[ Home ] [ Calendar ] [ Welcome to TPG ] [ flyingtorrey ] [ Competition Results ] [ Contact ] [ 2008 IHLGF ] [ Upcoming Contests ] [ Maps ] [ What's New ] [ Reference ] [ Membership ]

Radios

Everything You NEVER Wanted To Know About Radios !

Ian Hirschsohn

To most R/C flyers the Radio System is Black Magic. Most of the time understanding its workings is academic to flying a model, but when it fails or even just "glitches" this understanding can be critical. You don't have to know what is under the hood to drive a car. It may sound sexist but at least half the drivers don't; yet understanding the functioning of the engine's cooling system is key to predicting the consequences of a broken fan belt.

In this series of articles we will examine the basic operation of the Radio System. From this we will be able to differentiate many common radio "problems." Equally important we will be able to debunk several misconceptions, for example that FM is superior to AM or that PCM beats PPM.

Each system has its pros and cons, being informed will hopefully allow you to choose which best suits your needs. Just because a Rolls Royce is five times more than a Toyota Camry does not make it "better" and for most drivers it is probably worse. Radio price does NOT equate to performance. Independent laboratory tests by RCM expert George Steiner prove that some of the least expensive are among the best.

Receiver internals

Let us start with the Receiver System. Your home AM/FM stereo consists of 3 parts: Tuner, Audio Amplifier and Loudspeakers. Likewise the model Receiver System comprises the: Tuner, Decoder and Servos. In the same way that the tuner and audio amp are housed in the same receiver box, so the tuner and decoder are squeezed into the model's receiver. The home receiver tuner section selects the radio station, excluding all others, and passes the now "demodulated" music to the audio amp. Likewise the model receiver tuner targets just that frequency of the tuning crystal, removes the 72 MHz or 50 MHz "carrier" and passes the demodulated servo info on to the decoder. The decoder sorts out the individual servo signals and directs them to their respective plug. Key to note is that the audio amp does not know whether the music came from an AM or FM station and the speakers care even less; likewise the signal fed to the decoder is independent of whether the modulation was AM or FM. Also the position info fed to the servos is standardized so that almost any brand servo can operate with almost any brand model receiver (with the appropriate plugs), much as any brand 8 ohm speaker can be plugged into any brand stereo receiver.

How servo info is encoded

Let us begin by examining the way in which position info is fed to each servo. This is done by means of signal pulses that go from O volts (OFF) to about 3.3 volts (ON). The position is determined from length of time that the pulse is ON (mark) versus the time it is OFF (space). This pulse occupies a 2 millisecond (0.002 sec) time slot and is repeated 50 to 70 times per second( i.e., every 1/50 sec = 20 ms). During the first 1 ms of that 2 ms interval the signal is always full ON, actual mark/space is in the 1 - 2 ms part. Thus Neutral corresponds to 1 + 0.5 = 1.5 ms ON, full left (or right, depending on servo horn set up) is 1 + 0.0 = 1.0 ms ON and conversely full right (or left) is 1 + 1 = 2 ms ON. For the rest of the time the signal is OFF. Thus Neutral corresponds to 1.5 ms at 3.3 volts followed by 18.5 ms at O volts, repeated continuously.

The pulse info is fed to the servo down the white or yellow wire (with the black or brown wire as common = battery negative = O volts). The red or black+red wire provides positive voltage power to the servo and is generally connected in the receiver directly to the input battery red wire. (The exception is Battery Eliminator Circuit, BEC, receiver in electric power model receivers, which condition the voltage not to exceed 6 volts before feeding it to the servos). The operation of the servo itself will be described later, but the key points to note are:

• Servo position is independent of battery voltage and depends ONLY on pulse timing.

• The internal mechanism of a servo is independent of the position signal. It could use a box of monkeys with hand-cranks. For example as long as 1.5 ms ON with 18.5 ms OFF results in Neutral, the implementation wizardry is academic.

• The position signal (on the white wire) is relatively weak, the main driving power is drawn from the red and black.

The last point is insidious because it explains why long servo runs are susceptible to electrical noise, especially feedback noise crossing over from the red line (main power), usually on an electric power model. Any noise that fouls up the pulses causes the servo to behave erratically.

