Mildly technical question

[hirez]

What's the lowest radio frequency one could usefully transmit phone quality speech/music upon? I am ashamed to admit I have entirely forgotten what little wireless theory I knew.

Comments

[http://users.livejournal.com/_whitenoise/]

Phone-quality is about 4kHz. AM usually has a carrier more than 10x the maximum data frequency, so probably about 40kHz absoltue minimum.

IIRC: Long-wave went down to about 150kHz.

[hirez.livejournal.com]

Excellent. That clatter you heard was the noise of pieces falling conveniently together in a Tetris-like manner.

[aoakley.livejournal.com]

Note his point about longwave, then glance at the whopping antennas as you drive past Droitwich on the M5.

Anything lower than mediumwave and you'll need an antenna the size of... well, bigger than any land area you or any of your friends can afford (regardless of whether you mount horizontal or vertical - it still has to have guy ropes, and probably air traffic clearance at that height).

Mediumwave antennas will comfortably fit inside a field or on a ship. Hence why there are pirate mediumwave stations but not longwave stations. Sunshine Radio out of Ludlow operated an AM-MW TX from a field (and briefly from a raft in a lake, if anecdotes are to be believed), before they went legal and rented a proper TX off of Merlin like all the other ILR sell-outs.

Radio 4's wavelength is about one and a half kilometres, for example. It only gets longer as the frequency gets lower. The rough equation is "divide 300,000 by KHz to get metres", although I'm sure a physics student will be along in a moment to talk about the speed of light etc.

[sinibar.livejournal.com]

Well, I don't have any qualifications in physics but I do have a HNC in electronics, specialising in communications.

I can't actually remember anything as I've been doing software since I graduated, but I one thing I specifcally don't remember is any fudge factor for variance in the speed of light in wavelength calculations; the speed of propogation though air was assumed to be constant (at ground level).

(insert hazy memory may be talking bollocks disclaimer here)

The only problem I remember re variance in the speed of light was the fact that different frequencies propogate at different speeds^[1]. Fourier shows us that a complex waveform is in fact made up of many sine waves of different frequencies. If all these frequencies are travelling at different speeds when it gets to the other end your waveform is going to be right mullered.

One of my first software jobs was doing some predictive modelling on this to see if it was possible to deduce from a badly distored wave what it should have been based on the known distortion factors. Sadly I don't get to do anything quite so interesting any more.

[1] My work was actually done with optical fibre, but I believe it's a problem with microwave transmission as well.

[quercus.livejournal.com]

Mediumwave antennas will comfortably fit inside a field or on a ship

That's pushing it for "comfortably". Looking at pictures of the MV Mi Amigo with that vast top-heavy tower mast on top of it, I was always happy to be sitting in a tower block. It was easier for the Buzbies to show up outside, but at least Cantril Farm never threatened to capsize.

[aoakley.livejournal.com]

Ah, yes, but the Mi Amigo (and MV Communicator etc) had the antenna mounted vertically, instead of slung horizontally between two masts. Capital Radio (Holland) and Radio 270 had a horizonally slung antenna and fared much better in storms (although Capital Holland slipped her anchor for an entirely different reason - debts and lack of maintennance).

[jarkman.livejournal.com]

I'm highly intruiged - why do you need lowness ?

[hirez.livejournal.com]

The whole VLF and radio-at-audio-frequencies thing is strangely fascinating.

Actually, many odd corners of the wireless spectrum are oddly interesting. Has anyone tried using the output of numbers stations for the lottery?

[hazeii.livejournal.com]

It's a cycles thing.

Folk song, bicycles and number stations, expressed as hand-crafted steel (http://www.on-one.co.uk/index.php?module=announce&ANN_user_op=view&ANN_id=336).

I've got the previous incarnation, Il Pompino (italian for hand-job) and it's resonant at 11Mhz.

[deathboy.livejournal.com]

There's a very reasonable chance that I have this wrong, but I was vaguely aware that phone quality audio was typically considered 8khz sampling rate - which would tie to the 4khz you mention, because Nyquist theory says you need to have twice the transmission frequency of the highest audible frequency you want to represent, so that you can get a peak and a trough in.

I think.

No idea if that applies to transmission via radio waves, but if so, it would mean that there's a x2 on top of that 40khz.

[hirez.livejournal.com]

I suspect you're right. Still well within the cricket ground (I am English and we'll have no 'ballpark' malarkey here) and aiming to wallop the pill into the beer-tent.

[eljaydaly.livejournal.com]

You hate us! (Although I have no room to talk. I haven't seen the inside of a ballpark in lo these many years.)

[hirez.livejournal.com]

:p

[avocadovpx.livejournal.com]

JHR: "I am English"

Whoa. When did this happen?

I'm always the last to know.

[quercus.livejournal.com]

I am English
No you're not, you pasty-eating piskie. You can take your leather-on-willow ways back to Kernow with you.

