Sleuthing the source of a SARSAT signal?

It was August 11, 2006 - a late Friday afternoon - and I received a telephone call from Gordon, K7HFV and with him on the line was John, WA7UUJ with an unusual request:  The FCC's Denver field office had called him to ask if he - or someone he knew - would be willing look for a signal that NOAA had reported as having had appeared on several recent SARSAT passes.  If they could recruit some Salt Lake area locals to find this signal, it would save them a trip.

For some reason, I got called.

About the SARSAT system:

The SARSAT system was originally intended to provide a means for locating, via satellite triangulation, a signal emanating from a vehicle in distress - specifically, a ship at sea or a downed aircraft using an ELT.  Originally operating on 121.5 MHz, other frequencies such as 243.0 and those in the 406 MHz area had also come into use.  Today, the 406 MHz area frequencies are used almost exclusively, with the spacebourne capability of the other two frequencies having been phased out in 2009.

Note:  The 121.5 MHz frequency continues to be used, but only for short-range earth-based searches as many emergency beacon transmitters emit a much weaker, continuous-duty signal on this frequency to aid search parties as they close in.

Asked if I was interested in looking for the source of this signal, I agreed and John passed on a frequency - 406.2 MHz, some coordinates, and the number of a contact at the FCC.  Calling that number and talking to the contact I quickly determined that the coordinates given were not, in fact in decimal degrees, but rather in degrees, minutes, and seconds - and these coordinates, 40 degrees, 48 minutes North by 111 degrees, 49 minutes West - reportedly had a "circle of probability" of 1.1 km in terms of accuracy, placing the source of the signal 1.9 miles (about 3km) northeast of the "U" on the mountain north of Salt Lake, about halfway between Twin Peaks and Black Mountain.

It was reported that there had been several "hits" on this signal going back several days and since this territory was somewhat remote and not likely to have any power nearby so it was presumed that whatever it was that was causing the problem was likely to be battery-powered and would not persist for much longer.

Being that this frequency (406.2 MHz) is well-below the 70cm amateur band, Yagi antennas cut for that band were not likely to be useful in their directional characteristics.  Fortunately, Glen, WA7X, had previously acquired a circularly-polarized GOES downlink antenna surplus for only a few dollars with the intent of modifying it for 70cm use - but had not yet done so.  Since this antenna was designed to work at 402 MHz, it was still usable at 406 MHz and an additional feature was that the elements of this antenna could easily be removed, making it practical to carry while hiking.

That evening, after work I went over to Gordon's house.  He reported being able to hear a signal on his scanner near the frequency of interest and had already driven around a bit to see if he could make sense out of what was being heard as well as to scout out the trail head for the most likely access.  A point of confusion was that even though this signal went away when he removed the antenna, he couldn't hear it on any other receiver - and neither could I - so it seemed likely that it may be emanating from the radio itself or could possibly be an image response of some kind.

Sunset over Ensign Peak, as viewed from the ridge just below Twin Peaks.
Click on image for a larger version.

We piled into Gordon's Jeep and drove up through some fairly fancy neighborhoods until we arrived at the trailhead to Ensign Peak.  At this point Gordon was wondering where his scanner had gone and a fairly thorough search of his Jeep (assuming that such was possible) did not yield it.  Fearing the worst, Gordon drove back home while there was still some sunlight to look for it while I proceeded up the trail, talking occasionally with Glen who also relayed my transmissions when Gordon as I didn't have a simplex path to him from my location due to geography.  Once he arrived home he searched some more for the scanner and carefully retraced some of his steps to see if he could find it - as well as keeping a careful eye along the road:  No luck.  As best as he could determine he'd set the scanner on the roof of the Jeep and taken off, probably leaving the scanner on the road somewhere...

