Featured
Latest
Radio
148
Posts
10
Users
39
Reactions
35.7 K
Views
Topic starter
The 3.1 MHz 1st IF number sometimes used in point-to-point ISB receivers has been discussed previously, in this post:
The following is a recap, and an extension of the analysis.
The 3.1 MHz 1st IF was used in the GPO point-to-point SSB/ISB receiver of the late 1940s/early 1950s. The 2nd IF was 100 kHz, as would have been expected for this type of receiver at the time. The tuning range was 4 to 30 MHz. In respect of that, 3.1 MHz was a reasonable 1st IF, but one could ask why 3.1 MHz, and not another number in the vicinity of 3 MHz.
The answer is found in the corresponding SSB/ISB generation and transmission equipment developed by the GPO at the same time as part of the whole suite. The scheme was described in an IEE paper of 1947 (1).
SSB/ISB was generated at 100 kHz, logical considering the ready availability of suitable filters, the same as used at the receiving end. This was converted to an intermediate frequency, and then up to the transmission frequency in the range 4 to 30 MHz. For transmission frequencies above 10 MHz, the sideband positioning was maintained as generated. For transmission frequencies below 10 MHz, the sideband positioning was inverted, so that the upper sideband as generated became the lower sideband as transmitted. This was aligned with the then-prevailing convention.
This sideband inversion at 10 MHz was obtained by using infradyne conversion to the IF, and then infradyne conversion from the IF to transmission frequencies above 10 MHz, and supradyne conversion to transmission frequencies below 10 MHz. This also had the advantage of reducing the required range of the transmission frequency conversion oscillator.
With this arrangement, it may be seen that the IF needed to be below 3.33 MHz, the transmission frequency above 3.33 MHz (not a problem given the 4 MHz minimum), and the conversion oscillator frequency above 6.67 MHz.
Working through the 4 MHz minimum transmission frequency case results in an IF of 3.0 MHz and a conversion oscillator frequency of 7.0 MHz and upwards. 7 MHz converts 3 MHz to 4 MHz (supradyne) and 3 MHz to 10 MHz (infradyne). The required conversion oscillator range is 7 to 27 MHz. the 7 to 13 MHz part is used both for the 4 to 10 and 10 to 16 MHz transmission frequency ranges. The 13 to 27 MHz part is used for the 16 to 30 MHz transmission range.
That leads to the question as to why the IF was offset slightly from 3.0 to 3.1 MHz. That relates to the first upward conversion from 100 kHz at the transmission end. For relative simplicity and perhaps to minimize the possibility of spurs, it was desired to use the same crystal oscillator to generate both the 100 kHz required by the SSB/ISB generator, and the conversion oscillator frequency required to translate 100 kHz to the IF. That meant multiplying up from 100 kHz. The 3.0 MHz IF would have required an oscillator frequency of 2.9 MHz (100 kHz x 29) for infradyne conversion. That would have been an awkward multiplication at the time. On the other hand, an oscillator frequency of 3.0 MHz was much easier, as it could be derived from 100 kHz by three single-digit multiplications, namely x5, x3, and x2, all relatively easy to do. In turn that meant that the IF was offset by 100 kHz to became 3.1 MHz.
Thus, the required transmission conversion oscillator range became 6.9 to 26.9 MHz, used infradyne for the transmission frequency range 10 to 30 MHz, and of which 7.1 to 13.1 MHz was used supradyne for the transmission frequency range 4 to 10 MHz.
Probably for standardization reasons, essentially the same scheme was used in reverse at the receiving end, again providing sideband inversion for frequencies below 10 MHz. Hence the 3.1 MHz 1st IF. In this case the 3.0 MHz signal required for the 2nd conversion was generated separately to the 100 kHz signal required for demodulation. It [the 3 MHz signal] was subject both to fine tuning and AFC variations. Considered in isolation, there does not seem to have been a compelling reason why the receiver 1st IF was 3.1 and not 3.0 MHz; it was simply the result of standardization with the corresponding transmitter.
After compiling the foregoing but before posting it, I found some additional information in the 1956 October initial edition of the Marconi “Point-to-Point Telecommunications” journal, in an article “Linear Amplifier Transmitter”.
In discussing the ISB drive for that transmitter, it was said:
“The output from the ISB drive is usually at 3.1 Mc/s with sidebands consisting of modulation of the two channels respectively.
