Another Interesting Observation from ADS-B Watching

I used to think, perhaps like most people, that most commercial flights are flown to get from point A to point B. Sometimes, it might be a milk run, but that just means that it goes from point A to B to C to D. To that point, I’ve had the admittedly curious experience of flying a Southwest Airlines flight from Burbank (BUR) to El Paso (ELP) via Las Vegas (LAS) and Albuquerque (ABQ). It was a long day and a lot of take-offs and landings. (and I also got to see what I think was a Super Guppy at ELP!)

Obviously, more or less straight-line routes are predominant, following the invisible but well-marked flightways across the US, North America, and the rest of the globe. Just as obviously, I’ve been on a/c flights which had to make significant course adjustments due to weather, whether it was hurricanes, mesoscale thunderstorms, or extremely fast jet streams. These are all reasonable things to do.

In my rabidness with this new ADS-B monitoring, I discovered last week a route I’d never seen a commercial a/c take. A major US freight carrier’s flight entered my ADS-B monitoring space about 150 miles NNE of me, then executed a dogleg around the west side of the greater Phoenix area, then departed my coverage headed to the ESE toward El Paso. The only reason I even noticed it was that it suddenly popped up on my Virtual Radar Server (VRS) display and it was fascinating in the grand arc it traced around Arizona, yet never touching the ground.

This carrier apparently does this regularly, and someone on-line called it a “sweeper” route. The flight is scheduled to depart Denver (KDEN) and arrive Memphis (KMEM), a straight-line distance of only 872 miles. However, it generally takes a circuitous route, which on the day I saw it, was 2108 miles. Here’s a graphic of the same flight, even more extreme, with the route distance well over 2450 miles. The bit that’s in red is the portion that I can hear with my ADS-B receiver.

Interesting “question-mark” route across US

The purpose for the flight becomes apparent when it was called a sweeper. Just in case extra packages show up at the last minute at PHX, ELP, DFW or somewhere else along the route I suppose this flight can divert and grab those. If another plane has mechanical problems at one of the airports along this route, this flight can stop and grab the freight. If your mission is to guarantee delivery next day of someone’s letter or package, it costs you a lot in money and reputation to blame it on mechanical problems or a bin being already full.

Pretty interesting things to be learned with this ADS-B tracking. In the meantime, I’ve set up an alert at FlightAware for when this flight gets scheduled and I received my first alert just a little while ago.

ADS-B Coverage and Site Performance Update

Today my ADS-B receive site here at the house has moved into position 291 on the FlightAware Leader Board! Just a few days ago it was at around 350.

There’s only so many a/c that can overfly my general coverage area – Scottsdale Arizona is not Amsterdam, London or Paris. Heck, it’s not even the greater NYC area with its multiple major airports and sheer number of flights. In other words, there’s only so much improvement to be had. The site will never even crack the top 100.

In my search for high-ranking sites in the general US west I found the one which I think will be the one to chase. The site is near to the top or atop Abajo Peak (elev 11,365′) in far southeastern Utah. In addition to being a whopping big mountain, Abajo is one of several major laccoliths in the Four Corners / Colorado Plateau region and the tallest thing for 44 miles. There’s even a live webcam atop Abajo.

From atop Abajo, the visible horizon is below 0 degrees in nearly every direction, with the 44-miles-away Mounts Mellenthin and Peale the only pieces of earth that just barely break 0° (0 degrees, or horizontal). Check out Abajo Peak at HeyWhatsThat.

From atop Abajo, a good receiver will capture a fair amount of the commercial traffic crossing over the western US. An aircraft at 40k ft can be over 200 miles from Abajo and still be above the zero-degree horizon. Since the actual horizon at Abajo averages about -0.5 degrees, this adds another 40 miles or so of crows-flight range. And that’s still not including atmospheric refraction and the occasional edge diffraction. Likely that the Abajo Peak site can hear out to 250 mi / 400 km for an a/c at 40k ft.

