Beg pardon- it was actually the Geneva Frequency Plan of 1975, coming into force in November 1978, that put European MW channels onto integer multiples of 9kHz.
Thanks much for that. I had completely forgotten that the 1978 changes could have had a corresponding effect on IFs. Not just in Europe, either, as that was approximately when countries like New Zealand, Australia and Japan changed from 10 kHz to 9 kHz MW channelling.
Thus I suspect that the 450 kHz idea might have come from Japan, given that the OEMs then needed to accommodate 9 kHz channelling for its domestic and many export markets, and also 10 kHz for the all-important North American market.
I have had a look through the Ambit Catalogues 1, 2 and 3 (I do not have #4), as these were issued at about the time of the change and just after. None mention the 450 kHz IF or filters for it. Amongst the AM IFs and filters mentioned, 455 kHz was the most common, with some references to others, including but not limited to numbers such as 460 and 470 kHz. I don’t recall any discussion in Wireless World of the period to the effect that the 1978 changes required changes in British (or European) receiver IFs. And I think that some makers continued to use 470 kHz. That said, the British post-Copenhagen IFs of 422 and 470 kHz were not, as far as I know, universal for Europe, which probably explains the existence of 460, 468 kHz and so on. So it is possible that some parts of Europe needed different IFs after the 1978 changes.
Some data on hand for several Sansui tuners provides a clue. The TU-X1 “supertuner” of 1979-to-1986 was analogue-tuned and had a sophisticated AM section with PLL demodulation. It had a 455 kHz IF.
The TU-919 tuner, also of 1979-to-1985, was analogue-tuned but was also “digitally quartz locked”, with a digital was as well as an analogue readout. The data shows two alternative AM IFs, namely 450 and 455 kHz. Whether these were “either/or”, perhaps according to destination market, or sequential, with say a change from 455 to 450 kHz during the production run is not clear. But it does seem to have been an early use of 450 kHz by a Japanese OEM. The TU-919 used a Hitachi HA1197 AM IC.
The TU-S77AMX (released 1983) and TU-D99AMZ (1985) were both digitally synthesized, with 450 kHz IF. The former used a Sanyo LA1245 AM IC, the latter either the same or the LA1247.
The American OEMs, probably less concerned about export markets than the Japanese, on the face of it had less reason to change away from 455 kHz than did the Japanese, but 450 kHz was appearing as an IF by 1985 at least. Digital tuning may have been a part of this. The American “car radio” IF seemed to have moved from 260 and 262 to 262.5 kHz in the digital tuning age, although that is based upon casual observation.
Starting in 1982 December, the Australian magazine “Electronics Australia” ran a construction project for a hi-fi AM tuner. This had analogue tuning with a digital readout and a 455 kHz IF. Starting 1984 October, it ran another construction project for an add-on C-QUAM AM stereo decoder, for use with the above or other tuners. This was based upon the Motorola MC13020 IC. In the detailed commentary, after noting that the VCO in that IC ran at eight times IF, it was said: “Typically the IF will be 455kHz so the VCO runs at 3.64MHz. Note, however, that many synthesized tuners use a 450kHz IF and this requires a VCO frequency of 3.6MHz.”
So a possible pathway to 450 kHz might have been that the Japanese OEMs found this to be a convenient and workable number for both 10 kHz channelling and the new variant of 9 kHz channelling, which had the channel frequencies as integral multiples of 9 kHz. Then it was also found to be convenient for digital tuning purposes, which thus pushed it into the mainstream. Quite recently, somewhere I have seen that 450 kHz was included alongside 455 kHz in the appropriate American standard, but right now I can’t find it. The standard would have been ANSI/CEA-109-D of 2010, which I think was the first revision since CEA-REC109-C of 1955, although it was later renamed as ANSI/CEA-REC109-C.
Cheers,
Steve
Discussed upthread was the use of 1.4 MHz as a second IF in professional upconversion HF receivers, which practice may be dated back at least back to the Redifon R550 of early 1969.
The following item from Wireless World 1970 May, about some new Cathodeon IF filters, provides some additional background.
The subject Cathodeon filter range was aimed at the marine market, in anticipation of the then-forthcoming introduction of suppressed carrier SSB for marine R/T communications.
The range included 1.4 MHz USB, LSB and DSB filters, from which it may be inferred that Cathodeon saw this as at least a de facto standard that was likely to be chosen by a reasonable number of marine R/T equipment makers. One assumes that Cathodeon had reviewed the matter with its potential customers before proceeding with design and production.
