NRAO Home  >  Green Bank  |  Wiki Topic:    GB > Electronics > RfiGroup > ArchivalMonitoring

Changes | Index | Search | Go

Archival Spectrum Monitoring

• 1. Need
• 2. Feasibility
  • Requirements
  • Options
    • Front Ends
      • GBT RFI Monitoring Station
      • RFI Trailer Monitoring Station
      • GBT RCO's
    • Back Ends
      • Spectrum analyzer
      • GBT Spectrometer
    • Data Display
      • OASIS
      • IDL & Other astronomy display tools
  • Decisions
    • GBT-Based RFI Monitoring
    • GBT RFI Monitoring Station

1. Need

• Archival spectrum data would:
  • show how the spectral environment is changing over the years
  • enable us to see when sources began to show up so we could then correlate it with dated activity in the NRQZ and WVRAZ
  • in addition to the database of known sources, serve as a tool for observers in planning their observations
  • be a current resource to draw on when preparing spectrum summaries

2. Feasibility

Requirements

• Automated: monitors the spectrum on a regular or continuous basis without any operator intervention.
  • Is total automation required? Assuming that before each observing run with a particular receiver, that receiver needs to be checked for proper function, then part of this check could be viewing the total band pass of the receiver to ensure that the receiver baseline had not changed or at lease was what was expected and to see if there are any RFI signals present that would hamper observations. This could be done with the telescope at zenith and then archived and I think the process could easily be automated except for someone to run the "startup RFI/Rx checkout script". WesleySizemore
  • Let's consider the semi-automated option, where a human initiates the scan on some regular or semi-regular basis, then it runs as an automated process. The questions are:
    • Who initiates it and when? (Depends on what system we use)
    • If not someone from within the RFI group, then who? What is the feasibility of making this a standard part of RCO's? Ron:

Probably shouldn't be done before each observing run. My estimate is that, with our typical 35 observing session per week, we'll loose 3 hrs per week of Astronomy to such measurements. Except for the PF receivers, RCO's don't happen often enough for our other receivers (once every 2 years). Instead I suggest that this be done during Tcal measurements, which are scheduled once to twice a month and cover 1 to 3 receivers per month. These are bandscans, by definition and the RFI data comes as a byproduct that, up to now, we have been throwing away. We are able to cycle through measuring all receivers once to twice a year. And Tcal measurements are often already part of a RCO. But, Tcal measurements are at whatever Azimuth and Elevation we find calibrators at the time of the observing. What we need to discuss is whether it is best for the scientific staff or RFI group to take the Tcal data and pull out the RFI information. Whether Tcal measurements need extra telescope time to do a bandscan at a single Az and El. How to get the RFI information out of IDL.

My back-of-the-enevlope estimate suggest that, in comparison to your monitoring station, the GBT will have about 10 db less sensitivity to locally generated RFI. The weak sidelobes of the GBT that fall on the horizon (-70 dB at best at 1 GHz) are probably no match to a small but directed antenna. Thus, using telesscope time for a repeated, directed RFI survey may not be as useful as a survey with the monitoring stations. Ron Maddalena

Ron has a good point that doing spectrum surveys with the GBT receivers will take up too much telescope time. At best, it might be done on maintenance days, which would be hit and miss and require quite a bit of attention from someone to find appropriate times and initiate the scripts. RickFisher

Rick's measurements (paper attached) using the dipole array antenna, which has a gain of ~7dBi shows a 9dB Signal-to-Noise advantage of the monitoring station over the L-band sidelobes. The Omidirectional antennas on the station have a gain of 0dBi, which mean measurements using these (which would not require motion) still give us about 2dB advantage, and if we used the directional antennas (which do require motion and may be weather-restricted), with gains around 13dBi, we've got a huge advantage (15dB) over the GBT sidelobes. So your back-of-the envelope calculations have bourne out experimentally already. Rick used his own data collection system (a dual-channel, 12-bit A/D PC-card) for this. I was thinking (as one possible approach) that we might want to send the input from the monitoring station into the spectral processor, and use the existing data-handling tools of the GBT to archive it. I don't know how to do this, however, and would need to make an official project of it if we decide to go that route. CarlaBeaudet

monit_gbt_comparison.pdf: Rick compares GBTMS and GBT sidelobe sensitivity

One of the projects on my TO DO list is a head-to-head comparison of sensitivities from the spectrum analyzer and the software FFT technique. I did a quick look at this, which looked encouraging, but we need something more quantitative on a variety of RFI spectrum types. We can also compare simultaneous spectrum analyzer results taken at the RFI Trailer Monitoring Station and the GBT RFI Monitoring Station. And we should also do some more GBT receiver to GBT RFI Monitoring Station comparisons on signals other than DME.

