Commercial versions of the single-span concept?

[saddlestone-man]

Hello All.

I've been looking back at scans of the Single Span Receiver articles in 1934 issues of Wireless World. It's an interesting concept - tuning from 2,000m to 200m in one band, and was thought to be revolutionary at the time.

I know there were kits available of the W T Cocking designs described in the WW articles, but does anyone know if any radio manufacturers produced any commercial radios based on the concept at the time?

best regards ... Stef

[dave walsh]

Nothing commercial but I think a modern version of the same idea appeared [possibly in Practical Wireless] in perhaps the sixties or seventies? I only recall this because it seemed so unusual at the time that I thought it might even be a joke article article [along the lines of the Wireless World spoof on HI-Fi that appeared-with a giant system blowing out the windows and the deck having to be kept in the garden shed!]
Dave

[Radio Wrangler]

Weren't some of the 600 series Armstrong receivers of the seventies up-converters to an IF above medium wave so they tuned straight across longwave and medium wave in one band?

David

[Herald1360]

I didn't see a satisfactory explanation in the articles, but the small value oscillator variable capacitor and a fixed LPF on the input strongly suggest up conversion to a relatively high IF. I think the cap was 160pF max, andover about 10:1 cap swing needs to tune oscillator over a 1.5MHz range at least 1.6MHz above the input 150-1500 KHz band span.

The LPF on the input solves the long wave sideband cutting from too high Q in the front end tuning and eliminates second channel problems but leaves the problem of in band intermods- perhaps less likely in the relatively low power early '30s.

[ukcol]

Originally Posted by Herald1360

I didn't see a satisfactory explanation in the articles, but the small value oscillator variable capacitor and a fixed LPF on the input strongly suggest up conversion to a relatively high IF.

Correct, the IF is in fact 1.6Mhz. The single span piece is spread over several issues of WW which is probably why you missed it.

[Synchrodyne]

Originally Posted by Radio Wrangler

Weren't some of the 600 series Armstrong receivers of the seventies up-converters to an IF above medium wave so they tuned straight across longwave and medium wave in one band?

Indeed so - see:

http://www.audiomisc.co.uk/Armstrong/600/600page4.html.

Upconversion to 3.1 MHz, then down to 455 kHz.

Cheers,

[turretslug]

That is indeed an interesting approach, and not particularly extravagant of components or involved, either. It's not hard to understand why it wasn't more widely followed, though, really- most manufacturers (and consumers) would have been content with the thoroughly established self-oscillating mixer plus 2x IF amp format that was de rigeur, and bandswitching frequently only involved selecting appropriate ferrite rod windings and connecting additional capacitance across the MW oscillator coil for MW, as the capacitance swing was still sufficient to cover (150 + IF) to (260 + IF)-ish kHz. Nonetheless, it's a shame that innovation doesn't win out over the bottom line more frequently in any technology, tuning all the way across LW and MW is appealingly elegant, and may have introduced a few bored twiddlers to NDBs! I wonder if any sold in Finland, where Oulu 433Khz apparently broadcast as late as 1977.

I imagine that the designers didn't dare connect the external aerial terminal to the input of the low power, untuned bipolar pre-amp.

How much precedent was there for the LPF followed by up-conversion type scheme in relatively straightforward receivers? (i.e. not the likes of the RA17 etc. approach!). Marconi's circa WW2 CR300 used a RF LPF on its 15-85kHz band with conventional LO converting to 98kHz IF- this overcame the quandary for LF receiver designers of ever-narrower tuning spans as coverage went LF with conventional tuned RF stages.

[saddlestone-man]

Thanks for the very useful replies. Use of the single-span concept in the Armstrong 600 series is most interesting.

I'd like to locate the PW article in the 1970s or 80s if anyone can find it in their collections.

As far as I can tell, the original S-S articles started in the 23 March 1934 issue of Wireless World. The magazine was published weekly at the time, and further details, including full constructional details of an AC mains, a battery and an AC/DC version appeared in many issues up to the end of 1934.

Scans of all WWs from 1911-1986 can be found at:
http://www.americanradiohistory.com/...d_Magazine.htm

Scans of PWs from 1932-1989 can also be found on this website.

best regards ... Stef

[Synchrodyne]

Originally Posted by turretslug

I imagine that the designers didn't dare connect the external aerial terminal to the input of the low power, untuned bipolar pre-amp.

Yes, I think that that would have been a bridge too far. Something like a 2N3866 or 2N5109 in a feedback circuit would have been preferable. Towards the end of the 1970s, though, Delco (GM) had introduced its jfet/bipolar (small-signal) cascode circuit for use as an untuned RF amplifier for AM (MW) car radio receivers, and this seems to have been quite widely adopted, sometimes with tuning between the RF amplifier and the mixer, the latter usually within a multifunctional IC.