Also included was an RF bandpass filter covering the range 1.6 to 3.8 MHz. This, I should guess, was intended for the simpler R/T equipments used on coastal and fisihing vessels, etc., where coverage of the higher HF marine bands was not required. This filter also had 70 dB rejection at 1.4 MHz.
The logical deduction here is that 1.4 MHz was seen as a suitable IF for single-conversion receivers that covered the 1.6 to 3.8 MHz marine HF band, as well as the marine MF band below say 520 kHz. On the other hand, 1.6 MHz or thereabouts would not have been workable. Then also, 1.4 MHz was suitable for dual-conversion receivers where 1.6 MHz might otherwise have been used. So, one set of IF filters would cover both cases.
Confirmation of that deduction is provided in this item from WW 1970 October, announcing the introduction of a similar range of 1.4 MHz IF filters from Plessey, including one suitable for CW, as well as a 1.6 to 3.8 MHz bandpass filter with 1.4 MHz rejection.
So with 1.4 MHz established as at least a quasi-standard number and with a full range of professional-quality filters available, it is not so surprising that Racal, Plessey, Eddystone, Marconi and others also adopted it.
Cheers,
Steve
Here is another case where the products of an IF filter maker provide an historical reference point. Here it is Salford with 37.3 and 21.4 MHz filters, in a WW 1968 October item.
37.3 MHz was the 1st IF used by Plessey in its PR155 professional HF receiver. I am not sure, though, that it was a widely used IF.
The 21.4 MHz filter was mentioned in connection with HF receiver applications, but with a 30 kHz bandwidth, it looks more like a VHF/UHF receiver filter. At the time, 10.7 MHz filters for VHF/UHF communications receivers were usually available in 30, 15 and 7.5 kHz bandwidths or thereabouts. This is the earliest mention I have seen of the 21.4 MHz number.
Cheers,
Steve
30KHz is fine for a 1st IF in a dual conversion receiver. Even 200KHz wide first IF is fine on a dual conversion with 450 to 480 KHz second IF. The main purpose of the high first IF is for image rejection, secondarily intermodulation, so 30KHz is better than 200KHz for intermodulation performance / noise on the 2nd mixer.
It's only on a single conversion IF you need the 9KHz (or whatever) bandwidth.
The Sony ICF5900 uses the 10.7MHz FM IF filters as 1st IF on dual conversion for Shortwave. That became a common approach. The ICF2001D is also dual conversion, but I don't remember if it uses same filters for FM IF and AM 1st IF.
Some cable receivers and modems receive 80MHz to 870MHz. They use a quite wide 1.2GHz first IF. The second IF is often approximately 44MHz (USA TV IF I think) and before SDR versions was 2MHz, 4MHz, 6MHz or 8MHz wide. Europe usually is 8MHz DOCSIS modem/ Digital Cable and USA 6MHz (to be compatible with Analogue channels on same cable) but data only wireless linked systems using DOCSIS cable modem indoors and 2.5GHz, 3.5GHz, 6GHz, 10GHz or 12GHz outdoor full duplex radio sets, or 700MHz, 800MHz, 900MHz directly can be using 2MHz, 4MHz, 6MHz or 8MHz channels depending on country and band in use. Only the 10.5GHz band 8MHz version is much in use in UK and Ireland, but other bands used in Canada and Caribbean.
In respect of the 1st IF filter bandwidth for an HF receiver, I guess that it depends upon the desired performance level. The folks at Racal provided a succinct summary when writing in 1974 about their then-new high-performance RA1772 HF receiver, which was I think at least epoch-marking if not epoch-making.
“Although it is sometimes beneficial to frequency selection it is never advantageous to the receiver performance if the first i.f. bandwidth is wider than the final output bandwidth. The highest possible amount of single-signal and dynamic selectivity are required both of which are obtained if the bandwidth is made narrow as soon as possible. It can be arranged for all frequency selection processes to be made in the first mixer, with fixed frequency injection in the subsequent mixer(s), so that a narrow first i.f. filter can be used. This filter can also be a crystal type so that its bandwidth need only be wide enough to pass the widest i.f. bandwidth envisaged, normally ± 6kHz. This allows protection to subsequent stages against signals farther off-tune than 10kHz and considerable pro-tection at 20kHz off-tune. Having such protection we may concentrate on providing a very high linearity in the stages which are wide-band, particularly the first mixer and r.f. amplifier.”
As well as Racal (35.4 MHz 1st IF in the RA1772), others used a 12 kHz bandwidth 1st IF filter, examples being found in the Plessey PR155 (37.3 MHz 1st IF), Marconi Oceanic (75 MHz 1st IF) and JRC NRD-525 (70.455 MHz 1st IF) to name just one or two. For upconversion HF receivers, 1st IFs seemed to be very much an individual manufacturer choice.
