>>> 2020-12-15 weird wireless (PDF)

To start with, as a result of a brief and uncomfortable brush with fame I have acquired a whole lot more readers. As a result, I pledge to make a sincere effort to 1) pay a little more attention when I mash my way through aspell check on each post, 2) post more regularly, and 3) modify my Enterprise Content Management System so that it is slightly more intelligent than cating text files in between other text files. Because it's a good idea to only make promises you can keep, I have already done the third and hopefully text now flows more coherently on mobile browsers and/or exceptionally narrow monitors. I am also going to start the process of migrating the mailing list to a service that doesn't, amusingly, repeatedly bungle its handling of plaintext messages (HTML is somehow easier for MailChimp), but rest assured I will move over the subscriber list when I do that. I am still recovering from years of trauma with Gnu Mailman as a mail server admin, so mass email sending is not something I can approach easily.

I will not, however, make the ASCII art heading represent properly at narrower than 80 characters. Some things are simply going too far.

So after the topic of odd scanner regulation around cellular phones, I wanted to talk a little more about cellular phones. The thing is, I have decided not to try to explain the evolution of current cellphone standards because it is not something I have any expertise in and frankly every time I try to read up on it I get confused. The short version is that there are two general lineages of cellular standards that are (mostly incorrectly) referred to as CDMA and GSM, but as of LTE the two more or less merged, differentiating CDMA and GSM phones by what they fall back to when LTE is not available. Then 5G happened which somehow made the situation much more complicated, and is probably a government conspiracy to control our minds anyway.

On a serious note, one of the legacy cold war communications systems I am very interested in is GWEN, the Ground Wave Emergency Network. GWEN was canceled for a variety of reasons, one of which was up-swell of public opposition founded primarily in conspiracy theories. The result is that it's hard to do much research on GWEN today because you keep finding blogs about how cell towers are actually just disguised GWEN sites being used to beam radiation into our homes. GWEN itself has a slightly interesting legacy, it was canceled before it achieved full capability but a number of sites had already been built. The more coastal of those sites ended up being transferred to the Coast Guard which reused them as stations for their new Differential GPS network[1], so a large portion of DGPS sites were just disused GWEN sites. DGPS was replaced with NDGPS which itself has recently been decommissioned, in part because the FAA's Wide Area Augmentation System (WAAS) is generally superior and can also be used for maritime purposes. So the GWEN sites have now died two deaths.

So we can all hopefully agree to call that tangent a segue to the topic I really want to discuss: less-well-known terrestrial mobile communications systems. Basically, I want to take the family tree of pagers and cellular phones and call out a few lesser known members of the order.

Let's talk first about pagers. Pagers have largely died out, but I had the fortune of working in a niche industry for a bit such that I carried one around with me. Serious '80s drug dealer vibes. The basic idea of the pager is very simple: there is a radio transmitter somewhere, and a bunch of people carry around belt-pack pagers that beep at them when the radio transmitter sends a message intended for them. In its late forms, it was essentially a dedicated text-messaging infrastructure, although early pagers delivered no payload at all (only the fact that a message existed) and later pagers, many into the '90s, only delivered the phone number of the caller.

By the '90s though not only had alphanumeric pagers come about with full text message capability, two-way alphanumeric pagers had been introduced where it was possible to respond. Two-way pagers basically represent a weird mutation on the way to modern cellphones and so I won't discuss them too much, I'm more interested in the "pager" as distinguished by being a strictly one-way device. This is, for example, the reason that I sometimes jokingly refer to my phone as my pager: I detest typing on it so much that I often use it was a one-way device, responding to messages I receive later when I'm at a computer.

Pagers have gone through a number of different technical evolutions, but most modern pagers run a protocol called POCSAG. One of the reasons for widespread standardization on POCSAG is that it is not uncommon for institutions to operate their own private paging transmitters, so standardization is more or less required to make a sale to any of these users (which today probably represent most of the paging market). Understanding this requires commenting a bit on another huge way that pagers are differentiated from cellular networks.

Modern cellular phones (and really all cellular phones if you use the term strictly) employ a "presence" or "registration" system. Essentially, as you walk the mean streets of suburban Des Moines your cellular phone is involved in an ongoing dialog with base stations, and central systems at your provider continuously keep track of which base station your cellphone is in contact with. This way, whenever you get a call or text message or push notification, the system knows which base station it should use to transmit directly to your phone---it knows where you are.