By the time Gordon got back to the trailhead I'd arrived at a ridge to the west of Twin Peaks where I could see into City Creek Canyon.  Up there, for the first time, I noticed a fairly weak signal on 406.28 MHz that sounded either like a repeater blowing squelch noise or, more likely, a data transmission of some sort.  Getting out the Yagi, the signal bearing seemed to indicate that it was coming from the direction of Ensign Peak, so I decided to break out another receiver, a Kenwood TM-733A to get a second opinion and to help eliminate the possibility that it was an image or spur of some kind.  Confirming that my FT-530 HT was, in fact, really hearing a signal, I decided to do a lateral traverse of this ridge to double-check the bearing.  The destination was a minor peak about a half mile to the west and it was hoped that by taking several bearings along the way, I could determine if the bearing that I'd taken was, in fact, likely to be a "true" bearing and not merely a reflection.

While making the traverse I was in occasional contact with Gordon, who was well underway (sans scanner, unfortunately) and was keeping him apprised as to what I had found.  Upon arrival at the minor peak I checked the signal strength and bearing once again and it was, in fact, still coming from the direction of Ensign Peak and of about the same strength - a reading entirely contrary to what the satellite coordinates had seemed to indicate.  Armed with this information, I headed back to the saddle.

By the time I got to the saddle Gordon had been there for a few minutes and it was starting to get quite dark with the city lights having turned on a short time before.  By now I was wishing that I'd done a bit more research as to the passband of the SARSAT system and had a few specific other specific technical details as to how it worked:  For example, I was wondering if a signal at 406.28 MHz would be bothersome to a system that received at 406.20 MHz?  I also wished that I'd queried the contact at the FCC a bit more about any relevant details of this signal that he might have such as whether or not that it was modulated - and if so, what was the nature of this modulation?

The block diagram of a typical SARSAT transponder.  This is some of the information that I'd wished that I had with me on the first trip to look for the signal.  This diagram includes the older 121.5 and 243.0 MHz subsystems which are no longer used onboard the spacecraft.
Click on image for a larger version.

Being that this saddle was still a few kilometers west of the coordinates we decided to follow the roads and trails to the east towards Black Mountain while keeping an eye on the clouds, virga, and lightning that we were seeing to the south and west.  Following the trail we wound around to the back side of Twin Peaks and before too long, we split from the old jeep road and began to follow a foot/bicycle trail to the east, lighting our way via flashlight.  At some distance past Twin Peaks we started to break out from behind the peaks at about the same time that the wind, rain, and lighting started to pick up.  After a few lighting strikes that appeared to be much less than a mile away, we beat a hasty retreat to the only thing resembling shelter that we could find - a shallow ravine along the trail near some scrub oak with a few 10's of feet of ridge above us.  There, we hurriedly donned our ponchos and waited out the now-heavy rain, fierce wind and lightning.  Fortunately, this storm wasn't too long-lived and it moved off toward the east and after it had passed we stuffed our rain-soaked ponchos into our packs and, defying common sense, continued on our way in the direction of Black Mountain.

By this time, I was really starting to wonder about some of the technical details of the SARSAT system so I called Glen on the cell phone and requested that he see what he could find on the web.  Before long we found ourselves in a clear area on a ridge top a little more than a kilometer from the coordinates that I'd been given so we decided to see what we could hear.  At about this time Glen called back on the radio with the results of his quick research on the web.  He reported that the SARSAT satellite operated on a center frequency of 406.025 MHz and NOT 406.2 MHz.  He also reported that in this frequency range there operated "Personal Locator Beacons" (PLB's) that could be activated by someone in distress.  These PLBs operate with several watts of power and transmit a short data burst that contains not only a unique serial number with which a person, boat, or vehicle could be associated, but a GPS-derived location as well.