Presumably wider bandwidth filters were chosen on cost grounds for lower-priced, lower-performance receivers in the consumer class, or, as you say, because a given IF strip, typically 10.7 MHz, was required to handle both wideband FM and AM signals. Also, wider bandwidth was required where the 1st IF was variable in order to accommodate fine tuning or interpolation tuning.
Regarding 21.4 MHz, I can’t say that I have seen evidence of its use as an IF in HF receivers, but then I am hardly guilty of rigorous research in this regard. It would seem to be a bit of an awkward number though, as it is just below the lower edge of the 13 metre broadcast band, 21.45 to 21.75 MHz. On the other hand, 10.7 MHz fitted neatly between the 25 and 31 metre bands. My best guess is that 21.4 MHz, as a doubling of 10.7 MHz, was chosen to provide better image rejection in professional VHF equipment, particularly that which tuned beyond 200 MHz. The range of filters available for 21.4 MHz seems to parallel that for 10.7 MHz. The Eddystone 1990R VHF receiver of the 1970s, which could be equipped to tune from 25 to 500 MHz, had a 21.4 MHz IF. The available IF filters had bandwidths of 250, 30, 15 and 7.5 kHz. The last three were said to be respectively for 50, 25 and 12.5 kHz channeling.
The earlier Eddystone 990R VHF receiver, covering 27 to 240 MHz, had a 10.7 MHz IF. In a data sheet dated 1967 May it was said: “…bandwidths of 30 kHz and 200 kHz are provided as standard, the former employing a crystal filter, alternative filters being available to order.”
A data sheet of nearly a decade later, 1976 November, stated: “It can be supplied with crystal filter to suit 12.5, 25 or 50 kHz channel spacing, the latter [sic] being fitted as standard. A wide selectivity position of 200 kHz allows the receiver to be used as a monitor for wide-band FM transmissions.” In the tabulation of specifications, the standard IF bandwidths were quoted as 30 and 200 kHz, thus connecting the 30 kHz number with 50 kHz channeling et seq.
Anyway, it would appear that in the late 1960s, 30 kHz IF bandwidth (for 50 kHz channeling) was the norm for VHF communications, whereas by the late 1970s, narrower bandwidths for closer channeling had become important. If we assume that the 21.4 MHz IF came into use in VHF equipment in the late 1960s, then the 30 kHz bandwidth filter would likely have been the first crystal type made available. And it could be that Salford had this application in mind – amongst other possible uses – with its 1968 filter release.
Cheers,
Steve
I had been mildly curious (though hardly enough as to lose sleep) as to why the Pye (originally Rees-Mace) CAT619 maritime general coverage receiver (I think introduced around 1953, so presumably an early adopter of the ECH81 over the ECH42) had adopted 1.4MHz as 1st IF, rather than anything somewhat higher- especially considering its main 460kHz IF strip. However, it makes sense if you consider that, as a marine-orientated receiver, there would have been need to offer uninterrupted (and, hopefully, largely sprog-free) coverage from the top of the MF broadcast band across the lower HF bands. The established circa 1.6MHz IFs would, as said, have risked breakthrough in this environment and the chance of broadcast stations around 1.4MHz giving trouble was presumably felt to be low.
Whilst this example used LC filtering at 1.4MHz (a half-lattice crystal filter featuring at 460kHz), perhaps it showed in service that this frequency was a reasonably trouble free-spot to adopt crystal filtered IF technology later on- I think Redifon were a largely marine-market orientated outfit. As mentioned upthread, this frequency also suited uninterrupted receiver coverage between LF and MF broadcast bands.
Perhaps by time 21.4MHz filters came along, the electronics had advanced far enough that generating a synthesized LO in the 35-85MHz region was sufficiently economic (at least in professional terms!) that an elegant, effective and potential low-sprog solution for high dynamic range receiver front ends was a steep 30MHz low-pass filter with 35MHz+ IF, thus keeping front-end and IF strip out of each others' hair, so to speak, and 21.4MHz (or any other in-band frequency) would have involved more awkward work-arounds. I'd agree that 21.4MHz seems more VHF/UHF receiver orientated.
Presumably wider bandwidth filters were chosen on cost grounds for lower-priced, lower-performance receivers in the consumer class, or, as you say, because a given IF strip, typically 10.7 MHz, was required to handle both wideband FM and AM signals. Also, wider bandwidth was required where the 1st IF was variable in order to accommodate fine tuning or interpolation tuning.
Purely cost on non-professional gear, about 100th of cost, yet removes the plague of approx 910kHz image on HF, which is quite severe on some single conversion sets with no third RF tuning section from 1935 to the cheap Tesco World receiver today!
