[pskmail] Re: 300 Baud Rate in the USA

  • From: John Douyere <vk2eta@xxxxxxxxx>
  • To: pskmail@xxxxxxxxxxxxx
  • Date: Fri, 27 Jan 2012 11:20:11 +1100

Hello Dave,

I can see value and no downsides to faster (= wider) modes for VHF.

1. First with the "tough of FEC" version (I like that designation)  for
FLARQ/PSKmail, I believe we can get in that vicinity with what we have now.
I would try either 2 x PSK500R as in two PSK500R carriers in parallel. This
would deliver around 880wpm.

If we can fit a third carrier in the bandwidth then we can get to around
1320wpm, in the middle of your request.

Sorry, but my "tough of FEC" starts a 1/2 (= send everything twice) as I
don't know how to make a higher ratio yet, although they do exist out there.

Now for very good quality channels maybe multiple PSK500 (no FEC) could be
of value too, as with two carriers we would have 1600wpm and with three
2400wpm. If the error rate is reasonable this is of value under ARQ.

And there is no 300baud limit I believe for VHF transmissions.

I have done a few test of Pskmail over VHF but my focus was on which mode
to use for weak signals and therefore focussed on the THOR modes.

But I miss results of field tests of PSK modes over VHF FM.

Could you shed some light there? Ideally with results over direct links and
over repeaters if possible. The critical information is the rate (%) of
repeats under ARQ.

2. As far as the heavy FEC version is concerned it is really a trade-off I
believe between robustness and speed I believe.

I think it would really depend on the objective from the application point
of view. I see at least two different constraints:

A. Conditions of operation, meaning in particular the quality of the
coupling between the transceiver and the computer. I understand that one
valuable characteristic of MT-63 is the ability to still provide meaningful
data transfer with speaker to microphone coupling only, and accommodate
frequency misalignments of about 100Hz .

B. Delivering broadcast messages with a reasonable rate of success, and
avoiding the complexity of the ARQ principle. Here heavy FEC (MT-63 has
7/64 rate meaning we send the same data more than 8 times) is the name of
the game unfortunately.

So in summary, and from a practical point of view, I don't have a clear
view right now of how to achieve what you requested for the "heavy FEC"
mode without making some compromises.

Some improvement could be made by using the full 3 or 3.5Khz of bandwidth
instead of the current MT-63 2Khz and therefore deliver a proportional gain
is speed but nothing earth shattering IMHO.

Another direction could be something like the PC modes that have been
implemented in the alpha version of Fldigi, maybe with higher bandwidth and
baud rates which could deliver an improvement over MT-63 (no linearity
requirements but wide frequency mis-alignment tolerance, therefore less FEC
required and consequently a higher user data rate maybe).

But of course others may have some better understanding of the science than
I do and have some good ideas and they are welcomed.

So I am happy to put the multiple PSK500/R carriers in the list of projects
for the ARQ applications but I need some data regarding VHF FM tests with
current PSK modes.

Can you help with those?

Best regards, John (VK2ETA)

On Wed, Jan 25, 2012 at 1:44 PM, David Kleber <kb3fxi@xxxxxxxxx> wrote:

> I'm interested in testing Pskmail on VHF for local emergency
> communications.
> We have great success with MT63 2k long on FLDIGI and using FLMSG with
> form based messages including the embedded checksum figure for the message.
> This allows us to do a single transmission to multiple stations for
> broadcast type message bulletins to an unlimited number of receiving
> stations.
> I would love to see 2 new modes for VHF to take advantage of the wider
> bandwidths and cleaner channels that we typically operate on. How about a
> heavy FEC mode for uninterrupted broadcast message transmissions (like MT63
> 2k long), but will a little less decode latency and maybe 400-500 wpm
> throughput at 3 - 3.5 k bandwidth which would work fine on the average FM
> transceivers and/or repeaters.  And another mode at around the same
> bandwidth with maybe just a touch of FEC but very low latency for ack/nak
> systems like FLARQ and Pskmail with the ultimate goal of something in the
> area of up to 1000-1500 wpm.
> I know that no one will ever beat the Shannon limits, but is my wish list
> within the realm of possibility?
> -Dave, KB3FXI

Other related posts: