|
Roger, All
ROS HF /MF / Multi access ./ How did
we get here ?
Well as previously, the
mode was designed as a low
signal data mode optimised for LF/MF ,
the origins are routed in the early
days of the ROS project, when un-wittingly Jose
had accurately described the function of the
mode (then HF only) as , shall was say 'dispersed spectrum' ,
which triggered an all too well
documented response
In the midst of all this the EU
decided by the use of 'English' to define
the telegraphy channel maximum bandwidth for 500
as 100 Hz , which, by 'syntax error' resulted in
the authorisation of 'narrow data modes'
after a poll conducted on this
reflector over bandwidth allocation's I suggested
to Jose , that a 100 (99) Hz
narrow mode would be usable on the as then
,new, 500 allocations.
Followed then was quite a
interesting development path , being actually
able to use the experimental licence
to conduct 'experiments' , the mode was
coded as MF-1 and MF-7 , 7 giving a similar
data rate to PSK31 with approx 12/15
dB lower s/n performance. Initial tests had shown PSK31 to be
virtually useless on 500 , the MF-1 provided
reliable decodes some -2 to -3dB past
the point wspr stopped decoding (tested on a stable 500
Khz 350 mile day time path via delft) , where
was I think the initial QSO was with
Jim using MF-7 , resulted in the first
'long' data qso on 500 , which was
also decoded in France .
Only lately was it pointed
out, that the MF mode differed from the HF mode,
in that , 'dispersal' techniques are not used in
the data coding path, the reason being , s/n gain
can only be realised via bandwidth
and 100 hz is not sufficient for the technique to be
applied .. so If you observe a MF beacon,
the modulation 'pattern' is repeated.
The HF mode differs in
that frequency 'dispersion' is controlled by
randomization as part of
the data path , this affords improved s/n
performance over the HF > UHF path with
wider frequency tolerances , conventional
processing is also deployed, as well as FEC ..
these techniques re common to other data modes ,
but as this point 'we' enter the un-known as John
call's it !
So in part answer to
Roger's question , over multi access ,
There has been added a
modification to the code routine , which , by use
of the call sign, causes a unique locking sequence
to be generated, use of which may be
selected by the decode switch .
the result being , that all
stations may be decoded , but if the
call of a wanted / qso station is entered ,
then the software will only enter the lock
phase when that station is detected .
Use is made of the
overheads in the transmission system, affording the mode
the ability to withstand 'data'
collisions , hence more than one station may
simultaneously use the same frequency.. as once locked,
the digital recovery process , rejects 'non
locked' data as (HF) noise. Limitations exist as to signal
levels etc , but the skip distance
experienced on HF has enabled the system to
function quite well. both the 2k and 500 Hz
versions have the feature , the 500Hz
version suffering more from collisions but
still is quite functional
Multi access is therefore
not available to the MF mode, but as the
mode is 100 hz wide , its quite possible
to simply QSY , the reasons for the defined
channels are simply , that its not possible
to casually intercept a ros transmission,
you have to be on frequency and capture
the lock sequence for the frame to decode,
it is in effect a packet data system ,
coupled with the sub-audio and signal levels
below the waterfall resolution , invisible
ros as Mal coined it , defined channel
operation was the only way to ensure decodes
, wider AFC was possible but at the
expense of cpu loading, unlike most modes,
the audio frequency is 'fixed' mouse
click/tune is not possible
In passing,
The display is slightly
unique , the indicated levels are of engineering
significance , a limited user guide exists, the
author is the sole developer, there is
remote internet access built in, remote station
configuration via the internet for remote
beacon operation , a 'msn' internet
chat function is provided , cat function
for most radios , and a repeater function on band
or cross band . it is primarily a data
communication mode , unlike wspr which was
developed as a auto reporting beacon mode.
Personally In have suggested that
some HF frequency's for the MF mode be defined
as its performance in cross band qso
with 80 has been exceptional , however ...
one day may be .. but we all have
dials hi
I hope that fills in
the gap's to date !
73 -G..