While a high site is a great thing, the potential number of a/c is limited by population center locations, flight routes, number of flights, number of non-commercial airports, etc. Abajo doesn’t have a major US hub airport next door. Flights headed from Denver to nearly anywhere in the West will pass through, as will flights from ORD to SoCal airports, PHX, and LAS. Flights from DFW headed west to any locations above about Los Angeles will as well, but not the great majority of flights to SoCal, which is a lot of airports. Pretty much any intercontinental from the NE US to the Southwest will intersect that area. There are a bunch of small civil airports in the footprint, as well.

Let’s compare Abajo’s stats to the Scottsdale site’s stats for the past two weeks (I know, that’s not much of a sample set, but that’s all I have).

8-Feb-15 7-Feb-15 6-Feb-15 5-Feb-15 4-Feb-15 3-Feb-15 2-Feb-15 1-Feb-15 31-Jan-15 30-Jan-15 29-Jan-15 28-Jan-15 27-Jan-15 26-Jan-15
Sun Sat Fri Thu Wed Tue Mon Sun Sat Fri Thu Wed Tue Mon
Abajo position reports 189,086 215,723 251,825 259,477 245,641 250,943 242,716 203,284 232,858 269,162 269,226 229,558 224,085 233,383
Scottsdale position reports 206,155 223,556 269,480 268,903 243,494 239,195 223,117 133,820 175,748 158,466 241,074 234,657 203,885 209,868
Difference (%) 8% 4% 7% 4% -1% -5% -9% -52% -32% -70% -12% 2% -10% -11%
Abajo a/c count 1,910 1,939 2,211 2,218 2,021 2,103 2,100 1,762 2,013 2,277 2,240 1,928 1,843 1,964
Scottsdale a/c count 1,491 1,594 1,760 1,747 1,639 1,734 2,063 1,420 1,590 1,624 1,792 1,540 1,475 1,535
Difference (%) -28% -22% -26% -27% -23% -21% -2% -24% -27% -40% -25% -25% -25% -28%

What the above two weeks of data appears to show is that Scottsdale gets somewhere on the order of the same amount of position reports per day, but that Abajo sees a whole lot more a/c per day.

The Scottsdale best-case 30k ft+ range is to the ESE at just over 240 miles. (It may be just as good to the S, but there’s very few a/c flying down there due to the US/MX border.) However, I’d say the average 30k ft+ range is more like 150-160 mi due to the local hills and more distant mountains. In any case, that’s pretty amazing given that the antenna is 15′ off the ground and in a valley. A little of the performance is likely due to the desert climate and general lack of dense vegetation on exposed ridgelines.

I’ll keep working on improving the setup here at Scottsdale and see how the statistics line up with Abajo’s.

City-Wide (sort of) Wi-Fi Service

I spent much of last year in the boroughs of NYC, and a bit up the Hudson and Bronx river valleys, as well as some time along the rails in CT up to New Haven. What I started to notice was a great number of Wi-Fi access points called “CableWiFi” or similar. I also discovered that, in the right areas with enough signal strength, I could associate to these and get a login screen which asked me if I was a cable service subscriber for any of a number of cable ISPs. One of these was Cox, and as we’re a Cox subscriber here in Arizona, I eventually figured out which username/password to enter into the authentication page and BOOM I was on the Internet. Speeds were often (if not always) better than what I could get on the hotel Wi-Fi.

A number of large cable ISPs have gotten together and are rolling out in volume Wi-Fi access points in major metro areas. In the NYC greater-metro-tri-state area, there are apparently well over 250k APs now available. You can recognize them certainly by seeing SSIDs that are “CableWiFi” or similar, but also the physical devices hanging along the distribution cable on phone poles. As far as I can tell, it’s only in areas where there is above-ground ISP coaxial cable hanging on phone poles.

I grabbed a few shots of one hanging on the line near the house. Sorry about the poor resolution, all I have is my iPhone to snap pictures with. The boxes are Cisco APs, hardened outdoor devices in cast metal cases, with three antenna bumps on the base of the case. Here’s a few pictures.

