Fullscreen HTML5Med Res (??MB)HTML4

RADIO CONTROL BY BOB YOUNG How radio-controlled models can be lost through interference This month, we will continue with the comparison between AM and FM and examine some of the ramifications of the two sys­tems. One surprising result is the ease with which a model can be lost through interference. In the last November 1996 article I mentioned that we re­ceived many letters and phone calls about the Mk.22 R/C system and that the most common query was why 29MHz AM? As a result we went into an in-depth analysis of the relative merits of AM and FM and concluded that both systems were incorrectly named and that “FM” was greatly oversold against “AM”. In the end we demonstrated that the difference between the two sys­ tems was much less than commonly believed. The 29MHz discussion we left in abeyance and we will have to deal with that another time. We also dealt briefly with capture effect in FM models which in my Fig.1: the scope trace at the detector of an FM receiver running off a 6-channel transmitter and with a 7-channel trans­mitter interfering on the same frequency at a signal level of approximately 1:2. 74  Silicon Chip mind is a very serious issue, especially when we come to single conversion receivers operating on 36MHz. Let me ex­plain. In our discussion on capture effect in FM receivers we noted that capture is a phenomenon that occurs when an interfer­ing signal on the same fre­ quency exceeds the wanted signal by a small margin. The actual point at which capture occurs depends on the capture ratio of the receiver and may vary from 1dB (1.12:1) to a maximum of 3dB (1.41:1), whereas capture in AM receivers occurs with signal levels of 100:1 or more. Figs.1, 2 & 3 show the sequence of events leading to cap­ture of a radio control receiver by an interfering trans­ mitter. Fig.1 shows the scope trace at the detector of an FM receiver running off a 6-channel transmitter and with a 7-channel trans­mitter interfering on the same frequency at a signal level of approximately 1:2. Fig.2 shows the same Rx with the two transmit­ters at approximately equal level. Note the severe disruption of the signal. Fig.3 shows that the signal from the 6-channel transmitter has gone bye-bye and it’s hello to the 7-chan­nel transmitter signal. Control has now passed over completely to the inter­ fering transmitter, which has exceeded the 1:1 signal level ratio. At this point, the interfering trans­ mitter now has complete control and the model could easily be flown away. Indeed, when FM first made its appearance in England, the press there reported on a spate of incidents where models were flown away by pirate transmitters. “So what?”, I can hear the “experts” Fig.2: the same Rx as in Fig.1, with the two transmit­ters at approximately equal level. Note the severe disruption of the signal. saying. Nobody is going to be silly enough to fly two models on the same frequency and in years of flying FM in Australia, there has never been a recorded incident of a model being pirated away. Well let me tell you there is still a very definite risk of running into strong interference every time you fly on a field using both ends of the 36MHz frequency allocation. There may be no intention of deliberate interference or pirating but you could still lose your model. When you see how this interference and capture can easily take place you will see that there is still a strong ar­ gument for operation on the 29MHz band. Interference on 36MHz For some time now, there have been rumblings amongst the technically inclined R/C modellers about the pos­ sibility of transmitters spaced 455kHz apart causing interference with each other. The significance of 455kHz is that it is the intermediate frequency (IF) used in all R/C receivers. This has been rein­forced by the number of glitches experienced on some flying fields. There have also been rumblings about AM receivers not being satisfactory on 36MHz and this has been put down to inter­ference from harmonics arising from broadcast FM stations. The problem arises on the 36MHz band due to the fact that it consists of 59 spots spaced 10kHz apart in a 600kHz block. The 27MHz, 29MHz and 40MHz bands are only 300kHz wide or less and therefore there can be no trans­ mitters spaced 455kHz apart in these bands. Hence we are looking at some­ thing relatively recent from an R/C point of view. Thus at each end of the 36MHz band there are a number of frequencies which are spaced either 450 or 460kHz apart. All of this has lead to a deal of con­ fusion on exactly how to handle the situation. It particularly affects me because as the designer and supplier of the Silvertone Keyboard system of frequency control, I am responsible for arranging the keyboards for safe operation on 36MHz. Up till now I have always recom­ mended that where single conversion receivers are used, they should be on frequencies in the middle of the 36MHz band, while dual-conversion receivers could be used with frequen­ cies at each end of the band. In other words, use dual-conversion receivers on channels 601-614 (36.010MHz to 36.140MHz) and channels 646-659 (36.460MHz to 36.590MHz) and sin­ gle conversion receivers on channels 615-645. The rationale behind this is that by using single conver­sion receivers in a band only 300kHz wide, a 455kHz difference signal would not arise in the mixer and therefore no interfer­ ence would occur. For the double conversion receivers, the first IF is 10.7MHz and therefore the possibility of the 455kHz difference would not be a problem. That’s as I saw it, anyway. How wrong I was! All of this as­ sumes that the only two channels being affected were the overlapping pair of transmit­ters. The situation is complicated some­ what by the fact that 455kHz falls midway between two frequencies. As long as we use only 20kHz keys (2inch), the key width protects us from this compli­cation. What has forced me to rethink this problem