This episode is sponsored by Rohde & Schwarz. Check out their reviews of their new product line at AskAnEngineer.us, including a video review from friend of the show Shahriar from The Signal Path.
Welcome, Dr Adrian Tang of NASA’s Jet Propulsion Laboratory (JPL)!
Adrian is the only CMOS chip designer at JPL (though he has grad students)He got his PhD at UCLA working on ultra high frequency devices. Check out some of his projects on his personal website.JPL has instruments that can measure these high frequencies.Terahertz spectroscopyThere's a noise floor in the universe, but gases tend to emit at certain frequencies in the 50-500 GHz region.Resonance in the THz rangePressure has to be lowTrying to process as much bandwidth as possibleTry to get 3GHz of bandwidth out of devices.4000-8000 point FFTPower budget might only be 50W on a satellite.Was doing commercial CMOS previouslyHorn antenna + lensTempest D missionUsing cosmic background as a cal pointWhat is the output data?Time domain dataUsing the LO to remove the movement of the spacecraftDoppler propertiesHerschel telescopeReflector to focus it down into the feed is about a meterHas to be accurate to lambda over 20Goes into a receiver feedMixer needs an LOIF comes out of the mixer at 3-4 GHzTakes it through the IF chainAmps and filtersApply a window functionPolyphase filterCompute the FFT1 sample isn't enough to see somethingAverage over 1000s of samplesRosetta orbiter was put out by ESAThe more spectrum, the more chance to discover somethingRekTangLE balloon missionMeasuring NO2 at 570 GHzDid have a failureThis was one of the first CMOS chip to do atmospheric measurementIsotopes have different linesHeavy water and regular waterHeavy water is naturally occurringEverywhere you go, the ratio is comparable, in the solar systemMoreau on a jovian cometHow have things moving into CMOS?Active and Passive ReceiversNorthrop Grumman makes some of the only amplifiers that can operate up to 600 GHzPast guest Tony Long talked about (relatively) high frequencies and works at NGPassive is a mixer first instead of an amplifier firstSchottky mixerExpress sensitivity in terms of differenceHot electron bolemetric mixersExotic semiconductor junctionsWhat has improved?WindfreakPackaging is not simpleWaveguide is the only thing that will work at 100GThey do a lot of the fab in house at JPLMeasuring snow in the Sierras using Ku band radarSWE - Snow water equivalentSpectreRFBuilt in self testEverything on chip is the best caseUsing different process technologies65 nm process technology cannot work with 180 GHz signals.What is this stuff useful for?Short range 60 GhzKen CooperThreat detection, 680 GHz radar, 30 GHz bandwidthAnother balloon "whatsup" will go up soon. Read about it and other balloon launches (past and present) here.Astros - 8 px, multi THz telescopeGround penetrating radarUAV snow sensorCheck out more of Adrian's work on sadcircuitdesigner.comOr check out his papers on his "articles" page