General View, Cable Provider Wi-Fi AP (circled) hanging from Cable Messenger
Close-up View of the 3 Antennas on the Base of the Enclosure (note the F connector on the side coax)
Another View
Another View
Detail of Cable AP Wiring

This last photo gives some insight into how the Cisco AP is connected into the cable. Here, at least two coupler/splitters (left boxes) are visible, with the AP to the right. There are also 3 distribution coaxes visible, one coming from beneath the ground (the one exiting conduit on pole). There has to be power supplied as well for the AP to work, I don’t know the details but there’s likely a constant source of DC or AC impressed on the coax. The smaller grey coupler (most left) has a connection directly to the Cisco AP, and supplies both the backhaul connection and power via the thin coaxial jumper cable. If I had a camera with a real optical zoom lens, I could read the labels and perhaps decipher a bit more about the setup.

The service is pretty good. I have a Win 7 PC in the truck which has one of the Amped Wireless UA230A dual-band Wi-Fi transceivers installed, and the YAWCam application pulling images every 15 seconds from the Mobius ActionCam attached to the windshield. When the PC detects the Cox/CableWiFi connection, it associates and then can push an image or more to my main website page.

I’m slowly collecting position information data for CoxWiFi / CableWiFi APs in the greater PHX area, and will publish some maps soon. Until then, it’s possible to get that kind of information from Cox’s website.


Two Weeks of ADS-B Coverage

It’s been over two weeks since I installed my ADS-B setup, and it’s been a real performer and seems to be very reliable. As I’ve pointed out, I’ve continued to tweak the setup, first with a better antenna, then by raising the antenna a bit, then with a cheap satellite preamp.

An important thing I still don’t know is the absolute amount of air traffic on any one day. ADS-B transmitters are only on a limited number of civil a/c, so it doesn’t see all the non-equipped a/c that might be flying on that particular day. I’ve never looked in any detail at commercial traffic repeatability, so I can’t say for sure if flight XYZ123 between airport A and airport B flies  every day of the week. I don’t know if Monday is a busier flight day locally than is Tuesday. And with so many non-commercial airports within hearing range (KSDL, KDVT, KGEU, KFFZ, KCHD, KPAN, etc.) the number of small a/c equipped with ADS-B could change dramatically on any given day. KDVT has an especially large number of ADS-B-equipped a/c and they’re constantly flying loops around KDVT, so they provide a good number of position reports.

With all that said, when I saw the precipitous drop (34%) in location reports on the rainy day here on the 30th, I was concerned that my setup might have some issues that needed resolution. After doing a few checks, it appeared that things were nominal, so the next thing was to compare my station’s statistics with those of other stations near me. There aren’t any stations within 7 miles, but that was good enough to start as the rain was a southern-AZ-wide event, a good soaking, steady winter rain, which was over a very large area.

I took the best performing stations and a couple of typical stations for comparison. I also marked when I changed something in my setup to see if what I did actually mattered or if everyone else reflected some change that day which might indicate that my equipment changes had no particular effect.

Jon Adams ADS-B station vs other local ADS-B stations

The above graph (done in Excel) compares the JonAdams ADS-B receive site against two other solid performers (SN and DR) and two others down in the typical performance range (LY and PE). Raising the antenna (just noticed I drew it one day later than it should be) improved the performance from about 130k reports to about 150k reports. Adding the preamp was the real improvement, improving sensitivity and increasing the number of reports between 25 and 35%. I also noticed that my ultimate range improved as well. At least the preamp change to my station’s setup doesn’t correspond with significant changes in the other stations, so I’d say that the preamp was a big deal.

Another important thing, and why I started this article, was my concern over the rain having a specific deleterious effect on my station. But, it appears that the other high performers suffered very similar degradation. In my case, the drop was 34%. Comparing that to station SN, the drop there was also 34%. For station DR, the drop was 63%. For the other two typical stations, there was no particularly significant change in performance.

So, for now I will conclude that my ADS-B station has no unique problems not faced by other high performing stations. I remain interested in the cause of the drop, but have never had so much data at this particular frequency range that it’s going to take a little more Googling to figure out why and to what extent the rain should have an impact.

ADS-B Coverage Range Improvements


1090 MHz ADS-B Coverage from N7UV QTH

The above is a screenshot of the aggregated coverage over an approximate 120 hour period from last Saturday to now. The rings are spaced 50 km apart. As you can see, there are even a few a/c locations over 400 km away, which is incredible given that the house is in a bowl surrounded by hills or mountains in nearly all directions. For that one particular coverage radial SE into Sonora, that happens to be a very narrow clear shot (not including trees) from the house toward the horizon.