Pagers, excepting some weird late-model pager-ish things, don't have any such concept. The pager itself has no transmitter to advertise its whereabouts (this is a large portion of why pagers remain in use today). Instead, every page destined for a pager must be sent via every transmitter that that pager might be near.

You can imagine that this poses inherent scalability limitations on pagers. When you purchase pager service from a commercial provider, you generally have to specify if you want "city," "regional," or "nationwide" service. This really just determines whether they will transmit your pages in just your city, throughout a state or other regional area, or nationwide. Nationwide service is surprisingly expensive considering cellphone competition. Even then, paging transmitters tend to only be located in more urban areas and so coverage is poor compared to cellphones.

This limitation of pagers, though, is also an advantage. The simplicity of the total paging system (just encode messages and send them out a transmitter, no huge technology stack involved) encourages private paging systems. In my area, hospitals and universities operate private paging systems, and government facilities contract them out but still to a local, small-scale scheme that is effectively private. They're particularly popular with hospitals because an already-installed paging system is fairly cheap to maintain, it's guaranteed to work throughout your building if you put the transmitter on the roof (not something cellphones can always offer in large, byzantine hospitals), and as long as your staff live reasonably nearby their pagers will work at home as well[2].

So that's "what a pager is" at a bit of a technical level. More interesting to me are some pager-adjacent devices, such as the Motorola MINITOR. MINITORs are so popular with volunteer fire departments that you can pretty reliably identify volunteer firefighters by the MINITOR on their belts, although the nineteen bumper stickers and embroidered hat tend to give it away first.

So what is a MINITOR and how does it relate to a pager... this requires getting a little bit into radio systems and the concept of a coded squelch. Let's say that you are, example out of nowhere, a fire department. You have a VHF or UHF FM radio system that you use to communicate between dispatch and units. When dispatch receives an event they want to notify the units that should respond, but they don't want to wake up the entire department. One common way of achieving this is some manner of coded squelch. This is not the only application of coded squelches (they're often used just as a way to minimize false-positive squelch opens), but it's one of the most complex and interesting.

The idea is this: instead of a given radio just opening squelch (enabling the speaker basically) when it receives a carrier, the radio will only open the squelch when it receives a specific tone, series of tones, data packet, or other positive message that that radio is supposed to open squelch. By programming different tone sequences into different radios, the dispatcher can now "selective call" by transmitting only tone sequences to open squelch for the specific units they wish to contact.

There are two major coded squelch systems used in public safety (actually there are a ton but these are the two most widely seen on analog, non-trunking FM systems): two-tone, also called Selcall, and Motorola MDC. Two-tone is the format supported by MINITORs and probably the more common of the two because it has more cross-vendor support, but it's also much more primitive than MDC.

The concept of two-tone selective calling is very simple and you can probably guess from the name: Before a voice transmission, essentially as a preamble, the transmitter sends two tones in sequence, each for about a second. Yes, this takes a while, especially if calling multiple units, enough that it's not done on key-up like MDC or many other selective calling schemes. Instead, the dispatcher's radio console usually has a dedicated button that starts sending tones and they have to wait until it's good and ready before they talk. It's not uncommon to hear the dispatcher say something like "wait for tones" or "tones coming" to warn others that things will be tied up for a bit.

So how does this all relate to paging... the MINITOR and other devices like it are basically handheld radios with the entire transmit section removed. Instead, they are only receivers, and they are equipped with a two-tone decoder. So if you are, say, a volunteer firefighter, you can carry a MINITOR which continuously monitors the dispatch frequency but only opens squelch if it receives a two-tone sequence indicating that the dispatcher intends to activate a given group of volunteers. This is basically a paging system, but simply "built in" as a side feature of the two-way FM radio system.

I'll also mention MDC briefly. MDC is a more sophisticated system that uses a short FSK data packet as the selective calling preamble. This transmits quickly enough that the radios simply send it every time the PTT is pressed. This allows some more advanced features, for example, every time someone in the field transmits the dispatcher's console can tell them the ID of the radio that just transmitted. Auxiliary information in addition to addressing can also be send in the MDC preambles. MDC is also very popular in public safety and if you've spent much time with a scanner you'll probably recognize the sound of the MDC preamble. It's actually very common to mix-and-match these systems, for example, some fire departments use MDC but also send Selcall tones when dispatching, often specifically to trigger MINITORS.