Armed with this new information we started to look around in the 406.025 MHz range as well and sure enough, we heard a few bursts of what sounded like data transmissions.  Another interesting signal source seemed to appear at about 406.18 MHz - a bit off the SARSAT's center frequency - but since we weren't sure of the bandwidth of the SARSAT system, we couldn't readily discount it, either.  This latter signal proved to be problematic, however:  Although it could be heard on both the TM-733A and the FT-530, it seemed to resist attempts at determining a bearing.  The real clincher came about when I wielded the step attenuator and noted that adding just 3dB of attenuation seemed to cause the signal to go completely away:  It seemed to us that this indicated that this signal was likely an intermod product, but we were still somewhat puzzled by the fact that it was a clean, unmodulated carrier and we were wondering what signal sources would be likely to cause that?

At this location we had quite a good view of the Salt Lake Valley and over toward the Oquirrhs.  Not too long after arriving we were looking across the valley and our eyes caught a tremendously bright flash of light on the ground near the north-center of the valley.  What was striking about this flash was that not only did it seem to last the better part of a second, but it clearly had some obvious color to it and after disappearing for a few seconds, it appeared again, lasting for about the same amount of time.  Not having any clue what it was, we made several guesses, including the possibility that it was a malfunctioning power transformer.

This was just some of the electrical activity witnessed occurring over the Salt Lake valley that evening.
Click on image for a larger version.

Another thing that we could see from this spot was another thunderstorm moving in from the West but since this one was quite a ways to the south and since the skies above us were relatively clear, we weren't too worried.  As we sat there, I took advantage of the viewpoint and snapped several dozen short time exposures, hoping that I would be able to catch a half-decent lightning strike on at least one of them:  The results speak for themselves - see the picture above.

After a few minutes of listening, looking and taking pictures we decided that because it was approaching midnight and that the storm seemed to be inclined to approach us after all, it was time to go.  At this point, when putting gear back into my pack, that I'd realized that I couldn't find my transit compass:  I'd tried to find it a little bit earlier when we were trying to take bearings on the signals that we had heard, but it now seemed not to be buried in my pack after all.  All that I could figure was that in the haste to don the poncho in the rain I must have either dropped it or inadvertently taken it off, so on the way back we stopped at the place where we'd taken cover from the rain (after several back-and forth attempts to try to positively identify the exact location) and tried to find the compass by flashlight - but to no avail:  It seemed that the radio gods weren't satisfied with just Gordon's scanner, but they wanted my compass as a tribute as well!

The trip back down to the vehicle was mostly uneventful:  The only pause was a brief stop to attempt to take more lightning pictures - this time of a storm that was starting to appear over mid-southern Oquirrhs.  At about the time we got the the vehicle, it was sprinkling very lightly and the wind started to pick up, but by the time we got back to Gordon's house the wind was gusting fairly strongly and the lightning was much more intense and and accompanied by a hard rain.  We decided that we'd timed our retreat about right!

The next day I toyed with the idea of going back up to listen for the signal and to look for the compass but my plans were changed by the appearance of a strong thunderstorm that dropped marble-sized hail at my house.  Apparently Gordon toyed briefly with the idea as well, but he, too, was dissuaded by mother nature.

The next day (Sunday) was very nice, but I didn't feel inclined to tromp around in the 90 degree weather and I figured that the batteries running whatever it was that was generated the nuisance signal probably weren't likely to last 5 days, anyway...

A topographical map showing the hiking route (in blue) along with the precise satellite fix and the "circle of probability".Click on image for a larger version.

More reports of signals:

On Monday (8/14) I received another email from the FCC that not only reported that the signal was still there, but it had satellite "hits" going all the way back to the 3rd of August - 11 days before.   An interesting piece of information was that this signal usually seemed to appear in the late afternoon or early evenings - a possible clue of its origin.  Also included in the email was a new position fix:  40 deg, 48' 00" North and 111 deg, 54' 36" West.  This new fix was about 5 miles west of the original fix, indicating that it had either moved, or that the quality of the earlier fix wasn't quite as good as was believed.  By this time we'd also done enough research to know that the proper frequency to listen on or near was, in fact, 406.050 MHz, plus or minus 40-50 kHz.