Initially, my setup included a simple discone antenna already sitting up on the roof, but in a day I had built the “QST Special” ADS-B antenna and my number of location messages went up from about 35k per day to over 130k per day.

After a few days of that, I raised the QST antenna by 4′ to get it above the parapet wall that surrounds the roof and the number of position reports went up by about 10-15% to about 150k. Next, I found in the garage a cheap (cheap being $5) cable satellite preamp and put that in line with the antenna. The position reports skyrocketed from about 150k to well over 200k per day. In fact, yesterday’s summary shows 240k+ position reports, and about 300 per second at peak times of the day. Here’s a chart which shows the changes in performance as a function of time and setup.

ADS-B Stats from 15 Jan to 2100 hours 30 Jan

As can be seen, the performance continues to improve as the setup is further tweaked. Today (30 Jan) looks to be a bad day for reception. It has been raining lightly (about 0.03″/hr) since before midnight. I expect that at the current rate the station will receive only about 160k position reports today, which will be about 2/3 of yesterday’s performance.

My assumption right now is that the connected rain drops on the antenna radome are causing significant loss of signal at the antenna radome (the ABS pipe it’s covered with). It’s also possible that the added attenuation of the signals as they travel through the rain clouds and falling rain is a factor. Not sure which is more significant. At 1090 MHz, light rain shouldn’t be a really significant issue, but perhaps the PiAware software decoder is running right on the edge of performance so even a dB or so of added loss could have an impact.

While the rain is great, I look forward to a few weeks of dry skies so that I can get a solid number on the performance as a function of day of week.


OMG – ADS-B & RaspPI

Enough about non-wireless stuff for the time being. Today, after a few weeks of inaction, I set up my Raspberry Pi B+, the FlightAware PiAware ADS-B code image, and deployed my first ADS-B receive site! It was easier than cake. It was stupid simple. In fact, the most challenging thing was discovering I didn’t have a microSD card and so I had to drive over to Fry’s and grab one.

What’s ADS-B? Why, it’s “Automatic Dependent Surveillance – Broadcast”. many (and more every month) aircraft, private and commercial, have GPS and a 1080 MHz transmitter on board to send a position message every second. For air traffic controllers, this means much improved position/speed/heading/altitude resolution. For the rest of us, it’s like having our own window at the air traffic control center and watching blips on the radar moving at 400 – 500 mph.

More soon as I get things going.

Venus and Mercury

After a number of days of clouds and rain here, I walked out of the grocery store and was impressed to see two bright planets low in the western sky. One was obviously Venus, the other was a little orangey and initially I thought it might be Mars. However, a quick check on the Internet shows it to be Mercury!

Both planets stand out beautifully in the twilight sky, and with Mercury preceding its brighter sibling toward the western horizon. I pointed the N7UVCAM2 Roving Reporter toward the pair, and captured a few images. I will post them here as soon as I figure out what’s going on with WordPress.

What Good Is a Wireless IP Camera?

I recently installed a new weather station here at the homestead. With a weather station, it’s always nice to be able to have a picture of the sky to see what the sky looks like, to augment and bring to life the sparse yet functional gauges and dials.

The world of digital image capturing devices is a miasma of terminology, mis-terminology, ignorance and sometimes (at least so it seems) a bit of disingenousity. Go to eBay, type in “IP network camera outdoor”, and there’s at least a zillion, give or take, cameras available from no end of sources, mostly in Asia.

Not being really up on the latest and greatest for IP network cameras, all I really knew going in was what I wanted. The camera needed to be 1) able to withstand living outdoors, in the direct sun and rain; 2) at least 1080p vertical resolution; 3) as sensitive as practical, so I could see an image at night, and hopefully see a few stars; 4) cheap (hard to define); 5) have an included webserver so that with a browser it’d be possible to see the image, and be able to ftp an image on a scheduled (not event) basis to the WUnderground site where my Personal Weather Station data is displayed; and 6) straightforward to configure and tweak. I know, it’s a lot to ask, especially with 4).