One may wonder whether the reasonably widespread adoption of 1.4 MHz as a final IF for professional upconversion HF receivers, apparently from 1969, was in any way influenced by the fact that Rees Mace had used this number as a 1st IF for its CAT receiver, designed to an admiralty specification. By the way, that receiver was described in Wireless World (WW) 1954 July, p.333ff.
As previously mentioned, Redifon, with its R550, announced in WW 1969 March, p.136, looks to have been the first upconversion model to use the 1.4 kHz final IF, and soon thereafter this number was supported by the filter makers. Redifon was a major supplier of marine HF receivers, and perhaps unusually during the valve era, used common designs for marine and general-purpose models, whereas for example IMR and Marconi designed dedicated marine HF receivers. With Redifon, the commonality goes back at least to the R50, which was advertised at least as early as 1947 (WW 1947 November p.41), and which was described in WW 1949 July p.251ff. The marine version of the R50 was the R50M; the marine version of the R550 was the R551, announced in WW 1970 January p.41. So Redifon would have had marine applications in mind when it was designing the R550/R551, and like Rees Mace, would have wanted an IF that was outside of (below) the upper MF marine band in which vessels might transmit.
I don’t think that Racal was much concerned about the marine market – although it certainly had sold its RA17 receiver into marine applications, albeit for ancillary purposes. Rather it said that the choice of 1.4 MHz as final IF for its RA1772 was due to the ready availability of suitable filters; an example of using an established and generally suitable number rather than starting from scratch.
More-or-less a contemporary of the Redifon R550 was the Eddystone EC958, announced in WW 1969 July p.322 (and I understand earlier in other trade journals). Whereas previously Eddystone had stayed away from the marine HF receiver market (except for designing and initially building the IMR54), the EC958 was designed with marine applications in mind. The marine version was the EC958/5, also rebranded as the Marconi Nebula, which thus competed with Marconi’s own Apollo design. Eddystone had opted for a more traditional design rather than upconversion, albeit one that incorporated what was more-or-less a narrow-band Wadley loop. It was triple conversion with IFs of 1235 to 1335 kHz (tuneable), 250 kHz and 100 kHz. The tuneable 1st IF was well established in Eddystone practice, having been used in the valved 830 series, where the range was 1250 to 1450 kHz, 1350 kHz nominal, evidently chosen to allow its use on incoming frequencies down to 1.5 MHz.
In the case of the EC958, the tuneable IF was used on incoming frequencies down to 1.6 MHz. It was 1335 kHz nominal, tuneable in the downwards direction only. It looks as if it may have been a carryover from the 830; perhaps the 5 kHz offset was something to do with the Wadley Loop implementation. Anyway, it was in the right place, that is below 1.6 MHz, for the marine application. Also probably connected with the Wadley Loop was the need for an intermediate IF of 250 kHz, whereas the 830 had converted directly to the 100 kHz final IF from the 1st 1350 kHz (nominal) IF. 250 kHz looks to have been an ad hoc number, perhaps chosen on the one hand to keep the 2nd mixer local oscillator injection frequency – which was on the high side - below 1.6 kHz, and on the other to keep any images from the tuneable 1st IF far enough away, something that may have been more difficult with conversion from 1335 kHz (nominal) direct to 100 kHz.
But even 250 kHz had a precedent. Recently I read an article about the use of SSB in aircraft HF communications, in WW 1958 October, p.460ff. The transmitter and receiver block schematics shown had IFs of 250 kHz. I suspect that these block schematics were representative of Marconi equipment, given that the article was written by a Marconi staffer. Whether 250 kHz was a norm in aircraft HF practice I don’t know. The only other reference point I have is that of the earlier Marconi AD108 aircraft HF receiver, which was single-conversion with a 600 kHz IF.
All of the above-mentioned WW items and articles are available at: http://www.americanradiohistory.com/Wir ... gazine.htm.
Cheers,
Steve
The use of 21.4 MHz as a 1st IF in VHF receivers may go back earlier than I first thought, which was the late 1960s.
Wireless World 1966 April, in a survey of communications receivers, listed the Rohde & Schwarz ESM180 VHF receiver, covering 30 to 180 MHz, with IFs of 21.4 and 3.4 MHz.
It turns out that there is quite a bit of information about this model on the web, including an Instruction Book at: http://bee.mif.pg.gda.pl/ciasteczkowypo ... ESM180.PDF. Rmorg dates it from 1956, which seems reasonable considering that its valve complement included some Rimlocks.