“To facilitate sideband filtering, two stages of frequency changing are used in the course of this modulation process, one at 100 kc/s and the other at 3 Mc/s. The second stage results in two pairs of sidebands centred on 3.1 Mc’s and 2.9 Mc/s respectively, hut the latter is filtered out leaving a 3.1 Mc/s signal.
“The choice of 3.1 Mc/s is rather an arbitrary one. Whilst it fits in with the range of transmitted frequencies best suited for propagation (i.e. 4 to 27.5 Mc/s) it has no other special significance. Other frequencies of 6.2 Mc/s and 2.15 Mc/s are used on ISB drives where the radiated frequency is required to be less than 4 Mc/s.”
No information was provided in respect of the generation of the 6.2 and 2.15 MHz IFs, nor were any specific applications mentioned. Whether either was used in SSB (or other) receivers is unknown.
Whereas the GPO used the same 3.1 MHz 1st IF in its ISB receivers of the period as was used for ISB generation, viz 3.1 MHz, Marconi did not. The 1st IF was evidently chosen independently, with numbers such as 1.6 MHz (HR92/93 and HR22) and 2.6 MHz (HR21/23/24). Nonetheless, as recorded upthread, in the solid state era, Marconi did sometimes use the same IF structure for ISB generators and receivers.
As previously mentioned, an unusual use of 3.1 MHz as a 1st IF was in the AM section of the Armstrong 600 series hi-fi tuners and tuner-amplifiers of the early 1970s. These were anyway unusual in having an initial upconversion, with an aperiodic input, that covered the LW and MW broadcast bands in a single sweep. The 2nd IF was 455 kHz. In this case, possibly Armstrong derived the 3.1 MHz from 1st principles. But on the other hand, it might have looked for an existing quasi-standard number that was in the vicinity of where it wanted the 1st IF to be, and which was high enough to provide a small enough local oscillator swing that it could be realized with the standard varicaps of the time. 1.6 MHz was probably too low. And although the 10.7 MHz/455 kHz combination was well established in VHF communications practice (and later used for HF receivers), 10.7 MHz was already used for the FM IF strip. Somewhere around 3 MHz, with a 1.5:1 oscillator swing, evidently “fit the bill”.
(1) The Design of Transmitter Drives and Receivers for Single-Sideband Systems; W.J. Bray, H.G. Lillicrap and W.R.H. Lowry; 1947 March.
Cheers,
Steve
Posted : 05/09/2024 5:05 am
Topic starter
Here is an apparent oddity in respect of some Pye VHF radiotelephone receivers.
The R10 AM and R 10 FM receivers of c.1964 were intended to cover channels in the range 25 to 174 MHz. They were of the double conversion type, with a second IF of 455 kHz. The first IF though depended upon the frequency. It was stated as “10.7 MHz (6 MHz when carrier frequency is below 68 MHz.)”
The 10.7 MHz/455 kHz combination was fairly standard for this type of equipment. But the 6 MHz/455 kHz combination was unusual. It does not escape notice that 6 MHz was the intercarrier frequency for system I TV receivers, in 1964 fairly recently introduced, nor that 68 MHz was the upper edge of TV Band I. But these apparent alignments could have been simply coincidental.
The AM10FRX receiver had a similar IF arrangement.
Pye also had VHF RT receivers that covered the 68 to 174 MHz band only, presumably with 10.7 MHz IFs. Against that, the 25 to 68 MHz range might have been seen as an “add-on”.
In a general sense, it is difficult to impute a reason as to why 6 MHz would be preferred as a first IF over 10.7 MHz for frequencies below 68 MHz. On the other hand, it was probably satisfactory in respect of image rejection; no worse than was 10.7 MHz at 174 MHz, say. So, there might not have been a significant argument against its use other than that it was a departure from normal practice. But that still leaves the questions – why use 6 MHz in the first place, and why go to the complication of having a dual-frequency first IF strip?
One possibility was to minimize the number of local oscillator crystals required to provide the required band coverage. I think it was customary to use harmonics of crystal frequencies in this kind of receiver, obtained from multiplier stage(s) following the (crystal) local oscillator stage itself.
If we assume infradyne conversion, then the first conversion injection frequency range required for the 25 to 68 MHz range, with a first IF of 6 MHz, was 19 to 62 MHz.
