Daniel Mundall 07/11/17 16:37
I really resonate with your desire to just jump right into the
protocol its self, I've played around with many such ideas - it really
is fun to dream about what could be possible!
I think that the design of a TDMA protocol for Codec2 on VHF/UHF will
turn out to be an exercise in engineering compromise.
Samuel Hunt 07/11/17 21:52
Having had some experience with TDMA data protocols (not voice, just
data), my 2 pence on this is that I feel you are considering far too
fast a slot timing for most radios to cope with.
The Si5351 frequency synthesiser proposed for the SM2000, has a rated
frequency transition time of no more than 10 microseconds (datasheet).
From my experiences with AX25 packet radio, quite a few existing ham
VHF/UHF transceivers take up to 200 - 300 milliseconds to switch from RX
to TX (or back again).
I think that the choice of slot length and inter-slot guard time will be
a compromise. Long slot times favour channel efficiency (less time spent
on inter-slot guard times and synchronisation vectors), better forward
error correction, maximum range between user & repeater, and relaxed
RX-TX-RX timing requirements in user radios. Short slot times favour
low-latency communications and "responsiveness" - *how much latency is
acceptable?* how long should a user have to wait from pressing PTT, to
gaining the channel ("floor tone"?) and being able to talk?
For FDMA uplinks, I think that you would need mandatory automatic power
control for user radios to prevent adjacent uplink channel interference
at the repeater's RX, and uplink frequency accuracy would also be
important. Of course, without cavity filters, a 2m user radio cannot RX
the downlink from the repeater whilst transmitting 600kHz away on it's
FDMA uplink channel, so the user's uplink TX would have to be turned off
whenever that user is receiving a slot from the repeater.
Realistically, unless your existing radio is something like a Symek
TRX4S (fast RX-TX-RX timing), I think that you will probably need a new
radio for TDMA, eg a SM2000 http://www.rowetel.com/?p=4694
Other ways of making a low powered portable reciprocal with a high power
base is by improving the RX performance of the base (spatial diversity
RX, beamforming, etc).
Glen English 07/11/17 22:18
whatever you guys do, just remember to keep the SYMBOL rate BELOW 1500
symbols per second if you want the thing to work well in hilly/ echoey
terrain....
I understand that channel equalisation can be used to mitigate the
effects of multipath, allowing higher symbol rates to be used (CPU
processing is cheap): http://www.ti.com/lit/an/spra140/spra140.pdf
If appropriately chosen, the synchronisation vector/unique word in each
burst could be used as a training sequence for a channel equaliser (ala
GSM, DMR, etc).
As Samuel Hunt also mentioned 07/11/17 22:49, low symbol rates would
require higher-order modulations, which require a linear TX.
With reference to:
http://www.arrl.org/files/file/QEX_Next_Issue/SeptOct2017/FrankeTaylor.pdf
2kbps MSK in a 3 kHz bandwidth has a PAPR (peak to average power ratio)
of 1.28 (or 1.07 dB). Refactoring, I'd expect 10 kbps MSK in a 15 kHz RF
channel (or 5 kbps in 7.5 kHz) would be similar. MSK can be demodulated
(sub-optimally) with a discriminator, or coherently (use each burst's
sync vector as the reference for the coherent demodulator?).
Frequency diversity with say 200 kHz shifts would require additional
cavity filters at the repeater site, and whilst it might work ok on 70cm
and up, possibly having a 2m TX only 400 kHz away from the 2m RX on the
repeater site could pose a challenge. What might be more practical is to
have spatially diverse linked repeaters, with hand-off from one repeater
to another? (like various trunked mobile radio systems)
Glen English 07/11/17 22:59
only 1000 bps ish is wanted ?
what about DQPSK on OFDM (into the mic input) say 100 Hz bins, 400-3000
Hz, 64 pt FFT easy on STM32F4/F7.
A possible problem, is that David has stated that the SM2000 "...is
designed primarily for constant amplitude waveforms such as FSK, so does
not have an AGC.".
Also, high PAPR leads to poor mobile TX power efficiency. For example,
the "base" FDMDV modem has a PAPR of ~ 12 dB: http://www.rowetel.com/?p=2965
With the 700bps FDMDV variant, PAPR is reduced to ~ 7dB:
http://www.rowetel.com/?p=4108
This is not a problem for a 100W HF set with mains power, but might be
more of a problem for a 1W handheld running on an internal battery...
A question regarding OFDM with slow baud rates (eg 50 symbols/second),
is how precisely can you recover timing from a received synchronisation
vector?
73 de ZL2WRW
Ross Whenmouth
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Freetel-codec2 mailing list
Freetel-codec2@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
