Aero ‘Lectrics

BY JIM WEIR

We test the

ILS radios. oo boy, here we go again! In the last 10 years of this craziness, I’ve taken you down some long and winding paths, but this one is the doozy of all time. We are about to design a test box that will allow us to simulate a full-house ILS installation in our airplane as well as test the VOR installation. It is not going to be easy. I doubt more than 50 of you in the next year will actually make one or more of the circuits that we are going to describe in the next half-dozen articles. That’s right—six articles to get this whole system up and running. But that’s OK; these columns will survive for at least a dozen years in the archives; and the ILS/VOR system will survive far longer than that. And, those of you who actually do make one of these boxes will loan it to at least a dozen friends. GPS is sexy. Loran is bonehead simple and megawatt powerful. But the ancient and honorable ILS system is still the only thing that will get you down to an instrument approach of 200 and a half. GPS can do that, and someday it will. But today the ILS is the only thing that gets you down onto concrete when the clouds are in your way. The funny thing is that the ILS is inherently 10 times less accurate than GPS or Loran; ILS is based on 1940s technology that hasn’t been upgraded since it was invented. Be that as it may, it still gets us down to where our wheels are just a little bit above the pavement when the weather gets tough. There are three components to an ILS and one to a VOR. I’ll take you on a tour of the ILS this month—just so we know where we are headed in future columns. I’ll look at the VOR later.

H

ILLUSTRATION: JIM WEIR

This illustration presents a pictorial description of how the ILS localizer and glideslope work—similar to an automobile headlight beaming light in a particular direction.

A Little Local Color The first part of an ILS is what is called a localizer. The localizer is nothing more than a radio signal that tells you where you are relative to a straight line down the runway—it’s simply an electronic extension of the dotted lines down the center of the runway. To get an idea of how the localizer works, think about an automobile headlight that beams light in a particular direction. Now think of a yellow headlight at the end of the runway that is pointed slightly to the left of the centerline and think of a blue headlight pointed slightly to the right of the centerline. When you are down the middle, you will see the yellow and blue headlights equally. Off to the left the yellow headlight predominates, and off to the right the blue headlight predominates. If you had a yellow/blue meter on your instrument panel, you could tell whether you were left or

right of the centerline. Radio waves work just like light. You point one beam to the left side of the runway, another beam to the right side of the runway, and you can tell where you are relative to the runway centerline by measuring the strength of the two beams. Let’s stop talking in circles and instead discuss how the localizer really works. There is a ground transmitter that is channeled to a VHF frequency between 108.1 and 111.9 MHz on the odd 100 kHz. Channels such as 108.1 and 108.3—roughly 100 watts of output power split into two 50 watt signals. One of these splits is amplitude modulated at 90 Hz and is run to the beam antenna pointed to the left of the runway centerline; the other split is amplitude modulated at 150 Hz and is run to the beam antenna pointed to the right of the runway centerline. Get the picture? Tune your nav radio to 110.5 (for example, Marysville, K I T P L A N E S

J U L Y

2 0 0 4

55

Aero ‘Lectrics CONTINUED

California—MYV), and inside the radio there are circuits looking to detect the relative amplitudes of the 90- and 150-Hz signals picked up by the antenna on your airplane. If there is more 90- Hz signal than 150, the radio drives your navhead meter to the right (fly right). More 150, and the needle goes to the left. Actually, rather bonehead-simple amplitude tone modulation, but it works remarkably well and is very reliable. Then again, the same basic principle applies to touch-tone dialing, and that system has worked flawlessly for a few dozen years. So, for a localizer test we need a carrier wave on an odd-kilohertz frequency from 108.1 to 111.9 split into two parts—one part modulated by a 90-Hz signal and the other part modulated by a 150-Hz signal. We know, however, that if we have equal amounts of 90 Hz and 150 Hz that the localizer needle will be centered. But a good test unit will let us switch-select to deflect the meter at the half- and full-scale points. We’ll do a heuristic design (hammer, file, kick in the edges, weld shut and paint to match) to figure out what full-scale means in terms of DDM (difference in depth of modulation).

The Glideslope There is another component of the ILS that is at least as critical as the localizer, and that is the glideslope. Now there are many instrument students (including yours truly 40 years ago) that would swear that there is some tiny little gremlin inside that indicator that is playing chin-ups on the glideslope needle while you are trying to keep it somewhere near the center of the scale. Sorry, it just isn’t so. However, the glideslope is about five times more sensitive to per-foot errors than the localizer. The net result is that most of us chase the glideslope about five times more aggressively than we do the localizer. Be that as it may. The glideslope RF frequency is channeled to the localizer frequency by a scheme known only to some little gnome in Oklahoma City 56 K I T P L A N E S

J U L Y

2 0 0 4

who has long since retired. It greatly resembles how Japanese house street numbers are assigned—the first house constructed on the street is No. 1, the next house (perhaps across town) is No. 2, and so on. Glideslope frequencies are paired to localizer frequencies in the same manner; there is no rhyme or reason to how this pairing takes place. If you are really interested in the pairing frequencies, refer to the AIM Table 1-1-4 that shows each localizer frequency and its associated glideslope frequency. Note that you will never set your glideslope receiver manually to a frequency—you set the localizer frequency, and magic electrons (internal circuits) set your glideslope receiver to the proper channel. What is true is that the same modulation frequencies that we used for the localizer are also used for the glideslope. The 90-Hz beam says you are too high, and the 150-Hz beam says you are too low. When these two beams are received by the glideslope receiver, the 90-Hz circuits drive the meter down, and the 150Hz circuits drive the meter up. Again, bonehead-simple amplitude modulated signals, but inexpensive and reliable. It is worth noting as we leave the glideslope description that the transmitter power is a relatively low 5 watts. While you might be able to receive that 100watt localizer signal at 50 miles or so, the glideslope range at best is from 8-10 miles.

Marker Beacons One more system, and we can start to work on the electronics of this problem. We’d like some indication of rough distance from the runway, and we get that with the creaky old marker beacon system. Again, we have an amplitude modulated set of transmitters located a reasonably accurate distance from the runway. There can be up to three of these transmitters associated with the ILS— outer, middle and inner/fan marker transmitters. They all operate at the same 2watt 75.0-MHz RF (carrier) frequency, but they are modulated with different W W W . K I T P L A N E S . C O M

tones. The outer marker is a 400-Hz tone, the middle marker is a 1300-Hz tone, and the inner or fan marker is a 3000-Hz tone. These tones also light three different colored lights on the marker display—traditionally blue, amber and white, respectively. The outer marker (which may or may not be the final approach fix, by the way) is about 5 miles out, the middle marker is about half a mile out, and the inner marker is about 500 feet out. This inner marker is used mainly for a Category II airline type approach and in general doesn’t apply to those of us flying light single engines. However, the 3000-Hz tone is also used on what is called a fan marker, which can be used as a distance indicator on some approaches such as San Diego-Gillespie (LOC-D); these are few, far between and being eliminated one by one.

Our Plan of Action Dang, all these words, and as yet, not a single electron has been emitted in the cause of getting us some ILS test equipment! Patience, patience. This is going to be a long series and a complex one at that. Let’s take it one step at a time. I will write these ILS test box columns every other month so that we don’t have them jamming each other up. In the meantime, relax, build the boxes one step at a time, and let’s go for it! If you want to get a head start on the August column, go buy a 30-cent 32.8-kHz tuning-fork crystal, a couple of LM324 opamps, and the following 50-cent CMOS digital ICs: 4011 NAND, 4013 FF and 4040 Counter. You can get these parts from the usual gang of suspects: Mouser, Jameco and Digi-Key. The total cost should amount to less than $10. Jim Weir is the chief avioniker at RST Engineering. He answers avionics questions in the Internet newsgroup rec.aviation.homebuilt. Check out his web site at www.rstengr.com/kitplanes for previous articles and supplements.

FOUR STROKE POWER! THE • LIGHTPLANES • UL TRAINERS • GYROS • TRIKES • REPLICAS • P P C ’s HPower - LTD.

• • • • • •

700E - 60 H.P.