In wideband mode, it is single conversion with an IF of 21.4 MHz. In narrow-band mode, a second conversion to 3.4 MHz is used. There is a similar model of apparently the same vintage, the ESM300, which covered 85 to 300 MHz and had the same pair of IFs.
One may surmise that Rohde & Schwarz thought that the standard 10.7 MHz IF was not high enough to avoid image problems at the higher end of the tuning range, and simply doubled it to 21.4 MHz. Even then, the ESM180 had a four-gang front end. The 2nd IF of 3.4 MHz appears to be a very individual choice.
The 21.4 MHz IF transformers in the ESM180 were wideband; if the previously quoted Salford case is indicative, then it was not until the late 1960s that the filter makers offered package narrow-band 21.4 MHz units, say about a decade after they did the same for 10.7 MHz.
Cheers,
Steve
The lowest radio receiver IF that I have found to date is 18 kHz, used in the various Racal SSB and ISB outboard adaptors (RA63, RA98, RA121, etc.) intended for use with the RA17 series of receivers.
These took their feed from the 100 kHz final IF of the associated RA17 family receiver. Another conversion was necessary in order to facilitate the inclusion of an on-board AFC loop for pilot carrier transmissions, thus obviating the need for arranging for AFC control of one of the receiver oscillators.
So why 18 kHz? Here’s an attempt at post facto rationalization: Racal used multi-section LC low-pass and high-pass filters for sideband separation, and I imagine that these were easier to implement and had steeper slopes (on a per kHz basis, not a per octave basis) at lower frequencies. So the goal may have been to have a final IF that was as low as reasonably possible. Given that the input bandpass was 94 to 106 kHz, that band, and in fact somewhat beyond it, would need to be free of any IF harmonics. So 18 kHz, with 5th and 6th harmonics at 90 and 108 kHz respectively, was about as low as was workable. The 118 kHz oscillator frequency was also clear of any IF harmonics.
Also, for greater sensitivity, the AFC discriminator worked on the 5th harmonic of the 18 kHz IF, namely 90 kHz, which was another reason for the harmonics to be well clear of 100 kHz and the passband.
Cheers,
Steve
Some more on the 1.4 MHz IF number mentioned in earlier posts:
Redifon also used it for its fixed-tuned, single-conversion R499 receiver. This was announced in WW 1968 June (page 184), so it preceded the R550 and R551. Apparently it used essentially the same circuitry as those two, except that it had a crystal-controlled oscillator (up to 10 thereof) and fixed-tuned RF coilpacks. The latter must have provided sufficient RF selectivity that single-conversion provided adequate image rejection.
Conceivably in selecting 1.4 MHz, Redifon did look back to the Rees Mace precedent, and perhaps the original Rees Mace and Admiralty work-up was still available. One supposes that there was also some kind of industry informal agreement or understanding, perhaps also including the GPO, that 1.4 MHz was an appropriate number for marine receivers designed around the new SSB requirements. It seems unlikely that Cathodeon and Plessey would have developed standard ranges of 1.4 MHz IF filters without there being reasonable certainty about the market.
And although having marine origins, 1.4 MHz then moved into the wider professional receiver field, with the Racal RA1772 being an important marker.
Cheers,
Steve
It occurred to me that the designers of Eddystone's 910 receiver (late '50's/early '60's) had also come to the conclusion that the 1.4MHz region represented a good place to be for a marine receiver IF. This set could be viewed as a development and "professionalisation" of their 750 theme, apparently specifically aimed at the marine market and incorporating the 750's basic topology and much similar circuitry. Instead of fixed 1.62MHz first IF and 1.535MHz second LO for 85kHz IF, +-50kHz interpolation tuning was now adopted around 1.4MHz first IF with 1.265-1.365MHz second LO. The subsequent 830 receiver could be seen as a refinement of the 910 theme, now with +-100kHz interpolation IF centred on 1.35MHz, rather than 1.4MHz- perhaps this was to ensure adequate clearance between the consequent highest possible 1st IF excursion of 1.45MHz and the 1.5MHz lowest signal frequency of dual-conversion operation.
Thanks; that’s an interesting connection. So the Eddystone 910 was another arrow that pointed to the eventual standardization (de facto if not de jure) of the 1.4 MHz IF for marine receivers.
The lineage may be extended from the Eddystone 830 to the EC958 solid-state receiver. This was triple-conversion (on the HF bands above 1.6 MHz) with a 1st IF of 1335 kHz nominal, but tuneable down to 1225 kHz to facilitate high-stability operation above 1.6 MHz, for which a narrow-bandwidth drift-cancelling loop was used. It could be that Eddystone started with the 1350 kHz precedent of the 830, and adjusted it slightly to suit the high-stability circuitry. The Eddystone paper on this receiver did not go into detail, essentially stating just the main requirement which was for a 1st IF below 1.6 MHz.
