The novel 1632 confronted our fictional friends with the Maunder Minimum, a unique period in early modern times when solar activity stayed abnormally low for decades. Hams traditionally rely on ionospheric skip in the high frequency bands for long-distance communication at modest power levels with antennas that a private individual can put up, but that was elusive by the 1630s, if it was there at all. Perhaps the situation wasn't quite as difficult in actual history as it is portrayed in 1632 canon, but there was nobody using radio then to leave us historical records. With the equipment and antennas the eighteen original hams in Grantville had, anything much beyond twenty miles or so was severely limited by short daily openings in the lowest ham bands, covering limited spans of longitude. Rick Boatright's "Radio in the 1632 Universe" series beginning in the first Grantville Gazette and at the 1632 web site, www.1632.org, explains those problems in detail, and what the hams were able to do about it.

Despite all that, several stories and novels in later canon ("A Friend in Need," "Storm Signals," "In Remembrance," 1636: The Saxon Uprising) have relied on dependable radio communication across salt water for hundreds of miles at any time of day or night, with strong hints of plans to operate over much greater distances. So how was that possible? To answer that, we need to take a close look at radio propagation modes and the frequencies where the different modes do their best work.

There are many physical effects that can carry a radio signal around local obstructions and the curvature of the earth, some of them familiar to hams and some hardly mentioned at all in ham technical literature. Physicists, engineers, and technicians who study and use these phenomena refer to them as "propagation modes." Historically, a number of different modes have been put to work to communicate with ships at sea and the coastal stations that serve them.

In our fictional universe, Grantville arrives with no capability to fabricate radio parts, and nobody who undertstands the design and manufacture of tubes and transistors. Up-timers and down-timers together are forced to create an electronics industry from scratch. It gets on its feet in the years from 1633 to 1638 or so, building an electromechanical alternator transmitter in 1633 and the first all-new vacuum tube in late 1634. Some of the earliest radio equipment the embryonic industry will be able to put into production will turn out to be ideal for modes that work well across salt water. If the hams themselves overlook some of the possibilities, the scholars in the libraries will come across clues, and the engineers and scientists coming out of the private study groups and the new technical colleges will mathematically deduce the existence of some of the most useful modes from the equations of electromagnetic theory.

In this article we will examine four propagation modes that our fictional friends should be able to put to work during the Maunder Minimum, when ionization is weak in the upper ionosphere. We will discuss the frequency ranges where those modes work best, the variables that influence their performance, and the equipment needed to use them. Then we will work out the numbers for several receiving locations and radio paths of interest to our friends.

Because the focus of this article is marine radio and the point-to-point links that connect shore stations into worldwide message-handling nets, the calculations will be based on the requirements for reliable communication by Morse code at commercial speeds.

We'll look first at ground wave, which makes use of the electrical properties of salt water, and isn't affected by solar activity at all. We can calculate the ranges achievable with this mode with reasonable accuracy, using published data in the Radio Propagation Handbook.

Next, we will turn to earth/ionosphere waveguide mode, used for long-range communication in the very earliest days of radio. It does use the ionosphere, but at very low frequencies, where strong ionization isn't required. Historical data indicates that it functions well even with weak ionization. This mode, too, is highly stable, and mathematically predictable.

Following that, we will turn to night-time sky wave at 500 KHz and 150 KHz. While sky wave is highly variable, extensive data in Naval Shore Electronics Criteria: VLF, LF, and MF Communication Systems and in a BBC report on propagation at 150 KHz gives us reasonable confidence that even the quiet-sun ionosphere would support skip on these lower bands starting a couple hours after sunset. Anecdotally, the "NMO Report," reminiscing on the experiences of operating at the old Coast Guard station in Honolulu, says that the whole Pacific typically opened up on 500 KHz around 9 PM every night. That station operated for decades, through many sunspot cycles.

Finally, we will examine moonbounce, which means just what it says: long-distance communication by bouncing radio waves off the Moon, as described in the VHF/UHF Handbook. Like ground wave, it makes no use of the ionosphere. Unlike the other modes, it operates well above the frequencies where atmospheric radio noise is significant, and for that reason is the most reproducible and predictable of the four.

Each of these modes has its strengths and weaknesses for our purposes. Ground wave doesn't offer the longest ranges, but it is dependable, and its antennas and transmitters fit well on shipboard. Sky wave in the lower frequencies is useful on shipboard and has greater reach than ground wave, and offers considerably more range between shore stations where antenna space isn't as limited. Earth-ionosphere waveguide mode offers full-time worldwide communication, but requires truly colossal antenna structures occupying large tracts of land, and would probably not be affordable in money, material, or logistics within the first decade after the Ring of Fire—even though the transmitter technology is easily within reach from the earliest days. Moonbounce offers reliable scheduled communication to distant parts of the world, but requires higher frequency tubes than the electronics industry will be able to produce during its first few years. Its highly directional antennas require clockwork to track the Moon, and for that reason would be very difficult to use on shipboard.

This article will not cover propagation modes that are useful primarily at shorter ranges during the Maunder Minimum, valuable as they are for other purposes. We will also limit discussion to the marine and fixed services, and not cover the uses of these modes in other services, such as broadcasting or land mobile. Regretfully, we must also pass by communication satellites, ideal as they are for the purpose, because it's likely to take well over a century to conjure up the vast array of materials, technologies, and engineering specialties needed to lay the groundwork for such an undertaking.

Antenna theory and radio propagation physics are intensely mathematical subjects. Detailed analyses of how these things actually work take up entire books packed with dense mathematics, and still some matters have to be left to actual measurement. We will skip all that, and present only some simple calculations and conclusions relevant to the matter at hand: basic communication by Morse code across salt water.

 

Terminology

 

A few terms of interest, before we move on:

VLF: Very Low Frequency, 3 to 30 KHz, 100 kilometer down to 10 kilometer wavelengths

LF: Low Frequency, 30 to 300 KHz, 10 kilometers down to 1 kilometer

MF: Medium Frequency, 300 KHz to 3 MHz, 1 kilometer down to 100 meters

HF: High Frequency, 3 to 30 MHz, 100 meters down to 10 meters

VHF: Very High Frequency, 30 to 300 MHz, 10 meters down to 1 meter

UHF: Ultra High Frequency, 300 MHz to 3 GHz, 1 meter down to 100 millimeters

RF: Radio Frequency, radio waves or equivalent electrical signals of any frequency

MUF: Maximum Usable Frequency, the highest frequency at which a radio wave meeting the ionosphere is bent sufficiently to return to earth. MUF varies widely with solar activity and time of day. In the Maunder Minimum it is often below the lowest ham band.

Bandwidth: The width of a signal in terms of the range of frequencies it occupies, as a result of the information being transmitted; or the range of frequencies a particular filter or other circuit will pass

For example, a human voice containing audio frequencies up to 3 KHz will create a radio signal at least 3 KHz wide (or considerably wider, depending on the method of translating it to a radio signal). A single-sideband voice signal at a nominal frequency of 3980 KHz would actually occupy 3980 to 3983 KHz.

CW: Continuous Wave, the type of clean signal produced by vacuum tubes and RF alternator transmitters, as opposed to interrupted wave, produced by spark transmitters. Radio people usually mean Morse code transmission when they speak of CW, even though nearly all modern transmission methods are actually based on continuous waves. CW makes far more efficient use of power and bandwidth than interrupted wave.

Wavelength versus frequency: Wavelength is inversely proportional to frequency; in free space the product of the two is equal to the speed of light, approximately 3×108 meters per second.

Isotropic antenna: An antenna that radiates equally in all directions in three-dimensional space

It doesn't actually exist, but is a convenient mathematical reference for comparing real antennas.

Omnidirectional antenna: An antenna that radiates equally at all points of the compass

Directivity: The degree to which an antenna radiates or receives in a preferred direction

Horizontal and vertical directivity may or may not be the same.

Antenna pattern: A polar plot of the radiated power density versus azimuth or elevation angle

There are many kinds of antennas with many pattern shapes.

Aperture: The effective electrical capture area, expressed by the ratio of the antenna's received power to the incoming wavefront's power density

The aperture may be very different from the physical dimensions of the antenna elements.

Antenna gain: The ratio of the power density in the preferred direction to the power density of some reference antenna, such as an isotropic antenna or a half-wave dipole.

Gain may be expressed as a numerical ratio, but is usually expressed logarithmically, in db.

Loss: The ratio of the power or power density at the output of some lossy medium, such as a propagation path or a cable, to the power or power density at its input.

Loss may be expressed as a ratio, but is usually expressed logarithmically in db.

Decibels or db: A logarithmic way to express a gain or loss ratio P2/P1

G=10Log10(P2/P1)

The db form of expression is convenient for propagation calculations, because gains and losses in logarithmic form can be added up algebraically, instead of multiplying very large and small numbers. Gains are positive, losses are negative. Absolute power levels can be expressed as db relative to some reference level, such as a milliwatt or the thermodynamic noise floor of a reference antenna.

dbm: decibels relative to 1 milliwatt

Reciprocity: A theorem stating that an antenna's transmitting and receiving patterns are identical

S/N: Signal-to-noise ratio, usually expressed in db

OTL: Old Time Line or Original Time Line, real history up to April 2000

NTL: New Time Line, fictional history beginning in May 1631

 

Ground wave propagation

 

Ground wave was one of the earliest modes to be understood mathematically. It is an excellent workhorse for reliable two-way communication across medium distances on salt water, and it always works regardless of what the sun is doing. What it won't do is carry us across oceans at manageable power levels.

Let's look at how this mode works.

If a radio wave is traveling tangent to a smooth conductive surface, it induces electric currents in the conductive medium, which interact with the wave and cause it to become focused onto the surface and follow it. If the surface isn't perfectly conductive, and of course no part of the earth's surface is, kilometer by kilometer the medium absorbs a small fraction of the wave's power and converts it to heat. Propagation loss is built in.

Thus, there are two effects that decrease the signal strength as the wave travels further from its source. In free space, simple geometry and conservation of energy tell us that the area of the expanding spherical wavefront is proportional to the square of the radius (the distance traveled from the source), so that the power density (power per unit area) decreases proportionally to the inverse square of the distance. But the ground wave medium is not free space. Because the wave travels in a lossy medium, the power density decreases as a decaying exponential function. So even as the wave spreads out in space, it loses part of its power as it advances.

The practical result of these two mathematical functions multiplied together is a moderately hard limit to the distance at which any particular ground wave signal can be detected above the atmospheric background noise. Beyond that distance, the power required to push it further becomes unreasonable very quickly. The usable distance depends on the transmitter power, the frequency, the transmitting antenna's efficiency and ability to launch its power into the ground wave, the electrical conductivity of the surface the wave travels over, the receiving antenna's properties, the natural atmospheric noise power density at the receiving location, and the receiver bandwidth. We will take these factors one at a time.

There is a good deal of published information on the performance of ground wave propagation, particularly in the MF broadcast band. AM broadcasting stations use this mode for reliable daytime operation over land across hundreds of miles, using fairly high power levels. Performance across land depends heavily on the quality of the earth over which the signal passes; there are performance curves for a range of earth conditions.

The interesting thing about salt water, though, is that it's about 500 times as conductive as "good earth" and 5000 times as conductive as "poor earth." Ground wave propagation is uniquely suited to the needs of the marine radio service.

In the frequency allocation tables of our OTL world, a number of bands in different parts of the spectrum are assigned to the marine service, each suited to particular propagation modes. Those allocations are made for good reasons, and most of the reasons will apply even during the Maunder Minimum. It is reasonable to expect the NTL band allocations to resemble the ones we know, though the HF marine bands may be somewhat neglected until the skip comes back.

It turns out that three major practical considerations influence the choice of frequency for a particular communication requirement: path loss, natural noise, and antenna size as a function of wavelength. The bands of most interest for salt water ground wave are in the neighborhood of 150 KHz, 500 KHz, 2 MHz, and 4 MHz.

Ground loss increases with frequency, but noise decreases with frequency in nonlinear ways and varies wildly according to receiving location, time of day, and season.

The size of a transmitting antenna of any one type is proportional to the wavelength, and therefore inversely proportional to frequency. This becomes a major cost and construction obstacle at the lowest frequencies we'll consider here. The shortest straight antenna that will resonate by itself and operate efficiently without external additions is a quarter wavelength tall, mounted on top of a conductive ground plane consisting of 100 or so radial wires that reach out at least a quarter wavelength horizontally in every direction. Of course, the space available aboard a ship is much more limited than the space at a large shore station, and the sea itself must substitute for the much more conductive copper-wire ground plane, so this puts ships at a disadvantage if they want to transmit on the lower bands. Electrical power and space for large, heavy radio equipment also tend to be more limited aboard ships than on shore.

So the tradeoffs get complicated.

 

Earth/ionosphere waveguide propagation

 

At sufficiently low frequencies, the D layer of the ionosphere becomes a reflector instead of an absorber. This is important during the Maunder Minimum because the D layer, being generated by the sun's ultraviolet light instead of charged particles, is much less dependent on sunspot activity than the E and F layers responsible for HF sky wave reflection. Regardless of sunspots, at VLF a radio wave becomes trapped, bouncing back and forth between the conductive surface of the earth and the lower ionosphere, doing most of its traveling through the free space in between, with far lower loss than ground wave offers. Because the lower the frequency, the deeper the RF current penetrates into the ground and therefore the larger the cross-section of the current-carrying layer, ground quality becomes much less important at VLF, and the wave bounces off poor earth almost as readily as sea water.

The sweet spot for this mode is the 10 to 30 KHz VLF region. The attenuation is astonishingly low, in the range of 1.5 to 2 db per 1000 kilometers. Furthermore, the power density doesn't decrease with the inverse square of the distance as it does in free space, because once it gets far enough from the transmitter to fill the height between the earth and the ionosphere, only the horizontal width of the wavefront is able to increase further. But because the earth is spherical instead of flat, the perimeter of the wavefront ceases to increase as the wave gets a quarter of the way around the globe from its origin. After that, the wavefront perimeter shrinks, as it focuses down toward a point halfway around the world from the origin, and the power density suddenly increases again.

With the right station design and enough power, full-time worldwide communication is possible.

Unfortunately, the noise level is very high at such low frequencies, and the wavelength is far too large for conventional self-resonant antennas to be feasible. Even though the tallest feasible transmitting antennas are drastically shortened in the mathematical sense, and suffer in efficiency because of it, they're still enormous structures occupying large tracts of land, and require very high power transmitters to realize their full potential.

Furthermore, the practical problems of shortening reasonably efficient VLF transmitting antennas to dimensions even a wealthy national government can afford result in very narrow bandwidths, which limit communication to Morse code or slow-speed teletype. On the other hand, one of the simplest ways to reduce the received noise and so reduce the required transmitter power is to use a narrow-bandwidth filter in the receiver—which limits communication to Morse code or slow-speed teletype.

VLF receiving antennas, though, can be very small, because there is no need to make them efficient. Once the antenna picks up more noise from the natural atmospheric background than the receiver's front end generates, signal-to-noise ratio does  not benefit from making the antenna any bigger. For example, I have an electronic clock with a tiny internal loop antenna a couple of inches square, which is is able to receive daily time signals from a 60 KHz station 1800 miles away. There is a benefit to making the receiving antenna directional, though. A directional receiving antenna can increase the signal strength without increasing the noise; however, a directional VLF receiving antenna is typically a pole-supported wire 0.5 to 2 wavelengths long. At 30 KHz, a 2 wavelength wire is 20 kilometers long—another real-estate hog, feasible only for well-funded shore stations in sparsely populated places.

With this mode, shore stations can transmit one-way to distant ships, or they can communicate among themselves to provide a "fixed service" linking widely separated shore stations into a worldwide net. Some real-world stations in the 15 KHz region are used as much to communicate with the far side of the world as to reach ships in their own neighborhood.

The hams may know about ground wave, but they almost certainly know nothing of VLF waveguide mode. Possibly there may be some clues in the older encyclopedias, or the radio people may come across it eventually while exploring the spectrum.

Of all the modes that could potentially provide worldwide communication, this one is the least demanding of advanced technology. It can even be done without vacuum tubes. In our own history, specially designed AC generators were used in this band for decades as mechanically driven high-power RF sources. In fact, there is still one transmitter of this type preserved as a historical exhibit in Sweden, and it comes on the air once a year.

In our NTL, an MF alternator was developed in 1633 for the Lutheran broadcasting station WVOL in "Canst Thou Send Lightnings?" Once developed, that technology could be applied very quickly to high-power VLF Morse code transmitters.

The question of when in the NTL this mode could be put to work to bridge oceans and continents comes down to when it will become feasible and affordable to obtain the materials and do the large-scale construction for the enormous transmitting antennas, when sufficient cargo capacity will be available to ship the staggering quantities of material and construction equipment to their destinations, and when suitable sites can be secured amid the NTL political conditions. To get an idea of the scale of a no-compromises VLF transmitting antenna, consider NAA at Cutler, Maine. Each of the two antennas has a central radiating element 900 feet high, a wire top hat consisting of six diamond-shaped petals 3000 feet long supported by a ring of six towers at the ends and an inner ring of six towers supporting the wires at mid-span, and a set of buried radial ground wires running 3000 feet out from the base. The site is built on a peninsula surrounded by sea water on three sides, with the ground plane wires distributing the current into salt water at their ends. The tuned circuits that match the transmitters to the antennas fill good-sized buildings in themselves. With all that care in engineering and construction, the efficiency of these antennas is just 50%. Smaller antennas for VLF can be built, even without top hats, and they will radiate. Just not as efficiently.

One crucial difference between the earliest days of radio in the OTL versus the NTL is that in 1900 the steel industry had decades of growth behind it and was keeping up with demand, and there were railroads everywhere, 10,000 ton steamships capable of bringing massive quantities of construction material to any part of the world within weeks, and mature port facilities almost anyplace a major coastal station would be desired. None of that infrastructure is in place in our 1630s.

VLF stations could still be built in the NTL, but probably not big enough ones to cross oceans within the first decade. Still, calculation of the power required for radio paths of interest is worthwhile to show what can be done when the time comes, if our friends ultimately decide on that approach.

 

Sky wave propagation

 

Sky wave is the night-time workhorse in the 500 KHz marine band and the LF region, from a few hours after sunset until sunrise. It comes alive soon after sunset reaches the place in mid-hop where the wave needs to reflect off the ionosphere. At that time the D layer fades and the signal can pass through to the higher layers, where it gets bent back down toward the ground. As noted earlier, the absence of sky wave during the Maunder Minimum in the bands the hams know, which are all above 1.8 MHz, is a severe bottleneck to long-range communication, especially in the first few years of the NTL.

But Rick and I believe there's good reason to expect that the weaker ionization levels in the F layer during those decades would still be sufficient to get usable skip in the LF region and the lower end of MF. The navy manual, the NMO report, and the BBC measurements covering the declining years from maximum to minimum sunspot activity, all appear to bear this out. With MF or LF sky wave, reliable night-time transoceanic communication would be possible at manageable power levels. While these bands call for transmitting antennas far too large to fit in a ham's back yard, and hence the up-time hams have little or no experience with them, an LF or MF antenna and the land to build it on are affordable for a navy or a commercial radio communication service.

As we shall see, ground wave even at 150 KHz doesn't get us across the Atlantic, except at totally insane power levels. Earth-ionosphere waveguide mode can go right around the world, but only at such low frequencies that it may not be possible to construct adequate transmitting antennas for a good many years. Moonbounce is a possible alternative for transoceanic communication, but it requires more advanced tubes than Grantville's electronics industry will be able to produce in its earliest years.

Until then, night-time sky wave at 500 KHz and below looks like the best chance for predictable communication at ranges over 1000 kilometers. Published data shows a nightly opening lasting up to 8 hours. A single hop covers up to 2000 km; longer distances require multiple hops. When the intermediate ground reflection is off sea water, the losses are low enough to support paths of two or three hops in these bands.

