|
|
Thread Tools | Display Modes |
#31
|
|||
|
|||
"RD" == Rob Dekker writes:
RD I don't want to sound negative, but your plan is really not an RD alternative or even competition for the SKA. Other people already RD commented on many yet unresolved issues : RD [...] - Cost is not only in the receiving elements. Cost RD calculation would need to include the entire system. Indeed, there is considerable concern about the 12-m dishes proposed by the U.S. and Indian groups for exactly this reason. The 12-m dishes and associated electronics have costs that are similar to other proposals. There are some people, though, who worry that it will be impossible to process the data from an array composed of 12-m dishes because there won't be enough computing power, even in 2020. -- Lt. Lazio, HTML police | e-mail: No means no, stop rape. | http://patriot.net/%7Ejlazio/ sci.astro FAQ at http://sciastro.astronomy.net/sci.astro.html |
#32
|
|||
|
|||
Dr. H. Paul Shuch, executive director of the SETI League, sent a
response via email on this proposal which he has granted permission to post he ================================================== ========== Date : Tue, Feb 08 09:47 AM EST 2005 From : "Dr. H. Paul Shuch" To : Robert Clark Reply To : "Dr. H. Paul Shuch" Subject : Could DirecTV satellite dishes be used for the Square Kilometer Array - and a more radical proposal. Robert Clark wrote: Hello Dr. Shuch. The following was posted to sci.astro. Many thanks, Bob. This is an interesting thread, and I appreciate you sharing it with me. I can see two possible drawbacks to the proposal (and you have my permission to post my comments to the usnet group): (1) Fequency coverage. There is a practical limit to the frequency range over which a parabolic reflector can efficiently be used. The upper limit is set by surface smoothness (deviations from the parabolic curve should not exceed about 1/10 wavelength at the highest operating frequency). There is also a lower limit; the diameter of the dish should not be less than about 20 wavelengths at the lowest operating frequency. In the case of 0.5 meter diameter DBS dishes, this means the shortest practical operating wavelength is 2.5 cm, equating to a lowest practical operating frequency of 12 GHz. (I know, many hams use these dishes successfully at 10,368 MHz, but their efficiency is lower than you would expect, and the pattern not as sharp as it should be.) It is no coincidence that 1/2 meter diameter dishes are used for a satellite band that extends from 12 to 14 GHz! If used for SETI, we would lose access to most of the interesting microwave spectrum. (2) Spatial coverage. Antennas used for satellite TV reception spend their life pointed at communications satellites. Those satellites park in a specific arc (the Clarke, or geosynchronous orbit, belt). As viewed from Earth, at different latitudes, all antennas pointed at the Clarke belt subtend a narrow range of declinations (from about +10 to -10 degrees Dec.) So, even neglecting limited right ascension coverage (the subtended RA will vary as the Earth rotates on its axis), a network of DBS dishes doing parasitic SETI will miss about 90% of the sky. Now, we might get lucky, and find ETI lurking on or near the ecliptic. But out to maybe 1000 LY, stars are uniformly distributed in declination, with respect to Earth. And even more distant stars, such as those dense populations at the galactic centre, are at declinations very different from the Clarke satellite belt (the galactic centre is in Saggitarius, some -27 or so degrees in declination). Thus, unless we point those millions of dishes somewhere other than at the direct broadcast satellites, we'll probably miss ETI. Not to throw cold water on the subject; I would certainly encourage such a collaborative venture. Only, don't think it will solve all the SETI problems. Yours for SETI success, Paul -- H. Paul Shuch, Ph.D. Executive Director, The SETI League, Inc. URL http://www.setileague.org "We Know We're Not Alone!" ================================================== ========== Robert Clark wrote: A.) According to this article $1.4 billion dollars is earmarked to be spent on the Square Kilometer array with completion expected by 2015: Radio Astronomy Will Get a Boost With the Square Kilometer Array. http://www.universetoday.com/am/publ...a.html?1592004 As I stated below the primary extra components to be added to the DirecTV dishes to be used for radio astronomy are the feedhorn and the preamp. The feedhorn is simply a metal pipe of little cost. The SetiLeague site lists sellers of specially made preamps for SETI search at $150. But these are handmade. From the low