Dependence of Radar Emission Spectra onMeasurement Bandwidth and Implications for

Compliance with Emission Mask Criteria

ISART 2003 Symposium6 March 2003

Frank SandersChief, ITS Theory Division

NTIA/ITS.Tfsanders@its.bldrdoc.gov

303.497.5727

Institute for Telecommunication Sciences – Boulder, Colorado

Institute for Telecommunication Sciences – Boulder, Colorado

IntroductionEmissions: Radar transmitters produce measurable emissions faroutside the classical sinc2 spectrum. These are called out-of-bandand spurious.

Subtlety: Absolute power is measured at radar fundamentalfrequencies, but spectral power density is measured in the OOB andspurious regions. This leads to an interesting phenomenon…

Emission masks: Out-of-band (OOB) and spurious emission levels ofGovernment radars are regulated in the United States by the NTIARadar Spectrum Engineering Criteria (RSEC). Other masks may bespecified for non-Government radars and radars in foreign countries.

Specification: Emission masks typically specify suppression levelsin decibels relative to the measured peak power transmitted from aradar.

Institute for Telecommunication Sciences – Boulder, Colorado

Classical sinc2

line spectrumenvelope

out ofband(OOB)

out ofband(OOB)

Spuriousemissions

Spuriousemissions

-40 dBbandwidth

40 dB down X dB down(-80 <= X <= -60)

Institute for Telecommunication Sciences – Boulder, Colorado

The ProblemMeasurement Bandwidth: Emission spectra are convolved withmeasurement system response functions. These are typically the IFbandwidths of a spectrum analyzer, for example.

Question: How do OOB and spurious emissions vary with Bm?10 log, 20 log, or somewhere in between? Somewhat equivalentquestion is, how noise-like are OOB and spurious emissions?

Measured power at the radar fundamental varies as 20 log of themeasurement bandwidth, up to the point that measurement bandwidth(Bm) exceeds emission bandwidth (Be), (approximately 1/pulse width).

Measured power (density) in the OOB and spurious regionscontinues to increase when Bm > Be. Furthermore, the 20 logrelationship does not necessarily hold. Some references claim thatOOB and spurious emissions are “noise-like” and should thereforevary as 10 log (Bm).

Institute for Telecommunication Sciences – Boulder, Colorado

ImportancePossibility that corrections might need to be applied tocompliance criteria measurement results: If measured levels ofspurious and OOB emissions do NOT vary as 20 log (Bm), then somesort of correction might be required for the purposes of fitting anemission mask.

This issue has arisen in ITU-R Working Party 8B in connection withDraft New Recommendation M.1177, Recommended MeasurementProcedures for Radar Spurious Emissions.

For example, if OOB and spurious emissions vary as 10 log (Bm),then a correction factor of [(20 log)-(10 log)] might have to be appliedbetween those regions and the measured fundamental power.

(It’s also something that’s just worth knowing.)

Institute for Telecommunication Sciences – Boulder, Colorado

ApproachModelling of radar spurious emissions might have much to offer, butlevel of effort was beyond our immediate resources (especially time).

Instead, we decided to employ an empirical approach for ourintroductory effort.

Decision was made to measure emission spectrum of a single radar inmultiple bandwidths and then determine how measured OOB andspurious emissions were observed to vary with Bm.

More detailed studies may well follow, but this work represents a firstlook at this question in a systematic way.

Institute for Telecommunication Sciences – Boulder, Colorado

Setup: Radiated

(Software designer)

RSMS-3 van

Photo credit: John D. Ewan

Institute for Telecommunication Sciences – Boulder, Colorado

RSMS-3HardwareSystem 1 - 0-1 GHzSmall tower, YIG andvaractor preselectors.

System 2 – 1-18 GHzLarge tower, tower-top YIG& bandpass filters/preamps.

Computer control, GPS,noise diode calibration atantennas, radar pulseanalyzer, digitaloscilloscope, etc.

Institute for Telecommunication Sciences – Boulder, Colorado

Interior equipment bay of RSMS-3

RSMS-3 Hardware