 

Moonbounce propagation

 

This mode was never widely used in our world for marine or fixed-service communication. To make it work, both stations need a clear line-of-sight path to the Moon. This means the duration of the daily opening depends on their difference in longitude, and it moves approximately an hour later each day as the Moon proceeds along its orbit. By modern standards, this makes for an unattractive traffic scheduling problem for a commercial or military communication service. The U.S. Navy has made limited use of moonbounce at sea, but only for special purposes. On the other hand, for our NTL friends where the alternative is perhaps a couple of years to carry a letter to the far side of the world if it gets there at all, a sliding daily communication window may not look so bad.

One very attractive feature of moonbounce for our 17th century friends is the logistic simplicity of delivering the hardware to a remote part of the world. A complete station, electrical power source and all, will stow in the hold of a 200 ton sailing ship with plenty of room left over. Every colony expedition can bring along a disassembled long-range station, capable of communicating with the outside world. Most of these stations would be able to reach home directly; only the most distant parts of the Pacific would need a relay through an intermediate longitude.

 

Important coastal station sites

 

When we design a low or medium frequency radio network to deliver a commercial or military communication service, an acceptable level of performance is generally considered to be full-time operation at 25 words per minute. Failing that, the service should at least operate on a dependable daily net schedule.

To accomplish that, our calculations for transmitter power and antenna design must begin with the natural noise level at the receiving site. (See Signals, Noise, and Bandwidth, below.) Some of the desired places are obvious from canon alone; others can reasonably be extrapolated from canon and forum discussions. Having identified the receiving locations, we turn to the worldwide radio noise maps to calculate the required signal strength.

For best coupling to salt water ground wave paths, we want the transmitting antenna in a wet sea-side location with unobstructed views to the horizon in all the desired directions, on which to lay out a copper ground plane to distribute the antenna current. Exact site selection requires local surveys; here we'll just talk about approximate geographic locations.

The USE's closest political and economic associates consist of Scandinavia and the Netherlands. Where maritime commerce is concerned, that community's back yard consists of the Baltic, the North Sea, the English Channel, and the waters around Norway and Iceland. Those are the places where they'll want radio coverage first.

Eric Flint has reserved fictioneering in the New World for himself and the mainline novelists, and has allowed only a few stories to be set there. We don't know what's going to happen in the western hemisphere, but we know something is. Where trade and politics go, so goes commercial shipping and the navy; we can be reasonably sure there will be a need for rapid transatlantic communication sooner or later. We can make some educated guesses as to likely locations. Contingency planning will be under way very early; some signs of this are already in canon. There may also be things happening in the Mediterranean. There, Venice is a possible friendly power.

Let's begin with the Baltic net. With RF alternators, an LF or MF ground wave backbone could possibly be operating in time for the Baltic War in 1634.

Rügen: The Baltic net's home station. This is about as far east as the coast goes, while still falling comfortably within USE territory. The great circle path to Stockholm passes over land for a good part of the way, which is a problem, since Stockholm is the most important place in the Baltic that the USE needs to communicate with. But there's a salt water path to Gotland.

Stockholm: Of major importance for coordination with the Swedish military and government.

Gotland: Needed as a relay point to achieve continuous operation on ground wave, because the loss over the land portion of the Rügen-Stockholm path is too high without sky wave, and the shallow diffraction angle around the Kalmar peninsula can't be depended on for a useful way around. Also useful as a coastal station midway between Rügen and Stockholm.

Ingria: Important for trade and commerce with Muscovy, shipping in the Gulf of Finland, and naval operations in the area.

Kemi Island: Supposedly there is chromium ore there. An expedition has been searching for it since 1633. We have yet to hear the results, but support communication with Stockholm would be good to have.

Between these backbone sites, coastal stations will eventually be filled in to provide service to shipping and the ports themselves; also to communities near the coast. A station at the mouth of the Kymi River in southern Finland has been mentioned.

Next, we have the North Sea net.

Ritsenbuttel (modern Cuxhaven): The USE's home station in the net, at the mouth of the Elbe.

Harlingen: USE naval base, call sign WNP.

Vlissingen, Netherlands: A collaboration between the Netherlands government and the USE navy, call sign PBN. Besides covering the southwestern North Sea, its location provides a straight shot down the English Channel out into the Atlantic. Also has a good path to Ireland, with only a short land hop across southern England. These paths make it the likely home station for the Atlantic net, and of major strategic importance to the USE and Dutch navies. This will become a big station.

Stavanger, Norway: Actually on the coast a few miles west, call sign NOK. A collaboration between the USE and the Kalmar powers. Local service, and backbone relay service between coastal stations along the German and Dutch coasts whose direct paths are blocked by the conformation of the land.

Faeroe Islands: Can be considered part of both the North Sea net and the Atlantic net. Provides relay service to the Norwegian coast beyond Bergen, and to Iceland. Also offers regional coverage of the northernmost parts of the Atlantic.

This brings us to the Atlantic net itself. As noted above, it begins with Vlissingen. The ideal place for the next station in the net is Clear Island, Ireland. An alternate route from Ritsenbuttel through the Faeroes is less satisfactory, and lacks a shore station covering the western approaches to the English Channel.

Cape Clear, Ireland: May or may not get built, depending on how diplomacy and events in the British Isles turn out, but the USE navy is very interested in this site. It offers clear salt water paths well out into the Atlantic along the major sailing routes, and can reach across the Atlantic by sky wave to Cape Race at reasonable power levels. There was a hush-hush mission in 1635 to assess the site's possibilities, with favorable results ("A Friend in Need.")

Suriname: The European colony there is probably the most secure and defensible friendly territory anywhere in the western hemisphere, and it's within sky wave reach of Cape Race westbound, and Cape Clear eastbound. It's the logical place for a large, fixed installation capable of communicating between the West Indies and Europe. Unfortunately, the high noise level at that location requires strong incoming signals, presenting a challenge to the network planners.

Unspecified locations in the West Indies: There has been much discussion of natural resources and political goals in this region. There may be operations on both land and sea. Because of high noise levels and practical limitations on the size of temporary installations that could be erected in this part of the world, communication outside the region may need to be relayed through Suriname.

Tampa, Florida: Tampa Bay offers about the only decent sheltered anchorage anywhere along the southeast mainland. In the 1630s, the local area is completely uninhabited, though the Spanish would certainly be concerned if foreigners began operating there and putting up massive installations. It is, however, on the route to the oilfields of Texas, the mineral resources of Alabama, and the mouth of the Mississippi. The noise level is, if anything, even higher than in Suriname and the West Indies, but at least there's room to build big.

Cape Race, Newfoundland: Useful to the new Danish settlements in eastern Canada, and a likely relay point between Europe and Suriname. Also a good location for a coastal station along the main eastbound sailing route. It's much closer to Suriname and the West Indies than any other friendly territory, which is important because of the high noise levels in those areas.

 

Signals, noise, and bandwidth

 

The Radio Propagation Handbook contains worldwide radio noise maps and supporting frequency curves, from CCIR Report-322, 1963. These present the noise in the form of noise power density at the antenna terminals, relative to a theoretical thermodynamic noise floor, measured in power per unit bandwidth (watts per hertz). Surprisingly, there is no factor in this relationship for the antenna gain. What this implies is that because antenna gain is directly related to directivity, the received noise is constant in an environment of uniform noise coming from all directions, with noise rejected from the deaf side of the antenna compensated by extra noise received on the more sensitive side. There's probably a theorem about that, but it doesn't appear in the book.

The ground wave propagation curves, on the other hand, present the signal strength arriving at the receiving antenna in volts per meter. The signal power collected by the antenna depends on both signal strength and gain.

In order to calculate the signal-to-noise ratio presented to the receiver, it's necessary to put signals and noise into the same units. Since there's a simple relationship between signal strength in volts per meter and the wavefront's power density in watts per square meter, and because the effective aperture is well known for various types of receiving antenna, the received signal can easily be converted to watts. In this article I'll simply present the converted results for both noise and signal, in the form of db relative to 1 milliwatt (dbm) in a bandwidth of 100 Hz. Following that, I'll discuss design tradeoffs available to the station designers.

Receiver bandwidth is a critical parameter. Since the noise power delivered to the operator's headset is proportional to the bandwidth of the receiver's filters, digging a weak signal out of the noise requires that the filter be wide enough to pass the signal, and no wider. To transmit Morse code at 25 words per minute, which is the traditional speed required to pass the license examination for a commercial radiotelegraph operator, we need a rise time of about 10 milliseconds. That is, the transmitter must go from zero to full power in 10 mS each time the key is pressed, and return to zero in 10 mS when it's released. To reproduce that rise time in the operator's headset, the receiver's filter must have a bandwidth of at least 33 Hz. However, that's a bare minimum, and would make for a ringy-sounding signal. A filter 100 Hz wide would make for easier copying at 25 WPM, and allow a skilled operator with a typewriter to copy at as much as 50 WPM. 50 Hz would be a reasonably comfortable filter bandwidth for 25 WPM, and not pass a great deal more noise than the bare-minimum 33 Hz. A well-designed marine receiver should offer filters for both bandwidths. For simplicity of calculations, we'll use 100 Hz for noise bandwidth calculations, and note that the noise power could be reduced by half (3 db) when needed, by switching to the 50 Hz filter and slowing down the transmission. (That doesn't mean the filter passes a frequency of 50 Hz, it passes a band 50 Hz wide centered on the audio frequency the operator prefers for Morse code, for example 1 KHz. In that case a 50 Hz filter would pass 975 to 1025 Hz.) Any form of voice modulation would require much more bandwidth than this, hence much more power to overcome the noise.

The reason for choosing to express power levels in logarithmic form rather than absolute units is that propagation path losses and antenna gains are normally expressed in db, and this makes it simple to add up their contributions rather than multiplying very large and very small factors together.