Sent: Friday, September 02, 2011 9:37 PM
Subject: Re: LF: My thoughts on ROS
Well, I agree with GM4SLV's comments on ROS, although my
experience is very limited and RX only so far. I have successfully decoded
G0NDB, G4WGT and GM4SLV on 500kHz.
Like Mal and John, I also found the
user friendliness of the software lacking: the array of dials presented is
impressive but rather meaningless apart from the one on the left dealing with
audio level. The dials may be meaningful to some. I appreciate this is a "one
man job" and very credible too, but the UI does seem less user friendly than
WSPR for example.
The spots for other bands, are indeed annoying and
irrelevant. I wanted the spots for 500kHz to remain but these disappeared
soon after appearing. When I first fired up the software I had no idea
what should happen and there seemed little FAQ data to help understand
things. Did I miss the basic "getting started" user guide? Eventually
bleeps happened and signals appeared, but this was more by luck than
plan.
Unlike WSPR I am still confused about how several stations can
operate at the same in the same sub-band, although this may be my lack of
understanding. On 14MHz there seems to be 3 unique channels for ROS, so does
this mean on 500kHz there is just one? With WSPR there can be around 200
stations happily co-existing in just 200Hz of band over an hour or so because of
time and frequency separation.
In summary, I was pleased to give ROS a
successful go the other night (on RX) but I doubt I'll stick with it
either. Like John P-G I'd like to give some other weak signal FSK 2-way
communication modes a go on 500 and 136kHz.
73s
Roger
G3XBM
On 2 September 2011 20:51, John P-G <[email protected]>
wrote:
LF,
One of my objectives listed on my application for
a NoV for 500kHz was to
assist other UK stations in their own experiments,
acting as remote eyes
and ears, and it was in this role that I accepted
Graham's request to join
him and others using ROS.
The mode has had
a huge amount of press - negative and positive - and
being a one-person
development (and closed source) is certain to
be
controversial.
After installing the software on my NetBook
(Samsung NC10 - the only
Windows PC I had available) I was initially
bewildered by the look, feel
and configuration of the beast.
The
continuous appearance of information relating to bands quite unrelated
to
the band you've selected is annoying and, on the small screen of a
NetBook,
very distracting - taking up valuable window space.
The MF mode, with
it's 2 symbol rates, is nicely compact, in 100Hz
bandwidth, and the modem
seems very sensitive - often (on a quiet band)
giving 100% copy of signals
that were inaudible in the speaker and
invisible on a separate RX/waterfall
(the netbook screen is too small to
allow me to use the ROS
waterfall).
In the presence of lightning static crashes I found it less
sensitive,
often failing to lock on weak signals, but coping only with
those that
were both audible and visible.
As a QSO mode - yes it
probably does very well, but the user interface is
awful - little
documentation to get the casual user started - and the
continuously and
pointless spots of other bands is enough to drive one to
distraction - and
there is no way of storing a sequence of spots on the
band you're operating
on, which means you can't easily digest just who has
reported your last
transmission before it disappears, replaced by a spot
on 50MHz...
I
am always interested in digimodes for QSOs, not just for beacons, but
ROS
falls short, I'm afraid.
Mal's reported problems with both ROS and WSPR
decodes failing might be
more a problem with his computer - as most people
manage to decode WSPR
signals from the very weakest (just visible) at -30dB
up to the very
strongest (bright white trace) at +10dB or more. Failure on
stronger
signals is generally a timing issue - soundcard sample rate errors
- and
failure to decode in general is a PC clock timing issue - the PC
/MUST/ be
synchronised to UTC - via NTP/Dimension 4 or whatever. Window's
own
inbuilt "network time sync" just isn't good enough.
Now -
anyone want to try some other datamode (FSK only, Class-E amp) tests
this
weekend?
For tonight I've got a CW beacon running on 501.5kHz but would
be happy to
try a CW QSO if anyone hears me. Can't do x-band - no HF
antenna....
Regards,
John
GM4SLV
--
http://g3xbm-qrp.blogspot.com/
http://www.g3xbm.co.uk
http://www.youtube.com/user/g3xbm
https://sites.google.com/site/sub9khz/
|