You may be curious what that initial 1 ms ON period does. It is the servo "synchronization" pulse: it flags the servo electronics that the actual mark/space position follows. In other words the servo waits until it sees the signal voltage change from O to about 3.3 volts, counts off exactly 1 ms and then decodes the 1 ms position pulse. Thus the pulses do not have to repeat exactly 50 times per second. It also serves to indicate that the transmitter is alive and well because even with an all SPACE position pulse there is still the 1 ms ON preceding it. Devices such as downed aircraft locators use this to activate a beeper when transmission is switched off or lost - continuous OFF. (It is also handy for R/C bomb makers to arm their devices.)

Driving Multiple Servos - PPM

Why is the servo position pulse only 2 ms long with 18 ms of twiddling its thumbs before the next pulse? Well, this is where the other servos come in. Each servo is allocated a succeeding 2 ms slot so that potentially 20 / 2 ~ 10 servos could be driven independently. In other words the actual signal consists of a "pulse train" with the first 2 ms pulse sent to servo 1, the 2nd to servo 2 and so on. The pulse train, or "frame" is repeated every 20 ms = 50 times per second. The Decoder section of the receiver splits the fame into its component servo pulses and directs each to its corresponding servo.

This is somewhat analogous to the decoder in a stereo receiver which splits (decodes) the left and right speaker signals. Again we have a similar synchronization problem to the servo: which of the endless 2 ms pulses belongs to servo 1? In other words how do we detect the start of a frame? This is done by holding the signal OFF for at least 2 ms, which is longer than the longest possible legitimate servo pulse, following the last servo pulse. Thus the Decoder can recognize the start of a pulse train by waiting for a 2 ms OFF period; the first pulse that follows is servo 1. The 2 ms OFF period constitutes "synchronization pulse" (it can be longer than 2 ms and on a 4 channel transmitter may be 8 ms, or more). Moreover, since there is always a 1 ms ON pulse heading up every mark/space, the transition from OFF to ON following the 2 ms OFF provides an accurate "start of frame" indicator from which to count off the succeeding servo pulses.

This simple appending of the basic servo pulses into a frame is referred to as Proportional Pulse Modulation and (voila!) you now know what PPM means (some refer to PPM as Pulse Position Modulation). PPM may be simple but it is an elegant encoding and remarkably robust (resistant to error). Based on what we know about PPM we can deduce:

• External interference (e.g., a phone pager) longer than 2 ms will corrupt several servos, not just one. So a "glitch" due to an external signal will almost surely manifest itself in ALL servos acting erratically - simultaneously. (It is unlikely that the interference will repeat for exactly the same period at exactly 50 times/sec.)

• The maximum number of servos that PPM can drive simultaneously (i.e., that a frame can accommodate) 20-2 ms = 18 ms, with 2 ms per servo 18/2 = 9 servos. (which explains why 9 channel PPM radios are the end of the line.)

• Servo position is independent of signal strength, so distance from the model will not affect the controls unless the signal is lost.

• Even if one frame is corrupted, succeeding frames are quickly synchronized (by looking for the next 2 ms OFF) so the effect of a "glitch" is momentary.

• Since the frame start is flagged by a transition from 2 ms OFF to 1 ms ON, not by an exact 50 times/sec counter, PPM is resilient to differing "frame rates" which may be up to 70 frames/sec for 4 channel transmitters. A boon to cheap swap-meet fans, like yours truly, is that transmitters and receivers generally interchange quite well, within each brand (even across certain brands; we will discuss this later.)

• Different servo channel count transmitters and receivers can be intermixed, for example a 4 channel transmitter can drive a 6 channel receiver and vice-versa. The extra channel pulses are simply turned OFF on the transmitter or ignored by the receiver.
  