As it turns out, the modern SARSATs have two separate receive systems onboard for 406 MHz:  One of these simply receives and does a store-and-forward of the data packets from the PLB's.  The other part of the system is a passband-limited,  frequency conversion receiver that translates the spectrum centered on 406.050 MHz with a passband of about +- 40 kHz, converts it down to 170 kHz and then phase-modulates it (along with the converted baseband signals of the other frequencies) on a 1544.5 MHz carrier.  In this way the effects of the Doppler shift are almost exclusively confined to the receive system onboard the satellite itself owing to the fact that the satellite downlink was not a direct frequency translation.

Note:  Another mode of operation for the SARSAT is with the transponder centered on 406.025 MHz with a passband of +-10 kHz:  We do not know for certain what mode the SARSAT transponder(s) were in when the signal was received.

With this new information Gordon and I decided that we might, in fact, be able to find this thing.  To be sure, the new coordinates were much "friendlier" than the others as they put the signal source just off the east side of Victory Road - just above its junction with Beck Street - all in places where we were likely to able to drive.

When trying to locate a signal such as this it is helpful to have an all-mode receiver capable of operating in the frequency range.  While one would normally listen for a signal in FM, being able to listen for the same signal on a receiver with a BFO (e.g. an SSB-capable receiver) is extremely helpful for several reasons:  It is possible to detect far weaker signals using SSB than with FM alone, it is easier to determine the precise carrier frequency of the signal being sought when using an SSB receiver, and it is fairly easy to analyze the "signature" of the signal to determine something about its origin.  This last point can be particularly important as not only can one often identify one specific transmitter in the presence of others, but its observed frequency stability may provide a clue as to whether the signal being heard is one that is stable and on-frequency or more likely to be a free-running spurious emission of a malfunctioning device.

The blue cross at the center of this map shows the location of the "new" SARSAT fix obtained on 8/14.
The ACTUAL location of the transmitter turned out to be near the lower-left corner of this
map, at approximately the location of the first letter of the word "LINE".
Click on image for a larger version.

While the most likely candidate for all-mode use was my Yaesu FT-817, it does not cover below 420 MHz.

What to do?

An answer occurred to me:  Throw together a quick-and-dirty frequency converter.

Fortunately, John, K7JL, had onhand several UHF bandpass cavities and he was able to tune one of these well out of its original frequency range all the way down to 406 MHz so that it could function as a bandpass filter.  The next step was to provide a local oscillator reference and this was provided by my Schlumberger 4031 service monitor - a device that can produce an SSB-quality carrier and run from 12 volts.  As it turns out, I had some diode ring mixers handy, but they were rated only to 250 MHz.  A quick test revealed that they did, in fact, function reasonably well at 406 MHz, albeit with a few extra dB of insertion loss.  The final component needed was a good preamplifier, but I was able to quickly retune a hombrew 70cm GaAsFET that I had onhand to operate at 406 MHz.  Cabling everything together I set the service monitor to generate a 280 MHz signal for the local oscillator, tuned the FT-817 to 126 MHz and, using another service monitor, I checked the sensitivity of this lashup and found that, on FM its 12dB SINAD sensitivity was about -115 dBm - more than good enough for our purposes!

After heaving all of this equipment into the car (along with a pair of 100 amp-hour batteries to run the service monitor, which draws 8-10 amps) I set off and picked up Gordon from his house.  From there we drove through downtown Salt Lake and past the state Capitol building and as we climbed the hill we could hear some activity in the 406.025 MHz area that sounded like data transmissions.  We had heard what we thought had been similar-sounding signals on the previous Friday when we were overlooking the valley and had presumed that they might, in fact, be some Personal Locator Beacons transmitting.  The fact that there was an outdoor exhibitors' show in town seemed to make it even more likely to us that one or more of these transmitters was active.

