Thanks for the replies so far.
mnttech: I'm using amateur radio equipment that I've calibrated and know to be accurate to within 100Hz or so. It is very evident that the local Scottish information ground transmitter is way off the channel nominal centre frequency, as it would have to be to avoid audible heterodynes with other transmitters in the same service. The approach this side of the pond is that all transmitters are keyed simultaneously.
You're right that the term BFO has rather gone out of common use but the effect is the same - on CW it is used to produce an audible heterodyne, on SSB it is placed at zero beat to replace (insert) the missing carrier. I think the in vogue term for this is now Carrier Injection Oscillator. I've been playing with radio for 40+ years and I'm showing my age I suspect!
I don't think it's to do with the aircraft Rx being off frequency - it's a known fact that the ground Tx is deliberately off frequency. My question was how the magnitude of the delta from channel centre frequency relates to 25kHz channel spacing, Rx filter widths and the eventual onset of 8.33kHz channel spacing.
Do you happen to know what the bandwidth (say at the 6dB points) is for a standard 25kHz chanel aircraft Rx?
mono: Thanks for the link - yes that's the beast and it goes some way to explaining what's going on. I guess that the local Scottish Tx is two "offsets" away from the channel centre.
Perhaps what has happened is that it was until recently only one offset away. Now it's two offsets away, local people are reporting troubles. It still seems mighty odd to offset a Tx by 7.5kHz when the channel is only 25kHz wide. That assumes that the aircraft Rx filter is at least 20kHz wide to accommodate both sidebands of the offset transmission (up to 2.5kHz sideband), which leaves precious little guard space between adjacent channels.
Interesting stuff.
John.