Before launching into routine spectrum monitoring, let's try a few things to see what's most useful and most sensitive. From this experience we can make some better judgments on how to proceed. If you like, I can sketch out some test scenarios, and, hopefully, we can find some time to execute them. RickFisher

• Accessible Data: Data is displayed and retrievable via a web interface, perhaps in a similar manner to Rich Bradley’s SRBS page: http://www.nrao.edu/astrores/gbsrbs/index.shtml

Options

Front Ends

GBT RFI Monitoring Station

Operating range: 130 MHz - 3 GHz

Pros:

• Location on the GBT means:
  • Less prone to be dominated by nearby terrestrial sources
  • More apt to see interference experienced by GBT users

Cons:

• The Automated requirement may be occasionally foiled by icy weather.
• Does not cover the entire GBT operating range.

RFI Trailer Monitoring Station

Operating range: 100 MHz to 2.0 GHz (can easily be extended to 4 GHz)

Pros:

Cons:

• Does not cover the entire GBT operating range.
• The Automated requirement may be occasionally foiled by icy weather.
• [Would require attention to maintain calibration and functionality and we already have dedicated engineers maintaining the astronomical receivers.] main.CarlaBeaudet: (We are already maintining calibration and functionality on both this system and the GBT RFI monitoring station.)

GBT RCO's

Operating range: By definition matched to GBT operating ranges

Pros:

• Operating range covers all of GBT operating range
• GBT Spectrometer is the back end
• IDL already has lots of tools for displaying the data; it may be simply a matter of tapping into what exists
• [Any other system would require constant attention to maintain calibration and functionality and we already have dedicated engineers maintaining the astronomical receivers.] main.CarlaBeaudet: (We are already maintining calibration and functionality on both the trailer RFI monitoring station and the GBT RFI monitoring station.)

Cons:

• Probably can't fulfill the "automated" requirement because:
  • RCO's are neither regular nor continuous
  • ?RCO's are a largely manual process? Is this true? Could a routine be written, or is the spectrum already stored in a way where frequency is sky frequency, or can it be derived? Could this be a semi-automated process?
• Can't point at terrestrial sources; most signals of interest would come in through the sidelobes. At the very least, a few degrees of telescope motion would be required to ensure no nulls.

Back Ends

Spectrum analyzer

GBT Spectrometer

We can probably gain a factor of ten or more in sensitivity over a spectrum analyzer with A/D sampling of the monitor antenna output and doing the FFT's in software. This is currently available only on the GBT RFI Monitoring Station. Since software FFT's are considerably slower than real time, this covers less spectrum in a given time, but the spectrum noise goes down as the square root of the integration time for as long as you want. 100 seconds per spectrum is probably reasonable. The spectral resolution is also as good as you need with 1 kHz being a reasonable value. The A/D in the computer at the GBT RFI Monitor Station output in the GBT equipment room allows us to do about 6 MHz at a time. We could increase this to about 12 MHz, if warranted, with a new A/D. The A/D - FFT approach also lends itself to the detection of short pulses, like DME, radar, power line noise, and switch closures, which a spectrum analyzer doesn't handle well. RickFisher

• May need to coordinate use with observations/maintenance, which doesn't fulfill the "automated" requirement. (Unless there's plenty of spectrometer to go around??? Can we be a fixed window in the dynamic schedule? What about the lab spectrometer?
• May need to schedule manually each week at coordination meeting (Carla could schedule, Wes could execute.)

Data Display

Requirements:

• Need to have displays that astronomers can access.

OASIS

Pros:

• Can create pretty 3D displays

Cons:

• Must be used with spectrum analyzer
• Web page would display graphics; actual data would be stored elsewhere, and new graphics created manually on a regular basis and posted.

IDL & Other astronomy display tools

• Web page needs only instruct astronomers about where to find the data; they already know how to use the display tools to extract the info they're after.

Cons:

Decisions

Two Options, outlined below, will be pursued in parallel:

GBT-Based RFI Monitoring

• Front end: GBT Receivers, Back end: GBT Spectrometer, Data Display: IDL (RonMaddalena has taken the lead on this.)
• GbtBasedRfiArchivalMonitoring

GBT RFI Monitoring Station

• Front end: GBTRFIMS, Back end: Spectrum Analyzer OR A/D & FFT (TBD), Data Display: Python Utility
• GbtRfiMsArchivalMonitoring

A Single Web Page, linked off the RFI page, will give access to the information. Carla has created a 1st draft of this page.

-- CarlaBeaudet - 27 Sep 2007

Attachment:	Action:	Size:	Date:	Who:	Comment:
monit_gbt_comparison.pdf	action	4030763	20 Nov 2007 - 16:17	CarlaBeaudet	Rick compares GBTMS and GBT sidelobe sensitivity