122 LBS. COMPLETE LOW FUEL CONSUMPTION SUPER QUIET MUFFLERS 2 R E D U C T I O N R AT I O S PUSHER OR TRACTOR WARRANTY

P.O. Box 690 Ellington, CT 06029 860-875-8185 • Fax 860-870-5499

• HUNDREDS FLYING

www.HKS700E.com - SPECS • 3 VIEWS • INSTALLATIONS

Buy A Sell A

@

www.

, or ...On-line! .com

®

Our On-line classifieds feature: • Fast Placement • 2.3 million site hits per month • On-line ordering and editing

• Multiple on-line photos • Direct E-mail responses • Low Rates

K I T P L A N E S

J U L Y

2 0 0 4

57

Aero ‘Lectrics

BY JIM WEIR

How to build an

ILS/VOR test box, Part 2. oomph to generate a signal in the milliwatt range. U101A is connected in a conventional non-inverting amplifier configuration with Q101 as the emitter follower so that we can draw as many milliamps as we might need. D101 sets a reference 5 volts on the non-inverting input of U101A, and resistors R101 and R102 set the gain of the opamp at 1.70, which multiplies the 5-volt reference to an output of 8.5 volts. This is a good starting point, because this will back-bias D102 and keep the little battery from powering the circuit when we have 12 volts available from either a wall-wart supply or the aircraft battery.

Crystal Clear Schematic. If you need pinouts for any of the ICs in this article, simply Google on the part number and CMOS. For example, search for 4040 and CMOS for the 4040 divider. You will get several hundred hits to pick from. As far as I know, they are all correct.

f you take a two-month backstep to the July 2004 issue of KITPLANES®, you can see that to produce the signals needed for ramp testing the VOR and ILS receivers, there are a number of audio-type frequencies that you will have to generate. The most difficult signals to generate are the precise 30-Hz waves for the VOR tester. Not only do you have to generate the frequency precisely, but you also have to find a way to vary the phase of the 30-Hz waves to test at various bearings around the VOR dial.

I

The Power Supply I generally begin any design with ILLUSTRATIONS: JIM WEIR

the requirements of the power supply. I took a few things into consideration when determining these requirements: (1) I’m going to use CMOS (complementary metal oxide semiconductor) digital logic in the design, and (2) I want this little rascal to be able to run from an ordinary household wall outlet, from a 12-volt aircraft battery or from a 9-volt transistor radio battery. Those considerations taken into account, I think an operating voltage somewhere around 8.5 volts should be just fine. CMOS will run anywhere between 3-18 volts, and when we go to do the little RF oscillators later, 8 volts should give us more than enough

With the power supply in hand (at least for the moment), we can turn our attention to the generation of our 30-Hz tones. Some radios depend on these tones to be exactly 30 Hz and will give you errors if the tones are off, even by just a little. The only way I know to generate a stable, precise frequency is with a crystal oscillator. And since you can’t buy 30-Hz crystals, we are going to have to use some other frequency and manipulate our way to 30 Hz. The cheapest source of crystals I know of are manufactured by the billions every year—watches. As the tuning fork crystals used in watches, these little rascals are quite precise and are available for half a buck or so from the usual providers such as Mouser, Digikey and surplus stores. But a crystal by itself does not oscillate. Just like a bell by itself does not ring, we need a little hammer to

K I T P L A N E S

S E P T E M B E R

2 0 0 4

59

Aero ‘Lectrics CONTINUED

ding our crystal and make it ring also. U102A is such a hammer. In reality, it is an inverting buffer amplifier, and it starts out life hitting the crystal with a burst of noise when the power supply is first turned on. Shortly thereafter (nanoseconds) the crystal will start to ring, feeding this ring signal into the input of U102A. The output hits the crystal with its own frequency (32,768 Hz), and we are off to the races with our Pierce-type oscillator circuit. The nice thing about a Pierce is that it is rock stable, changes practically none with voltage or temperature and is inherently self-starting. The device is actually called a 4049 and is of the CMOS family that I mentioned before. Here’s the result of this month’s work.