So, for example, a power increase of a factor of 10 is +10 db, and a decrease in power by a factor of 100 is -20 db. If a certain propagation path has a loss of -130 db, the transmitting antenna has a gain of +4.3 db and the receiving antenna has a gain of +5.2 db, we simply add up these contributions and determine that the combined effect is a loss of -120.5 db from the transmitter's output terminals to the receiver's input. If, for example, the transmitter's output is 10 watts or +40 dbm, we can immediately calculate the receiver's input power as -80.5 dbm, and compare that to the noise within the receiving filter's bandwidth.

 

Atmospheric noise in a 100 Hz bandwidth at selected locations

 

At the frequencies we're considering for all but moonbounce, the earth's atmosphere is the dominant RF noise source. The noise is generated mainly by thunderstorms, primarily in the tropics and in some continental interiors. Lightning bolts act as both the transmitter and the antenna, generating brief high-energy wideband pulses, rather than statistically uniform "white noise." Advanced marine receivers can take advantage of this to clip off the high-amplitude pulses, somewhat reducing the noise energy presented to the filter. That level of electronic sophistication may arrive in the late 1630s.

Precisely because these bands offer the best long-distance propagation, they deliver the most atmospheric noise to distant parts of the world. And because loss on long-distance sky wave paths is lowest at night, when D layer attenuation is weakest, noise levels are higher at night in most places. Of course, local thunderstorms also generate intense noise, so there can be local radio blackouts. To actually determine the typical noise level at a particular receiving location, we must turn to the published radio noise maps.

Each entry below is calculated from the CCIR noise density map for summer, midnight to 4 AM, which is typically the noisiest time of day and season.

To a good first approximation, these can be considered "worst case" values. A signal that can be heard above that level of noise should get through at any time. The numbers are somewhat rough, having been read off a small reproduced map with a coarse grid.

Typically, the noise level is somewhat quieter or a good deal quieter for roughly half of the day, from 8 AM to 4 PM more-or-less. Unfortunately, there is no good rule of thumb for estimating the difference; each case must be worked out separately. Also unfortunately, the spread between noisiest and quietest times of day is narrowest at the lower frequencies, where the propagation losses are lowest and the longest ranges are possible.

 

Location

30 KHz

150 KHz

500 KHz

2 MHz

4 MHz

North Germany, Netherlands, Southern Scandinavia, Ingria 50°-60°N, 0°-30°E

-11 dbm

-42 dbm

-69 dbm

-87 dbm

-94 dbm

Cape Clear 51°N, 10°W

-18 dbm

-52 dbm

-79 dbm

-97 dbm

-103 dbm

Faeroe Islands 62°N, 7°W

-21 dbm

-60 dbm

-86 dbm

-107 dbm

-108 dbm

Suriname, West Indies 6°N, 55°W

-6 dbm

-32 dbm

-52 dbm

-70 dbm

-78 dbm

Tampa 28°N, 82°W

-6 dbm

-32 dbm

-52 dbm

-70 dbm

-78 dbm

Cape Race 47°N, 53°W

-12 dbm

-48 dbm

-71 dbm

-90 dbm

-98 dbm

Mediterranean 35°-45°N, 5°W-35°E

-9 dbm

-40 dbm

-67 dbm

-85 dbm

-92 dbm

Table 1: Noise levels at selected locations

 

Mid-ocean is somewhat quieter than the USE's home waters, by as much as 10 db at 500 KHz. The Azores, for instance, would be a good place to listen to distant signals, if anyone could get the use of a station site there. Actually, the Azores and the Cape Verde Islands would be good locations for shore stations to serve shipping along the westbound leg of the North Atlantic square rigger route. Bermuda would be a desirable location for a shore station between Suriname and Cape Race. The political status of those places hasn't been clarified in canon as yet.

 

Basic land transmitting antennas

 

The published ground wave signal strength curves are based on a "short vertical antenna installed over a good ground radial system." This will be our starting point for path loss calculations. In terms of antenna engineering, this means a vertical radiator 1/8 wavelength long or less, with an approximately uniform current distribution along its length. The ground radial system is a large number of wires spreading out from the base for at least a quarter wavelength in every direction; they may be buried or elevated. The transmitter appears electrically as a voltage source, connected between the ground plane and the base of the vertical radiating element. One common feature of large shore stations in the lower bands is a capacitive top hat, which is an array of wires stretching away horizontally from the top of the antenna in a symmetrical pattern, supported by additional towers located at a distance from the vertical antenna. This gives the current a place to go when it reaches the top of the antenna, helping to keep the current uniform along the length of the radiator. It also lowers the self-resonant frequency, simplifying the matching circuitry at the base, which is needed to transfer the RF power efficiently.

This is the basic low frequency transmitting antenna, used when the wavelength is too great for a full-size antenna to be technically or economically feasible. It produces an omnidirectional doughnut-shaped pattern, with its strongest signal toward the horizon in all directions, falling off with increasing elevation angle to near-zero straight up.

Compared to an isotropic antenna, it has a gain of +1.76 db. This means that the signal in the strongest direction (toward the horizon) is 1.76 db stronger than the same transmitter power would produce, if fed to an ideal isotropic antenna. The gain of real antennas is most commonly specified relative to an isotropic antenna.

While the directional pattern of a short vertical doesn't vary much with length, and therefore the gain (and the aperture) is essentially constant, getting it to operate efficiently becomes more and more difficult, the shorter it gets in terms of wavelength. The practical problems of design and construction become quite challenging at the lowest frequencies.

The basic full-size vertical antenna is a quarter of a wavelength tall, mounted on the same ground plane of quarter wavelength radials. The quarter wave vertical is self-resonant, without the need for a top hat or a tuning circuit, although a passive circuit of inductors and capacitors may be needed to match its impedance to the transmitter. It has a lower radiation angle than the short antenna and therefore a higher gain toward the horizon: 5.14 db over an isotropic antenna, and therefore 3.38 db over a short vertical. High efficiency is much easier to achieve as well, with the full-size antenna.

The highest gain available in a one-piece vertical antenna occurs at a height of 0.64 wavelength. This is conventionally called a 5/8 wave vertical. It has a further 3 db gain over a quarter wave vertical. Since it's not self-resonant, it requires the minor complication of external tuning as well as a matching circuit. There is no point in making it taller than that unless it's to be broken up in mid-span with additional electrical components or phasing stubs, a more complex design known as a vertical collinear.

So which bands will force our NTL friends to rely on short verticals, and which ones will allow them to build full-size antennas which have higher gain and are easier to drive? That depends on whether the antenna is on a roomy shore site or on a ship, how tall a tower is affordable, and how tall the ship's masts are.

We know that the Great Stone Radio Tower in Grantville is something over 100 meters tall, but that was an expensive one-of-a-kind project, for deception purposes. The tallest radio towers in the OTL United States are typically about 500 meters high (the world record is a little over 800 meters), but that's with unlimited supplies of steel. Steel production is far short of demand in the NTL 1630s, but wood is comparatively easy to get, and early modern builders are intimately familiar with it. Based on historical wood lattice towers, let's say the practical limit in the early NTL is a guyed wood lattice tower 150 meters high. The actual current-carrying radiator would be a set of four to eight vertical wires, symmetrically mounted around the outside of the supporting tower and connected together at the base. If the base of the tower is mounted on insulators, it shouldn't be necessary to insulate the wires from the tower structure.

150 meters is the height of a full-size quarter-wave vertical at 600 meters, or 500 KHz. It would be 1/8 wavelength at 250 KHz, so it's a short vertical at all lower frequencies and requires a top hat for best performance in those lower bands.

A quarter wave vertical would be 37.5 meters high at 2 MHz, or just under 19 meters at 4 MHz.

A 5/8 wave antenna would be 96 meters high at 2 MHz, or 48 meters high at 4 MHz. These are certainly feasible on land.

With some antenna engineering tricks (stubs and tuned traps) a single tower can be constructed to operate as a quarter wave vertical on 500 KHz, a 5/8 wave vertical on 2 MHz, and a high-gain vertical collinear on 4 MHz, all simultaneously. A set of four carefully spaced towers might be driven as a steerable phased array on 500 KHz with a gain of 11 db in the preferred direction, and a short vertical on 150 KHz. The USE navy's engineers might master such tricks by perhaps 1640.

Any of these can be installed with buried or elevated ground plane wires, so as not to obstruct farming or grazing.

 

Basic shipboard transmitting antennas

 

The USE navy's smallest vessels are the Wild class courier schooners, with a mainmast standing about 19 meters above the waterline. The ground plane is the ocean itself, connected via metal contact plates on the outside of the hull. Using standard rigging techniques, it wouldn't be difficult to extend a non-structural metal radiator another 10 meters or so above the masthead, giving a practical limit of around 30 meters. So the mast height alone is sufficient for a full-size quarter-wave vertical at 4 MHz. It would just fall into the definition of a short vertical at 2 MHz, and could not be extended to a full quarter wavelength. The designers might opt for the extended masthead radiator anyway, to get a somewhat improved vertical pattern and easier matching at 2 MHz.

The proposed Fearless class escort schooner is nearly twice the size of the Wild class; its mainmast is tall enough for a quarter wave vertical at 2 MHz, or a 5/8 wave vertical at 4 MHz with a modest extension. Most down-time ocean-going ships would have masts this tall.

On all the lower marine bands, no ships are likely to have masts tall enough for anything but a short vertical. Few vessels of any era have masts much taller than 60 meters.

A minimal top hat can be installed on a three-masted schooner or ship, by running the antenna up the mainmast, and running cages of parallel wires from the top of the antenna to the tips of the foremast and mizzenmast. (While a single set of top loading wires can be run between the mastheads of a two-masted vessel, it would waste some of the power by radiating a horizontally polarized signal, because the horizontal current wouldn't be cancelled out by an equal current going in the opposite direction.)

Beyond the fixed vertical antennas that could be carried on the masts, a vessel could temporarily send up a full-length wire supported on a kite or balloon, when the weather is favorable. Marconi used kites for his transatlantic experiments in 1901.