number of circuit elements, I estimate that if mass produced at the millions of items level the price could come down in the range of $10 each. This site states the rate of growth of satellite TV subscribers is at 4 million per year: Satellite TV Basics. http://www.satellitetv-hq.com/hqguid...tv-basics.html This means the *receivers* for the SKA system if using these satellite dishes could be installed in 1 year at an extra cost of $10 x 4,000,000 = $40 million dollars. Note that the feedhorn and the preamp could be attached at manufacture would not cost more on installation labor. The installation is already paid for by the satellite TV subscribers. To pay for the extra cost for the added equipment you could simply add $1 extra per month to the subscriber rate. Then these 4 million, .5 meter wide dishes would have a total collecting area of a disk 2000 x .5 m = 1000m = 1km wide, the total collecting area expected for the SKA. Moreover this would have the advantage that an additional square kilometer of collecting area would automatically be added every year over several years going by the present growth rate. You could also attach the extra equipment to the 25 million satellite systems already installed in perhaps 4 or 5 years. The number of satellite TV subscribers worldwide was 60 million in 2003 and is expected to grow to 100 million by 2008: Digital Satellite TV Platforms Continue to Gain Subscribers, and Profits are on the Rise. http://www.instat.com/press.asp?ID=1171&sku=IN0401236MB If this many .5 meter antennas were networked together, they would have the collecting area of a single antenna 10,000 x .5 m = 5 km wide. Note that the idea of using over 50 million separate, stationary elements is one of the proposals being considered for the SKA architectu Aperture Array (AA) http://www.skatelescope.org/pages/design_nl.htm This method of keeping the receiving antennas fixed while detection directions are determined electronically is called the phased array approach and has the advantage that many separate targets can be observed simultaneously. It also has the advantage that interfering local signals can be suppressed. However, the Aperture Array has antennas close together in a predetermined configuration with the positions precisely determined. How could this work for the randomly positioned satellite dishes? Methods of differential GPS and carrier phase synthesis now have the capability of determining position to within millimeters. The method compares the GPS signal between a precisely known site and an unknown site to locate the unknown site to within centimeters. Then a comparison is made in the actual phase of the signals received at the two sites to locate the unknown site to within millimeters: CARRIER-PHASE TRACKING "Carrier-phase tracking provides for a more accurate range resolution due to the short wavelength (about 19 centimeters for L1 and 24 centimeters for L2) and the ability of a receiver to resolve the carrier phase down to about 2 millimeters. This technique has primary application to engineering, topographic, and geodetic surveying and may be employed with either static or kinematic surveys. There are several techniques that use the carrier phase to determine a station's position. These include static, rapid-static, kinematic, stop-and-go kinematic, pseudokinematic, and on-the-fly (OTF) kinematic/Table 8-4 lists these techniques and their required components, applications, and accuracies." http://cartome.org/FM3-34/Chapter8.htm This should be sufficient for keeping the signals for the millions of antennas in phase up to perhaps 3 cm wavelength, 10 Ghz frequency. Timing synchronization can be obtained by synchronizing from the common signal received by the dishes from the satellite. The Argus telescope at Ohio State University (this is different from Project Argus operated by The Seti League) may provide a model for how sensitive such a system can be operating from noisy populated areas: Newsgroups: sci.astro.seti From: Bob Dixon Date: Wed, 19 May 2004 12:49:02 -0400 Subject: The Argus Telescope http://groups-beta.google.com/group/...df1f29f629f20/ Argus Expands the Search For Life. By Daniel Sorid posted: 03:30 pm ET 09 June 2000 http://www.space.com/searchforlife/s...us_000609.html Note though that the continent wide satellite dish system will have an advantage over Argus in dismissing unwanted signals in that such signals would only be detected by a local group of antennas not the continent wide system. In mentioning an estimated price for this system, I emphasized the estimate was for the *receiving* part of the system. But of course for such a system of separate