I understand that during the late 1960s, there were some sets being sold that could be described as digital/analogue hybrids - but I don;t understand why.
I always thought that a single channel set operated with a signal that was either on or off - be it carrier or tone. If you turned the signal on and off rapidly, you could create a 'pulse-proportional; effect at the model - Galloping Ghost is a common name for this. But the signal is either on or off.
If you combined and varied various aspects of the signal - typically the tone(s) carried, you would get a 'multi' system. But each signal is still on/off.
If you varied the sigbal aspects proportionately - perhaps tone frequency and amplitude - you would have an 'analogue' proportional system. And if you used the Spreng/Brook system of sending a pulsed 'frame' with a sync pause you had a 'digital' system.
But the systems described as hybrid seem to be doing something different. Perhaps it is to do with the way the signal is processed to drive the servo? It's this that I can't understand, and don't seem to be able to find a straightforward description on the web... Can anyone here help?
Strictly speaking, a digital system sends a data stream that contains a numerical value, usually between 0 and 4095, or in really old digital sets 0 to 254, these being the ranges of 16 bit and 8 bits respectively. The early 'digital proportional systems were nothing of the sort because they sent an analogue representation of the servo position which was obtained by modulating the carrier in some way. (either amplitude A.M. or frequency F.M) The 'digital' part was claimed as the different servo positions were sent serially (one after the other....hence the 'hybrid'). These days, with the advent of microprocessor control which have digital to analogue and analogue to digital converters built in the stick position feeds an analogue voltage level into the processors analogue to digital converter which spits out a digital value which is easy to manipulate for mixing and stuff and these numerical values are arranged in a string and sent out to the Rx as a digital data stream. I would guess most , if not all sets these days are true digital sets.
That is my understanding. I'm happy to be corrected.
You only ever need two tools....WD40 and duct tape.
If it doesn't move when it should use the WD40 and if it moves and it shouldn't use the tape.
Small nitpick. 8-bit range is 0 to 255, and 16-bit range is 0 to 65535.
Most of the "Digital A.M." sets weren't really digital, nor were they A.M. in the normal sense. The carrier wave was simply switched on and off by the modulation - in that sense, it was more like Morse code being transmitted "Carrier Wave" in radio amateur terms. I suppose you could call it amplitude modulation, where the only possible modulation values were 100% or 0%, which is very different to what we normally mean by A.M.
The carrier was normally present at 100% most of the time, with short 'off' pulses mixed in. It was the timing of those off pulses that carried the channel information. Each servo position was set by the time gap between two adjacent pulses (usually the leading edge of the pulses). Each channel usually had a centre time gap of about 0.0015 seconds and a range of approximately 0.001 to 0.002 seconds. If the set was, say, a four channel set, then after the pulse for the fourth channel there would be a big gap of .010 seconds or more before the first pulse of the next 'frame'. This big gap was called a "sync pulse" and allowed the receiver to know when a new frame of information was starting. Note that because it was the time gaps between pulses that encoded the servo information, a four channel transmitter would transmit frames of five pulses per frame. In the same way that you need five fence posts to hold up four fence panels. Often the whole frame was transmitted 50 times per second, which, with the standard, up-to-two-millisecond pulse intervals and the need for a sync pulse, meant that only eight channels could be reliably transmitted. This is one reason why there were few radios with more than eight channels.
When "Digital F.M." became the norm, the encoding system was usually exactly the same. The only difference was that instead of transmitting a carrier wave, and switching it off briefly for the pulses, the transmitter always transmitted a 100% wave, but that wave changed slightly in frequency (often by about 5 kHz) when it was transmitting a "pulse".
Some "digital" transmitters encoded some extra information by varying the width of the pulses, as well as the gaps between them, but these were fairly rare.
Note that everything described, so far, was really an analogue system, in that the time spans of the pulses could have any values within the range: in principle, the resolution was infinite - though of course in practice it wasn't. This is why the term "digital" shouldn't really have been applied to them. "Digital", back then, meant that the signal was being abruptly switched, rather than varied gradually. Nowadays, we think of digital as meaning the transmission of numbers, rather than smoothly varying analogue values.
When true digital transmitters came along, the same frequency modulated pulses were still used, but now those pulses were used to encode numbers, so servo positions were encoded as a number, often 512, or 1024, or 2048 possible numbers in the range (these are 9, 10, and 11-bit numbers). Because the "digital" name had already been used for the earlier sets, the manufacturers had to use some new terms to describe the true digital sets - so they started using names like "PCM", and "computer", to describe them.
I'll leave it to someone else to describe how some of the earlier "true analogue" sets worked. These really did vary the modulation of the A.M. or F.M. signal to carry the necessary information - more like how voice and music is transmitted on broadcast A.M. and F.M. radio stations. For model control, these early analogue sets were notoriously unreliable, as well as being more complicated and more expensive than the later "digital" sets.
All the proportional sets on the market these days are "feedback" systems. In these, the receiver feeds a signal to the servo wich is proportional to the stick position. In the servo, this is compared with an internally generated signal, which is proportional to the servo position. The servo the drives to make the two match.
All the first generation proportional systems were truly analogue in operation. For example, the original FlightLink was based on the single channel Galloping Ghost principle. The mark/space ratio represented the rudder signal, the rate represented the elevator and full on or off drove the throttle servo. Obviously you lost control of the rudder and elevator controls, which centralised, while the throttle was being operated. Not really very desirable!
Early American sets, like Space Control and the British RCS Tetraplex managed to get this up to four fully proportional channels by switching between two alternating "tones". The frequency of these tones provided two further channels.
