HP 8510, User Manual

Page: 15 / 73

The pulsed-RF net w ork analyzer system allo ws y ou to select either the normal precision 10
kHz IF bandwidth or the new wideband 3 MHz IF bandwidth. The wide IF and detection
bandwidth allo ws testing using pulses as short as 1 microsecond but with accuracy comparable
to traditional non-pulsed measuremen ts.
These system comp onen ts may b e rac k-moun ted or arranged on a desktop.
Theory of Operation
A simplied blo c k diagram of the system is sho wn in Figure 2-1. One syn thesizer pro vides
the test signal stimulus to the RF input of the test set and the other pro vides the LO signal
to the four frequency con v erters (only t w o are sho wn). The LO source is alwa ys tuned 20
MHz ab o v e the test signal source. The standard in ternal phaselo c k tec hnique is not used;
instead, a common 10 MHz frequency reference is used for b oth of these sources and the
in ternal sample selection and timing logic in the HP 8510. These sources are considered to b e
coheren t, th us generating the correct 20 MHz rst IF and the correct clo c k frequency for the
reference and test sync hronous detectors. This eliminates the need to use the reference signal
for receiv er phaselo c k and allo ws all reference and test signals to b e pulsed, thereb y making
fully error-corrected 2-p ort, pulsed-RF S-parameter measuremen ts p ossible.
Figure 2-1. Simplified Pulsed-RF Netw ork Analyzer Block Diagram
Using a wider bandwidth in b oth the reference and the test IF, sync hronous detectors that
op erate at 20 MHz, and accurate measuremen t sample timing circuits allo w exible analysis
of device resp onse during the pulse. With the sources and test set con trolled b y the HP 8510
o v er the 8510 system bus, and with all necessary pulse generation and measuremen t timing
signals generated in ternally from a common 10 MHz reference, the HP 85108 is a complete
pulsed-RF stimulus/resp onse test system. One pulse of a user-sp ecied width is measured at
eac h data p oint and the measuremen t is sync hronized so that it is made at a certain kno wn
time in the pulse. The stimulus dut y cycle can b e predicted for a given instrumen t state,
but the actual pulse rep etition p erio d dep ends up on the curren t domain, cal t yp e, a v eraging,
sw eep time, and pulse width selections. F or this reason, if y our device is sensitive to dut y
cycle, refer to General Timing Information.
2-2 System Description