Selective calling systems in public safety are often also used to trigger outdoor warning systems such as sirens, which are of course one of my favorite things. A surprising number of outdoor sirens used in tornado-prone areas, for college campus public safety, etc. are just equipped with a radio receiver monitoring a dispatch frequency for a specific selective call. This can interact in amusing ways with "mixed" selective calling. I used to work on an Air Force base with a fairly modern Federal Signal outdoor warning system. When it played Reveille and Retreat each day it sounded fine, but when they tested the emergency sirens one day a week you actually heard MDC and then Selcall tones over the speakers before the siren. My assumption is that regularly scheduled events like Reveille were played via a Federal Signal digital system while emergency alerts went out over some force protection dispatch frequency, and the "siren" speakers opened squelch in reaction to some Federal Signal-specific preamble that was sent before the preambles used for mobile radios. As another anecdote, the US military has the charming habit of referring to all outdoor warning systems as "Giant Voice," which was the brand name of a long-discontinued Altec Lansing system that had been very popular with DOD. Other siren systems are triggered using telephone leased-lines, and of course on modern systems there are options for cellular or other more advanced data radio protocols.

There are also a number of other selective calling systems in use. Another example I am aware of is a proposal among amateur radio groups called "long tone zero," which suggests that persons experiencing an emergency should tune to a nearby repeater and transmit a DTMF zero for several seconds. The idea is that other radio amateurs who wish to be helpful but not have their ears glued to their radios (or more likely be woken up at night) can set up a software or hardware detector for the zero digit and essentially use it as a selective-calling scheme, with their radio (presumably with the volume cranked to eleven) only opening squelch upon receiving a long-tone-zero. It's a clever idea but to my knowledge not one that is widely enough implemented to be particularly useful. Of course selective calling is also widely used to open the squelch on repeaters but I find that less interesting.

A similar scheme that is oddly well-known to the public ear is employed by the Emergency Alert System and NOAA All-Hazards Radio. Those "emergency weather radio" receivers you buy at the store from the likes of Midland monitor an All-Hazards Radio frequency but only open squelch when they receive a preamble indicating that there is an emergency notification. Historically this was based on a simple dual-tone scheme (the tones that are now used as the emergency alert ringtone on most cellphones), but nowadays a digital scheme is used that allows the radio to know the type of alert and area it applies to. This is actually how EAS messaging is triggered on many television and radio stations as well. I will devote a whole post some time to the history of the Emergency Alert System in its various outdated and modern versions, because it's really pretty interesting---and frankly I am amazed that incidents of unauthorized triggering of EAS are not more common, as the measures in place to prevent it are not particularly sophisticated.

So that's one pager-adjacent thing. Let's talk a bit about a different pager adjacent thing, and one that I know less about because it's more proprietary and less frequently heard in the modern world: taxi and freight dispatch.

Several manufacturers used to build taxi dispatch systems that allowed for individually addressed text messages to specific receivers installed in cabs. This is basically a pager system but using a larger display, and the systems were almost always two-way and allowed the cab driver to at least send a response that they were on the way. The system in use and its technical details tended to vary by area and it's hard for me to say too much in general about them, other than that they have been wholly replaced today by cellular systems.

A system similar to taxi dispatch systems is ubiquitous in the freight trucking industry, but is far more standardized. Qualcomm Omnitracs is an integrated hardware and service product that places a small computer in the cab of a truck which both reports telemetry and exchanges text messages between the driver and dispatcher. The system has been bought and sold (I don't think Qualcomm even owns it any more) and has been moved from technology to technology over years, but for most of its lifespan it has relied on a proprietary satellite network. This gives it the advantage of being more reliable in between urban areas than cellular, although the fact that the system is now available in a cellular variant shows that this advantage is getting slimmer. It's also the reason that a great many semi tractors feature a big goofy radome, usually mounted behind the roof fairing. You just don't see that kind of antenna on vehicles very often. Like most satellite communications networks, Omnitracs relies on the messages being small and infrequent (very low duty cycle) to make the service affordable to operate.