receivers, it is just as important to combine and process the signals. In the thread for DirecTV being used for SETI, someone mentioned you might need to transfer 1 Gbps from each antenna for detections at 12 Ghz. I seem to recall that analog signals can be tranferred in greater density than digital signals. Perhaps the signal received by each antenna can be transmitted in analog form with a stamp indicating its location and time of origin. For examples of the data density required we could look at some examples of systems of separate antennas that have been used to give combined signals in *real time*: Astronomers Demonstrate a Global Internet Telescope. Date Released: Friday, October 08, 2004. http://www.spaceref.com/news/viewpr.html?pid=15251 This produced data of 32 Mbits/second for each telescope for observations at 1.6 Ghz. So at 16 Ghz perhaps 320 Mbps might be expected for each antenna. The data was sent over a high-speed internet network available to universities that operates at gigabits per second. Within a few years, the data transfer rate is expected to reach tens of gigabits per second. And: Prototype SETI Antenna Array Will Help Radio Astronomers Too. Date Released: Wednesday, June 07, 2000. http://www.spaceref.com/news/viewpr.html?pid=1992 This is of the Argus telescope at Ohio State University. The 64 antennas here detect signals from 400 to 2000 Mhz. The antennas together produce 2.56 gigabytes per second, or 20.48 gigabits per second. So each telescope produces 320 Mbps. This article states that no physical connection could economically carry that much data over distance however this was written in 2000. Ultra wideband technology (mentioned below) now has that capability. For processing the data for the proposed SKA system, I expect the distributed computing system used by Set@Home to be used, wherein millions of computers take part in the calculations. As for how the data can be sent by the individual antennas, there are a few possible ways the signals could be combined. 1.)DirecTV offers a two-way broadband satellite internet service called DirecWay. This allows signals to be sent from the home antennas back up to the transmitting satellite. However, this system currently has only a 100,000 subscribers in place. I want to use the millions of subscribers using the satellite TV systems. I think a minor low-cost modification of the current TV antennas would also allow them to transmit to the satellites used for broadband internet service. (I don't think the satellites used for TV service can be used to receive signals.) 2.)Another possibility for transferring the data from each antenna might be to use military satellites currently used for surveillance on radio transmissions, perhaps using satellites that were decommissioned and are no longer used for sensitive military tasks. 3.)Possibly the techniques used with amateur packet radio could be used. Here radio links are used to setup data networks analogously to how the internet sets up data transfer networks using the TCP/IP protocols: N6GN's Microwave Link Page http://www.sonic.net/~n6gn/uwavelink/uwv.html INEXPENSIVE MULTI-MEGABAUD MICROWAVE DATA LINK http://www.sonic.net/~n6gn/hr89/uwvarticle.html 4.)Ultra wideband (UWB) promises gigabit data transfers over both cable and wireless connections and should be available this year (2005): New chipset promises gigabit broadband on cable and wireless. Rupert Goodwins ZDNet UK May 11, 2004, 15:20 GMT http://news.zdnet.co.uk/communicatio...9154271,00.htm Ultrawideband in 2005, but only in America Rupert Goodwins ZDNet UK February 19, 2004, 09:45 GMT http://news.zdnet.co.uk/communicatio...9146644,00.htm Ultrawideband: Wireless Whoopee. 08:34 AM Oct. 09, 2004 PT "SAN FRANCISCO -- Think of it as Wi-Fi on steroids. On its way to U.S. living rooms and maybe even automobiles is a new type of high-speed wireless connection that promises downloaded data rates of up to 1 gigabit per second -- roughly 18.5 times the speed of Wi-Fi -- to personal computers and other devices. "This ultrawideband technology, which could become available in the next two years, also allows the devices to send data upstream to a network at 480 megabits per second." http://www.wired.com/news/technology...w=wn_tophead_3 5.)Some public utilities now collect their meter readings from radio transmitters attached to their meters. The data is collected by receiver on utility poles and then transmitted to a central site. This method could be adapted to work for collecting the data from the separate antennas. 6.) The above methods would require that the data transmissions be on specified frequencies that will not be used for detections. However, another method might not have this limitation: Broadband Over Power Lines? 