The Details CMOS is a popular logic family, and one of the great virtues of CMOS is that it draws only microamperes of current to do its job. Actually, the 4049 is what is called a HEX inverting buffer because there are six individual buffers inside of each IC package. Let’s use one of those other five buffers (U102B) to provide some isolation from the rest of the circuit to the crystal. We will call the output of this buffer the 32-kHz Clock. And just because I suspect that we may need it somewhere else down the line, let’s use one of the remaining four buffers to invert the clock. We will call this output the /32 kHz Clock (the / symbol means inverted or not). For reasons that I will explain later, the first thing I want to do with this accurate clock signal is divide it in frequency by 273. This will give us an output signal of 120 Hz. U103 is such a divider, and again, it is a CMOS device—a 4040. Each of the Q outputs is a power of two. Q8, for example, is a divide by of 28 or 256. Similarly, Q4 is a divide by 16, and Q0 is a divide by one. D103-D105 act as an adder so that 60 K I T P L A N E S

S E P T E M B E R

the IC is reset to start counting all over when 256+16+1 (273) happens. The signal at the cathodes of these diodes is the input frequency (32 kHz) divided by 273, or 120 Hz. R107 and C104 simply make a 10-microsecond pulse stretcher so that I can see the 120-Hz signal on my oscilloscope. Make sure you understand what I am doing here, because to get the localizer, glideslope and marker beacon tones in future articles I am going to use the 32-kHz clock and simply choose the divide-by ratio to get the audio tones for these tests. I am also going to use this technique to get the 9960-Hz subcarrier for the VOR reference channel. So, what we’ve got feeding the Cl (Clock) inputs of the next part of this circuit (U104A and U104B) is a pulse 10 microseconds wide with a frequency of 120 Hz. U104 is (once more) a CMOS device called a 4013 flip-flop. There are two of them in a single package, and they are arranged in what is called a quadrature generator, which is some2 0 0 4

times named a two-bit shift register. The function of this quadrature generator is to divide the input frequency by four and simultaneously give out the four cardinal angles (bearings) of 0°, 90°, 180° and 270°. Thus, we wind up with a 30-Hz square wave at these four angles. Now the engineer in my right brain fights with the pilot in my left brain. The engineer says that you can never have too much test equipment capability. And wouldn’t it be nice to have the angles in 1° steps? The pilot says that to make the circuit that complex would require at least as much circuitry as we already have, and the odds of needing to check our VOR at anything other than these four headings is slim to none. The pilot wins. We can have any of the four bearings that we just generated.

The Saga Continues I’m going to stop here and let you cobble up this little circuit and test it. In the upcoming November issue, we will make the 9960-Hz reference channel, a couple of filters to make our square waves into sine waves and probably the RF generator/modulator as well. Then in January, we’ll go for the localizer and glideslope generator. We’ll wrap the whole thing up next March with the marker beacon generator. During the “in-between” months, we will busy ourselves with a little antenna work, perhaps some more work on the Ferry Box from last month and a few more goodies I’ve got rattling around in my head. Jim Weir is the chief avioniker at RST Engineering. He answers avionics questions in the Internet newsgroup rec. aviation.homebuilt. Check out his web site at www.rst-engr.com/kitplanes for previous articles and supplements. W W W . K I T P L A N E S . C O M

Aero ‘Lectrics

Builders’ Marketplace CONTINUED

HIRTH AIRCRAFT ENGINES-15-110HP

How to build an

ILS/VOR test box, Part 3. * Sales, Service

and Repairs Hirth Repair Parts * One Year Warranty, 1,000 hour TBO * BlueMax 2 Cycle Aviation Oil * Genuine

Recreational Power Engineering 5479 E. Co. Rd. 38,Tiffin, OH 44883 Phone 800-583-3306 Tech. Support (419) 585-7002 Fax (419) 585-6004 www.recpower.com

Last month’s VOR generator is shown on the left; this month’s work is displayed on the right.

ops—my goof! In the September issue, I told you how to build the first part of the VOR generator but neglected to explain exactly why we needed the signals we needed. My bad. Here ya go:

O

VOR for Dummies While GPS is the darling of the current navigation world and Loran the backup to GPS, there are still a couple of thousand VORs out there chugging out their 24/7 signals on a reliable basis. There is talk of the VOR system going away in 15 or 20 years, but my guess is there will be at least a few dozen that make it into the next century...especially those that have a non-precision approach to a major airport associated with them. Let’s face it, given 100 grand or so, any engineer worth his or her salt could jam the GPS and render it unusable over several thousand square miles. VOR isn’t quite so easy to jam. I’m going to gloss over the intimate details of how the VOR works. If you want to get deep into it, I recommend visiting www.navfltsm.addr.com. The VOR generates a single RF 68 K I T P L A N E S