The EC958 was designed with marine main receiver requirements in mind, including the new SSB rules, although the base version did not meet those in full. The marine version was the EC958/5 (Marconi Nebula).
The EC958 2nd and 3rd IFs were 250 kHz and 100 kHz respectively, the latter being a well-established standard. 250 kHz might have been an ad hoc number chosen to suit drift-cancelling loop implementation. But there is one datapoint that suggests that it [250 kHz] might have been used for aviation HF SSB receivers. In WW 1958 October there was an article entitled “Single Sideband Aircraft Communication”. This included a diagram of receiver upper sideband and carrier extraction filters, with 250 kHz carrier frequency (page attached). There was no commentary on the IF itself, so whether it was representative of a production receiver or just a “for instance” is unknown.
Cheers,
Steve
Revisiting the Domestic FM receiver standard IF of 10.7 MHz, I mentioned upthread that the earliest mention that I have found was in AWV Radiotronics 1946 May, pertinent page attached:
My best estimate is that the 10.7 MHz proposal was made late in 1945. There appears to have been a quick uptake by the industry. For example, GE used it in its X-415 domestic receiver (which covered both the old and new FM Bands) which appears to date from c.1946. Hallicrafters also used 10.7 MHz on its SX-42. Also from 1946. But there were some exceptions. Zenith used 8.3 MHz on its 8C20 AM-FM chassis, which covered both M bands. But it appears to have switched to 10.7 MHz soon thereafter. Philco used 9.1 MHz for a few years before changing to 10.7 MHz. The 9.1 MHz era included receivers with its own locked-oscillator limiter-demodulator, although the latter appears to have had a relatively short production life.
Post facto it is difficult to impute rationales for the 8.3 and 9.1 MHz numbers. Both had in-band images, but presumably RF selectivity was considered adequate to minimize their effect, given that three-gang front ends were typically used in those days.
Another non-standard number was the 14.1 MHz chosen for the BBC VHF-AM/VHF-FM Comparator receiver designed and built in 1949 by R.N. Fitton (Ambassador Radio). Here though the reason was provided in the IET paper on this receiver, as follows:
“Continuous tuning over the range from 87.5 Mc/s to 95 Mc/s was decided upon as more satisfactory from the point of view of flexibility. A choice of intermediate frequency had to be made which would permit the observance of a 40.db rejection of image without involving the complexity of variable tuning of the r.f. circuits. A pass band of 7.5 Mc/s from 87.5 Mc/s to 95 Mc/s was needed for the pre-frequency-changer circuits and this is discussed in greater detail in the section on the r.f. amplifier. Experience indicated that the intermediate frequency could be somewhere between 9.6 Mc/s and 16.6 Mc/s; detector harmonic interference within the r.f. pass band could be avoided by using any of the following frequencies: 9.6, 10.6, 12.1, 14.1, 16.6 Mc/s. The lower frequencies are dangerous since they only approximate to the r.f. pass band and image rejection is consequently jeopardized. At the same time, the higher the selected frequency the more difficult becomes the oscillator drift situation, should the oscillator operate at the high-frequency side of the signal.
“In view of these facts, the first item to be designed was the radio frequency circuits. As a result of the curves obtained for these circuits, the lowest safe intermediate frequency to provide the 40-db image rejection figure is 14.1 Mc/s, and this was therefore adopted.”
In this case the key to the IF choice was the desire to use an essentially aperiodic front end, which resulted in additional constraints. That knowledge prompted another look at the Bush VHF54 case. Here a 19.5 MHz IF was used, notwithstanding that at the time, the UK setmakers were generally using 10.7 MHz. Given that 19.5 MHz was a common UK TV sound IF (SIF) at the time, it would be easy to explain the Bush choice as being made simply and only for commonality with its concurrent TV receivers. But a look at the VHF54 schematic shows that it had an aperiodic front end, with both the RF amplifier and mixer valve inputs being broad-band. Whilst an untuned RF amplifier input was common in UK practice, an untuned mixer input was not.
Quite possibly then the fully aperiodic front end was why Bush needed to go to a higher IF than the standard 10.7 MHz. With a 12.5 MHz tuning range (87.5 to 100 MHz), an IF reasonably beyond 12.5 MHz was required to avoid any in-band and near in-band images. Avoiding in-band IF harmonics would have been more difficult. The ranges 16.67 to 17.5 and 20.0 to 21.875 MHz present themselves as possibilities. The former may well have been too low for adequate image rejection. Once the higher range was seen was seen desirable, Bush may have elected to compromise a little and go with the established 19.5 MHz SIF number. That put the IF 5th harmonic at 97.5 MHz, probably considered not too much of a problem given that there would be no UK FM transmitters above around 96 MHz for quite a few years to come. There was some method in Bush’s choice beyond simple harmonization with the TV SIF case.