01:15 PM Feb. 09, 2003 PT "ST. LOUIS -- Coming to a home or office near you could be an electric Internet: high-speed Web access via ubiquitous power lines, of all things, making every electrical outlet an always-on Web connection." http://www.wired.com/news/technology...,57605,00.html This is a new technique already being tested in small markets to provide interent service over power lines. The speed of transmission can be ramped up to 1 gigabits per second using ulta wideband technology. B.)This last leads me to another proposal for large scale separated antennas for radio astronomy: using the electrical wiring in households as radio antennas. Here's a post to rec.radio.amateur.antenna discussing this: ================================================== ======================== From: Ed Hare, W1RFI ) Subject: ISO info about using house wiring as a TV antenna Newsgroups: rec.radio.amateur.antenna Date: 2000-12-29 15:33:06 PST Richard Friday wrote in message ... I know this post might be off-topic but could find no other newsgroup that had "antenna" in its name. I'd be most appreciative if someone could point to a more suitable discussion or other source of information. I've seen advertised a device that claims to allow you to use the electrical wiring of your house as a tv antenna. You plug this device into an outlet, and then use connections it provides as your tv antenna. I'm trying to find out if this actually works but have not been able to find any reviews. This device will receive some signals. However, house electrical wiring is not a very good VHF antenna system for a couple of reasons: First, it is very difficult to predict the direction that the house wiring will best receive from. It is quite likely that the antenna pattern will have all sorts of peaks and nulls, sort of as if you had a rotatable TV antenna that was pointing in several directions at once. This may not pick up much of the TV signal you want to pick up or may have multiple responses, resulting in ghosts. Also, an electrical power line can be a very noisy place. All sorts of electronic devices on the line, from power-line equipment itself to every motor or power supply plugged in near you may create noise that will interfere with the signal you want to receive. If you have no other antenna choice, that device may be useable, but I don't think it will work as well as a good set of "rabbit ears" on top of your TV. Ed Hare, W1RFI ================================================== ======================== The disadvantage of the electrical wiring going in several different directions may actually be an advantage in regards to a SETI search since you would want the detections to be omnidirectional. If there are 100,000,000 homes which average 10 meters across then this would result in a collecting area of 10,000 x 10 meters = 100 km across. Since you would want to include large commercial establishments, the size would actually be larger than this. Bob Clark wrote in message roups.com... I was interested to read this on the Seti League web site: __________________________________________________ ________ Parasitic SETI Dear Dr. SETI: As a Satellite dish owner and a strong interest in SETI, I was wondering if anything is available to allow the home satellite dish owner to 'search' when he is not watching TV. I do a bit of programing and would love to help make it so home dish owners could do this. Is it possible? What would it take? Does the dish have to follow a spot or can it sweep the sky from a fixed position? If this is possible it could add a million listeners to the system. Bill T. The Doctor Responds: Absolutely, Bill! Parasitic SETI with a home satellite TV dish is not only feasilble, it's widely practiced. A second feedhorn and preamp assembly are mounted next to the C-band horn/LNB at the apex of the dish (see Figure 2 of this article). This assembly feeds the rest of a SETI system (see our online Tech Manual). You can then sweep out the sky, as described here. And yes, a million participants would be nice, but our goal is a more modest 5000 stations. __________________________________________________ ________ http://www.setileague.org/askdr/parasite.htm I believe they are referring to the 6 ft. backyard type antennas, judging from the linked images on the page. But could the roof mounted DirectTV and Dish Network type antennas be used for SETI? The mentioned extra equipment are an extra feedhorn and a preamplifier. The feedhorn can made cheaply but the preamp seems expensive. If these preamps were mass produced for