Perhaps one of the more intriguing uses for radio broadcasting is spying. Covert radio stations broadcasting coded number or letter groups have been the interest of SWL and others for years. The Conet Project sought to gather several of these recordings and make a CD out of them. What they ended up with is a rather spooky 4 CD set of various spy numbers stations throughout the world conducting their business which dates back to 1997. Since that date, samples of those recordings have been used in several movies and by recording artists.
While stationed on Guam doing important work for the government, we would often come across these numbers stations in the late 1980s. An East German numbers station was only 1 KHz away from one of our working frequencies, thus around 2 am local time, an East German lady would regale me with half an hour of five number groups in German, which is slightly off frequency, was utterly delightful. We knew where it came from because of this and others like it.
Sort of like that. That recording sounds like a computer-generated voice, ours was a real operator that would occasionally screw up.
The two most famous numbers stations are Cherry Ripe and Lincolnshire Poacher, so named after the songs they use for interval signals.
Both are allegedly off the air now, surely replaced with something else. When pressed as to the purpose of these stations, a British information minister replied “They are what you think they are.”
The numbered groups are coded groups meant to be received by agents in the field. The use of unique interval music helps to identify the broadcast. Once those field agents have written down the coded groups, they use a one-time pad to decode them. The one-time pad is then destroyed. In all, it makes for a system that is almost unbreakable by any currently known analytical system.
In March 2003 on the eve of the Iraq war, also known as Operation Iraqi Freedom, NY radio station WFMU played the entire 4 CD set, appropriately freaking everybody out.
By now, you have heard of the great solar flare of 1859 that set telegraph wires afire across the US and Europe. This phenomenon was due to a large electromagnetic pulse from the earth’s magnetosphere interacting with the particles from a Coronal Mass Ejection (CME) caused by the solar flare. The long stretches of wire suspended above the earth acted like a large generator winding and cutting through the magnetic field, which in turn caused voltage (Electromagnetic force or EMF), which in turn, caused the fires.
In 1859 our understanding of electricity and magnetics is not what it is today, thus, the fires were likely caused by overloaded conductors without means to bypass excess electrical energy to ground.
Schematic of Earth’s magnetosphere, courtesy NASA
Fast forward to today. We are in the upswing of solar cycle 24, which is expected to peak in May of 2013. NASA predicts that this will be the lowest sunspot peak since 1907. That does not necessarily mean the coast is clear. Over the last 11 years since solar cycle 23 peaked, computers and electronic automation have proliferated exponentially, becoming the norm. Things like GPS not only guide clueless travelers where to turn but also sync up all those cellular telephone transmitters with timing signals. IP networks, SCADA, telephone networks, and so on all run on some form of CPU. Newer Energy Star appliances like toasters and refrigerators also have CPUs. That technology has yet to experience a large electromagnetic pulse (EMP) in the real world. Things could indeed get hairy if a moderate to large class X solar flare generated a CME that was polarized correctly to interact with Earth’s magnetic field and cause damage at ground level.
Solar flares and CME are slow-moving events, with 1-2 days warning before the effects of a CME reach Earth. One can stay apprised of solar flares and other solar activity by subscribing to NOAA Space Weather Prediction Center’s email service.
HEMP Mechanism for 400 kM high altitude burst
Of greater concern is other sources of EMP like high altitude nuclear explosions (HEMP). Those types of events, while rare, can happen. The good news is, the defense mechanisms for solar flares, high-altitude nuclear bursts and lightning-induced EMP are the same. They are effective grounding, shielding, filtering and surge suppression.1 Of the three EMP scenarios, high altitude nuclear burst has the tightest design spec, so creating a building that incorporates the ideas in MIL-STD-188-125-1 specifications is a good start.
The question becomes, is all of this really necessary. It depends on how important it is to radio station ownership to remain on the air during such an event. Based on historical information and global geopolitics, the probabilities of such an occurrence are:
Lightning strike – 1:1 Any radio station that has a tall structure, particularly a steel tower, will get struck by lightning, perhaps several times per year depending on the region.
Large Class X solar flare resulting in damaging CME – 1:21 Since 1859, there have been seven solar flare events that have disrupted communications or power systems on Earth. This is a bit misleading since 6 of the 7 events have occurred in the last 22 years, making the real probability more like 1:3.6
High Altitude EMP – 1:30 Based on seventeen high altitude tests carried out by the US and USSR in 1962, the growing nuclear proliferation and a June 2005 Reuters article “Experts warn of substantial risk of WMD attack” in which the author stipulates a thirty percent risk of a nuclear attack of any type in the period of 2010 to 2015.