has been a host of discussion about the existing keyboard and its shortcomings in dealing with the new MAAA 10kHz frequency alloca­ tions. When I sat down to write the instructions for the new keyboard I thought that the 455kHz difference was not really a problem. But when I thought about it in detail I realised that maybe I was coming at it from the wrong angle. I needed to get the facts, so I warmed up the old spectrum analyser, signal generator and CRO and went to it. All of this of course was one day before the magazine deadline (as usual, I can hear Leo muttering). One hour later I finished my refresher course on mixer theory and realised we had all been thinking inside the square. What I rediscovered is this: Any pair of transmitters separated by 450kHz or 460kHz will generate a very high level of signal in the mixers of single conversion receivers, AM or FM! This will happen in every receiv­ er operating on that flying field, regardless of frequency! In other words one overlapping pair of transmitters will interfere with all 59 receivers operating on the 36MHz band. Now this is pretty startling stuff and needs some explana­tion but it is really quite simple. First of all, the receivers operating in the 36MHz band (or any other band for that matter) are wide open to all frequencies in that band. That means that a single conversion receiver which may have a crystal in the centre of the band still receives all the transmitted frequencies across that band – nothing too radical here, so far. What happens is that normally all of the difference fre­quencies between February 1997  75 Fig.3: here, the signal from the 6-channel transmitter has gone bye-bye and it’s hello to the 7-chan­nel transmitter signal. Control has now passed over completely to the interfering 7-channel transmitter, which has exceeded the 1:1 signal level ratio the incoming frequencies and the local oscilla­tor (crystal) frequency appear at the output of the mixer. How­ev­er, the IF amplifier is a very narrow filter which will only pass a 455kHz difference signal to the receiv­ er detector. That is why we change both the transmitter and receiver crystals when we change frequencies, so that the difference between the two is 455kHz. But if we also have two other trans­ mitter frequencies on the band which differ by 455kHz, they will be picked up by the front end of the receiver, will be fed through to the mixer and the same difference frequency will automatically appear at the output. So now we have an apparently legitimate signal at the output of the mixer which is very much an inter­ fering signal. Whether this becomes a problem or not depends on its strength in com­ parison to the wanted control signal. And this is where capture effect comes into its own. Capture effect usually works in our favour and tends to lock out the 455kHz interference in all but the worst cases. Thus, it is very difficult to simulate the problem in a simple three transmitter field test. The danger arises mainly in situations where the trans­mit­ter radiation patterns sudden­ ly favour the interfering pair. Now the effects on the flying field of all of this have yet to be verified and extensive testing needs to be put in train immediately. In practice, the lev­ SILICON CHIP BINDERS These binders will protect your copies of SILICON CHIP.  Heavy board covers with 2-tone green vinyl covering  Each binder holds up to 14 issues  SILICON CHIP logo printed in gold-coloured lettering on spine & cover Price: $A14.95 each (incl. postage in Aust). NZ & PNG orders please add $A5 each for p&p. To order, just fill in & mail the order form in this issue to: Silicon Chip Publications, PO Box 139, Collaroy 2097; Or phone (02) 9979 5644 & quote your credit card details or fax (02) 9979 6503. 76  Silicon Chip el of interference will vary depending on a whole host of factors including capture effect, mixer compression, re­ ceiver bandwidth, oscillator injection levels, relative signal level ratios and PC board leakage. The most probable effect is random interference showing up in the form of brief glitches as models move in and out of trans­mitter radiation patterns. Add to this the random nature of the pairing on any one flying field on any one day. Some days the club would have some overlapping transmitters, some days none, some days a large number. The more overlapping transmitters that are transmitting at any one time, the higher the level of 455kHz gener­ ated in the mixers of all receivers on that field. The effect is cumulative and impossible to predict. Now you can see why capture is so important. Nobody would be silly enough to fly on the same frequency but we are acciden­tally generating the same frequency every day on flying fields all over Australia, wherever overlapping pairs of transmitters are allowed to operate. The testing I have carried out to date is brief and incom­plete. I simulated the 3-transmitter scenario by removing the crystal from a single conversion receiver. I could then work the servos from my modulated signal generator using a second unmodu­ lated trans­ mitter 455kHz away to supply the mixing signal. In this mode, I could achieve the equivalent of normal re­ceiver sensitiv­ ity, depending on the relative strengths of the incoming signals. With a crystal in the receiver (any frequency) and no carrier from the wanted transmitter, this effect was still pres­ ent but diminished somewhat, probably due to mixer compression. I did not test with a carrier because capture would confuse the issue and it is here that extensive testing needs to be done. To reiterate, wherever single conver­ sion receivers are in use, transmitters overlapping by 450kHz or 460kHz must not be used. The foregoing is yet another reason for me to continue to push for 29MHz AM. It is simple, cheap and reliable. It is free of the complications and expense of 36MHz FM and is by far the most cost-effective system for SC sports fliers.