this purpose could their per item cost be brought under $50? I'm envisionig a government agency such as NSF, or a scientically interested billionaire, paying satellite TV companies to attach this extra equipment to their satellite dishes. Say $100 million is earmarked for the program. Then you would want the extra cost to be under $100 for each dish for say 1,000,000 subscribers. Judging from the diagram in the online Tech Manual linked to on the page, the other equipment should be doable by the equipment that comes with the satellite TV system. Computer processing would be done separately at a central location. If you had a 1,000,000 of these .5 meter wide antennas it would have the detection sensitivy of a single antenna 500 meters wide. Bob Clark |
#33
|
|||
|
|||
"Joseph Lazio" wrote in message ... "RD" == Rob Dekker writes: RD I don't want to sound negative, but your plan is really not an RD alternative or even competition for the SKA. Other people already RD commented on many yet unresolved issues : RD [...] - Cost is not only in the receiving elements. Cost RD calculation would need to include the entire system. Indeed, there is considerable concern about the 12-m dishes proposed by the U.S. and Indian groups for exactly this reason. The 12-m dishes and associated electronics have costs that are similar to other proposals. There are some people, though, who worry that it will be impossible to process the data from an array composed of 12-m dishes because there won't be enough computing power, even in 2020. Mmm. That would surprise me (that there would not be enough computing power). The amount of computing power to create (compute) a phased single beam from an array roughly linear with the amount of elements to be 'phase-correlated'. The ATA is planned to have 350 elements (of 6m dish). The SKA would plan to have 3300 elements (of 12m dishes) . So that is only a factor of 10 more elements than the ATA. In terms of Moore's law, a factor 10 in computing power is achieved in roughly 2.5 years. This means that if SKA is completed 2 1/2 years after ATA, the system w.r.t. computing power would cost the same as the same system on a full operational ATA. ATA is planned to have cost about $30-40M upon competion, so I don't see that the computing power available for the SKA would be a limiting factor (either technology or cost). Rob -- Lt. Lazio, HTML police | e-mail: No means no, stop rape. | http://patriot.net/%7Ejlazio/ sci.astro FAQ at http://sciastro.astronomy.net/sci.astro.html |
#34
|
|||
|
|||
In article ,
Rob Dekker wrote: So that is only a factor of 10 more elements than the ATA. But a factor of 100 more baselines. |
#35
|
|||
|
|||
Rob Dekker wrote: "Joseph Lazio" wrote in message ... "RD" == Rob Dekker writes: RD I don't want to sound negative, but your plan is really not an RD alternative or even competition for the SKA. Other people already RD commented on many yet unresolved issues : RD [...] - Cost is not only in the receiving elements. Cost RD calculation would need to include the entire system. Indeed, there is considerable concern about the 12-m dishes proposed by the U.S. and Indian groups for exactly this reason. The 12-m dishes and associated electronics have costs that are similar to other proposals. There are some people, though, who worry that it will be impossible to process the data from an array composed of 12-m dishes because there won't be enough computing power, even in 2020. Mmm. That would surprise me (that there would not be enough computing power). The amount of computing power to create (compute) a phased single beam from an array roughly linear with the amount of elements to be 'phase-correlated'. The ATA is planned to have 350 elements (of 6m dish). The SKA would plan to have 3300 elements (of 12m dishes) . So that is only a factor of 10 more elements than the ATA. In terms of Moore's law, a factor 10 in computing power is achieved in roughly 2.5 years. This means that if SKA is completed 2 1/2 years after ATA, the system w.r.t. computing power would cost the same as the same system on a full operational ATA. ATA is planned to have cost about $30-40M upon competion, so I don't see that the computing power available for the SKA would be a limiting factor (either technology or cost). Rob Also one of the planned architectures would use 50 million separate antennas: Aperture Array (AA) "One of the options of SKA telescope is to use phased arrays with over 50 million receiving elements with a mixed RF/digital adaptive beamformer." http://www.skatelescope.org/pages/design_nl.htm The phrase "mixed RF/digital adaptive beamformer" leads me to believe some of the correlation is to be done using the analog signals from each antenna. This is made simpler with the antennas in a central site located right next to eah other. Also, the number of PC's signed up for Seti@Home is now in the millions. The number of PC's that exist now exceeds 1 billion with 1.5 billion expected by 2010. One could make an argument that the SKA will be an important part of planetary security since one of it's purposes is to seek NEO's and potential Earth impactors. Then one could say "induce" most of these PC owners to install Seti@Home-like software to run in the background on these PC's. Bob Clark |