And for the 68 to 174 MHz range, with a 10.7 MHz first IF, it was 57.3 to 163.3 MHz.
There is close to a 3-to-1 relationship between those injection frequency ranges. If one starts with 57.3 MHz for a 10.7 MHz IF at 68 MHz, then one third of 57.3 MHz is 19.1 MHz, which with a 25 MHz signal would give an IF of 5.9 MHz. If that was the pathway, then evidently a slight adjustment was preferred to give an IF of 6.0 MHz, perhaps simply because that was an “established” IF number. Anyway, with an injection frequency range of 19 to 62 MHz for the 25 to 68 MHz band, tripling a major part of that range, namely 19.1 to 54.3 MHz, would provide the injection frequency for the 68 to 174 MHz band. The 19 to 62 MHz range could be fundamentals or harmonics, or perhaps a mix of both, such as 19 to 38 MHz as fundamentals and 38 to 62 MHz as second harmonics. That is pure speculation though. More likely I think there were other, less obvious reasons for the use of 6 MHz.
In the early days of double conversion VHF RT receivers with 10.7 MHz and 455 kHz IFs, it was usual to do the major part of the IF bandshaping at 455 kHz, in some cases with crystal filters, with a “roofing” filter at 10.7 MHz, sometimes of the LC type. But once a standard range of 10.7 MHz narrow band crystal filters became available at acceptable economics, then the IF bandshaping, including adjacent channel rejection, was done at 10.7 MHz, with simpler filtering at 455 kHz. Which approach was used in the Pye receivers with two first IFs is unknown. I imagine that 6 MHz narrow band crystal filters would have been special items. On the other hand, as Cathodeon was a Pye company, supply of special filters is unlikely to have been a problem in this case.
Cheers,
Steve
Posted : 05/09/2024 6:48 am
Topic starter
I have found some more information on the Pye AM10D.
Firstly, here is the information on the crystal frequencies required for the various bands:
Doing the arithmetic results in the following:
The range of required crystal frequencies was 24.00 to 54.43 MHz. Had the 10.7 MHz first IF been used throughout, with the same conversion orientation, then the range would have been 21.80 to 54.43 MHz.
I don’t see that there was a convincing argument in that data for the use of two different first IFs. Thus, the reason for the 6 MHz choice for the lower bands must lie elsewhere.
Evidently the receiver could be configured for one of nine bands A through J, as shown above. Thus, the first IF would be either 6 MHz or 10.7 MHz; there was no case where both were required.
In this case, the IF bandpass was determined by a crystal filter in the 455 kHz second IF strip, as shown in the block schematic:
The first IF strip used L-C filtering, so there was no major issue with its being 6 MHz for some versions and 10.7 MHz for others.
The second conversion was primarily supradyne, with infradyne used where conflicts might otherwise arise.
Anyway, we are still left with the question – why was there a “non-standard” 6 MHz first If for the lower VHF bands?
Cheers,
Steve
Posted : 18/09/2024 12:47 am
Topic starter
In contrast to the Pye AM10D is the Bantam HP1 FM, which used the 10.7 MHz first IF for all VHF bands.
Firstly, here is the information on the crystal frequencies required for the various bands:
And the arithmetic:
The range of required crystal frequencies was 25.77 to 54.43 MHz.
Here, the IF bandpass was determined by a crystal filter in the 10.7 MHz first IF strip.
As with the AM10D, the second conversion was essentially supradyne, with infradyne used where conflicts might otherwise arise.
Cheers,
Steve
Posted : 18/09/2024 12:54 am
Topic starter
Mentioned upthread was the Marconi HR120 HF receiver as being of the “converter” type, with block” downconversion of 1 MHz bands to a broad first IF spectrum. See:
Essentially the same scheme was also used for the Marconi HR28 SSB receiver, intended primarily for semi-mobile applications, and I think also used for some shipboard installations. More information for this model is available than for the HR120. The HR28 tuned over a slightly smaller range, 2.1 to 27 MHz, as compared with 2.1 to 30 MHz for the HR120. Here is the HR28 block schematic:
Effectively, each 1 MHz block of spectrum from 2 to 27 MHz was tuned and then downconverted to a corresponding 2 to 1 MHz block as the first IF. The conversion was supradyne for all bands, using a crystal-controlled oscillator.