 

Basic receiving antennas

 

While transmitting antennas are required to be efficient in order to deliver enough signal strength at the receiving site to overcome the atmospheric noise at that location, no such requirement applies to receiving antennas at the frequencies we're discussing here. Shortening a vertical antenna to tiny fractions of a wavelength, or making a small loop antenna, will reduce the received signal, but it will also reduce the received noise by the same proportion. As long as the noise coming from the antenna is much stronger than the unavoidable noise generated in the receiver itself, the signal-to-noise ratio remains unchanged. That requirement is trivially easy to meet at the frequencies we're examining here. (At higher frequencies where the atmosphere is much quieter, receiver noise dominates, and engineers put much more effort into optimizing receiving antennas and receiver front ends.)

Initially, most shore stations will probably use their transmitting antennas to receive. However, there are important benefits to separating the transmitter site from the receiver site by a few kilometers. A mature marine shore station typically has an assortment of receiving antennas of different types and orientations, usually connected to multiple receivers in the same band. The operators are located at the receiver site, and key the transmitter remotely. Among other benefits, this eliminates the need for mechanical antenna switching between transmitter and receiver, allowing the operator to hear distant ships whenever the key is up between dots and dashes; this is known as "full break-in" capability.

Basic receiving antennas are mostly omnidirectional. Small single-site shore stations would use greatly shortened antennas in the 30 KHz, 150 KHz, and 500 KHz bands, and perhaps 1/4 wave verticals in the 2 MHz and 4 MHz bands.

There would also be some rotatable shielded loop antennas on the roof of the radio shack, for direction-finding.

Ships are unlikely to install separate receiving antennas, other than rotatable direction-finding loops.

Directivity matters when correcting signal-to-noise ratio for receiving antenna gain using the path loss curves based on short verticals at both ends, because it increases the signal without increasing the noise. Receiving antenna efficiency can be ignored in these bands; even a very inefficient antenna will still deliver more atmospheric noise than the receiver generates internally.

 

Required signal-to-noise ratio

 

The Radio Propagation Handbook contains a table of recommended signal-to-noise ratios relative to the noise in a 1 Hz bandwidth, for different grades of service with several types of modulation used in commercial service. We'll confine our analysis to Morse code, because that requires much less bandwidth than any form of voice communication, and so demands much less transmitter power than any alternative available in the first decades of the NTL. For hand-sent Morse code the handbook gives +36 db for "operator-to-operator" service. But commercial Morse code implies a typical receiver bandwidth of roughly 100 Hz to handle a 25 WPM keying rate, not 1 Hz, so the noise power passing through the filter is 100 times greater, or 20 db stronger. Hence, the signal-to-noise ratio in the actual bandwidth needed to achieve that grade of service is +16 db. A good operator could copy through a somewhat worse signal-to-noise ratio, but it would be tiring and probably result in errors and dropouts.

The same table recommends +45 db above the noise in a 1 Hz bandwidth for "good commercial service," which would be +25 db above our calculated levels.

We'll use +16 db as the criterion for our basic power / path loss / antenna gain calculations.

 

Baseline case: LF/MF ground wave and VLF waveguide modes between fixed stations

 

As a starting point for further analysis, the calculations underlying the following table are based on the following conditions:

 

  • On the 30 KHz and 150 KHz bands the transmitting antenna is a well-constructed short vertical. Only shore stations transmit on these bands.
  • On the 500 KHz band, shore stations transmit with a quarter-wave vertical.
  • On the 2 MHz and 4 MHz bands, shore stations transmit with a 5/8 wave vertical.
  • The transmitting antennas are used for receiving.
  • Path losses on the 30 KHz band are derived from the equation for the earth-ionosphere waveguide mode ("WVG"). All other path losses are derived from the curves for the ground wave mode ("GND") at the applicable frequency; except as noted, all ground wave paths are over salt water all the way.
  • The criterion for successful communication is a +16 db signal-to-noise ratio in a 100 Hz bandwidth at the receiver terminals, after all gains and losses are accounted for.

 

Where the ends of a path are far enough apart to be at different noise levels, the direction is toward the noisier location, since that's the worst case.

Since the published propagation curves show the signal strength in terms of db relative to 1 microvolt per meter versus distance from a 1 KW transmitter using a short vertical antenna, we must first correct for the gain of the assumed transmitting antenna, then convert field strength to the power the receiving antenna will pick up at that wavelength, so that it can be compared to the received noise expressed in terms of power. For each path we will show signal strength directly from the curves, then the power received at that wavelength taking into account the transmitting and receiving antenna gains, the noise level, the resulting signal-to-noise ratio at that location, and finally the power needed to achieve a signal-to-noise ratio of +16 db during the hours of highest noise.

 

­Path

Value, Units

30 KHz
WVG

150 KHz
GND

500 KHz
GND

2 MHz
GND

4 MHz
GND

Rügen to/from Gotland, 472 km

E1KW, db µV/m

+59

+48

+47

+38

+30

Psignal, dbm

+14

-11

-18

-31

-45

Pnoise, dbm

-11

-42

-69

-87

-94

S/N, db

+25

+31

+51

+56

+49

P16db, W

125

32

0.32

0.1

0.5

Gotland to/from Stockholm, 203 km

E1KW, db µV/m

+63

+60

+59

+57

+55

Psignal, dbm

+18

+1

-6

-12

-20

Pnoise, dbm

-11

-42

-69

-87

-94

S/N, db

+29

+43

+63

+75

+74

P16db, W

50

2

0.02

0.002

0.002

Lübeck to/from Stockholm assuming "poor earth", 758 km

E1KW, db µV/m

+57

+18

-25

Off chart

Off chart

Psignal, dbm

+12

-41

-90

 

 

Pnoise, dbm

-11

-42

-69

-87

-94

S/N, db

+23

+1

-21

 

 

P16db, W

200

32K

5M

 

 

Vlissingen to/from Stavanger, 848 km

E1KW, db µV/m

+56

+40

+27

+12

-1

Psignal, dbm

+11

-19

-38

-57

-76

Pnoise, dbm

-11

-42

-69

-87

-94

S/N, db

+22

+23

+31

+30

+18

P16db, W

250

200

32

40

630

Cape Clear to Vlissingen, 907 km; about 225 km over land assuming "poor earth"

E1KW, db µV/m

+56

+33

+16*

+6*

Off chart

Psignal, dbm

+11

-26

-49*

-63*

 

Pnoise, dbm

-11

-42

-69

-87

-94

S/N, db

+22

+16

+20*

+24*

 

P16db, W

250

1K

400*

160*

 

Cape Clear to Cape Race, 3181 km

E1KW, db µV/m

+47

-27

Off chart

Off chart

Off chart

Psignal, dbm

+2

-86

 

 

 

Pnoise, dbm

-12 dbm

-48

-71

-90

-98

S/N, db

+14

-38

 

 

 

P16db, W

1.5K

250M

 

 

 

Vlissingen to Suriname, 7375 km

E1KW, db µV/m

+37

Off chart

Off chart

Off chart

Off chart

Psignal, dbm

-8

 

 

 

 

Pnoise, dbm

-6

-32

-52

-70

-78

S/N, db

-2

 

 

 

 

P16db, W

63K

 

 

 

 

Suriname to Vlissingen, 7375 km

E1KW, db µV/m

+37

Off chart

Off chart

Off chart

Off chart

Psignal, dbm

-8

 

 

 

 

Pnoise, dbm

-11

-42

-69

-87

-94

S/N, db

+3

 

 

 

 

P16db, W

20K

 

 

 

 

Cape Clear to Suriname, 6572 km

E1KW, db µV/m

+39

Off chart

Off chart

Off chart

Off chart

Psignal, dbm

-6

 

 

 

 

Pnoise, dbm

-6

-32

-52

-70

-78

S/N, db

0

 

 

 

 

P16db, W

40K

 

 

 

 

Suriname to Cape Clear, 6572 km

E1KW, db µV/m

+39

Off chart

Off chart

Off chart

Off chart

Psignal, dbm

-6

 

 

 

 

Pnoise, dbm

-18

-52

-79

-97

-103

S/N, db

+12

 

 

 

 

P16db, W

2.5K

 

 

 

 

Cape Race to Suriname, 4530 km

E1KW, db µV/m

+43

-55

Off chart

Off chart

Off chart

Psignal, dbm

-2

-114

 

 

 

Pnoise, dbm

-6

-32

-52

-70

-78

S/N, db

+4

-82

 

 

 

P16db, W

16K

6.3×1012

 

 

 

Suriname to Cape Race, 4530 km

E1KW, db µV/m

+43

-55

Off chart

Off chart

Off chart

Psignal, dbm

-2

-114

 

 

 

Pnoise, dbm

-12

-48

-71

-90

-98

S/N, db

+10

-66

 

 

 

P16db, W

4K

160×109

 

 

 

Table 2: Baseline case for LF/MF ground wave and VLF waveguide paths

* These mixed-path estimates are uncertain due to extrapolation outside the chart boundaries.

 

Ground wave ranges of ship transmitters

 

Some of those numbers are eye-openers. But before we proceed to various practicalities and ways of overcoming the difficulties, it's also of interest to run some calculations backwards, to see how far ships could be heard above the noise. They're more limited in antenna space and transmitter power than shore stations.

Starting in 1632, Grantville's newly founded electronics industry began building "tuna can" transmitters for the army, using transistors scavenged from dead up-time consumer electronics. These could probably generate a quarter watt or so, up to at least 4 MHz. They would be available in time for the founding of the USE navy, and as we shall see, would be far from useless in local waters.

Alternators wouldn't be very practical aboard small vessels, especially sailing vessels, so we'll ignore them for the moment.

In the vacuum tube timeline, it's estimated that a 25 watt transmitting tube good up to perhaps 10 MHz could go into pilot production toward the end of 1635. That would be suitable for a ship's transmitter running off batteries charged by a masthead wind generator, as depicted in "A Friend in Need." A simple 250 watt tube could be available not too long afterward; four of these in a push-pull/parallel combination would make possible a 1 kilowatt transmitter, which would be practical in ships with engines.

So let's look at how far a marine transmitter could be heard with these three power levels.