PCM

PCM (Pulse Code Modulation) is to PPM what CDs are to phonograph records. Phono disks are "analog" i.e., the needle movement is continuous within the wavy spiral. CDs on the other hand record the music amplitude as numbers at a rate of 44,100 numbers per second on each stereo channel. As we have seen, PPM encodes servo position as a mark/space ratio in a 1 - 2 ms interval 50 times/sec. The mark/space proportion can take any value from O to 100% and is therefore "analog." PCM encodes servo position as a number much as a computer modem sends data across phone lines. Thus each 2 ms time slot in the 20 to 23ms PCM frame contains a number, first for servo 1, then servo 2, and so on. The numbers are "binary" (power of 2) so in the same way as 1 decimal digit represents numbers 0-1, 2 digits 0-99 and so on, 1 binary digit is 0-1,2 digits 0-3, 3 digits 0-7, etc. For PCM the number of binary digits, or bits, may be 8 (0-255), 9 (0-511) or 10 (0-1023) depending on the manufacturer. So for example a Futaba PCM 1024 may send O for full left, 511 for neutral and 1023 for full right. The actual number of bits and synchronization codes (similar to the 1 ms ON in PPM and 2 ms frame synch) are proprietary to each manufacturer so, unlike PPM, PCM transmitters and receiver are often not interchangeable within a brand, let alone across vendors. This "digital encoding" limits the number of servo positions (1023 steps in the above example), but this is not significant.

The advantage of PCM is that the numbers are exact so, for example, 255 always represents neutral (in 9 bit) whereas a sloppy PPM transmitter may send a 1.45 ms neutral pulse instead of 1.5 ms. Also noise may corrupt a PPM pulse causing a false servo mark/space reading. In a number based system extra "check" numbers can be added to correct false data values. This is done on CDs and gives them phenomenal noise immunity.

Since PCM receivers drive standard analog servos, those neat numbers have to be translated into proportional servo pulses anyway. So the benefit is purely in precise transmitter/receiver communication.

PCM does, unfortunately, have a serious Achilles heel weakness. Even minimal atmospheric or external noise can foul up those wonderful intricate binary numbers beyond any correction. In that case the receiver is up a creek without a paddle. The simple PPM pulses may be corrupted but some information generally gets through. The choice is between NO control (PCM) and some control (PPM). Most R/C flyers would prefer having some control even if erratic. PCM systems boast a "fail-safe" mode which causes all servos to assume a position pre-programmed by the flyer if the signal cannot be decoded. For example all servos to neutral except slight rudder may salvage a plane waltzing off into the blue yonder, but if a straight-wing aileron ship is in a banked turn you have the Zen pleasure of watching it spiral into the ground. Bottom line:

• Unless you want an edge for close-in precision aerobatics or racing, avoid PCM. You will save money and your hair.

• PCM may not be a good idea for thermal duration, where distance invites signal corruption (fail-safe may, however, be a help for polyhedral ships).

• Absolutely precise servo positioning is a debatable benefit for most gliders.
  

AM vs. FM

PPM and PCM define the manner in which servo position is encoded into an electrical signal. If we could connect the flyer to the plane with a long wire this is all we would need. You think this is a joke? Many anti-tank missiles do exactly that: R/C advertises the position of the launcher and invites a return favor.

AM (Amplitude Modulation) and FM (Frequency Modulation) specify the manner in which PPM or PCM are impressed on the bass, or carrier, radio wave i.e., how it is "modulated." Many modelers equate AM to interference prone AM stations and FM to crisp, clear FM broadcast stations. BS, and I don't mean a university degree. R/C AM is NOT AM and FM is NOT FM!

True Amplitude Modulation implies that servo position is proportional to amplitude analogous to music loudness in broadcast AM, so full left might be maximum carrier strength, neutral mid-strength and full right minimum strength.

In R/C "AM" there are only two states -- FULL signal and NO signal, nothing in between. R/C AM is more PM. Seriously, a more descriptive term would be Pulse Modulation, but that looks like PPM sans a P. How about AM being BM -- Binary Modulation.

Likewise, true Frequency Modulation implies servo position proportional to the carrier frequency moved up or down a little, so full left on 72.070 MHz carrier (channel 14) would be 72.075 (carrier + 5 Khz), neutral at 72.070 and full right 72.695 (carrier - 5 Khz). In R/C "FM" there is only full modulation i.e., carrier shifted by 5 Khz or NO modulation (carrier only). R/C FM is identical to wireless data transmission "frequency shift keying" (FSK), at least the computer people don't pretend to be FM.