#36
|
|||
|
|||
"RD" == Rob Dekker writes:
RD I don't want to sound negative, but your plan is really not an RD alternative or even competition for the SKA. Other people already RD commented on many yet unresolved issues : RD [...] - Cost is not only in the receiving elements. Cost RD calculation would need to include the entire system. In article , Joseph Lazio writes: Indeed, there is considerable concern about the 12-m dishes proposed by the U.S. and Indian groups for exactly this reason. The 12-m dishes and associated electronics have costs that are similar to other proposals. There are some people, though, who worry that it will be impossible to process the data from an array composed of 12-m dishes because there won't be enough computing power, even in 2020. Joe is an expert and can correct me if I have something wrong here, but I think it might be useful to point out some of the qualitative issues. For an interferometer, the point-source sensitivity depends only on collecting area, not on the number of antennas. That would favor large numbers of small antennas, but... well, there's always a "but," isn't there? First of all, in order to calibrate the interferometer, you have to detect bright point sources ("calibrators") _on every baseline_. That means with every pair of antennas. And you have to do it faster than the atmospheric phase changes, which is typically tens of minutes, depending on frequency. This sets a minimum size for your antennas, typically several meters depending on operating frequency. Second, receivers aren't free, and every antenna has to have one. For a major facility such as the SKA, you want to work at multiple frequencies, so every antenna needs several receivers. I haven't priced receivers lately, but I'd guess prices in the 10's to 100's of thousand dollars might be in the ballpark. This pushes you to larger antennas, but larger antennas cost more per unit collecting area than smaller ones. One estimate I've seen is that the antenna cost goes as the 2.5 power of diameter. This isn't a law of nature, but the practical number won't be too far from that. If the receiver price is A, and the antenna price is B*d^2.5, the price per unit area goes like A/d^2 + B*d^0.5 This function rises at large and small values of d and has a minimum at some intermediate value. That cost minimum is the size you want to make your antennas, independent of how many of them you buy. Of course receiver performance and cost aren't fixed numbers. You could buy cheap receivers, accept poor performance, and make up for it by buying lots more antennas and receivers. This is another optimization problem, but in all the studies I've seen, it turns out best to pay more for good receivers. Observing time goes as the square of system temperature, and having crappy receivers makes the calibrator problem worse. A third issue is the one RD and Joe mention: correlator cost. Cost ought to go roughly linearly with number of baselines, which for N antennas is N*(N-1)/2. This pushes towards fewer but larger antennas. A fourth issue is primary beam size. The size of the field of view you can map at one time with a single receiver on each antenna goes inversely as the size of the antennas. This favors small antennas and more of them. There are also technical issues such as sidelobes and shielding. You don't want any part of the antenna beam to "see" the ground. I am no expert but would expect these issues to favor large antennas or at least ones larger than a few meters in size. As you can see, the tradeoffs are complicated. I expect the interferometer designers know about cheap, mass-produced antennas and will use them if they represent the best solution. It seems to me unlikely that they do. -- Steve Willner Phone 617-495-7123 Cambridge, MA 02138 USA (Please email your reply if you want to be sure I see it; include a valid Reply-To address to receive an acknowledgement. Commercial email may be sent to your ISP.) |
#37
|
|||
|
|||
Thanks for a good overview of trade-offs for phased arrays Steve !