The following table assumes that:

 

  • On the 500 KHz band, transmission is with a short vertical.
  • On the 2 MHz band, the ship is large enough to mount a quarter wave vertical; for example, a Fearless class 100-foot escort schooner. Smaller vessels can use a short vertical, but we won't make that calculation.
  • On the 4 MHz band, all ships transmit with a quarter wave vertical. Higher vertical directivity could actually be detrimental during severe rolling and pitching.
  • Shore stations receive with their transmitting antennas, which have higher gain than ship antennas. It's more important for a ship to reach a shore station than distant ships.

 

Ereqd here is the calculated signal strength in db relative to 1 microvolt per meter, required to produce a +16 db signal-to-noise ratio in a 100 Hz bandwidth at the stated regional noise level, at the given frequency using the given antenna combination. With that, we can apply the ground wave curves to determine the maximum range at the stated power.

 

Region

Value, Units

500 KHz
GW

2 MHz
GW

4 MHz
GW

Baltic and North Sea

Pnoise, dbm

-69

-87

-94

Ereqd, db µV/m

+13

+4

+3

D1KW, km

1230

1050

960

D25W, km

840

625

640

D250mW, km

420

480

420

Suriname and eastern Caribbean

Pnoise, dbm

-52

-70

-78

Ereqd, db µV/m

+30

+21

+19

D1KW, km

820

760

640

D25W, km

485

540

450

D250mW, km

125

240

150

Mediteranean

Pnoise, dbm

-67

-85

-92

Ereqd, db µV/m

+15

+6

+5

D1KW, km

1200

950

950

D25W, km

800

750

620

D250mW, km

225

460

390

Table 3: Ship MF/HF ground wave ranges

 

There are some interesting observations here. We see that at the longer distances, where propagation loss limits the range, lower frequencies give the longest ground wave distances. At shorter distances, where atmospheric noise limits the range, higher frequencies deliver longer distances. And power requirements go up rapidly as distance increases; this is why modest power levels remain useful at moderate ranges.

 

Improving the situation

 

There are several variables station designers and operators could manipulate to improve a link, if the signal-to-noise ratio falls somewhat short of the minimum for successful communication. Some of these are easier to implement than others, and some could be put into effect sooner than others.

Suitable selection of the band is obvious, of course. More powerful transmitters, and the generating systems to run them, will come along in time. On links where full-time service isn't essential, ground wave and waveguide modes will face less noise in the daytime, especially in high latitudes; sky wave will benefit from vastly better propagation at night. Material, space, and technical manpower permitting, higher gain transmitting antennas will sometimes be enough to turn a marginal signal into solid copy. Highly directional receiving antennas can produce a considerably greater improvement than anything that can be done with the transmitting antenna at these low frequencies, and without the need for tall structures. If all else is insufficient, narrowing the receiver bandwidth and sending code much more slowly may sometimes be enough to make the difference.

These variables can be traded off against each other, of course. For example, a better transmitting antenna may make it unnecessary to install a high-power transmitter and find fuel to run its power plant. Or, if 500 KHz gives significantly lower ground wave loss than 2 MHz, is it better to tolerate the poorer gain and lower efficiency of the shortened 500 KHz shipboard antenna? In some cases it may be.

 

Time of day

 

As pointed out earlier, the baseline analysis was done with the noisiest time of day, to get an idea of what it would take to make communication work around the clock. Around 1 MHz, the noise can be as much as 20 db lower for several hours a day than it is during the early part of the night. A 20 db improvement won't radically increase the range, but it represents a 100:1 power reduction, enough to make an unreasonable power requirement reasonable, for some of the intermediate range cases. Unfortunately, that difference is much less in the lower bands where propagation loss is lowest, precisely because propagation of noise from distant sources benefits from that same low attenuation.

 

Directional transmitting antennas

 

It's possible to build more complex transmitting antennas on land, that use their power more efficiently by concentrating it in desired directions. Using a 5/8 wave vertical instead of a 1/4 wave, for instance, lowers the radiation angle and directs more of the power down on the horizon, where it couples to the ground wave more efficiently while still radiating omnidirectionally. It's also possible to stack antenna elements vertically to get more vertical directivity without giving up an omnidirectional pattern in the horizontal plane, but those configurations would be difficult to build much below 4 MHz.

By using more than one vertical antenna element, an antenna array can be given a horizontal pattern that concentrates the power in certain directions. Two towers spaced 1/4 wavelength apart and driven 1/4 wavelength out of phase will double the power density in the favored direction while canceling it 180 degrees away in the opposite direction. The power density is unchanged at the intermediate 90 degree directions. This would give a gain of +3 db relative to a single tower in the favored direction, and avoid wasting power in the inland direction. Three or four towers arranged in in-line or rectangular layouts can deliver a variety of horizontal patterns, depending on the geometry and how the phases of the currents driving the different towers are adjusted. The gain can be as much as +6 db relative to a single tower.

Another possible arrangement is an in-line array consisting of a single driven element and several non-driven parasitic elements; this is called a vertical Yagi-Uda array. It could be affordable at 2 MHz or higher, and deliver +8 db or more in a fixed direction.

 

Directional receiving antennas

 

The classic directional receiving antenna for MF down to VLF bands is the Beverage wave antenna, developed in the early 1920s. It consists of a horizontal wire supported on a line of poles, running above the earth from the receiver toward the oncoming signal, for a distance of 1/2 to 2 wavelengths. A common variant is two wires, one near ground level and one directly overhead at the top of the pole line. It behaves electrically as a leaky transmission line, magnetically coupled to the traveling wave. It receives in both endwise directions. Noise arriving from the undesired direction would reflect off the far end of the line and come back toward the receiver, except that the far end of the line is terminated with a resistor to absorb that unwanted power and dissipate it as heat. Dr. Beverage's experiments found, and later mathematical analysis confirmed, that performance doesn't improve at lengths greater than 2 wavelengths due to the difference in wave propagation speed on the line versus wave speed in free space.

Published papers I've found don't give numerical values for the improvement in received signal-to-noise ratio with this type of antenna. One ham reports an improvement over an omnidirectional antenna of +11 db. Dr. Beverage built arrays of as many as four elements across, spaced far enough apart not to overlap their apertures, and phase-combined at the receiver, giving an additional improvement of up to +6 db over a single wave antenna element. So we can estimate a possible improvement of +17 db over an omnidirectional receiving antenna at a large commercial receiving site.

The attractive thing about a wave antenna is that it doesn't involve tall towers or other large structures. It's a pole line, not much different from a power or telegraph line, except that it requires only one wire and a good ground connection at each end. Even at 150 KHz, it would be no more than 4 kilometers long. A single Beverage antenna could be installed at any coastal station at a very early date.

 

Improved case: LF/MF ground wave and VLF waveguide modes along some challenging paths

 

Now let's look at what might be done for a few of the more difficult long-haul point-to-point cases, by applying the possible improvements. Except as noted, we'll go with:

  • Four-bay Beverage receiving antennas at shore stations, theoretical S/N improvement +17 db
  • Two element cardioid transmitting arrays at shore stations on 150 KHz and up, +3 db gain
  • 50 Hz receive bandwidth, -3 db noise
  • Daytime noise levels fom the maps

 

Path

Value, Units

30 KHz
WVG

150 KHz
GND

500 KHz
GND

Cape Clear to Cape Race, 3181 km

E1KW, db µV/m

+47

-24

Off chart

Psignal, dbm

+17

-66

 

Pnoise, dbm

-13

-50

-81

S/N, db

+30

-16

 

P16db, W

40

1.6M

 

Vlissingen to Suriname, 7375 km

E1KW, db µV/m

+37

Off chart

Off chart

Psignal, dbm

+7

 

 

Pnoise, dbm

-8

-32

-52

S/N, db

+15

 

 

P16db, W

1.25K

 

 

Cape Clear to Suriname, 6572 km

E1KW, db µV/m

+39

Off chart

Off chart

Psignal, dbm

+9

 

 

Pnoise, dbm

-8

-32

-52

S/N, db

+1

 

 

P16db, W

800

 

 

Cape Race to Suriname, 4530 km

E1KW, db µV/m

+43

-52

Off chart

Psignal, dbm

+13

-95

 

Pnoise, dbm

-8

-32

-52

S/N, db

+21

-63

 

P16db, W

320

80×109

 

Table 4: Improved case for LF/MF ground wave and VLF waveguide paths

 

What these examples tell us is that getting across the Atlantic by ground wave is at the bare edge of possible, from Cape Clear to Cape Race on 150 KHz. In our own time, European broadcast stations in the 1 to 2 megawatt range do exist on that band. Doing it in the early NTL is highly unlikely; it could take quite a while to develop alternators to that power level, and tubes capable of that much power in a reasonable-size array are likely to be a good deal further off.

Getting by ground wave from there down to the West Indies, where we have reason to suspect there will be action in some impending novel, requires power levels that can be calculated, but are beyond absurd, even with fairly extreme antenna arrays.

Even a relay through Bermuda, assuming it were politically possible to secure a site there for a big station, wouldn't do the job at sane power levels, because the noise level is so high around the Caribbean.

Compared to that, VLF waveguide mode looks very attractive on the face of the numbers. A link directly from Vlissingen to Suriname requires only a modest 1.25 KW of radiated power at 30 KHz, assuming optimum antennas on both ends. And there lies the fly in the ointment. Even an austere transmitting antenna for that band requires a minimum of three towers 300 meters or so tall, to support a minimal top hat with the antenna wire hanging from the center. Four towers in a square with a radiating tower at the center would be a lot more like it. Towers that tall are almost certainly beyond the practical limits of wooden construction, which means they must wait for steel to become affordable, and a construction industry used to working with it.