Driving past the Capitol building the signal that we'd been hearing wavered a bit, but became fairly consistent as we went down the Victory road, and it sounded very much like what we'd heard on Friday.  Near the bottom of the hill - just above the junction with Beck Street - I pulled off the road into a gravel parking lot.  A quick check revealed that we were just a few 10's of meters south of the precise coordinates relayed to us based on the most recent satellite data.  Getting out the Yagi, we tuned around and listened.

After a few minutes of thinking about what we were hearing, it occurred to us that we were hearing only one transmission at this location.  While it sounded just like data bursts, we also noted that these transmissions were not of consistent length and it sounded as though they were being truncated randomly.  Switching to SSB, we also noted that the signal that we were hearing was fairly unstable, having several kHz of observable "chirp" when it keyed up.  From this evidence, we concluded that this was most likely our suspect:  We surmised that it could still be a PLB, but one that was malfunctioning but this seemed to be unlikely as such a device would likely be operating from a battery that should have long been exhausted.  The other obvious possibility was that it could be an existing, malfunctioning transmitter throwing out spurious transmissions.

Wielding the Yagi we determined that its bearing was west-southwest from our current position so we set out to find another vantage point from which we could hear this signal.  Wanting to guard against the likely possibility of being misled by a reflection Gordon suggested somewhere above the capitol so we drove in that direction to find a good overlook.  Driving behind the capitol we happened to drop behind the ridge, effectively shielding us from anything to the west and we noticed that the signal that we were listening for disappeared, indicating that it was not likely to be either to the south or east of us.  After a few more minutes of driving we found a nice observation point at the end of the road, just below Ensign Peak, and parked.

Judging by the sizable crowd up there we concluded that this was a popular spot as it commanded a spectacular view to the south and west and once we started waving the Yagi around we not only got some stares, but also an indication that the signal was quite strong, yielding a bearing that was almost due west.  As we were getting ready to depart several people asked what we were doing and (nearly correctly) had surmised that we were tracking animals.  "Close!" we told them as we explained what it was that we were really doing.  Interestingly, we again noticed a very bright flash of light on the ground in a midvalley location - just like the one that we'd seen on Friday:  We still don't know what it was!

Armed with the new bearing we wound our way down from the foothills and headed west on 6th North.  As we dropped, the signal became quite weak for a while and it wasn't until we got to about 20th West that it started to pick up again.  Soon, we turned south, finding ourselves passing the Executive Airport terminal and the local FAA offices and we parked for a moment to find another bearing:  From here, the signal appeared to be coming from the northwest - but it wasn't particularly strong.

Needing another clear location, we decided to head toward the International Center on the west side of the International Airport in the hopes that we'd be able to get a counterbearing.  En route, we lost the signal, but we assumed that the path to the source had been obstructed but once we got to the International Center we still could not hear it.  What we could hear, however, seemed to a very occasional data transmission on the frequency (perhaps once per minute) that sounded different from the one that we were seeking.  A quick listen on SSB revealed that, unlike our suspect, its carrier frequency was quite stable.

At this point we decided to go back to where he'd heard the signal before - in the general area of the Executive terminal.  Returning and parking in an empty lot we listened for a few minutes and didn't hear anything other than the "occasional" transmission that we'd noted previously.  I then surmised that being a spur, it may have drifted off-frequency a bit and eventually found it about 30 kHz higher than it had been when we'd first heard it earlier in the evening.  I then decided that since it was likely a spur, it would seem reasonable that the frequency on which the malfunctioning transmitter operated was likely to be nearby, so I decided to start tuning around on the TM-733 in my Jeep while monitoring and tracking the unstable suspected signal on the FT-817.

