Historic AM audio quality

[Ted Kendall]

Fair enough - but off-tuning wasn't always done with the requisite discretion...

[martin.m]

I have an early 1950s HMV radio (only the chassis because it came from a radiogram which was scrapped). The tone control is a rotary switch and one of the positions increases the I.F. bandwidth for a better treble response. The effect is noticeable even on today's AM broadcasts. I think it works by connecting a resistor across one of the I.F. transformer primary windings. I don't know the model number but it featured a louvred tuning scale.

[stevehertz]

Originally Posted by Ted Kendall

Fair enough - but off-tuning wasn't always done with the requisite discretion...

Oh for sure Ted, I know exactly what you mean. Some people are able to stand large amounts of distortion so long as the 'tone is right'. I'm not one of those!

[kalee20]

Originally Posted by stevehertz

There's no laws about tuning for best enjoyment factor... So, for a few reasons, you don't have to 'centre tune'.

True. Tuning 'bang on' is the most elegant, but isn't always the 'best.'

Some radios even have stagger-tuned IFT's to give a more broadband response. Obviously, you can't tune one of these exactly to the incoming frequency, because at least one of the IFT's is then going to be working on its upper slope and one on its lower slope. But with the wider passband response, overall sound quality is definitely improved!

Many moons ago when but a teenager I had a radio that mechanically moved the coupling twixt IF circuits for bandwidth control.

[stevehertz]

I've got a couple of tuners that automatically adjust bandwidth (and other parameters) dependent on the signal strength on FM. These are the Onkyo 9090 and the Pioneer F-91.

[colourking]

Originally Posted by kalee20

True. Tuning 'bang on' is the most elegant, but isn't always the 'best.'

detuning slightly with AM was often of benefit to reduce interference which was only on one side of the signal, be it an adjacent station, or switch-mode power supply etc harmonics which may have fallen on one edge of the bandwidth.
So a nuisance that digital tuned receivers usually have been 9kHz (or 10kHz across the Atlantic ) steps only on MW with no option to fine-tune in 1kHz steps ( a facility that DXers would also demand).

[Hartley118]

We've established that MW AM transmission bandwidth in the 1950s could extend to 10kHz, possibly as far as 15kHz from Brookmans Park, depending on the PO line quality.

The question remains as to when the current ruinous 5kHz 'brick wall' filters were installed.

I dug out the attached frequency response curve is of the IBA specified low-pass filter, published in the IBA Technical Review 5 in 1974 on Technical Standards for Independent Local Radio. That's an alarmingly steep HF cutoff. Unless there's some phase correction, the group delay around cutoff frequency must be about half an hour! .

Does anyone know when these steep filters actually began to be specified?

Martin

[Radio Wrangler]

Long time since I saw that spec, but at the time I wondered about the ringing due to the rather square corner. It must have been around the time of the re-jigging of the channel spacing.

If it's so much more important that nothing spills into adjacent channels that it doesn't matter that it becomes unlistenable to intended listeners, then the transmitter might just as well be turned off.

THere's such a reduction in demand for frequency space, that we could just revert?

David

['LIVEWIRE?']

If, by re-jigging the channel spacing, you mean the general changes to MW & LW, so that every channel is exactly 9kHz from the next & previous one, this happened in 1979, IIRC.

[Scott37]

I think it was 1978.

275 and 285
275 and 285
We're on a new wave band
And still the best in the land

[Synchrodyne]

Originally Posted by Hartley118

Does anyone know when these steep filters actually began to be specified?

See this thread - https://vintage-radio.net/forum/showthread.php?t=120815 – posts #21 & 22.

The BBC FL4/55 5 kHz low-pass filter was used from c.1973. The background work was reported in 1972, and the filter specification was dated 1973.

Cheers,

[arjoll]

If anyone's interested, when I was archiving a whole lot of old reel to reel tapes a few years ago, I found some off-air audio.

This is the only recording I know much about - it was Foveaux Radio (1224), 12/8/89 based on the news bulletin, and it would have been me recording it. Stereo four track, so recorded on dad's Akai 4000DB, and I'd have used the line out of a National T100-D radio. There's over two hours - double play tape at 3.75ips - but only a few minutes on that link. The quality is everything I remember from Foveaux Radio - strong signal, more bandwidth than 4ZA but always an annoying hum.

When I was copying tapes I also did a whole pile of my late father-in-law's - he was in the merchant navy in the 1960s, and I found 4ZA Invercargill, various BBC programmes (at time time I tagged them as Family Favourites, breakfast housewives, Navy Lark, Requests, BBC Light 1967) and Caroline - the UK one, not 3ZC Timaru .

Unfortunately those are very low bitrate MP3s I uploaded back before we had ADSL2, and it'd take a while for me to fish through all the original WAVs to find them again, but if there's interest I could probably do that sometime! The original tapes weren't great - various random brands, some at 1.875ips (played at 3.75ips, digitised at 88.2 kHz then changed down to 44.1 to correct the speed), and it sounds like he just used a microphone to record them.

[SteveCG]

Did Quad sell an LW/MW/SW tuner in the styling of the 33 control unit and 303 power amp (ie late 60s /early 70s) which had a selectable IF bandwidth? AM III comes to mind as a name.

[Synchrodyne]

Yes, it was the AM3, released I think in 1969 and discontinued sometime during 1973 after around 2000 were built.

The AM3 was basically a self-powered version of the AMII in a case that was very similar to that which was used for the FMII. (The latter was introduced late 1967 or early 1968 to match the Quad 33 control unit, and replaced by the FM3 in 1971.)