Now, a 150 meter tower can be made to radiate on 30 KHz – just not efficiently. Recall that NAA, with thirteen towers per antenna, 900 feet tall, with the most sophisticated engineering and construction available in the mid-twentieth century, was able to achieve an efficiency of only 50%. A single 150 meter tower would lack a top hat, which means its current distribution would be far from optimum for a low radiation angle, and the electrical properties of such a short radiating element would be far from optimum for conversion of electrical power into a radio wave. Most of it would end up as heat in the ohmic resistance of the tower and the ground system. With a central tower to support multiple wires to carry the RF current with miminum resistance, and four surrounding towers to support a small top hat, we might possibly get an efficiency in the rangte of 5 to 10%. If it's no worse than 5%, Vlissingen to Suriname would be possible with 25 KW, not an unreasonable power level for a VLF alternator after a few years of development. That makes it possible to stay in touch with the European colony in Suriname without the uncertainty of getting the use of the Cape Clear site, or relying on putting in a station at the new Danish colony in Newfoundland — as desirable as those stations would be for regional service.

This, however, only takes care of the fixed point-to-point links, and possibly shore-to-ship downlinks. A 30 KHz transmitting antenna with enough efficiency to be useful simply isn't going to fit aboard a ship.

And so we come to sky wave.

 

MF and LF sky wave paths

 

There's a good deal less published data for sky wave in these bands than ground wave and waveguide mode. The sky wave data in the Radio Propagation Handbook is for 5 MHz and higher, where common carrier services operated. The navy manual has sky wave data expressed in the form of path loss between a pair of shortened verticals for 100 and 200 KHz; we could interpolate between them for 150 KHz, and extrapolate to 500 KHz.

For 500 KHz, the NMO report does tell us that the Honolulu MF station was able to provide reliable service to a broad swath of the Pacific; the most distant ships mentioned, off Japan, were 6500 km away. The navy manual shows that as a three-hop path.

We have to infer the typical power levels used and the antennas, since they weren't stated in that paper. The Maritime Radio Historical Society's restored coastal station KFS at Point Reyes, California operates 5 KW transmitters on 500 KHz and several HF bands. The restored Liberty ship S.S. Jeremiah O'Brien has a 200 watt transmitter on 500 KHz. The Queen Mary, by contrast, carried a 1.5 KW transmitter for 500 KHz. Even on twentieth century steamships, a set of radio masts wouldn't be much more than 50 meters above the water, because these ships needed to pass under harbor bridges. So we'd expect a shortened vertical less than 1/10 wavelength high, with a minimal fore-and-aft top hat strung between two or three masts. We should expect such a shipboard antenna to perform less efficiently than an ideal shortened vertical, because of the fairly minimal top hat and because of the need to use the ocean as the ground plane instead of a much more conductive set of radial copper-coated wires. A penalty of -3 db at the transmitting end might be in the right ballpark.

Let's begin by comparing what the published data predicts for 100 KHz and 200 KHz, received in Honolulu, with the NMO report. Unlike the regions we've been studying for early activity in the NTL, the central Pacific is fairly quiet at night, so it's a rather favorable case.

In the following table, Psignal assumes a transmitter power of 1 KW and a transmitting antenna loss of -3 db relative to a standard shortened vertical. The path loss numbers for 500 KHz are extrapolated, and so are uncertain.

 

Path

Value, Units

100 KHz
SKY

150 KHz
SKY

200 KHz
SKY

500 KHz
SKY

Japan to Honolulu, 6500 km, 3 hops

Apath, db

-97

-110?

-123

150?

Psignal, dbm

-40

-53?

-66

-90?

Pnoise, dbm

-46

-52

-61

-84

S/N, db

+6

-1?

-5

-6?

P16db, W

10K

50K?

125K

160K?

Table 5: Baseline case for Japan-Honolulu LF-MF sky wave

 

Clearly, some help is needed from the receiving antenna. Coastal stations do have large collections of directional receiving antennas. For quad Beverage antennas, these 1 KW S/N ratios would improve to +23 db, +16 db, +12 db, and +11 db, respectively. This still doesn't entirely square with the anecdotal NMO report, but it's not wildly in disagreement. It does imply that a direct jump across the Atlantic into the West Indies, with the high noise level there, would be overly optimistic. Let's look at Cape Clear to Cape Race to Suriname again, this time by night-time sky wave. These links are point-to-point between coastal stations, so we assume standard shortened verticals at 150 KHz and below, and a quarter-wave vertical at 500 KHz.

 

Path

Value, Units

100 KHz
SKY

150 KHz
SKY

200 KHz
SKY

500 KHz
SKY

Cape Clear to Cape Race, 3181 km, 2 hops

Apath, db

-80

-90?

-100

-120?

Psignal, dbm

-20

-30?

-40

-57?

Pnoise, dbm

-35

-44

-51

-81

S/N, db

+15

+14?

+11

+24?

P16db, W

1.2K

1.6K?

3.2K

160?

Cape Race to Suriname, 4530 km, 2 hops

Apath, db

-88

-98?

-107

-127?

Psignal, dbm

-28

-38?

-47

-64?

Pnoise, dbm

-25

-32

-39

-52

S/N, db

-3

-6?

-8

-12?

P16db, W

80K

160K?

250K

630K?

Table 6: Baseline case for LF/MF Atlantic sky wave paths

 

The numbers for the first leg, across the Atlantic to Cape Race, look fairly reasonable for alternators that could be built in the mid 1630s. The second leg is a problem; 160 KW is getting close to the largest RF alternators built in the twentieth century. Let's add a quad Beverage antenna at the Suriname station, aimed at Cape Race. The S/N increases by 17 db, and we get:

 

Path

Value, Units

100 KHz
SKY

150 KHz
SKY

200 KHz
SKY

500 KHz
SKY

Cape Race to Suriname, 4530 km, 2 hops

Apath, db

-88

-98?

-107

-127?

Psignal, dbm

-28

-38?

-47

-64?

Pnoise, dbm

-25

-32

-39

-52

S/N, db

+14

+11?

+9

+5?

P16db, W

1.6K

3.2K?

5K

12.6K?

Table 7: Improved case for LF/MF Cape Race-Suriname sky wave path

 

That looks doable. A single Beverage antenna could reasonably be traded off against 6 db more transmitter power; 10 kilowatts isn't unreasonable for an RF alternator.

 

VHF and UHF moonbounce

 

Unlike earth-ionosphere waveguide mode, the hams not only know about moonbounce, they know how to do it. Although the odds aren't high that any of Grantville's eighteen up-time hams have ever visited a moonbounce station, it's well-documented in the ham literature, and just about everyone has read about it or attended a presentation at a ham convention.

Unlike all the other modes we've examined, there aren't any comparative route calculations to do in moonbounce. All routes are of practically the same length, through the same medium, with the same path losses. The main difference between paths is that some pairs of stations are at practically the same longitude and would have daily openings lasting twelve hours or so, while others are too far apart in longitude to see the moon at the same time for more than brief intervals; thus, there would be a need for store-and-forward relaying through stations at intermediate longitudes.

There have been successful moonbounce contacts on every VHF and UHF ham band, from 50 MHz to 2.45 GHz. Atmospheric noise is virtually nonexistent at these frequencies; receiver noise and galactic noise dominate the S/N budget. According to the VHF/UHF Handbook, the bare minimum needed to make a marginal CW contact with a +3 db S/N ratio on the 144 MHz band and exchange call signs is a really good low-noise receiver, about 500 watts of transmitter power, and an array of four Yagi antennas of 6 meter boom length each. That's not a large structure at all.

Of course, that's far from adequate for a reliable commercial communication service. To raise that to the +16 db S/N desired for routine message-handling, we need to pick up another +13 db. Quadrupling the size of the antenna array to sixteen Yagi structures is good for +6 db in each direction, and increasing the transmitter power by +1 db to 600 watts would nominally get us there. Above 400 MHz a medium-size dish would replace the Yagi array and its complicated feed system.

However, there are practical problems. There are always practical problems in engineering.

The two most difficult components to provide would be the power amplifier tubes for the transmitter and the low-noise preamp tube for the receiver.

For the transmitting tube, there are at least examples in Grantville for the new electronics industry to reverse-engineer. The 6146 was a very popular ham transmitter tube from the late 1940s through the 1960s; there should be a number of them in Grantville. Some of them are probably still in service. The 6146 data sheet rates this tube in CW service at 70 watts at 60 MHz or 35 watts at 175 MHz. Thus, an amplifier built around an array of several 6146s can generate the required RF power in the VHF bands. There's reason to hope that Grantville's electronics industry will assign a high priority to the cloning of VHF tubes in the late 1630s, possibly around 1637 to 1639.

UHF power tubes are likely to be more of a technical challenge. There could be an example available for study of something like the 4X150A, good for 150 watts at 500 MHz, but it's less likely; therefore they'll probably come later.

The preamplifier tube for the receiver's low-noise front end is more problematical. Hams have relied on solid-state VHF and UHF preamps for decades; they're quieter than tubes. Our friends won't be developing any kind of transistors before the early 1640s, let alone super-low-noise gallium arsenide field-effect transistors. In the 1950s and early 1960s a favorite low-noise receiving tube for weak-signal VHF work around 150 MHz was the Western Electric 417A. Unfortunately, that was never a common tube, and there may not be any examples in Grantville. A more common receiving tube offering relatively good performance was the 6CW4. It was designed for TV receivers; there might possibly be examples in old TV sets. If not, this leaves Grantville's newly educated electronics engineers with two choices: either try to develop a suitable tube from first principles, or use more common types of pentodes that they do have examples of, and compensate for the higher noise by building bigger receiving antennas.

Of course, the high-gain antenna arrays needed to overcome the extreme path loss of a trip to the Moon and back are very directional; that's why they work. The antenna must track the Moon fairly accurately during the daily opening. Rather than steer the antenna manually, a commercial station would rely on a clock drive. Pre-RoF clockwork would be accurate enough for this purpose, but Grantville brings in examples of much more accurate mechanical clockwork, and the knowledge to duplicate it. This works fine for a fixed station, but tracking the moon aboard a rolling, pitching ship would have to wait for our intrepid engineers to master gyro-referenced servomechanisms, if the antenna array could even fit between the masts. So moonbounce would be used in the fixed service, linking the coastal stations to the rest of the world, leaving ground wave and MF sky wave to handle the actual ship communication.