BREMA’s endorsement of the 10.7 MHz FM number was noted upthread as being covered in a Wireless World (WW) 1958 January item:
That included a recommendation for oscillator-low (infradyne) operation. The BREMA endorsement evidently predated that WW item though, as shown by these WW items from early 1956:
That was an interesting exchange, with a well-judged rejoinder from the BREMA side. The protagonist. G.H. Russell, associated the origins of 10.7 MHz with the development of the ratio detector, but in fact it predated that.
Cheers,
Steve
The CCIR Plenary Meeting receiver IF reports mentioned in the Television Receiver Intermediate Frequencies thread ( https://www.radios-tv.co.uk/community/black-white-tvs/television-receiver-intermediate-frequencies/) also made some mention of domestic radio receiver IFs, although not in the same detail as for the TV case. Although they bring little that is new to the discussion, for the record it is worth noting the salient points.
Report #41 from CCIR 1956 Warsaw included the following:
(a) The principal factors contributing to the production of undesired responses of superheterodyne receivers are inadequate image and intermediate-frequency response ratios and the generation of intermediate-frequency and frequency-change oscillator harmonics. Measured values of the undesired responses occurring in receivers of various types are given in Doc. No. 157.
(b) In the case of long-wave, medium-wave and short-wave sound-broadcast receivers, the only method of improvement not requiring an undue increase in cost is the choice of a suitable value of intermediate-frequency; no single value of intermediate-frequency is completely satisfactory for all parts of the European zone. Intermediate frequencies in the range of 420 to 475 kc/s are commonly used.
Some elaboration of (b) was provided in Report #98 from CCIR 1959 Los Angeles:
2.2 In the case of long-wave, medium-wave and short-wave sound broadcast receivers, the only method of improvement not requiring an undue increase in cost is the choice of a suitable value of intermediate-frequency; no single value of intermediate-frequency is completely satisfactory for all parts of the European zone. Intermediate frequencies in the range of 420 to 475 kc/s are commonly used. No specified value for the intermediate frequency can be recommended, because it will be of advantage to be able to avoid interference by convenient choice of intermediate frequencies according to the different situations with respect to powerful transmitters operating in these bands. Self generated beats and whistles can be avoided by conventional technical means with any of the above-mentioned values. For receivers of high quality the problem need not be considered, as an adequate IF and image rejection is always assured.
This report also covered the FM receiver case in a straightforward way:
2.3 For domestic frequency modulation receivers, the intermediate frequency of 10.7 Mc/s, normally used, is satisfactory provided that radiation by the receiver on the IF and its harmonics and at the frequency of the local oscillator and its harmonics, is sufficiently reduced.
Essentially the same statements in respect of AM and FM receivers were repeated in 1963 Geneva Report #184 and 1966 Oslo Report #184.
In Report #184-1 from New Delhi, 1970, the AM receiver wording was changed:
1.2 For long-wave, medium-wave and short-wave sound broadcast receivers, the only method of improvement not requiring an undue increase in cost is the choice of a suitable value of intermediate frequency.
No single value is satisfactory for the European zone.
Doc. X/161, Oslo, 1966, gives a sample calculation of interference which may be necessary in determining the correct intermediate frequency for standard-type domestic receivers in current use.
Although the choice of an intermediate-frequency value is not vitally important for high-quality receivers, it is for low-cost standard-type receivers, in order both to avoid interference and to reduce their cost by designing several series with a single intermediate frequency value.
An addition though was a table showing the standard AM and FM radio receiver IFs applicable in France. These were the subject of SCART documents.
10.7 MHz for FM receivers was hardly surprising.
That both 455 and 480 kHz were listed for AM receivers suggests that (1) the ubiquity of the American 455 kHz number, and its use by for example the Japanese setmakers, meant that it could hardly be excluded; and (2), 480 kHz was the best choice for LF and MF reception in France under the reception conditions then existing.
Report #184-2 from 1974 Geneva did not include any material changes.
But Report #184-3 from 1978 Kyoto included a somewhat more expansive AM receiver table:
The UK numbers were interesting, given that the original BREMA post-Copenhagen numbers were 422 and 470 kHz. Somewhere along the way there must have been a rethink that introduced the 460 kHz number.
I think that 468 kHz was also commonly used in Western Germany.