"Steve Willner" wrote in message ... [....] Joe is an expert and can correct me if I have something wrong here, but I think it might be useful to point out some of the qualitative issues. [....] A third issue is the one RD and Joe mention: correlator cost. Cost ought to go roughly linearly with number of baselines, which for N antennas is N*(N-1)/2. This pushes towards fewer but larger antennas. I'm not a total expert in this field, but know enough to be dangerous.. If I'm not mistaken, to phase correlate N elements, you do not need to phase-correlate all N^2 baselines, but (after time/phase-adjusting each of the individual N signals to intended beam direction), you can simply create a correlator tree of 2-1 phase correlators. The tree would then consist of N correlators, which counts for linear complexity of computation power. I could be off with this, but I seem to remember this from college (long ago). Question for Joe : Which part of the SKA system would be the bottleneck in computational effort, and create a problem even in 2020 ? Rob |
#38
|
|||
|
|||
|
#39
|
|||
|
|||
Matt Giwer wrote: wrote: I was interested to read this on the Seti League web site: __________________________________________________ ________ Parasitic SETI Dear Dr. SETI: As a Satellite dish owner and a strong interest in SETI, I was wondering if anything is available to allow the home satellite dish owner to 'search' when he is not watching TV. I do a bit of programing and would love to help make it so home dish owners could do this. Is it possible? What would it take? Does the dish have to follow a spot or can it sweep the sky from a fixed position? If this is possible it could add a million listeners to the system. I talked about this months ago. Forming a beam, antenna gain, requires knowledge of the phase angle differences between the sensors. The bottom line is you cannot aggregate dishes without both 1) relative location of all dishes to each other to a fraction of a wavelength 1a) data transmission phase lag = 0 after compensation 2) tolerances in the receiver phase being to equally high tolerances It is feasable if the combination of 1 and 2 do not result in more than about 1/36th wavelength error. The more it deviates from that the less useful. 1/36th would cover about a 10 degree circle in the sky. And that would indicate the array gain, roughly log(180^2/10^2). At 1/4 wavelength error it is log(180^2/90^2) and a 1/2 wavelength error it is an antenna gain of log(1). Maybe when the European GPS goes up and if it can be used seamlessly with the present US system it might be possible to get accuracies of a few inches which is way to great. But even if perfect sending the date to be aggregated depends upon the delays along the way from sender to inches in difference in the cable and fiber lines being known to the fraction of a wavelength. If installers do not have the exact measure of the length of cable and fiber used it all goes to ****. And then there is the impossible job of calibrating them all. -- Security at presidential press conferences is so good a man can use a fake name and pass the background check. Or are there no background checks? -- The Iron Webmaster, 3380 Methods of fixing your position within millimeters using GPS exist as long as you can make a comparison to the GPS signals received at a site whose position is known within millimeters. This precisely known site can be kilometers away. This is used in surveying for example: CARRIER-PHASE TRACKING "Carrier-phase tracking provides for a more accurate range resolution due to the short wavelength (about 19 centimeters for L1 and 24 centimeters for L2) and the ability of a receiver to resolve the carrier phase down to about 2 millimeters. This technique has primary application to engineering, topographic, and geodetic surveying and may be employed with either static or kinematic surveys. There are several techniques that use the carrier phase to determine a station's position. These include static, rapid-static, kinematic, stop-and-go kinematic, pseudokinematic, and on-the-fly (OTF) kinematic/Table 8-4 lists these techniques and their required components, applications, and accuracies." http://cartome.org/FM3-34/Chapter8.htm This would suffice for centimeter wavelengths. Bob Clark |