Once the key components are available, design of the rest of the equipment requires only ham-level knowledge, and there's plenty of circuit design information in the literature known to exist in Grantville. The NUS and USE military services trained hundreds of radio operators in the early days of the effort to repel the marauding mercenary armies; a good many of them would have become experienced electronics technicians by the time VHF tubes appear.

All of this suggests a time frame of around 1638 to 1640 when it might be possible to build a working moonbounce station for the first time.

It's possible that not all moonbounce stations would be the same. Major stations might have much larger antenna arrays than sites in obscure parts of the world, with better receiving preamps and more powerful transmitters. Small stations with insufficient antenna gain and transmitter power to work each other could nevertheless be able to work hub stations such as Magdeburg or Suriname. This is a common situation with ham moonbounce stations today. Thus, the moonbounce net might not be truly peer-to-peer in its connectivity. This could affect story plotlines.

The cost of moonbounce stations might begin to come down in the early to mid 1640s, when the NTL electronics industry makes its first inroads into semiconductors. Vacuum tubes can never achieve the same noise levels as solid-state preamps, nor can they be cooled to cryogenic temperatures as semiconductors can. In OTL, there were parametric amplifiers based on special types of diodes before there were high-performance VHF and UHF transistors. They were covered in the ham literature of the 1950s and 1960s, which is very likely available in Grantville. Reducing the receiver noise to near the theoretical minimum improves the S/N ratio, so that less antenna gain is needed, saving on both structures and site costs.

 

Pushing beyond the boundaries

 

Up to now, we've been analyzing the possibilities in terms of commercial and military radio as they existed in the first three quarters of the twentieth century. A commercial overseas point-to-point circuit had to guarantee full-time service with outages of no more than a few minutes a year; otherwise it wouldn't make money. A marine shore station's most important duty was to hear an SOS day or night before the ship in distress sank and to get help on the way. So coastal stations were built big. My understanding of military communication grew from my experience as an electronics technician in Strategic Air Command, where a get-off-the-ground order had to set off alarms at every SAC base in the world within seconds, with no excuses for failure or delay, ever.

An engineer faced with requirements like that combines the most pessimistic estimates of every variable and calculates equipment and siting requirements from there—what we call worst-case design.

But two of our authors have presented plots that can tolerate far more austere grades of service, and can even build suspense around communication that barely functions, and then only intermittently. If we re-analyze some of the more challenging cases with a requirement that it must be guaranteed to work at least sometimes, rather than be guaranteed to function all the time, we may find ourselves getting through with much less powerful transmitters, available at an early date, much smaller antennas that can be set up quickly for temporary use, or even improvised, and by making the operators work a whole lot harder.

Hams will tolerate propagation modes so sporadic that it might take ten years to get a contact with a rare location for an operating award. Our colonists and sailors wouldn't be willing to go quite that far, but story situations have been proposed in which a station's operators struggle for a week to pass a 100-word message over a long, difficult path without the help of a relay station along the way.

So what assumptions do we change?

First of all, we calculate according to best case, rather than worst case. We pick the time of day during the sky wave opening when noise is typically weakest, rather than strongest. Rather than add the typical short-term variation to the average noise level, we subtract it. Now we get random openings of a few minutes or even a few seconds at a time, when the signal pokes its head out of the background hiss.

We design for a much worse signal-to-noise ratio. Where a good operator might be able to sustain 25 words per minute with a low error rate at +16 db, we might drop to a barely copyable +6 db, and accept that there are lots of miscopied letters and operators get tired fast. We get fragments, not the whole message in one unbroken string. So we transmit the message over and over for long periods, starting each repetition at precise intervals; a clerk at the receiving end can patiently piece together the random fragments from the many incomplete copies, and eventually recover the text.

We can slow down transmission to 5 or 10 words per minute, and use abnormally narrow filters to reduce the noise.

If there's plenty of time for advance planning, the expedition might be equipped with the very latest receivers, equipped with automatic noise blankers; the ham literature doesn't say just how much they reduce atmospheric impulse noise, but -3 db wouldn't be an excessively optimistic guess.

For a really long path, the transmitting antenna might be operated well below its design frequency, as a short vertical rather than a full-size vertical, trading off the poorer vertical pattern for much lower propagation loss. For example, a 500 KHz quarter wave vertical might be driven at 200 KHz, with temporary top hat wires running down at an angle to masts set up at a distance. In the field, where it's impractical to construct anything much taller than a ship's mast, kites or balloons might be used to loft a long vertical wire as a temporary transmitting antenna. Similarly, it may be impractical or impossible to put up a temporary Beverage antenna for receiving, but a loop antenna a few feet across will deliver at least some directivity.

Combining advance planning with lowering the operating frequency, a low frequency alternator might be configured for high power output, by stringing multiple off-the-shelf rotors on a single shaft and wiring them in series.

 

A specific case

One path that was specifically requested was Vlissingen to "someplace in the West Indies." Let's run a few numbers and see what it looks like. We'll go with:

  • Baseline transmit power 1 KW
  • Quarter wave vertical on 500 KHz, short verticals on lower bands
  • Loop antennas for receiving; +3 db S/N improvement over omnidirectional antenna
  • 8 PM noise map, subtract standard deviation
  • Success criterion is +6 db S/N in 25 Hz bandwidth; 5 WPM sending speed

 

Path

Value, Units

100 KHz
SKY

150 KHz
SKY

200 KHz
SKY

500 KHz
SKY

Vlissingen to West Indies, 6900 km, 4 hops

Apath, db

-103

-120?

-136

-160?

Psignal, dbm

-43

-60?

-76

-97?

Pnoise, dbm

-50

-58

-66

-85

S/N, db

+10

+1?

-7

-9?

P6db, W

400

3.2K?

20K

32K?

Table 8: A long-haul LF/MF Atlantic sky wave path, marginal communication criteria

 

Summing up

 

The NTL world's marine and fixed communication capabilities will grow over time, as the technology and manufacturing mature.

Local communication around the USE's home waters at distances up to 1000 km isn't too difficult. With modest power and reasonable shipboard antennas, ground wave communication will get the job done at any time of day or night. This could be well-established, at least for the allied navies, by 1635.

When commerce and naval operations start to reach overseas, things become more difficult. Propagation losses add up fast as the distance increases. VLF could reach anywhere in the world, but that requires construction on a scale that isn't likely to be possible in the first decade of the NTL, for a great many reasons beginning with steel production. Also, it will never be able to offer high-speed transmission or large numbers of channels. Night-time sky wave can span station-to-station paths of 5000 km or so, without exceeding the limits of wooden towers or the sizes of RF alternators that might be built in the mid 1630s. That means relay sites would need to be secured in friendly territory to deliver commercial-level service across the Atlantic and support shipping and naval operations in the areas where we can expect things to happen in the next few novels.

There is already canon involving Danish settlements in Newfoundland and Cape Breton Island. A station at Cape Clear is far less certain, but a relay through the Danish-controlled Faeroe or Orkney Islands would be an alternate way to reach Cape Race, though it would do nothing for shipping in the western approaches to the English Channel. Either way, by 1636 or so, an Atlantic net linking the European colonies to Europe would be within the bounds of possibility. There would be considerable incentive to consider diplomatic or military efforts (or massive bribery) to secure the use of additional sites such as Bermuda, the Azores, and the Cape Verde Islands to provide something close to continuous sky-wave radio coverage along the major North Atlantic sailing ship routes.

Beyond that, service to the Mauritius colony, the East Indies, and the Pacific will have to wait for VLF or moonbounce. 1638 is probably the earliest date moonbounce could begin to happen, followed by sailing time to transport the stations to the more distant parts of the world. VLF, if it ever comes, is likely to be a few years later, though it's the only solution potentially capable of offering world-wide full-time radiotelegraph service.

With all these modes used together, the world will start looking smaller in the NTL's second decade.

 

References

 

American Radio Relay League The ARRL Antenna Book, 13th ed. Newington, CT, 1974

American Radio Relay League, Radio Society of Great Britain VHF/UHF Handbook ISBN: 9781-9050-8631-3 Chapter 30 "Space Communications" http://physics.princeton.edu/pulsar/K1JT/Hbk_2010_Ch30_EME.pdf http://physics.princeton.edu/pulsar/K1JT/EME_2010_Hbk.pdf

Belrose, J.S. et al "Beverage Antennas for Amateur Communications" QST Magazine, January 1983 p.22 http://nrcdxas.org/articles/Beverage0183.pdf

British Broadcasting Corporation, Research Department "Low-frequency sky-wave propagation to distances of about 2000 km", Report No. 1971/8 http://downloads.bbc.co.uk/rd/pubs/reports/1971-08.pdf

Boatright, Rick "Radio FAQ Part 3: RF Environment" http://www.1632.org/1632tech/faqs/radio-rfe.html

Department of the Navy, Naval Electronic Systems Command Naval Shore Electronics Criteria: VLF, LF, and MF Communication Systems Washington: U. S. Government Printing Office, 1972 FSN 0280-901-1000 http://www.navy-radio.com/manuals/0101-1xx/0101_113-00.pdf through -08.pdf

Kraus, John D. Antennas New York: McGraw-Hill, 1950 ISBN 07-035410-3

Maritime Radio Historical Society "Reports from NMO" http://radiomarine.org/gallery/show?keyword=pt3&panel=pab1_8#pab1_8

Navy Radio web pages "NAA Cutler Maine – Navy VLF Transmitter Site" http://www.navy-radio.com/commsta/cutler.htm

Payne, Craig Principles of Naval Weapon Systems Annapolis, MD: Naval Institute Press, 2006 USBN 1-59114-658-5 books.google.com/books?isbn=1591146585

Saveskie, Peter N. Radio Propagation Handbook Blue Ridge Summit, PA: Tab Books, 1980 ISBN 0-8306-9949-X, ISBN 0-8306-1146-0 pbk.

W8JI "How Low-noise Receiving Antennas Really Work" http://www.w8ji.com/receiving.htm

****

Art Director's Note: I would like to thank Jack for finding many of the images for his article, much appreciated!Garrett