Fourteen crystals were required to cover the 25 bands, as follows:
The 1 MHz block first IF was in turn tuned (in synchrony with the RF tuning) and then downconverted to a 100 kHz second IF. This conversion was supradyne. The 100 kHz section followed conventional point-to-point SSB practice, with three channels (USB, LSB and carrier) and an AFC system that operated on the second oscillator. The 100 kHz second IF was also available for feeding a telegraph unit.
The fact that the second oscillator range was 1.1 to 2.1 MHz limited the lowest band coverage to 2.1 to 3 rather than 2 to 3 MHz. A 2 MHz input signal would have produced a conflicting 2 MHz first IF. A 2.1 MHz input signal produced a 1.9 MHz first IF, requiring a 2.0 MHz second oscillator frequency, this separation evidently being sufficient.
Notwithstanding the 1 MHz wide first IF band, the receiver was narrow band throughout. Both the RF section and the second IF were tuned in synchronism. The tuning direction was the same for all bands, which was not the case for some “converter” type receivers, such as the Eddystone 880. see: https://www.radios-tv.co.uk/community/radio/radio-receiver-intermediate-frequencies/paged/6/#post-112910.
The HR120, which tuned to 30 MHz, would have required three more first oscillator frequencies, namely 29, 30 and 31 MHz. In turn that probably required two extra crystals, at 14.5 and 15.5 MHz, with 30 MHz possibly obtained by tripling the 10 MHz crystal output. Thus, the crystal total would have been 17, as compared with the 10 required for the Eddystone 880.
Cheers,
Steve
Posted : 20/09/2024 2:35 am
Topic starter
Below is the block schematic of a marine SSB/ISB receiver. The make and model are unknown, but Marconi is a strong possibility. It is taken from the 1966 book “Marine Radio Manual” by Danielson and Mayoh. It used worked examples from several makers, with Marconi usually the leading choice and overall dominant. But this example was not identified.
Although marine RT in general was not swung over from AM to SSB until the 1970s, SSB/ISB had been used for auxiliary/secondary purposes since c.1949, initially for international telephone calls from passenger vessels. As well as dedicated equipment, it would appear that point-to-point and general purpose receivers were used for shipboard SSB/ISB reception in the pre-1970 period. This example, although a dedicated marine receiver, was evidently based upon point-to-point practice.
Frequency coverage was 3.6 to 23 MHz, encompassing the 6 through 22 MHz marine HF bands. Unusually, two separate RF amplifiers were used, one covering the range 3.6 to 10 MHz, and the other 10 to 23 MHz. This was connected with the customary sideband inversion at 10 MHz, and the desire to have the A and B sidebands both on the same side in the IF strip. This could be done by using a mix of infradyne and supradyne conversion, one for the range below and the other for the range above 10 MHz. Supradyne below and infradyne above 10 MHz was the logical choice, giving the lowest range of local oscillator frequencies, with some overlap for the two bands.
A first IF of 3.2 MHz, probably about as close to the lower edge of the tuned range was reasonable to go, would have allowed full overlap of the oscillator ranges, with the required ranges being 6.8 to 13.2 2.85MHz for the lower band and 6.8 to 19.8 MHz for the upper band. However, the 6.4 MHz second harmonic of 3.2 MHz fell in the 6 MHz marine band, so was almost certainly outruled. The 3.1 MHz first IF number established in point-to-point practice had a second harmonic that fell right on the edge of same band, so was similarly contra-indicated. In the downwards direction, 2.85 MHz was the upper edge of the 2 MHz marine band, so something between that and 3.1 MHz was indicated. 2.9 MHz thus presented itself as a suitable choice. The second conversion from 2.9 MHz to 100 Hz could be done using a 3 MHz crystal oscillator, the same as would be used for conversion from 3.1 MHz to 100 Hz. So, the 2.9 MHz choice did involve an element of established practice, being the “image” of 3.1 MHz as it were.
The use of a somewhat lower first IF was probably not desirable anyway, but in this case, it would have needed to be below 1.6 MHz to be clear of any marine bands, and thus not so good in respect of image rejection performance.
With the 2.9 MHz first IF, the required oscillator ranges were 6.5 to 12.9 and 7.1 to 20.1 MHz, so there was still substantial overlap.
The AFC loop worked on the first oscillators, perhaps because the second oscillator was crystal-controlled. Point-to-point receiver practice typically had the AFC loop working on the second oscillator.