[hazeii.livejournal.com]

Oh dear, one of my speshul interests...

What range are you thinking of? For short range (a few feet) you don't even need to modulate - to make a "transmitter", just replace your speakers with a high wattage resistor and use the speaker wires as the TX aerial. The "receiver" can then just be an audio amplifier with a loop of wire attached to the signal input. Try looking up "low frequency receiving aerials" and "inductive loops".

For more range you will need to modulate - i.e. provide a carrier for your signal. The absolute theoretical minimum (the Nyquist limit) carrier frequency is twice the highest signal frequency of interest - the entire phone system samples at 8Khz (for a maximum signal frequency of 4Khz, though actually it really only passes up to 3Khz or so to keep the cost of filters down).

The trouble with any low radio frequency though is it's very hard to make effective aerials (esp. transmitting) because the wavelength is so long - for example, the Rugby time signal is sent at 60Khz which makes the wavelength 5,000 metres - and they use some absolutely bonkers big aerials to transmit it (if you've ever been past Rugby, you can't have missed them).

The other issue with low frequencies is the amount of interference you'll pick up - switching power supplies, fluorescents, motors, dimmers, cars, military submarine comms, time signals, Loran (that one's a killer) and so on.

For all those reasons the minimum I'd suggest is 150Khz - at that frequency you can use a ferrite rod to receive, a loop to transmit and it's pretty quiet (LW runs from 153Khz to 279Khz, stations spaced 9Khz apart hence R4 at 198Khz, there are strong stations on 162Khz and 171Khz though).

If you want to experiment, you can use a pair of soundcards with at least 44Khz sampling, and do it all in software. To create the "transmit" signal, create a sine wave at the desired carrier frequency and multiply each sample with the corresponding signal value at the transmitter - at the other end, either just measure the amplitude or multiply again by the same carrier frequency to recover the signal). Your max. carrier frequency is half the sample rate (Nyquist again) so you're limited to 22Khz with 44Khz sampling. Plenty of sound cards now do 96Khz though (so 48Khz max carrier) and if you lay down extra pennies you can get ones that sample at 192Khz (with those, you can indeed pick up the Rugby MSF time signal with a piece of wire attached to the input).

Finally, you could consider optical links - you can make a simple FM transmitter with infra-red LEDs and a 555 timer, and a simple receiver using something like a 565. That works great at 500Khz for in-room use, and no interference worries.

[quercus.livejournal.com]

You can modulate phone audio onto DC. As a convenient proof of this I offer ADSL and its ability to transmit 8M of data over phone wires. The only _possible_ explanation for this is the use of Demonology, not Nyquist. If ADSL works, it must therefore be possible to get phones over damp string and a bit of twanging. Q-E-fnor-D.

Phone audio is defined as 300Hz to 3.3kHz and used to be limited to such, back in the days of analogue multiplexing onto trunks. This is why from the mid '70s or so a local line was often audibly better quality than a good trunk call. In the mid '80s, the increased use of digital multiplexing meant that actual quality of calls was far beyond this (about 8kHz tops, AFAIR) and the punters began to complain if they were getting just what they'd paid for and no more (a neat example of Machiavellian principles in marketing, where giving the customer a better deal simply turns into a shifted expectation and increased complaints).

For AM radio, about a factor of 10 is a reasonable bet for minimum carrier to bandwidth ratio. The problem here isn't modulation capacity, it's the width of the sidebands. The the higher modulation frequency, the wider your transmitted bandwidth used and the more chance of overlap with stations on adjacent channels. Usual practice is to low-pass filter the audio so as to keep within the _allocated_ bandwidth (plus guard bands). This is a regulatory limit, not a technical one. It also varies between bands - MW AM broadcast has higher bandwidth than SW, contrary to the technical limitations, just because the market demands it and so the channel spacings have to be stretched further apart.

Obviously SSB halves this spacing (not quite, as the guard bands remain the same size) and tends to give better results anyway, so is favoured for anything that has an "operator" at the Rx to keep re-tuning the BFO.

LW is limited by the performance of ferrite rod antennae in a domestic radiogram (and 1930's ferrites at that). Not much point in a commercial broadcast that you can't receive without building permanent gubbenry in the back garden. Below it _could_ still broadcast speech, but the people who want to use it want better data rates, so it's pretty much all RTTY (teletypes) rather than voice. Some of the kit does do both though, it's just not often switched to VOX.

Phone-quality audio _can_ be broadcast on about 15kHz radio. You need a farking big loop antenna (lots of multicore cable wrapped in a loop about the diameter of a fat caver, which is who's usually using it). The propagation of this underground also depends on liberal libations of beer to the trolls and underground demons of B fields. Transmitter overlap is limited by low power and low-mounted Tx, so the fact that it almost is "one station per band" operation isn't a limit. It's impractical to build a long-wire antenna for this band, so no-one except the spooks gets to have distant propagation of it.