#40
|
|||
|
|||
"RD" == Rob Dekker writes:
RD Thanks for a good overview of trade-offs for phased arrays Steve ! RD "Steve Willner" wrote in message RD ... [....] Joe is an expert and can correct me if I have something wrong here, but I think it might be useful to point out some of the qualitative issues. RD [....] Steve's post certainly captured most of the essential details. In particular, the US proposal for the SKA involves 12-m dishes because, with current knowledge, these are thought to be at the minimum in the "cost curve." That is, a 12-m dish provides a good balance between the costs of obtaining antennas and obtaining receivers. A third issue is the one RD and Joe mention: correlator cost. Cost ought to go roughly linearly with number of baselines, which for N antennas is N*(N-1)/2. This pushes towards fewer but larger antennas. RD I'm not a total expert in this field, but know enough to be RD dangerous.. RD If I'm not mistaken, to phase correlate N elements, you do not RD need to phase-correlate all N^2 baselines, but (after RD time/phase-adjusting each of the individual N signals to intended RD beam direction), you can simply create a correlator tree of 2-1 RD phase correlators. The tree would then consist of N correlators, RD which counts for linear complexity of computation power. No, one does have to multiply the signals from all N antennas together, which is roughly an N^2 problem. The issue in radio astronomical interferometry is that the "visibility" or "correlation coefficient" is a measure of the Fourier transform of the sky brightness. Every unique pair of antennas gives one "visibility," so one wants to obtain all possible visibilities from the N antennas. RD Question for Joe : Which part of the SKA system would be the RD bottleneck in computational effort, and create a problem even in RD 2020 ? Actually, the correlator, as difficult as it is, may not be the limiting factor for the SKA computation. Some people think they know today how to build an SKA-level correlator. The most prominent concern has been how does one make an image? The SKA is envisioned as having a large field of view (e.g., 1 deg^2 at 1 GHz, and some people are concerned that that's not large enough). In order to make high-quality images, one probably has to image that entire field of view with something approaching 0.1 arcsecond resolution. Thus, a single image would be something like 10,000 x 10,000 pixels in size, and that's only for a single frequency. One probably has to utilize multiple frequencies simultaneously, in order to defeat some other effects. Moreover, current processing of radio interferometric images is an iterative process, in which one makes an image, corrects it for certain effects, makes a new image, improves the corrections, makes a new image, .... There's a memo floating around discussing all of these aspects of processing. IIRC, the conclusion was that the processing power required to make high-dynamic range images scaled as the dish diameter (which sets the field of view) as something like D^6 or D^8. Thus, all other things being equal, just changing from the 25-m VLA dishes to the proposed 12-m US SKA dishes would increase computation by a factor of 64 to 256, and that's even without taking into account the much larger number of antennas. I should say that not everybody agrees with this analysis, but it is recognized that the computational aspects of the SKA may be significant. -- Lt. Lazio, HTML police | e-mail: No means no, stop rape. | http://patriot.net/%7Ejlazio/ sci.astro FAQ at http://sciastro.astronomy.net/sci.astro.html |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Maybe you sci.astro.amateur and sci.astro readers can explain this | Sam Wormley | Astronomy Misc | 16 | July 2nd 04 10:17 PM |
Satellite Tracking | Pete | UK Astronomy | 0 | June 21st 04 09:11 PM |
Type I supernovae due to planetary impacts? | Robert Clark | Astronomy Misc | 2 | January 20th 04 07:59 AM |
UK Will Build First Satellite To Study Wind From Space | Ron Baalke | Misc | 0 | November 20th 03 04:05 PM |
Successful Launch for Boeing-Built Galaxy XIII/Horizons-1 Satellite | Gene Nygaard | Policy | 0 | October 6th 03 05:24 PM |