One thing that is missing in the above list, and some might wonder why, is the choice of wired vs wireless for the connectivity. After all, it’s the “WirelessJon” website! For this particular installation, it really wasn’t much of a decision to make. A camera takes a significant amount of power to operate (several watts at least). A wireless camera takes at least as much power to operate as an equivalent wired camera. Since a wireless camera would need a source of power, likely the house mains, a cable would still have to be run to supply the power to the camera. If I’m going to the trouble to run wire, I might as well kill two birds with one stone and use that wire to supply connectivity as well. Also, my desired camera might be streaming video on occasion, and that could end up being a significant load on my home wireless network. Finally, ensuring that the wireless coverage from the inside of the house would be good enough for robust connectivity to gadgets on the roof would mean that there’d be a constant unknown as to whether or not a problem in viewing the camera’s image was caused by RF propagation or by some hardware issue. So, like I said, the decision to go wired was pretty straightforward. I have another reason which made it simpler to go this way, which I’ll discuss in a future post on rooftop routers.

As can be imagined, many of the requirements listed above are mutually incompatible. And by “cheap”, I was aiming for somethiing under $100 delivered. Nonetheless, I started my search through Google and eBay as well, working to figure out some of the more esoteric terms (RTSP, DNC, AE, etc.) and attempting to validate things like FTP server and web browser functionality and configurability. A simple thing that I wanted was a scheduled capture on a regular, timed basis – something that doesn’t seem to be offered in most cameras listed on eBay – most offered appear to target the surveillance video market, and the cameras trigger video based upon events like movement in the field of view, or whatever. There were some challenges to figure out what many of the cameras actually did.

Another minor tricky thing was camera sensitivity to light; remember in my wish list I wanted a camera that could even see stars (not just the Sun). I’ve always had in my mind something that I call the PoleCam, which is a camera trained at the north celestial pole and capable of watching the current pole star, Polaris, in its daily orbit around the north pole, Yep, it’s true, Polaris isn’t really at the celestial pole. It’s close (about 3/4 of a degree away, close enough for government work), but it’s not stationary. The ideal camera would allow me to see the movement of Polaris as well. But, I digress: the main purpose of this camera is weather observation.

Cameras come with two different types of image sensor, CCD or CMOS, and a variety of lenses with different angular fields of view (FOV). Some cameras have varifocal lenses, where the “zoom” factor of the image can be changed to accommodate. For me, CCD is the only way to go for improving low-light sensitivity; compare a CCD imager with its equivalent CMOS imager and you’re sure to find that the CCD imager provides a higher quality image with better low-light performance. As far as lenses, it seemed that there were generally 4 mm, 8, mm and 12 mm used on the cameras for sale. Since I’d never bought one of these cameras before, I didn’t know how difficult it’d be to change the lens, so I decided to go with an 8 mm lens which should have about a 40 degree horizontal FOV.

A tricky thing I discovered (I know, I was born yesterday) was that some of the eBay ads had confusing, conflicting, or flat incorrect information about the product offered. Doing searches which included “CCD” didn’t always return only cameras with CCD imagers – in fact, some sites would use the term “CCD” in several places when selling a CMOS camera. Also, attempting to figure out if a camera had a webserver built-in, and whether that webserver could respond successfully to queries that came in on Chrome, or Firefox, or whatever, not just IE, is quite challenging.

Suffice to say, I did find a camera which appeared, at least in theory, to meet many (but not all) of my desirements. The camera I discovered appeared to have the following features: 1) outdoor operation, with an included sunshade; 2) 1920 x 1080 resolution; 3) CCD imager for good low-light sensitivity (hard to tell until it’s in your hands, though); 4) $70 delivered; and 5) built-in webserver, FTP server, etc. The things it didn’t appear to have included: 1) power over Ethernet (PoE), so I’d have to do that externally; 2) no idea of what the software was or how to configure the camera for my desired operating method; no idea if the camera was any good at all given the low price, and many other unknowns which wouldn’t be answered until I had one in my hands.

So, I placed the order on eBay and awaited delivery of my amazing new $70 super webcam. More soon on what I found out when I opened the box!


Commentary and insight on many topics, sometimes even about wireless

WP Twitter Auto Publish Powered By :