Report #184-3 was repeated in the 1982 Geneva and 1986 Dubrovnik meetings, and cancelled at the 1990 Dusseldorf meeting.
Cheers,
Steve P.
In respect of domestic AM receiver IFs, and interesting situation arose in Japan, where in the later 1950s and early 1960s, before the adoption of FM stereo broadcasting, there was some stereo broadcasting using pairs of AM transmitters. Thus the industry offered dual AM receiving equipment, including tuner-amplifiers that included two AM tuners.
An example, namely the National SA-32, is illustrated at this site: http://www.japanradiomuseum.jp/FMstereo-e.html#?????SA-32?.
Included was the comment: “To avoid interference, each channel had different IF frequencies.” Not surprising, really, and the interference possibility was probably worse if the two tuners shared an external antenna. Unfortunately, the actual IFs are not stated.
An inspection of the picture shows that the left-hand tuner had SW as well as MW coverage, but that the right-hand tuner covered MW only, there being no need to duplicate SW coverage. It would be logical then that the left-hand tuner had the higher IF. Perhaps it used the standard 455 kHz IF, whilst the right-hand tuner may have had something like 260 kHz, an older AM receiver number that lingered on in car radio applications. That assumes that the 455 and 260 kHz numbers could cohabit without major problems. Otherwise, the lower IF may have been an ad hoc number. To some extent what was required would have depended upon the frequency separations of the AM transmitter pairs actually used. Although potential problems with a given pair might be averted by choice of which tuner was used for which station. For example, if the transmitter pair frequencies were say 620 and 810 kHz, then using the 455 kHz IF for the first and the 260 kHz IF for the second could be problematical, in that the respective LO frequencies would be very close, at 1075 and 1070 kHz. Reversal of the IFs would result in LO frequencies of 880 and 1265 kHz, with no obvious problems.
Obtaining reasonably similar IF bandpasses and so AF responses for each stereo channel when each had a different IF probably required some care.
Cheers,
Steve
Another IF that would appear to justify inclusion in this list is 9.72 MHz. It appears to have been used in British and perhaps European marine and aviation VHF receivers in the late 1940s and into the 1950s. Examples were the STC STR.9-X (aviation) and “Receiver 62H” (marine). It was also found in the BBC HR/12 outside broadcast wireless microphone receiver, which was described as being based upon the EMI Type 1250 receiver.
The UK MoD/RAF Receiver 62H was an aviation VHF-AM receiver, covering the range 100 to 156 MHz, and intended for ground station or shipboard use. It was the successor to the functionally similar R1392 series, the main difference being that the 62H had an IF of 9.72 MHz, double that of the 4.86 MHz of the R1392. The latter was a WWII receiver. The 62H might have been from WWII, or perhaps it was from the late 1940s. Presumably the doubling of the IF was done to obtain better image rejection, without upsetting whatever IF harmonic interference considerations led to the original choice of 4.86 MHz.
So at least the immediate origin of the unusual 9.72 MHz IF number is now known. But that just moves the question back to “why 4.86 MHz?”
Cheers,
Steve
Not an answer but a suggestion about the 4.86Mhz IF. The aircraft would have had other electronics on board and just a thought that this IF was chosen so not to interfere with other equipment.
Excellent research Steve.
Frank
Thanks for that. One possibility that came to mind following your comment is that it may have been desired to avoid having the IF or its significant harmonics fall into one of the HF aviation bands.
Let’s say that at the time, an IF of around 5 MHz was determined as being a reasonable trade-off between image rejection and obtaining adequate IF strip gain and selectivity with the technology of the day.
I haven’t checked as to what were the aviation HF allocations in the WWII period, but I have used the ITU 1947 Atlantic City allocations as a proxy.
5 MHz itself, and its multiples up to 25 MHz were standard frequencies, so needed to be avoided. An IF slightly higher than 5 MHz would have put harmonics into the 10 and 15 MHz aviation HF bands, which started just above 10 and 15 MHz respectively. So that meant that the IF should be below about 4.99 MHz.
The 23.2 to 23.35 MHz HF aviation band indicated that the fifth harmonic of the IF should be above 23.35 MHz, meaning an IF of above 4.67 MHz. So now we have a “possible” band of 4.67 to 4.99 MHz.
There was also an aviation HF band that ran from 4.65 to 4.75 MHz worldwide, and which was extended to 4.85 MHz in Region 1 only. So preferably the IF needed to be above 4.85 MHz, narrowing the available range to 4.85 to 4.99 MHz. If this reasoning applied, then it looks as if 4.86 MHz was chosen to be just above the 4.85 MHz minimum.
Speculation admittedly, but the numbers do appear to “add up”.
Cheers,
Steve
