On Sat, Dec 13, 2014 at 10:50 PM, David Rowe <da...@rowetel.com> wrote: > I've started taking a look at modems for VHF FreeDV, starting with 1200 > bit/s FSK over FM: > > http://www.rowetel.com/blog/?p=3799
> Take a look at GMSK. Make sure you're looking at a sufficiently complex GMSK decoder, when I looked at this a few years ago the out of the box GMSK decoders (e.g. in gnuradio) were not coherent, didn't uses viterbi to jointly decode symbols ... etc. and seemed to leave a lot of performance on the floor. > Cop the performance hit and use BEL202 FSK. Maybe, the almost complete absence of inexpensive raidos with SSB for VHF/UHF is a barrier but on the other hand devices like bladerf and hackrf have made direct SDR a lot more reasonable... and 7dB covers a _lot_ of power budget difference. A dual-band 50w mobile FM radio costs about as much as a blade rf and a 1w amplifier for it... and that 1w is likely a lot more linear, the bladerf VCTCXO is probably a lot more stable, so I wouldn't even be shocked if the lower power better processing gain pair actually won out in performance. More importantly, that 7dB loss puts the digital mode at a strong starting disadvantage over regular FM... if the initial overhead is enough to lose most of ability to cut-through interference and tolerate weak signals why use digital? There are extra benefits, e.g. being able to carry metadata,.. but by themselves I don't know if they're that interesting. Also running in existing NFM probably greatly limits the ability to be spectrally efficient, even if you're able to hack a waveform that ends up being spectrally narrow, on the RX side you're still going to have your wide NFM demodulator lose its mind if there is a nearby signal (it'll just lock to the wrong one, most likely), so you won't be able to pack in more users in the same spectrum even if your signal is narrower... and I think thats a major selling point for digital in VHF (in UHF things like metadata and multipath handling are more important, and spectral efficiency less so). Plus, hardware that has direct access to the spectrum is in a position to support experimentation with techniques like CDMA (interesting again because more concurrency is useful in VHF); and coherent multiple antennas. Multiple inputs is VERY interesting for VHF/UHF due to multipath. Since amplifier considerations already demand a constant amplitude modulation scheme, a receiver can use the prior knowledge that the original signal was constant envelope to blindly recover the delay lines to coherently use multiple antennas... this scheme doesn't require the phase between the antennas be known, so it's sufficient to make the receivers run off a common clock (and bladerf and USRP b2xx have refclk inputs). Simple squelch voting scheme FM receivers exist and are widely used but I don't know if I've seen any commercial phase-coherent multiple input receivers targeting the ham. I think you could make a repeater with a large amount of multiple-input processing gain for relatively modest cost (as far as repeaters go), e.g. with 4 spaced dual polarization antennas feeding four USRP B210s and a beefy i7 for the DSP. Maybe building something like that commercially isn't interesting; but since it's COTS parts, it would be a fun project for an amateur radio club. ... but it's not at all as interesting if you're not talking a modulation scheme that can 'scales' by adding a lot of receiver intelligence. ------------------------------------------------------------------------------ Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server from Actuate! Instantly Supercharge Your Business Reports and Dashboards with Interactivity, Sharing, Native Excel Exports, App Integration & more Get technology previously reserved for billion-dollar corporations, FREE http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk _______________________________________________ Freetel-codec2 mailing list Freetel-codec2@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freetel-codec2