At the receiver, all the signals were converted into a varying voltage, to be fed to the servo. The servo compared this to the voltage from the feedback pot, and drove in a direction to make the two equal.
There were a number of problems with these systems. If the difference between the input and the output on the servo was small, the "error" voltage might not be sufficient to make the servo move! It was extremely difficult to implement more than four channels, and even four proved too much for some manufacturers. Kraft abandoned their analogue system after years of development without ever marketing a single set! Also the update rate to the servos was very slow and sluggish due to the need have at least ten cycles of audio for the discriminators to act on, even at the highest rate of transmission and the minimum mark/space ratio.
In practice, they also suffered from trim drift with varying temperatures. On the plus side, the interference rejection was good!
The breakthrough came when Doug Spreng "invented" the pulse tracking amplifier in the servo. I say "invented", because similar systems had been in use for adjusting satellite dishes by the military, and were meant to be secret! Spreng managed to make it practical for model applications.
The Spreng system replaced the varying voltage with a varying pulse width. This overcame nearly all the problems with analogue systems, at the expense of being more prone to interference.
The beauty of it was that all the signals were either "on" or "off" (0 or 1), hence being regarded as "digital". In truth they were just a different form of analogue. Because no audio "tones" were involved, the data repetition rate could be much faster. The drive to the motor was always "on" or "off", with only the mark/space ratio varying for small errors. This meant that the servo developed full torque even on the smallest of errors. Also, because every channel operated in exactly the same way, you only needed to design one circuit, which could be repeated for every channel.
True analogue systems disappeared quite rapidly after "digital" came along. They couldn't compete on either performance or price.
Even today, most servos operate on the Spreng principle. There are no systems that are truly digital from end to end. So called "digital" servos simply use speed controller technology to drive the servo motor at a higher frequency than the transmission rate of the original signal.
What has changed is how the signal is encoded. Following the switch from AM to FM on 27/35 MHz, it became practical to transmit the data as a PCM code, where a binary number is used to represent the desired servo position. This also enabled error checking to take place. The receiver now required a much more sophisticated decoder, but also offered the opportunity to implement a "fail-safe" if the data was corrupted.
The down-side was that due to the narrow bandwidth available on 27/35 MHz, combined with the need for error checking before the signal could be passed to the servo, there was a noticeable lag between the signal being sent and the servo responding!
Various methods of "creative accounting" managed to reduce, but never entirely eliminate this! And at the end of the day, the signal sent to the servo was good old Spreng pulse width!
The advent of 2.4 GHz solved the bandwidth problem. Most systems use between 200KHz and 2MHz of bandwidth, and the response is once again almost instantaneous.
One or two companies have experimented with a truly digital signal to the servo, but these have never been popular enough to usurp the Spreng system. This is largely because of "vendor lock-in". They will only work with their own servos and receivers, whereas the Spreng based servos can be swapped between manufacturers without problems.
Even those so-called fully digital systems still rely on an analogue feedback device - usually a pot - to provide the error signal. Until someone comes up with a micro-miniature optical encoder to replace the feedback pot (and stick pot!), we will never have a fully digital system!
Pete mentioned the RCS Tetraplex in his explanation above, thought it would be useful to share with you guys what it looks like with my example, (which I am so fortunate to have found) from my collection:
You will note the almost military look to the set, it certainly has a 1950s old electronics smell!
Last edited by stuart mackay on 15 Mar 2025, 15:24, edited 1 time in total.
As mentioned by Pete above, here is a somewhat used and abused Series 1 Flight Link from my collection. It is a 2+1 system with a faster frame rate than the Galloping Ghost systems of the time. It was designed to be operated as a "cuddle box" single stick unit with the throttle controls mounted on the right hand face and be operated by the pilot's left hand.
The lower pic shows Mike Kitchen's pristine example showing subtle differences in the stick bezel to my example. I believe Mike's is a very early example, while mine shows serial number 105. I think this illustrates the cottage industry style of manufacture and was probably made in Idris Francis back bedroom while he was at University!?!
And to complete the trio of radios, Pete mentioned, here is a pic of what the Space Control transmitter looked like. Thanks to my friend Gen Hirose from Japan who supplied the photos of his example from his extensive vintage RC collection!
Fascinating stuff chaps. This discussion answers many of the questions I've had over the years about how our R/C systems worked. The whole notion of whether something is/was digital or not has always made me wonder about so called digital servos today. How can something be truly digital when the output is clearly analogue? You can assign a servo a range of numbers for each position of a servo arm (each degree perhaps), but the servo would still look essentially analogue. After all that is exactly what control surfaces are. Obviously as the sampling rate increases (Say each step is now every 0.1 of a degree) the output starts to look even more analogue (proportional). This resolution is obviously fundamental to the digital to analogue conversion (and back) in these systems.
This reached bit of a climax when CDs came in for recorded music. The deal being that if we analyse the electrical analogue signal from a microphone which is recording some music and assign a number to its instantaneous amplitude every, say 0.1 of a second it will give a pretty rough equivalent to the true analogue signal. By increasing the sampling rate to much shorter time slices the analogue to digital conversion quality improves dramatically. Digital audio on CDs was sampled at 44,000 times per second (44Khz) and the quality was deemed to be as good as (or better than) the old analogue recording/playback. This brought out the HiFi buffs in droves claiming that their music could never sound as good on CD as it did on their vinyl records with the old stylus, cartridge and turntable combination. Personally I think the deterioration of the quality of music after this era has more to do with the music than the quality of digital /analogue conversion. But I digress.
Julian
Can anyone here help?