The Futaba, Hitec and Tower Hobbies camp are NEGATIVE shift (carrier - 5 Khz)= ON. The Airtronics/ JR camp are POSITIVE shift (carrier + 5 kHz)= ON. Carrier only is OFF for both. For example an Airtronics PPM frame, with all servos at Neutral, would start with 72.075 MHz (ON) for 1.5 m then 72.070 MHz (OFF) for 0.5 ms repeated for 2 - 9 servos and wraps up with 2 ms at 72.070 (OFF) synch.

From this we can see that R/C "AM" is not fundamentally inferior to R/C "FM" - it just uses one frequency instead of FM's two. Indeed a properly tuned AM radio is theoretically superior to its FM counterpart because it can only be glitched on ONE frequency whereas a glitch on either FM frequency will knock the FM receiver out.

Unfortunately AM gets a bad rap because the older AM radios were not very selectively tuned (pre 1991), there being fewer channels those days so channel separation was not the 20 Khz of today. So an older, mistuned, AM transmitter may have "side splatter i.e., emit an image on a neighboring channel. Also older AM receivers may pick up adjacent channels. Post 1991 (yellow sticker) AM transmitter/receivers are at least as good, if not better than many FM counterparts. (I use several "garage sale" bargain AM radios regularly at Torrey Pines, Poway and elsewhere and have yet to detect an interference glitch, or cause one. Likewise for the up-to-date AM radios I have seen.) The only possible advantage of FM is that one of the 2 frequencies (carrier and carrier +/- 5 Khz) HAS TO BE PRESENT, otherwise transmission is lost; whereas in AM no carrier just means OFF. It does not appear that any present day PPM radios make use of this (e.g., for fail safe servo positioning).

Bottom line:

• Most AM transmitter/receivers are compatible.

• Futaba, Hitec and Tower Hobbies (Futaba knock-off) FM are compatible (negative shift). Airtronics and JR FM are compatible (positive shift).

• If you want to save cash and weight consider the Hitec 2 servo channel AM (Focus 2). Also the Futaba 2DR 2 channel - $50 from Tower Hobbies with transmitter, receiver, 2 servos, battery box and switch (the 2 stick mode 1 transmitter is incompatible with standard, single stick, mode 2 and therefore almost useless).

• Current model AM receivers (usually 2 servo channel) work excellently with 4 channel AM transmitters (e.g., Futaba Attack 4 and Conquest AM) widely available at swap-meets and garage sales typically for $10-$15. If you are going to use them at the slope or places with other flyers, it would be neighborly to make sure they have a yellow AMA sticker or, better yet, have them tuned by an R/C radio shop.

• Several high-end FM computer transmitters support negative AND positive shift interchangeably (e.g., Futaba 8UAP, Airtronics Stylus and Hitec Prism 7X). These can control models with Futaba/Hitec and Airtronics/JR FM receivers equally well. Swap-meeters take note: the Hitec Prism 7X is a boon because for little more than most 6 channel radios it can operate almost every Futaba/Airtronics/JR/etc. FM PPM receiver bargain (that works). I love mine.
  

Dual Conversion

The signal from the R/C transmitter drops off rapidly with distance from the model and needs to be amplified many thousand times by the model receiver.

Amplifying the 72 MHz high frequency signal is tricky because it is so high that the coils, capacitors, wire lengths, transistors, etc. are critical. To make things worse, noise is generally endemic at high frequency. On the other hand, low frequency signals amplify very well with little induced noise or loss of quality. For example the milliwatt signal of Madonna bawling into a microphone can be amplified to fill Qualcomm Stadium with minimal loss in Fidelity (but who would notice?).