What I particularly like about the Omnitracs system (which seems to be widely referred to by truckers as Qualcomm regardless of who owns it now) is that the long near-monopoly it enjoyed, and probably its relationship to a big engineering operation in Qualcomm, lead to some very high quality hardware design compared to what we expect from communications devices today. The system was always designed to be usable on the road, and featured a dash-mounted remote control and speech synthesizer and recognition (to hear and reply to messages) long before these became highly usable on cellphones. The system also integrates secondary features like engine performance management and even guided pre-trip checklists. It's an example of what can be achieved if you really put hardware and software engineering expertise into solving a specific problem, which has become uncommon now that the software industry has realized it is cheaper (at the expense of user experience, productivity, etc) to solve all problems by taping iPads to things. And that's the direction that freight dispatch is increasingly going today, "integrated products" that consist of a low-end Android tablet in a dashboard mount running some barely stable app that is mostly just a WebView. And taxi dispatch barely even exists now because silicon valley replaced the entire taxi with an iPhone app, which if you think about it is kind of amazing and also depressing.

These two topics get very close to the world of mobile data terminals, and that's what I'll talk about next. MDTs are car-mounted computers often used by first responders and utility crews, and while nowadays "MDT" usually just means a Panasonic Toughbook with an LTE modem (maybe for a municipal LTE network), historically it referred to much more interesting systems that paired a Panasonic Toughbook with a VHF/UHF data modem that relied on some eccentric protocols and software stacks. One thing has never changed: Panasonic Toughbooks are way overpriced, even on the government surplus market, which is why I still don't have one to take apart.

So I'll talk a bit about MDTs and the protocols they use next, since in many ways they're more the ancestors of our modern smartphones than actual phones.

So, is there any big conclusion we can draw from looking at these largely "pre-cellular" (but still present today) wireless systems? I don't know. On the one hand, in some ways these confirm one of my theses that increasing commodification of software and hardware tends to make technology solutions less fit for purpose rather than more. That is, technology devices today are better only in a certain way, and worse in other ways in that increasing abstraction, complexity, and unification of design tends to eliminate features which are specific to a given application (everything is an iOS and/or Android app now, and half of those are really just websites) and increase complexity for users (what was once a truck dispatch system is now an Android tablet with all the ways that can go wrong).

At the same time, these effects tend to drive the cost of these devices down. So you might say that everything from semi-truck dispatch to restaurant POS (a favorite example of mine) is now more available but less fit for purpose. This is one of the big themes of my philosophy, and is basically what I mean when I say "computers are bad," so I hope to explore it more in this blog newsletter thing. So next time, let's try to look at mobile data terminals and dispatch systems under that framework---how is it that they have become cheaper and more available, but at the same time have gotten worse at the purpose they're intended for? But mostly we'll talk about some old radio data protocols, because those are what I love.

Postscript: now that I have 100+ email subscribers and probably as many as ten people who somehow still use RSS (I don't know, I don't really have any analytics because I'm both lazy and ethically concerned about delivering any Javascript whatsoever[3]), I'd love feedback. What do people find most interesting about my rambles? What do they want to hear more about? You can always email me at me@computer.rip, or hit me up on Matrix at @jesse:waffle.tech. If you send me an email I like enough I'll throw it in here sometime like an old-fashioned letter to the editor. Like The Economist, if you do not begin it "SIR -" I will edit it so that people think you did.

[1] Differential GPS is an interesting technique where a site with a known location (e.g. by conventional survey techniques) runs a GPS receiver and then broadcasts the error between the GPS fix and the known good location. The nature of GPS is that error tends to be somewhat consistent over a geographical region, e.g. due to orbital perturbations, so other GPS users in the area can apply the reverse of the error calculated by the DGPS site and cancel out a good portion of the systematic error. The FAA WAAS system was designed to enable RNAV GPS approaches, basically aircraft instrument operations by GPS. The main innovation of WAAS over DGPS/NDGPS is that the correction messages are actually sent back to orbit to be broadcast by satellites and so are available throughout North America.

[2] A huge downside to this is that POCSAG lacks any kind of security scheme. I have personally found more than one hospital campus transmitting patient name, DOB, clinical information, and occasionally SSN over an unencrypted POCSAG system. My understanding is that, from a legal and regulatory perspective, this is basically an accepted practice right now. Maybe congress will pass legislation against POCSAG decoders.

[3] I am not Richard M. Stallman, although I did once spend much more time than I am comfortable with in his presence. I'm more like Jaron Lanier, I guess. My ethical concern about delivering Javascript is not that it removes your control of your personal computer or whatever, but rather that it feels like the gateway drug to pivoting to a social network and creating a Data Science department. When it comes to modern webdev, "DARE To Say No to SPAs and Websockets." Oh god I'm going to design a T-shirt.