EMP Theory
High altitude nuclear burst EMP has three components; the fast component (20/550 ns pulse) is an electromagnetic shock-wave, the medium-speed component (1.5/5000 μs pulse), and the slow component (0.2/25 s pulse) resulting from the expansion of the explosion’s fireball in the Earth’s magnetic field.1 Compare that to a lightning strike, which typically has a 1.8 µs rise time. That means the first pulse frequency is from about 72 to 200 MHz, the second pulse frequency is from about 800 Hz to 2.5 MHz and the third pulse is basically DC and affects mostly long wires. Thus, any shielding, grounding and suppression needs to consider the highest frequency down to about 10 KHz.2
The rise time to frequency comparison is an important consideration in the design and construction of grounding systems. Grounds need to present the least amount of inductance possible. This means using solid, not stranded wire, keeping bends to a minimum and where required, use long radiuses. Bond all junctions by welding, exothermic welding or soldering. Use a single point ground system. Zone grounding should also be employed. The definition of zone grounding is concentric grounding areas connected to each other by a single low-inductance ground conductor.2 The idea of isolating grounds by use of a ferrite suppressor may seem odd, however, if there is a separate RF ground, tied together with the building ground using wide copper strap laid on the floor, this will minimize ground system “reception” of incoming EMP.2,3
Zone grounding diagram
Shielding means surrounding the protected equipment with a conductive material such as copper plate, aluminum plate, copper mesh, aluminum mesh, brass mesh, steel plate or steel mesh.1,2 There are advantages and disadvantages to each. Seams should be welded to prevent “leaks.” Doors need to have finger stock or metal compression gaskets to ensure proper sealing of opening. Other openings for ventilation, cable ingress, etc. need to be “significantly” less than one wavelength. MIL-STD-188-1 gives the allowable opening size of 10 x 10 cm or 3.9 x 3.9 inches. If openings in the shield become greater than approximately wavelength/6 meters (at 250 Mhz, about 8 inches), significant fields can penetrate to the interior.2
Suppression deals with those connections to the areas outside of the facility. These include incoming electrical service, data service and RF transmission lines to and from antennas. Since the fast component of the HEMP falls within the VHF spectrum, FM broadcast installations are particularly vulnerable. Suppression devices for incoming AC power are readily available from commercial sources and are well proven. LEA makes a series surge suppressor that uses a combination of fast acting silicone diodes, MOVs, and an LC filter made up of series inductance and parallel capacitance to ground. The LEA DYNA family series surge protectors have a system response time of less than one nanosecond and are tested to greater than 1,000 operations.4 The response time depends on a good, non-inductive ground connection.
LEA DYNA systems series surge protector
Suppression for incoming RF and data cables is more difficult because the normal operating frequencies fall within the HEMP rise time frequency. Incoming data at a transmitter site usually consists of a DS-1 circuit however, larger capacity circuits are sometimes used. Fiber optic cables are immune from HEMP as they have no metal conductors. Copper data lines must have a data line surge suppressor between the TELCO demark and the CSU.
RF cables must have their shields ground to the zone 0 ground, then go through a ferrite toroid to add inductance to the outer shield and isolate it from the zone 1 ground. After the ferrite toroid, a gas discharge type inline surge suppressor should be used. These come in a variety of configurations, frequency band and power levels. It is best to keep the suppressor rating as close to the peak carrier power as possible, affording the most protection to the transmission equipment.
Design and implementation of EMP hardened facilities
Of the four strategies for mitigating HEMP; Grounding, Suppression, filtering and shielding, shielding is the hardest and most expensive to implement. Good grounding should be included in any good radio station design same as suppression and filtering.
