GPS industry rages: LightSquared 4G network would "defy" laws of physics | Hacker News Reader

GPS industry rages: LightSquared 4G network would "defy" laws of physics

arstechnica.com

24 pointsjoelhaus28 comments

One thing not mentioned is LightSquared misuse of the spectrum. The spectrum in question is the L1 band, which was set aside for satellite communications. These satellite transmitted signals appear weak from the ground.

A loophole in the FCC rules allows the use of ground based stations to fill-in weak areas of satellite coverage. This is called Auxiliary Terrestrial Component (ATC). LightSquared's plan is to use the ATC loophole to build out a terrestrial ground-based network with very little satellite usage. They are essentially getting valuable ground-base spectrum at the low price of satellite spectrum.

LightSquared blaming of GPS receiver manufactures is a bit ridicules. The GPS unit manufactures weren't expecting out-of-spec usage of the neighboring L1 band. The L1 band was set aside for satellite signals, not terrestrial. Somehow none of the press is pointing this out.

Here's a good write-up http://www.directionsmag.com/articles/whats-the-deal-with-th...

You can find the report that LightSquared submitted in the right-hand column of the page at [0]. I'm reading it now to see if their suggestion to use filters (which was very vague in the linked article) really has merit. The "Save our GPS" consortium surely doesn't think it does.

[0] http://www.pnt.gov/interference/lightsquared/

The proposed workaround is to start using the lower 10MHZ (1526-1536) of their allocated range immediately, while cooperating with GPS device manufacturers on strategies that will allow them to use the the remainder (1536-1559Mhz) in the future. The GPS devices operate at 1575Mhz.

I'm reading the Recommendations portion of the report. It is surprisingly light on technical details. LightSquared's claim seems to be that there is only a conflict on the "edge" of one of their licensed signal bands, primarily because a small subset of GPS devices (mostly survey, mining, etc. use devices) are improperly accepting signal in adjacent bands, either accidentally or intentionally in order to carry a little additional custom data. (?)

So far, after reading most of the Recommendtions report and without being knowledgeable in this specific area, I'm inclined to side with LightSquared. Groups advocating protection of wireless spectrum are frequently hysterical when it comes to anything that might affect their signal. (e.g., the hams' strong and vocal objection to various proposals to carry telecom signals over power lines.)

I've glanced at the same doc, got a slightly different take from it. The higher precision devices are actually receiving data in LightSquared's licensed band (by having more receive bandwidth, they get better accuracy), on purpose. This is not the GPS devices' band to use, so IMO tough luck on them. (If you like analogies, it's like carving out a block of unassigned IPs for yourself, and then complaining when someone is actually assigned those IPs and wants to use them).

The one band they tested is the one that is further away, and therefore easier for devices to reject. They claim it's not a problem for "over 99% of devices" when they are at full power there. They concede that operation the the higher frequency band (that LightSquared has licensed) would cause trouble for "a significant portion of legacy devices."

The root cause of most of the problem (not the part where ultra-precision receivers use bands that aren't theirs) is that the GPS receivers expected things in neighbouring bands to remain quiet, when there was no guarantee that such would happen. This assumption made for easier and cheaper designs for them, and now all this hardware is out there that is susceptible to the interference.

Right. I might've done a bad job describing my take on the report, because I agree with all that.

It's also worth noting that LightSquared is offering to provide circuitry -- which they claim to have already developed -- to any device manufacturers that want to filter LightSquared's signal.

Regardless, I'm not yet seeing a single thing anywhere in this part of the report that would lead me to conclude that they're proposing something that "defies physics".

The "defies physics" part would be a filter that is physically impossible to implement. As you need sharper roll-off of your filter (more rejection of out-of-bands, and sooner), you start making sacrifices and attenuating the signals in your passband as well. GPS already needs very sensitive receivers, and attenuating the signals more than they already are could make them unusable. ("Oh!" you say, "just put more amplifiers behind the filter!" Well, no, at that point the signal is inderminable from thermal noise, which you will amplify just the same. And you can't put one in front of the filter, because then the LightSquared signals don't get filtered out and they interfere...)

Without seeing the specs they are being asked to hit, I can't speak to the challenges of realizing these filters.

And just to be clear, there is no way for LightSquared to use their upper band and not interfere with the ultra-precision stuff that's effectively squatting there. There's no happy compromise in that situation, but as I implied earlier, I have little sympathy for the guys who weren't playing by the rules.

Ok, because the implied situation in the article is the LightSquared is pooping all over the GPS spectrum and then selling licences to manufactures for their spiffy new GPS service which happens to work much better than old GPS now.

Nope, more like GPS guys built a luxury spa in the middle of a farming area, and are complaining that LightSquared is pooping all over the field they [edit: they=LightSquared] just purchased upwind, which has been wildflowers that spa patrons could trespass through on on illicit trails until then. And LightSquared is saying "guys if you make this simple wall on your own property, the poop smell won't blow that far."

And to bring my analogy one step further, the GPS guys are saying "that wall is too difficult to build." [I cannot speak to how hard that wall is to build.]