Back in the 1930s a crazy nut, but a genius, Armstrong (father of modem radio) discovered that if the receiver injected a constant signal of its own the result was a new signal at the difference (and sum) of the two with the same modulated information. For example if a received signal of 72.5 MHz is mixed with an artificial constant signal of 72.0 MHz the result is a signal of 72.5 - 72.0 = 0.5 MHz containing the identical pulse train as if the transmitter had used 0.5 MHz to start with. You didn't want to know, but the affect is called Superheterodyne mixing, or Superhet for short (us old timers may recall the advertising hype for Superhets some 40-50 years ago). 0.5 MHz is far more easily amplified than 72.5 MHz and almost all modem receivers, including R/C receivers are Superhets. The difference, or "Intermediate Frequency" (IF) for R/C receivers is around 455 Khz (0.455 MHz) which is the same as that for most broadcast receivers (surprise!) Actually it meant that the original R/C receivers could use the same cheap IF amplifier components as most radios.

Virtually all broadcast receivers have a single IF frequency i.e., they are "single frequency conversion" (or simply Single Conversion). This is fine because broadcast stations are typically separated by 200 Khz (e.g., FM stations at 95.1, 95.3, 95.5 ....) so 455 Khz can easily sort out adjacent stations (provided one is not in LA and the other in San Diego). It was also OK for R/C until 1991 when there were only 7 channels in the 72 MHz band with a 100 MHz separation. Imagine only a handful of channels total at Torrey Pines with the glider traffic of today! So in 1991 the AMA reduced channel separation to 20 Khz allowing 1.0/0.002 = 50 channels between 72 and 73 MHz. Now adjacent channels produce IFs of 435, 455 and 475 Khz which could cause a neighboring channel transmitter to glitch an older receiver if the IF component tuning is sloppy. Current production techniques produce acceptably tight tuning and AM receivers (which are limited to signal ON/OFF, remember) typically limit their IF span to less than 5 Khz each way i.e., anything out of about 450 460 Khz is lopped off drastically thereby eliminating the 435 and 475 interference. So AM is happy as a clam with 20 Khz channel separation. The problem is with FM.

Remember FM, or more precisely FSK (Frequency Shift Keying), uses the carrier frequency for OFF and a shift of 5 Khz for ON (+5 Khz for Airtronics camp, -5 Khz for Futaba camp). Consider an Airtronics transmitter on channel 14 (72.070 MHz carrier) sending 72.070 MHz during OFF and 72.075 MHz for ON. So a Single Conversion FM receiver would generate an internal signal of 72.070 + 0.455 = 72.525 MHz. This results in an IF of 455 Khz for OFF and 455 + 5 = 460Khz for ON. Now someone shows up with a Futaba transmitter on channel 60 (72.990 MHz) and happily bleeps out 72.985 MHz for ON. The channel 14 receiver, not knowing any better, mixes the 72.985 with its internal 72.525 and comes up with 72.985 - 72.525 = 460 Khz which is its ON!! So ... crash. The Superhet mixer produces the difference of the internal and external signals, irrespective of which is which; by adding special filters and other wizardry it is possible to sort out the components, but this adds to the direct cost and production labor for tuning. A more common technique is to 2 stage the IF, ala a rocket booster. First a 10.7 MHz IF is produced and amplified, then a 455 Khz IF. The 10.7 MHz IF knocks out "in-band" interference from your pal's transmitter. This "Dual Conversion" process uses a 2nd crystal for the 10.7 MHz to 455 MHz conversion. So for example a channel 14 Dual Conversion receiver has a 1st crystal of 72.070 - 10.7 = 61.370 MHz and a 2nd crystal of 10.7 - 0.455 = 10.245 MHz.

Dual Conversion is no panacea. It adds cost, weight, more things to go wrong and an extra crystal to crack in a crash. JR has taken note of this and their receivers stick with Single Conversion FM, using clever engineering design in their patented ABCBW circuit (Anti-Blocking Cross-modulation and Windowing).

Bottom line:

• Dual Conversion is needed for FM to eliminate in-band interference from other transmitter.

• AM receivers don't need Dual Conversion which explains why they are generally cheaper, lighter and more robust.

• The frequency of the receiver crystal is NOT the channel frequency e.g., the receiver crystal for channel 14 (72.070 MHz) is NOT 72.070 MHz but displaced by the IF frequency. The exact IF frequency is vendor dependant so DON'T MIX CRYSTALS FOR DIFFERENT VENDORS. (e.g., don't plug a Futaba crystal into a Hitec receiver and expect it to work.)

