ADL5960ACCZ-R2 datasheet(19/32 Pages) AD

19 / 32 page

Data Sheet

ADL5960

THEORY OF OPERATION

analog.com

Rev. A | 19 of 32

CALIBRATION AND ERROR CORRECTION

The accuracy of S-parameter measurements is prone to a variety

of errors introduced by the hardware in vector network analyzers.

Systematic errors, those that are repeatable and predictable, can

be eliminated from the measurement results using calibration and

error correction techniques. The types of systematic errors that can

be eliminated (to a certain extent) include the following:

►

Impedance mismatches

►

Gain and insertion loss differences between channels

►

Crosstalk between channels

Various error correction and calibration techniques have been de-

veloped over the years (see, for example, D. K. Rytting, "Network

Analyzer Accuracy Overview," 58th ARFTG Conference Digest,

2001, pp. 1-13, doi: 10.1109/ARFTG.2001.327486) that differ in

their effectiveness to eliminate certain errors, complexity, and cali-

bration standards used. The general principals that are applied,

however, are the same for all. The systematic errors are determined

by measuring the VNA response for a set of DUTs with accurately

known S-parameters—the calibration standards. Comparing the

measured S-parameters with the known S-parameters allows calcu-

lation of the error coefficients. These error coefficients, combined

into what is often referred to as an error adapter, can then be

used in a postprocessing step to calculate the error corrected

S-parameters from the measured S-parameters of an unknown

DUT. Figure 55 illustrates the concept.

Figure 55. Concept of Error Correction in VNAs

Different error correction techniques require the use of different

sets of calibration standards. Some of the most well known include

short, open, load, thru (SOLT), thru, reflect, line (TRL), and thru, re-

flect, match (TRM). The standards within a set are chosen to have

significantly different S-parameters, such that they span a large

area on the Smith chart. The calibration standards themselves, par-

ticularly the ones with extreme S-parameter values such as short

and open standards, are typically not completely ideal either and

usually include a manufacturer supplied model for the S-parameters

vs. frequency. In general, calibration standards can be subdivided

into the two following categories:

►

One-port calibration standards, measured at each VNA port

►

Two-port calibration standards, measured for each combination

of VNA ports

The calibration procedure for a VNA built from ADL5960 devices is

similar to that for any other VNA system. For maximum accuracy,

it is important to note that the frequencies used. Additionally, the

ADL5960 SPI gain, bandwidth, and frequency multiplier/divider

settings used during calibration must exactly match the settings

used during an actual DUT measurement. Calibrate the VNA for

multiple different configurations if the settings are expected to be

different during an actual measurement because gain and other

settings are subject to device-to-device and channel-to-channel

spread (mismatch).

Although S-parameters are the result of a power ratio calculation

and, in principle, independent of absolute power levels, it is often

important to accurately control the RF source power vs. frequency

during a measurement. When the DUT exhibits nonlinearity across

the applied RF power levels, such as semiconductor devices, a

change in power level can result in a change of the DUT behavior

that affects the measurement accuracy. The ADL5960 forward IF

channel can be used to monitor the power levels in the RF signal

path, particularly when the port is terminated with 50 Ω, such as

during calibration of the system with a load standard. The conver-

sion gain of the ADL5960 itself does exhibit roll-off vs. frequency

as well, which must be taken into account to achieve an accurate

power measurement.

One-Port Calibration

The calibration procedure for a one-port S-parameter measure-

ment can be explained using the flow diagram in Figure 56

(see also D. K. Rytting, "Network Analyzer Accuracy Overview,"

58th ARFTG Conference Digest, 2001, pp. 1-13, doi: 10.1109/

ARFTG.2001.327486). The directional coupler and error model

together describe the operation of a practical VNA. The incident

wave, a0, and reflected wave, b0, represent the forward and reverse

power measured by the VNA. When using the ADL5960, these

vectors are obtained from the IF outputs, IFFx and IFRx. The

actual power incident on the load is represented by a1, whereas

b1 represents the actual power reflected by the load. An error-free

VNA measures a1 and b1.

ADL5960ACCZ-R2 数据表(PDF) 19 Page - Analog Devices

ADL5960ACCZ-R2 数据表(HTML) 19 Page - Analog Devices

Html Pages

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

数据表下载