There were some changes as compared to the AMII. The latter had AF bandwidths of 5 and 12 kHz, whereas the AM3 had 3.5 and 12 kHz. The “native” narrow AF bandwidth of the AMII circuitry was 3.5 kHz, but there was a post-detector LC tilt circuit that extended this to 5 kHz. This used the same “L” as was used for the 9/10 kHz adjacent channel notch circuit. In the AM3, this narrow bandwidth extension facility was deleted, so the narrow AF bandwidth stayed at 3.5 kHz. Possibly this change reflected the more severe reception conditions that existed in 1969 as compared with 1960 when the AMII was released.

Cheers,

[SteveCG]

Thanks for the AM 3 info Synchrodyne.

[Synchrodyne]

No problem!

To add a bit more, the “Acoustical AM” of either late 1953 or early 1954, designed by Geoffrey Horn, and which matched the QCII control unit (itself released in 1953 September) had “native” bandwidths of 5 and 12 kHz. The Quad AMII of 1960 was designed by John Collinson, in part responding to customer feedback on the AM. It was something of a “textbook” job, one small criticism being that it did not completely eliminate differential distortion.

I heard two reasons for the demise of the AM3. One, from the then NZ Quad agent in Wellington, in 1974, was that the production rate of under 500 per year was too low for Quad to sustain. That was credible; Quad’s overall production rate at the time was probably somewhat above that of typical of cottage industry. Also, at the time that the AM3 was discontinued, it may still have had a big backlog of orders for the FM3, so needed the production capacity for these. Quad had been quite late in introducing a solid-state FM tuner, so that when the FM3 appeared, there was a lot of pent-up demand. I vaguely recall that for quite a while there was a six-month waiting period.

The second reason was given by PJW himself at one of the late 1970s London hi-fi shows – I forget which one. I asked the question, and the answer was that it was too problematical – users in situations where they could not get good AM reception for various reasons blamed the tuner, not the situation. So I suspect that it was a combination of factors. In hindsight one could see that something like the AM3 or a solid-state successor thereto might better have been handled by a smaller, low-production entity of the true cottage variety, someone like Motion Electronics, who specialized in TV sound tuners for quite a few years from 1971.


Cheers,

[kalee20]

Originally Posted by Synchrodyne

The Quad AMII of 1960 was designed by John Collinson, in part responding to customer feedback on the AM. It was something of a “textbook” job, one small criticism being that it did not completely eliminate differential distortion.

What's differential distortion, Steve?

[Nickthedentist]

Originally Posted by Synchrodyne

...designed by Geoffrey Horn...

Well I never! A chap who lived about 1/2 a mile from me in this part of Oxford, and who had a well-known radio/hifi shop which survived (albeit as a B&O centre) until last year.

He died in 2009: https://reader.exacteditions.com/issues/34226/page/9

[Synchrodyne]

Originally Posted by kalee20

What's differential distortion, Steve?

Langford-Smith described differential distortion a lot better than I can. From p.1074 of the Classic Edition:

“One type of distortion often encountered is caused by delayed a.v.c systems at the point where the a.v.c. diode just begins to conduct. This form of distortion is called differential distortion and may be kept to low values by making the delay voltage small.”

As I understand it, when the agc diode starts to conduct, it affects only one side of the modulation envelope and the resultant distortion is reflected back into the IF signal path. With a small delay, this happens for relatively low signal strengths that are probably below those normally used for programme listening.

The Quad AMII/AM3 used quite a small delay, somewhere around 2 or 3 volts I think. Another approach was to have no delay bias on the agc diode, but use an extra diode following it to effect the delay. (I think that this approach was found on some 405-line TV AM sound channels, where say a crystal diode was used for detection, and half of a double diode for agc delay, with the other half used as the noise limiter.)

A perhaps more elegant approach was to isolate the agc system by using an IF sidechain to feed to agc diode. This was a signature feature of the Chapman valved AM (and AM/FM) tuners, using an EBF80. Chapman claimed that its AM tuners thus had no perceptible differential distortion. Another advantage of the sidechain approach was that full agc bias could be applied to the final IF stage without running into modulation rise distortion. In the Quad AMII/AM3 case, modulation rise distortion was minimized by applying fractional agc voltage (about a quarter) to the IF stage. These had a single high-gain IF stage, neutralized to minimize any passband tilting, which would be undesirable on a wideband tuner. The EBF89 was also neutralized against interaction between the IF pentode and the detector diode.

Langford-Smith provided further discussion of agc differential distortion on p.1107 of the Classic Edition, with modulation rise distortion, and the fractional agc means of minimizing it, covered on p.1114. Diode-pentode neutralization was covered in pp.1066-67.

Probably the fractional agc approach was available for the Quad AMII/AM3 because it had an RF stage, allowing for two fully controlled and one partially controlled stages, presumably enough for an adequate agc curve. For receivers without an RF stage and one IF stage, fractional agc on the IF stage may not have been feasible. The earlier Acoustical AM, without an RF stage, had full agc on its IF stage, but the latter had negative feedback (by means of an unbypassed cathode resistor) which I think would have helped. The basic problem is that at high signal strengths, the IF stage is biased so far back that it is operating on the curved portion of its slope, thus producing in-band third order effects that manifest themselves as modulation rise. In a way, an IF stage is fighting itself under these conditions, trying to develop a high enough agc voltage to tame the high signal strength but also trying to reduce its own gain.


Cheers,