Pertinent to this thread, Motorola showed that kind of RF amplifier, fully aperiodic, in the data sheet for its MC13030 dual-conversion AM receiver IC, both for an MF-only case and a 5 to 10 MHz HF case. Both had a 1st IF of 10.7 MHz and a 2nd IF of 455 kHz, but it is evident from the text that these numbers were not fixed. There would seem to be no reason why the MC13030 could not have been used as the basis for a single-span LF-MF receiver with aperiodic RF amplifier with a 1st IF below 10.7 MHz, and sufficiently far above the top end of the MF band to allow adequate IF rejection from the RF LPF.


Cheers,

[turretslug]

A 9MHz crystal filter could have been an option for a single-span, single-conversion LF/MF receiver- I believe that the original motivation for these was amateur radio use where a single unswitched 5-5.5MHz LO could be designed for good stability and give coverage of both 3.5-4.0 MHz and 14.0-14.5 MHz. Probably most of these were of narrow bandwidth for SSB use, though- possibly there were wider versions for NBFM once the use of 9 MHz had spread to 28 MHz recievers. Narrow roofing filters were also made at 10.695 MHz for CB use, corresponding with a 10.240 MHz crystal for combined synthesizer and 2nd LO (to 455 kHz).

I wonder how much of the single-span and high up-conversion techniques found their way into professional LF/MF receivers such as those used for maritime use and aviation ADF? The latter often had a simple-as-possible control panel with thumbwheel/knob-and-Veeder-type selection and, with the maturing of the semiconductor era, a sophisticated-yet-compact receiver in the avionics bay.

[Synchrodyne]

The BBC RC3/10 LF/MF monitoring receiver of the 1970s might fit the professional category. It had a crystal-controlled local oscillator and a pre-tuned front end that could be set for any LF or MF channel. There was upconversion to 10.7 MHz followed by a standard 12 kHz bandwidth crystal filter. The IF was at 10.7 MHz, with no down-conversion, and included PLL fully synchronous demodulation.



I imagine that the basic design could have been segued into a single-span LF/MF tuneable receiver, perhaps without an RF amplifier, but with a balanced mixer and a variable frequency oscillator. Different standard filters could have been used as required, e.g. 30 kHz bandwidth for wideband AM, 7.5 kHz for communications quality, etc. IC3 (an MC1351) could easily have been configured as a quasi-synchronous demodulator, with the PLL facility either deleted (and with it IC2, another MC1351) or retaining it with switching from quasi-synchronous to PLL after tuning and capture.

Another professional example was the monitoring receiver developed for the Belar AM Stereo broadcasting system for use during the American NAMSRC (National AM Stereo Radio Committee) trials of the later 1970s.



This receiver had no RF preselection, but instead a 2 MHz low-pass filter at the mixer input, with upconversion to a 10.7 MHz 1st IF. The local oscillator input to the mixer was from a synthesizer. There was a second conversion down to a 260 kHz 2nd IF (a quite common AM car radio IF in the USA). Although, for US use, this receiver covered only the MF band, extension downwards to the LF band without bandswitching should have been possible. That might have required particular attention to screening between the RF input and the 2nd IF, or perhaps moving the 2nd IF up to 455 kHz. Whilst this AM stereo system was approved by the FCC (along with four others), Belar dropped out of the “competition” quite early on, and so its receiver concept was not developed commercially. Although the Motorola MC13030 IC could be seen as being a later development of it.

I am not sure what bandwidth range the standard 9 MHz filters covered. But the standard 10.7 MHz range covered “soup-to-nuts” so was probably low-hanging fruit for designers who wanted to use a 1st IF in this vicinity for upconversion LF/MF receivers. Or perhaps they chose 10.7 MHz because of economic filter availability and because it was not otherwise inimical to the design objectives. On the other hand, Armstrong evidently adopted the approach of the lowest 1st IF reasonably possible. 3.1 MHz would have allowed adequate IF rejection from the input LPF without being too close to a harmonic of 455 kHz and with none of its own harmonics closely related to other important frequencies such as 3.58, 4.43, 9 and 10.7 MHz. (Having a harmonic – that might possibly escape - close to one of these was probably worse than having a harmonic right on one of them.)


Cheers,

[G6Tanuki]

Choice of 'high' IF for these sorts of receivers [and for HF/VHF receivers in general] can be fun: 9MHz was popular because it allowed you to build a receiver using a single 5 to 5.5MHz VFO that then tuned both the 3.5-4.0MHz and 14.0-14.5MHz amateur bands merely by switching front-end filters. Downside was that one band tuned 'backwards' so to speak.