Grounding. Grounding for a transmitter site must include an outside ring ground around the periphery of the building.2 This is bonded to several ground rods installed at regular intervals. The ground is brought into the building and all coax shields, electrical service entrance, TELCO equipment, suppression equipment and safety grounds are connected to it. This forms the zone 0 ground. One conductor then goes to the zone 1 ground which is the transmitters and racks. Any other conductors at any potential that go from zone 0 to zone 1 are bypassed at EMP frequencies by use of ferrite toroids or other high mu ferrous metal filters. Inductors may need to be bonded to ground to prevent saturation. At studio locations, the building electrical safety ground should be evaluated for adequacy. Additional grounding may need to be installed depending on effectiveness of existing ground. Any outside antennas, supporting structures, satellite dishes and generators need to be bonded together and grounded to the building electrical grounding system. Studios and engineering rack rooms need to be bonded to the ground using star topology. Facilities that are not adequately grounded should be retrofitted.
Suppression and filtering. Good surge suppression and filtering should be a part of all transmitter and studio site designs. Hanging a few MOVs off of the service panel is not enough to prevent damage to a facility. All incoming lines from the street; electric, telephone, and cable need to have surge suppression connected and be bonded to a low inductance path to ground. The only exception to this is fiber optic, which is immune to the effects of EMP. Additional layers of filtering for sensitive, mission-critical computer systems such as FERUPS, shielded category wiring that is properly installed, etc. Facilities that do not have adequate suppression can be retrofitted.
Shielding. Shielding is the most expensive, time-consuming, and difficult to install correctly. The High Altitude nuclear test Starfish Prime in July of 1962 produced a field of 5600 V/m in Honolulu, some 1300 KM away from the blast. Building a shielded structure against those intense magnetic and electrical fields is very difficult. Attenuating the field through layers of shielding is the most effective means provided the distance between the shields is wavelength/6 or more (about 27 inches at 72 MHz) to prevent coupling.2 For example, using a concrete structure with steel mesh creates 35-40 dB of attenuation in zone 0, and a well-designed transmitter with good RF shielding in its cabinet design creates a shield for zone 1. Equipment racks can also be used to create shielding zones by using copper or brass mesh with good metal-to-metal contact around the front and back doors. At studio locations, engineering rack rooms should have copper or brass mesh embedded in the wall structure to create a shield. This will create a safe area to locate computer file servers, routers, switches, STL gear, satellite receivers, and the like equipment. Layered shielding with the use of metal, gasketed door will improve shield performance. Retrofitting shielding is more difficult to accomplish than grounding and suppression. It is best done in new construction. There are many different ways to accomplish even moderate shielding, which may serve well for lightning and solar flare-induced EMP.
From personal experience, investing an extra $10-20K in grounding and suppression at a lightning prone transmitter site in Florida solved all of the issues at that site. Prior to installing the ring ground and bonding, the transmitter was knocked off the air several times per year. Since the work was done in 2005, there has not been one lightning related outage at that site.
References: 1. Protection Technology Group, System Approach to EMP Mitigation, February 2011 2. US Army Corps of Engineers, Engineering and Design – Electromagnetic Pulse (EMP) and Tempest Protection for Facilities EP 1110-3-2, December 31, 1990 3. US Department of Defense, HIGH-ALTITUDE ELECTROMAGNETIC PULSE (HEMP) PROTECTION FOR GROUND-BASED C4I FACILITIES PERFORMING CRITICAL,TIME-URGENT MISSIONS, MIL-STD-188-125-1A, February 15, 1994 4. Protection Technology Group, Modular Hybrid Series Connected Surge Protection Device LEA DS21 data sheet, 2010
I had the opportunity to work on one of these recently, thought I’d post a few observations. The transmitter itself comes in three parts, the FM100 which serves as the exciter and driver, the PA2000, which holds the RF amplifiers and combiner and the PS2000 with supplies the DC voltages to run the PA.
Crown FM2000A transmitter running at half power
That configuration has some advantages and disadvantages. First, it takes up much more rack space than the comparably powered Nautel VS2.5. Second, because the unit does not come with slide-out rack rails, each part needs to be removed from the rack for servicing, which makes things a little difficult when working alone as the PS2000 weighs quite a bit. As far as the rest of the design, the PA2000 is very modular, all of the PA modules, controller card, fuse board, and RF combiner easily come out of the chassis for service.
Crown FM2000A top cover removed
This unit had been in service at WBEC in Pittsfield, MA for an undetermined amount of time. As such, there was quite a bit of dirt and bugs inside the PA chassis. I used an air blower to clean everything out. Checked the fans for bad bearings, checked all RF connections for signs of overheating, etc. I also cleaned out the power supply and rinsed all of the air filters.