That the second mixer was of the balanced type was not surprising, given the proximity of the incoming (2.9 MHz) and oscillator (3.0 MHz) frequencies. The use of a balanced first mixer was unusual, though. Perhaps it was done because the shipboard RF environment was typically seen as being rather difficult.
The general maritime HF RT move to SSB, decided in the later 1960s and implemented in the 1970s, saw major changes in marine main receiver architecture and IF structures. Several examples were covered upthread, such as the Marconi Apollo and Nebula, and the Redifon R551. Whether any marine-specific receivers were developed for the non-primary SSB/ISB traffic I do not know. (ISB might have been phased out anyway, along with AM, although I am not sure about that.) The then-new generation of compact point-to-point (e.g. Marconi Hydrus) and single-box general purpose receivers might have filled the bill, as it were, so that there is not a separate discussion to be had about IF choices.
Cheers,
Steve
Posted : 22/09/2024 2:49 am
Topic starter
Frequency coverage was 3.6 to 23 MHz, encompassing the 6 through 22 MHz marine HF bands.
That should have been 4 through 22 MHz marine HF bands.
Posted : 22/09/2024 3:27 am
Topic starter
The marine SSB receiver discussed in the preceding post is difficult to place as a Marconi model, based upon available information although given the diversity of Marconi’s output it cannot be outruled.
In 1964, Marconi Marine (Mimco) introduced a marine SSB receiver known as the Pennant. This was intended to work with and was dependent upon the Crusader marine SSB transmitter introduced at the same time. The Pennant was automatically tuned to the designated receive channel that was paired with the chosen transmit channel. It was said to obtain its first and third oscillator frequencies from the Crusader. The latter had synthesized frequency generation, possibly, or perhaps probably very similar to that used for the MWT MST point-to-point transmitters. I have yet to find any detailed information about the Pennant, although a 100 kHz final IF seems likely.
The SSB receiver at interest had its own oscillators, so it was clearly not the Pennant. Before the Pennant, Mimco had used the HR22 receiver (from the Marconi (MWT) point-to-point range) in marine SSB installations, although it was also known to have installed the Racal RA17 (or RA117) , presumably with outboard SSB/ISB adaptor, for this purpose. To recap, the HR22 tuned 2 to 32 MHz and was double conversion, with IFs of 1.6 MHz and 100 kHz. It could be said to have been somewhat of a CR150/6 (or perhaps CR150/5) derivative with features from the larger HR21 in the point-to-point series. The Racal RA117 with SSB adaptor would have formed a five-conversion receiver with IFs of 40± MHz, 2.5± MHz, 1.6 MHz, 100 kHz and 18 kHz.
Danielson and Mayoh made the following general comment about marine SSB receivers:
“In some equipments the aerial feeds into a frequency changer—with some adjustable pre-selection—and this is followed by a superheterodyne receiver of fairly conventional design. This method is good because: (1) It gives good frequency stability, since after the first frequency changer only a single range of frequency, of about 1 Mc/s or less, need be covered. (2) Frequency control can be by crystal. (3) Very good re-setting accuracy is obtained (to better than 1 kc/s). (4) Selectivity is good over all ranges. (5) The equipment is relatively easy to operate even by personnel who are not highly skilled.”
That description could have applied to the previously mentioned Marconi HR28 receiver, which is understood to have had some marine applications. The detailed description of the “unknown” receiver was preceded by the comment “..the arrangements are conventional”.
A further search had shown that the “structure: of that receiver, including the use of two RF sections, the 2.9 MHz, 100 kHz IF combination, and AFC acting on the first oscillator, parallels that used by Western Electric for an earlier point-to-point SSB/ISB receiver.
Given that STC had connections with Western Electric, possibly the receiver described by Danielson and Mayoh was an STC model. In that case, the 2.9 MHz first IF could have been a “house” norm, and one that happened to be a good fit to the specific marine requirements. STC had claimed the first marine SSB installation, on the passenger vessel ‘Caronia’, in 1949.
I have not been able to find any details of the STC RX9 receiver.
As an aside, the shipboard environment could include a wide diversity of receiver IFs in close proximity, with the main and reserve RT and WT equipment, SSB/ISB equipment, broadcast radio receivers and multistandard TV receiver head-ends.
Cheers,
Steve
Posted : 05/10/2024 4:10 am
Page 8 / 8
Prev
