known" standard at the desired calibration plane. For S-parameter cals, the known standards are reflections and terminations. For these Scalar calibrations, the standard required is the ability to set or measure a known power. We can then compare the measurement of the known to the known value and come up with the path loss or Gain.
In practical terms, for typical cable or fixturing schemes. A preliminary calibration must be done to get the "ability to set or measure a known power" at the DUT reference plane.
Consider this fixture:
Here we have three paths:
1.) RF7DutRF7 is a straight thru path. This can be cal'd with SOL standards
2.) RF6DutRF6 has a 30dB pad.
3.) RF3DutRF3 has an amplifier.
Once the RF7 path is calibrated, with the addition of a DutRF6 to DutRF7 thru, it can then be used as a known stimulus to measure the RF6 path. Then with the addition of a DutRF3 to DutRF7 thru, RF7 can be a known receiver to calibrate the Amplifier's gain.
Cal Boards for Thru's
Now to develop the testplan:
1. ) First we define the fixture, be sure to make the cal type RF
2.) The example measurement panel:
a.) The basic measurement is the output wave b2 (output port is RF6) minus the input wave a1 (input port is RF7).
b.) RF 7 was already calibrated, therefore it was known.
c.) Remember the cal for the path PathRF6DutRf6 must be reset. And it is only reset when the plan is compiled!
d.) The loss is calculated (in dB), has it's frequency added to the array, then is sorted by frequency and saved as a Local Variable (Path Loss). It is important for the loss to be in log. The add freq calculation expects the data in log.
3.) The next panel calculates the cal data and writes the cal.
a.) Path Loss is brought back as a local variable.
b.) The spline fit is done.
c.) Then the data is written to the variable PathRf6DutRf6.
The same method can then be implemented for the amplifier gain path, being careful of the RX atten, IF Gain, Changing the input and output ports etc.
Like all calibration plans, it makes sense to write a validate plan to see if the calibration was successful.
Here is a validate plan panel.
1.) The Reset Calibration button has been removed!
2.) There is no need to attach frequency information to the data save button, that is done automatically.
3.) The Spline fit and data write panel has been removed.
4.) The value should be 0dB loss if all is well.
Here is a copy of this cal plan, validate plan and fixture definition.