Why 10.7MHz came into use I don't know. 21.4MHz (twice 10.7) was also popular at one time in commercial two-way radios but had the disadvantage that it was also a shortwave broadcast-band and several 'base station' receivers using this IF were troubled by breakthrough into the IF - a couple of hundred feet of coax running up a tower to your UHF antenna acts as a good longwire on the HF bands!

Part of me wonders if it would have been possible to design a MW/LW single-span receiver by using the fundamental of the LO on longwave and the 2nd harmonic on medium-wave. Time to play with some spreadsheets!

[Synchrodyne]

I don’t think that you could do that with a “high” 1st IF, that is above 1.6 MHz. But a 300 kHz IF, just above the top of the LF band, might work. For a tuned frequency of 500 kHz, an oscillator frequency of 800 kHz is required. If that is the second harmonic, then the fundamental, 400 kHz, would tune a signal of 100 kHz.

A signal of 1700 kHz would require an oscillator 2nd harmonic of 2000 kHz, so a fundamental of 1000 kHz. So, the oscillator range is 400 to 1000 kHz. And at 1000 kHz, the fundamental would tune 700 kHz. Hence the two tuning ranges would overlap, 100 to 700 and 500 to 1700 kHz. The “flip” between fundamental and 2nd harmonic could be arranged to happen just below the bottom end of the MF band, and the continuous tuned band would span 100 to 1750 kHz, although reception around 300 kHz would probably be a bit murky.

Cheers,

[Radio Wrangler]

As you move the IF, good selectivity becomes progressively harder to obtai. Higher Q inductors are needed, alignment gets fiddlier and it all drifts more. You get forced into crystal filters at some point and you can now make broadcast-width filters up to several tens of MHz, but the price takes a jump and the filters tend to be all-in-one-block which isn't quite as good as being distributed down the IF amp in terms of dynamic range.

Also, the LO moves up and drift becomes proportionately more significant. There is also a proportionate increase in phase noise, though that's probably not much of a worry in a broadcast receiver.

The big three japanese amateur radio manufacturers took to up-converting 0-30MHz like ducks to water. They offered gapless general coverage receivers in their transceivers. At their price point and in their era, crystal filters and synthesisers were affordable.

Since then there have been a few transceivers which have gone against this pattern, using lower IFs to bring down synth frequencies (and phase noise because it does count in these applications) Bigger chunks of quartz in the filters improves big signal handling, and they're only roofing/antialias filters now ahead of DSP. These sets like the Elecraft K3 are seen as about the highest performance out there because of their lower IF.

For a long/medium wave broadcast set with LC filtering, you can do it other ways but the trad 455kHz IF architecture is close to optimum bang for the buck.

David

[Synchrodyne]

I did look at the 450 kHz IF single-span case. A 500 kHz signal would require an oscillator 2nd harmonic of 950 kHz, so a fundamental of 475 kHz. 1700 kHz would require 2150 kHz, fundamental 1075 kHz. The “LF” tuning range would then be 25 to 625 kHz.

My starting point was to assume that 150 kHz and 500 kHz were respectively the lowest required LF and MF band tuned frequencies using, again respectively the oscillator fundamental and 2nd harmonic. Thus, if f is the IF, and n is the oscillator fundamental required to tune 150 kHz, we need f such that:

f = n – 150
f = 2n - 500

So that:

2n – 500 = n – 150

And n = 350 kHz.

This gives an IF of (350 – 150) = 200 kHz, right in the middle of the LF band, where I think it would be unacceptable. Hence my choice of 300 kHz, just above the top of the band, for the first worked example, on the basis that this was about the lowest workable, and that as one went higher, the LF tuning spread went well beyond the boundaries of the band.

Further delving into receiver IF choice in a general sense, as interesting as it may be, would I think take us off-topic for this thread. It did come up a while back on the old VRAT forum, now an archive (see: http://www.forum.radios-tv.co.uk/vie...php?f=3&t=3833). That was largely a recording of the “what” with just a little of the “why” plus some speculation and deduction.

But it’s fair to say, in respect of the Armstrong 600-series, which is a single-span receiver, that its 1st IF of 3.1 MHz was not completely a standalone choice, as it had been seen in some earlier HF receivers, such as the Mullard GRF552. Whether that affected Armstrong’s choice is unknown, but one may see that if something in its vicinity was thought to be a satisfactory choice, there could be some leaning towards a number already in use. Its 2nd IF was the “standard” 455 kHz.


Cheers,