Crown FM2000A front cover off
My other minor complaint is the power adjust pot is under the front cover. When making adjustments and such, the LED display indicates operating constants based on a little LED light next to the display. The legend is on the cover, which has been removed to adjust the power. Minor thing, but slightly annoying, nonetheless.
There are four RF modules in the PA2000, each one generating 500 watts. This particular transmitter has a bad device in PA3. When the transmitter is running the DC fault LED flashes and the PA3 reading shows no current. The device is a BLF278, which is a fairly common, inexpensive RF MOSFET. According to the factory tech, they can be replaced in the field provided one can solder. After replacement, there is no special tune-up or anything needed as the module is wide-band.
Crown PA2000 500 watt RF module
The four modules are combined and then sent to the RF output filter which has the low pass harmonic filter and directional coupler.
Crown PS2000 output filter
It is a pretty simple transmitter, no bells or whistles or fancy things like IP connectivity. Overall, it seems to be well-made, robust, modular, and efficient. The remote control interface is via DB-25 connector on the back of the PA2000.
I did not get a chance to hear it on the air, I was just cleaning and testing the RF sections. The exciter is an FM100 transmitter, which I had to change frequencies on. I found that to be self-explanatory.
It would be fun to compare this to some of the other broadband FM amplifiers like PTEK and Armstrong.
Series excited AM towers require some way to get standard AC across the base insulator to the tower lights if tower lights are required. While many new AM towers do not have base insulators, through the use of a folded unipole, it is still a very popular design and has several technical advantages.
There are two methods for getting 60 Hz AC from zero RF potential to an excited tower:
Tower lighting choke
Austin Ring transformer
LBA Group TC-300 tower lighting choke courtesy LBA Group, Inc
When to use it depends on the tower and the RF potential on the base of the tower. For towers that are under 140 electrical degrees (RF) and carrier power levels up to 100 KW, a lighting choke works well. They are simple and less expensive than an isolation transformer. They can be installed inside the ATU cabinet or placed in their own weatherproof enclosure as required. Tower lighting chokes will add series impedance to the base of the tower and needs to be compensated for by adding capacitance to the circuit. This will become more pronounced at the lower end of the band, where, if one is not careful, RF from the tower can be coupled to the transmitter building’s AC mains, which is very undesirable.
Tower lighting chokes generally consist of three separate windings, one for the beacon, one for the side lights, and one for neutral. Their inductance is typically in the 800-1000 µH at 1 MHz region. They can be stacked to increase their peak voltage handling capacity:
LBA Group tower strobe light choke courtesy LBA Group, Inc
Peak voltage is determined by the base impedance and carrier power + modulation. On any AM station these days, a 150% peak modulation figure should be used (125% modulation allowed by FCC rules plus a 25% safety factor). For example, station B has a base impedance of 50Ω (typical 90° guyed tower) and a carrier power of 50 KW. The peak modulation power will be 600 KW. Thus, the peak voltage will be Epeak = √Ppeak x R, or Epeak = √600,000 watts x 50 ohms or 5,477 volts. With higher base impedances, the base current goes down but the base voltage goes up. A typical 140° tower will have a base impedance of 760Ω. Thus the peak base voltage for a 50KW carrier power modulating at 150% will be 21,354 volts. This is the worst case scenario, as few installations are designed that way and every tower impedance is different than the theoretical self impedances given.
For towers over 140 electrical degrees, it is better to use an isolation transformer because of the RF peak voltage/peak current conditions at the base of towers that are electrically tall. The ring transformer design minimizes stray inductance or capacitance at the base of such towers. Austin Insulators (previously Austin Decca) makes a variety of tower base ring isolation transformers. These have varying input and output voltages.
Diagram of typical Austin Ring transformer courtesy Austin Insulators, Inc
I have seen these at many locations over the years. They are rugged and add only a small bit of capacitive reactance to the base of a typical tower. They also completely isolate the building AC mains from the tower. For very high-power installations, Austin has the A-9600, which was designed for the Navy VLF transmitter towers where base peak RF voltages can run 200,000 volts or more:
Voltage drop is another consideration in tower lighting design. Long runs from the transmitter building to the tower should be on heavy gauge wire and at 230 volts if possible. FAA Circular AC 150/5345-43F “Specification for obstruction lighting equipment” advises that the input voltages for incandescent lighting systems vary by not more than ±3%. Additional tower lighting and painting information can be found in FAA Circular AC 70-7460-1K.