N O V E M B E R

2 0 0 4

frequency and then splits this RF energy into two parts. The first half of this energy is fed to a rotating antenna. The most common form of VOR has an electronic rotation of a horizontally polarized beam antenna going round and round at 1800 rpm. In addition, the second half of the RF energy is broadcast in all horizontal directions by an omnidirectional antenna. Note that the first signal can also be said to be a 30-Hz signal (1800 rpm=30 rps). The omnidirectional signal is amplitude modulated by a 9960Hz signal. This 9960-Hz signal (sometimes referred to as a subcarrier) is itself frequency modulated by a 30-Hz sine wave. The deviation of the 9960 FM signal is +/- 480 Hz. Here’s the deal. The two 30-Hz signals (the one from the rotating antenna and the one from the FM’d subcarrier) are arranged so that they are precisely in phase as the beam passes through 0° (due north). They are then 90° out of phase due east of the VOR, 180° due south of the VOR and so on. By a clever manipulation of a calibrated phase shifter on your instruW W W . K I T P L A N E S . C O M

BY JIM WEIR

ment panel (sometimes called an OBS or OmniBearing Selector), you can shift the phases so that the signals are precisely in phase, zero the omni needle and read your bearing to or from the VOR transmitter.

Creating the Signals OK. Now that we know what we need, let’s get to it and create these signals. We know we need 30 Hz (and

Here’s the schematic of this month’s circuitry.

The composite VOR modulation out shows equal amounts of 9960 subcarrier and 30-Hz variable signals.

The 9960-Hz signal (also called a subcarrier), is FM’d by the 30-Hz reference signal.

Here’s what the 30-Hz variable signal looks like after being filtered. ILLUSTRATIONS: JIM WEIR

designed a little generator in September that lets us pick any one of four phases of this signal), and we know we need 9960 Hz. Because we already have the 30 Hz as a square wave, a simple opamp circuit (U101A and U101B) will let us filter this square wave into the desired sine waves. A simple digital phaselock loop (U102) with an additional filter (U101C) at the output to get a 9960-Hz FM’d sine wave shouldn’t be all that hard, either. The 30-Hz filters need to be matched, and you will see that one of the filters (U101B) has a small tuning potentiometer associated with it. We don’t care about the amplitude of the two filter outputs as much as we do about the relative phase of the two outputs. If the same 30-Hz square wave is fed to both the filter inputs, the outputs are adjusted to bring the sine waves into perfect 0° phase correlation. We will do a heuristic (hammer shut, kick in the edges, file flat, weld closed, paint to match) adjustment after the whole shooting match is done using a known good VOR receiver. The reference subcarrier will also be pretty easy to generate—a digital phaselock loop (U102) locked to the 30Hz reference square wave that we just talked about. If you do the math, you

find that a divide-by-332 in the feedback loop (U103) of the phaselock loop will indeed generate a 9960 square wave...and give us the capability of frequency modulating (R118) it in the bargain. (Just another basic Weir two-fer.) OK…332 is made up of a 256, a 64, an 8 and a 4. (Which, you may recall, is 2(8)+2(6)+2(3)+2(2), or Q8 and Q6 and Q4 and Q2 on our 4040 digital divider U103.) Hey, this is going to be fun!

Part 4 The next installment of this little project (in the January 2005 issue) will take the VOR modulation that we generated this month, make a crystal oscillator on a specified VOR frequency and combine the two into a true VOR tester. Once we get this little gem working, it should be relatively simple to generate the LOC frequencies, an ident tone and keep on proceeding. Then we do the glideslope and the marker tester, and that winds this whole project down. Jim Weir is the chief avioniker at RST Engineering. He answers avionics questions in the Internet newsgroup rec.aviation.homebuilt. Check out his web site at www.rst-engr.com/kitplanes for previous articles and supplements.

K I T P L A N E S

N O V E M B E R

2 0 0 4

69