After a few minutes of tuning I found a signal on 410.075 that sounded very much like our suspect and after comparing its appearance and modulation with that of the spur, they appeared to match exactly.  For this signal the beam heading was due south from our location, but there was a problem:  Because that the antenna was designed for operation on 402 MHz, we weren't sure if it would function properly and yield a usable bearing that far off frequency.  Fortunately, we had a nearby-frequency signal with which we could calibrate our antenna:  A backhaul feed of the main National Weather Service transmitter on 410.575 MHz.  The problem was that we didn't know where this transmitter was located, so I called John, K7JL.  He wasn't sure of its exactly location, either, so he looked it up in the telephone book and we determined that it was just a few blocks south of us - a location that correlated with our bearing.  At this point it also occurred to us that the bearing of our suspect signal was about the same as the NWS signal.

Getting back in the car we drove a few blocks south and stopped in a parking lot adjacent to the NWS office and observed that the 410.075 MHz signal was extremely strong with a bearing pointing into a side yard that, in the dark, appeared to have a number of antenna-like structures.  To be certain, we drove around to the south side of the building and again, we resolved a bearing in their side yard.  At this latter location, I looked at this signal - and its spur - on the service monitor's spectrum analyzer and determined that the 406 MHz component was 38 dB below that of the 410.075 MHz one:  If the transmitter was running at just a couple of watts, this would imply that the power in the spur was a few milliwatts - more than enough to be heard around town and even by an orbiting spacecraft!

John, who was still on the phone, happened to know  the technician at the NWS responsible for their transmitters and left a message indicating that there may be a problem requiring his attention.  At this point we were satisfied that we'd uncovered the source of the signal and went home.  After all of this we were somewhat surprised that our snooping around government buildings, after dark, wielding odd-looking gear hadn't seemed to have attracted any attention!

The next morning the technician from the NWS returned the call and John explained what had been happening and what had been found:  The tech immediately arranged to have the transmitters looked at and, as a future preventative measure, installed bandpass cavities.  As it turned out, this transmitter was part of  the ASOS - the Automated Surface Observing System - which is used for gathering close-in meteorological data at airports.  Interestingly, the FCC's NOAA contact pointed out that, several years ago, they had a spate of similar interference issues at airports caused by these same-model transmitters in which a weak 400 MHz-area spur had caused some interference with some operations.  He also then mentioned that this was the first time that a "non-airport" transmitter had caused some problems, but this NOAA facility, which was adjacent to the airport, had escaped prior scrutiny since its transmitter wasn't actually on airport property like the other errant transmitters.

The irony did not escape us that this signal was generated from an NOAA facility and that it was, in fact, NOAA that had reported this signal to the FCC in the first place!

Interestingly, the latter satellite fix (the one indicating a site near Beck Street) was about 5km northeast of the actual location.  This error is understandable when one considers the nature of the signal:  It was likely fairly weak on the satellite.  It also was quite unstable in frequency, a factor that would tend to increase the uncertainty when trying to make a Doppler fix on the location.  Furthermore, being that the satellites used to determine these sorts of location fixes are in very high inclination (being nearly in polar orbit) they would have their best accuracy when determining latitude:  The accurate determination of longitude is a bit more problematic as its uncertainty relies on observations over multiple passes and/or through different satellites, hence the fairly close agreement on latitude, but the much larger longitude error.  (Note:  This system is "reverse" of the way the old TRANSIT (link) satellite system operated.)

The observation that this signal seemed to be only present in evening hours also made sense:  Only when the outside temperature was in the mid 90's Fahrenheit did it seem to drift into the SARSAT passband.  This temperature peak would, in fact, occur in the late afternoon and as it cooled off, it would drift back up in frequency, out of the transponder's passband - possibly disappearing altogether from the transmitter's spectra.

A few days after the transmitter's servicing another email arrived reporting that since the work had been done, no further SARSAT hits had been noted.  An email from the FCC indicated that they were pleased that they did not have to make a special trip to Salt Lake just to find this signal and offered to take us to the Red Iguana next time they were in town.

End

P.S.  The FCC folks still haven't taken us to the Red Iguana!

[End]

This page stolen from ka7oei.blogspot.com