Does anyone have any real worthwhile info of this entire issue? All I've seen is the few incomprehensible technical arguments and then plenty of over-popularised articles like this one. I think it's obvious that LS can't go ahead if these problems are real, but the whole debate seems to lack much real detail.

Lightsquared has a license to broadcast in the 1525-1559 MHz band. Since their transmitters are on the ground, the signals are powerful.

GPS is 1559—1610 MHz. These signals are very weak, since the transmitters are in orbit.

The pissing match is that even though on paper the two ranges don't overlap, in practice the Lightsquared transmissions tend to stomp all over the lower end of the GPS range, which make various interested parties unhappy.

There's additional humor in that while it's certainly possible to get around interference like this, in practice that's dual-use technology which can be used to evade military GPS jamming.

If anybody was going to be outraged by this, wouldn't it be the FAA and the Us Military?

Sure they are. The enemy might send a guided missile through the window of the office next to the general's. What a disaster!

Fun fact: (most) civilian GPS receivers stop working when they're moving faster than 515m/s, are more than 18,000m above sea level, or are accelerating faster than 4g. (39.2m/s^2)

If you're thinking that these operational constraints sound a lot like the performance envelope of a guided missile, then gosh, you're right. IIRC, unlocked GPS recievers are classified as MTCR category II dual-use missile related components, with the usual suite of import restrictions.

EDIT: Looked it up[1] for... some reason, and yep, category II, item 11, section A.3:

  b. Designed or modified for airborne applications and having any of the following: 
    1. Capable of providing navigation information at speeds in excess of 600 m/s; 
    2. Employing decryption, designed or modified for military or 
    governmental services, to gain access to GNSS secure signal/data; or
    3. Being specially designed to employ anti-jam features (e.g. null steering 
    antenna or electronically steerable antenna) to function in an 
    environment of active or passive countermeasures.

So, in fact, it's just the speed restriction. The numbers I have are probably from some American arms-control law, not a MTCR restriction.

1: http://www.mtcr.info/english/MTCR-April2011-Technical-Annex....

It seems that the cost of fabbing your own gps chip would be peanuts compared to the cost of even one missile. Given the raw GPS data its pretty trivial to do all location processing in software.

Due to my current job, I've had to become familiar with my country's export restrictions on some tech. I've since thought it would be great fun to start a string of projects with the goal of meeting those requirements that make them export-restricted. If nothing else, I know they are (at least mildly) challenging goals, and concrete specs to meet so that I know I've accomplished the goals.

You wouldn't need to fab your own chip. Plenty of FPGAs would work. In fact, you could make do with a decent speed ADC and a bit of software.

It should be noted that there are numerous civilian GPS receivers that do not have the altitude limit or have it much higher than 18km. Many are purpose built for high altitude balloon payloads.

I was researching GPS receivers for an high altitude balloon and most receivers stop working at more than 18km OR at 500m/s.

Some manufacturers misinterpreted the restriction. Instead of OR it should be AND.

If you look around high altitude balloon forums, they usually maintain a list of good gps chips(stop working at 500m/s AND 18km instead of 500m/s OR 18km).

    *3. Being specially designed to employ anti-jam features (e.g. null steering 
    antenna or electronically steerable antenna) to function in an 
    environment of active or passive countermeasures.*

I find this one slightly annoying. So no matter what, you can't connect an electronically steerable antenna to a GPS receiver and export the resulting device?

Sure you can export it. Category II items aren't banned, they're just controlled.

So you can export it, it'll just be really expensive to do so, and they won't want you to sell them to terrorists.

Seriously, we drop bombs targeted using this technology. I can't imagine that the military would let anything diminish the accuracy of existing systems.

I dont think (and hope) it's that simple to disarm the GPS on guidance missles. Else what's stopping our "enemies" from just building these transmition towers near their bases?

It is that simple, which is why the bombs just use GPS to get close to the target, then dead reckon[1] the rest of the way.

[1]http://en.wikipedia.org/wiki/Dead_reckoning

The presumption is that more expensive equipment based on GPS (aircraft navigation, etc) have the appropriate "filters" in place to prevent interference.

If LightSquared is to be believed, the problem is that commodity GPS makers did not take adequate precautions against interference.

There's probably a grain of truth to both positions. Dealing with RF interference is something of a black art, so given that LightSquared hasn't actually deployed their network yet, it's not surprising that existing GPS devices don't filter out those signals perfectly. At the same time, LightSquared seems to have assumed that they could avoid interfering with the GPS band to a greater degree than they were ultimately able to deliver.

Of course, it's also a question of money - from the GPS manufacturers' perspective, they will have to assume all the costs while LightSquared derives the benefits. From LightSquared's perspective, they paid to acquire their spectrum while the GPS manufacturers get access to government infrastructure for free and are effectively blocking access to a significant chunk of spectrum just to reduce their per-unit costs.

Military GPS operates at a different frequency.

I believe that military GPS units utilize both the L1 band (civilian) and L2 band (military only) to be able to calculate ionosphere delays.

Reply on news.ycombinator.com