Maybe I am wrong on what Andy was saying, but I got the impression that 
he did not want criticism of a specific hardware device, to wit:

"The comment
summary is that the product is good hardware device for most common
digital modes but may not work with ALE and AMTOR ."

I would hope that discussion of on-going timing issues is not avoided 
since it really is rather important for those of us involved in digital 
modes. The idea being to separate fact from fiction so that we have 
enough information to make an informed decision.

Can you truncate the beginning of an ARQ transmission and get good 
throughput? I am skeptical that you would not notice this when 
attempting to use a given mode. When the checksum is calculated, it will 
throw out the packet unless that packet is complete, or uses advanced 
concepts such as memory ARQ to "build" a good packet our of several 
damaged ones. But you would immediately see that throughput is much too 
low, wouldn't you?

I asked myself, whether or not our technology impacts any current modes, 
even relatively fast ones, and there does not seem to be much, if any, 
impact of commonly used sound card modes. With the testing that I have 
done, I have not seen any noticeable difference, whether using VOX 
keying of a rig, VOX keying with detached hardware, software rig 
control, or hardware PTT keying. I have not measured the keying time, 
since I am not sure that I have that capability, but I can tell if 
signals are not getting through if I know exactly what was sent and 
exactly what was received. If there was a discrepancy, I could determine 
if something was truncated, at least with text.

This morning I replicated a test with the RFSM2400 software sound card 
modem using several types of base modulation on a clear frequency. This 
software does not work with Microsoft Vista, thus I borrowed my wife's 
laptop that we use for public service/emergency use when mobile/portable 
since it is a Microsoft XP machine. It was connected to an outboard VOX 
keying interface device to my ICOM IC-7000 (simulating portable use) and 
for what would normally be my home station rig which is an ICOM 756 Pro 
II using an emachines XP tower interfaced through a low cost hardware 
interface that has both audio and an optoisolator PTT circuit. This 
allows me to simulate two nearby stations sending and receiving and 
perhaps at least as importantly, using different keying methods.

The RFSM2400 modem sends a beep tone with each transmission similar to 
what Rudd had mentioned earlier as a suggestion for the beginning of the 
transmission and I think that I was able to hear this on the receiving 
end for all data packets.

The following RFSM 2400 test results are delineated as HP Laptop and 
emachines tower to distinguish the two rig setups sending a standardized 
text document:

6 METER SSB - non-standard .3 - 2.7 kHz bandwidth

emachines tower - sent 20 secs @ 578 bps
HP Laptop - received 14 secs @ 825 bps

Although I did not expect the wider standard bandwidth to work, it seems 
to work well

6 METER SSB - standard .3 - 3.3 kHz bandwidth

emachines tower - sent 17 secs @ 675 bps
HP Laptop - received 12 secs @ 960 bps

I then tried 6 METER FM

HP Laptop - sent 13 secs @ 910 bps
emachines tower - received 8 secs @ 1485 bps

and switched with the emachines tower sending the same file

emachines tower - sent 17 seconds @ 678 bps
HP Laptop - received 12 secs @ 965 bps


Notes and additional comments:

- the programs indicated about +12 to +21 dB S/N

- the time and bps was taken from the program's information upon 
completion of the file transfer. The calculation is done differently for 
the TX vs the RX station, but I would consider the RX to be the number 
to use.

- for short files (my standard Gettysburgh Address)  which RFSM2400 
shows as 1493 bytes  but is just over 1400 according to my count, 
including spaces and punctuation),

- initially the software selected the short 500 protocol

- longer files will attempt longer protocols such as 3200 Long , 
(probably based on conditions) but I stayed with relatively short files.

- The audio bandwidth is very likely wider than with SSB, thus the 
improved throughput on FM.

- I am not sure about the difference in throughput depending upon which 
rig is selected as the sending station, but I did not run multiple 
trials. Since the HP Laptop which is connected to the ICOM 7000 had 
almost the same TX and RX, could this may be the limiting link in the 
chain?

- The CPS rate is faster than 1/10 of the bps throughput, so the fastest 
speed . With compression, this could nearly be doubled.

- Although I do not plan on using this modem in the future, due to 
discontinuance of further development,  it was interesting to see the 
throughput potential under pristine conditions. Also, I have tested it 
on HF, with previous reports  and it is was very limited due to not 
having a fall back position for weak signals. This is a MIL-STD 188-110 
design so it could have this feature added. The newer RFSM8000 modem has 
a fee associated with the product, so I have not continued further 
testing due to the availability of other programs that work better on HF 
and can still have good speeds on VHF, although not in the 1000 wpm 
range that can be possible with very good signals with RFSM2400.

73,

Rick, KV9U








Rud Merriam wrote:
How come you are discussing a topic the moderator asked us to drop?

I will repeat my assertion that VOX operation can be done without any
changes to the protocol.
All that is needed is for the IMPLEMENTATION that supports the protocol,
i.e. the software, to generate a tone. The tone triggers the VOX either in
the rig interface or the transmitter. During the keying period of the
transmitter the tone continues. After the keying delay expires the
IMPLEMENTATION starts sending the actual protocol.
The receive might hear some milliseconds of the tone before it begins to
hear the protocol. No changes needed there, either.

Nothing changes in the protocol. Nothing is required of the protocol.
Nothing of the protocol is damaged. Nothing of the protocol is thrown away.
Actually the tone could be useful as a marker. The receiver could use it to
tune into the transmitted signal.

 
Rud Merriam K5RUD ARES AEC Montgomery County, TX
http://TheHamNetwork.net


-----Original Message-----
From: [email protected] [mailto:[EMAIL PROTECTED] On
Behalf Of expeditionradio
Sent: Friday, August 29, 2008 2:06 AM
To: [email protected]
Subject: [digitalradio] Fast ARQ Hardware


It has been argued by several in this group that the first part of a 
digital mode transmission may be deleted by faulty transmit hardware 
without any problems in the reception on the other end. In other words, 
the first part of a transmission may be thrown away or discarded, and 
the message will still get through.  
73 Bonnie VR2/KQ6XA

Reply via email to