• Transmitters have no Superhet mixing so transmitter and receiver crystals CANNOT BE INTERCHANGED (transmitter xtals also use "5th overtone" so they are a totally different animal).
  

AM/FM considerations

As we have seen, the narrow separation of R/C channels makes tuning of the internal amplifier circuits important. It is more critical for the transmitter because the receiver is so narrowly tuned that any transmitter drift results in loss of driving signal, particularly when the model is far away. The strength of the transmitter signal is reduced by 100,000 at a distance of 360 ft. i.e., the typical 0.5 watt transmitter signal is 5 microwatts when received at that distance - if everything were tuned perfectly (50 dB free air attenuation @ 360 ft). Bad tuning and extreme temperature cause "drift" in the amplifier tuning so that the actual power reaching the transmitter antenna at the crystal frequency may be significantly attenuated.

For example if the Radio Frequency (RF) final stage amplifier for a channel 14 transmitter were to drift from 72.070 MHz because you left the transmitter in the sun so the coils and/or capacitors got hot, the actual power reaching the antenna at 72.070 MHz would be substantially less than the rated 0.5 watts. You may not be aware of this until the model is a dot in the sky and continues to become a smaller dot in spite of all your frustrated wiggling of the sticks.

It is physically impossible to tune the transmitter RF amplifier to all possible channel frequencies even just in the 72 - 73 MHz band. Typically the manufacturer tunes it to just the crystal supplied e.g., if the box says channel 20, the transmitter was almost certainly tuned to that frequency (72.190 MHz).

If you replace the crystal with some other channel e.g., channel 35 (72.490 MHz) the RF amplifier will be off by 0.3 MHz end the power output substantially diminished. One or two channels either side e.g., 19 or 21 is generally not too significant, but if you need more than that the transmitter should be sent to an R/C radio shop for retuning. (Note that xtal/amplifier mistuning does NOT apply to receivers because their amplifiers are tuned to the IF frequency which is identical for all channels.) Mismatched transmitter amplifiers are only possible with cheaper transmitters which have directly replaceable crystals e.g., Futaba Conquest/Skybport/6XA. Airtronics Vanguard, Hitec Flash/Focus, etc. More expensive transmitters e.g., Futaba OUAF/OUAP, Airtronics Infinity, Hitec Prism have modules which plug into rectangular cavities in the rear of the transmitter. These modules contain not only the crystal, but the RF amplifier and all other frequency dependent whirly-gigs. Thus modules can be swapped for any frequency while maintaining precise frequency alignment -- even across bands e.g., 72 MHz and 50 MHz bands. Because they contain so many components, these modules are substantially more expensive than simple crystals, typically $35 - $50 vs. $10 - $25 for crystals. However, if you value your model, their precision is worth the price. The Hitec Prism 7X at $225 is good value for a module based computer radio.

An interesting feature of module based transmitters is that all of the RF electronics is contained in that rectangular capsule so an FM transmitter can be switched to AM simply by plugging in the vendors AM module. Thus an up-to-date Futaba 8UAP or 8UAF computer transmitter can drive an AM receiver by plugging in a TP72AM AM module, of the appropriate frequency. This may seem a backwards combination, but the Futaba RJ112JE 2 channel AM receiver is lighter, smaller and more rugged than any FM micro receiver and a perfect candidate for hand launch, plus it costs under $30. (Hitec also has a 2 channel AM RX for their Focus 2.) V-tail and Zagi pilots take note - all the V-tail, Elevon, sub-trim, dual-rates, exponential and other smarts of the computer transmitter work on the cheap AM receiver.

With the migration from AM to FM, AM receivers are pretty much limited to 4 servo channels, so AM is not viable if you need 6 or 8 servos.

Bottom line:

• Don't leave your transmitter (which is usually blade plastic) in the sun for any length of time.

• Avoid a black fuselage color scheme, to limit receiver IF amplifier temperature drift.

• Don't switch crystals on the transmitter by more than 2 channels up/down.

• If you can afford them, use module based transmitters.

• Consider buying an AM module for your Futaba or Hitec computer transmitter if you are a hand launch or combat aficionado.