[pskmail] Re: Pskmail mode comparison.
- From: Pär Crusefalk <per@xxxxxxxxxxxx>
- To: pskmail@xxxxxxxxxxxxx
- Date: Wed, 27 Dec 2006 17:08:41 +0100
Excellent info Rein.
I dont have the same extensive test experience but I have been watching
your tests "live" on my web monitor page. It seems to me that almost
every APRS pos beacon is acked by my server when we run PSK125. It
didn't use to be that way when using PSK63 and gMFSK. It could actually
be the work of several factors here, fldigi has a much faster AFC than
gMFSK and that helps get acquisition of the signal faster and with less
starting errors.
I guess we should say that we tried Dominoex with a "manual" qso as well
as some link attempts on 80 meters. I got a rather poor first impression
of the mode and we had serious problems getting even a single aprs pos
message through (we did in the end but not many).
In all I agree that PSK125 works as well as and in fact probably better
than PSK63. I will change the scan here to only use PSK125 and wait for
a better implementation of Dominoex.
73 de Per, sm0rwo
tis 2006-12-26 klockan 21:39 +0100 skrev Rein Couperus:
> Pskmail mode comparison.
> ==================
>
> I have been testing Fldigi extensively during daytime. I was especially
> interested in a comparison of the new modes which have become available.
> Servers were PI4TUE (20 km) and SM0RWO (1200 km). Tests were run December
> 10-22, running DominoEX 8 and 11, and PSK63 and PSK125. The tests were done
> by downloading the same quick brown fox file using different modes, making
> the tests comparable.
>
> Channel conditions were good in most cases. With PI4TUE I have a steady S7,
> with some multi-path caused by reflection against airplanes (...). The SM0RWO
> server is S4-S9+, with fast, deep QSB (S0) in the late afternoon.
>
> I had expected a lot from DominoEX, but it looks like the implementation in
> Fldigi does not hold its promise. PSK125 was better in all cases. Frequency
> sensitivity was worse than PSK125 and the number of undamaged blocks was a
> lot less than PSK125. In fact, it was sometimes possible to set up a
> connection, but even when the conditions were very good I never got a
> complete frame through.Fldigi has a DominoEX implementation which is a bit
> simpler than the original, and Dave will improve it. But until then, DominoEX
> in its Fldigi implementation is certainly no improvement over PSK.
>
> The comparison between PSK63 and PSK125 was very interesting. Although PSK63
> has a theorectical advantage over PSK125 of 3 dB, the application band where
> that advantage really materializes is very narrow. Only on an empty
> frequency, with marginal S/N ratio and NO qsb, you can actually cash in on
> it. The key is the block length. As soon as there is QRM or QSB, the 50% time
> it takes to get a block through to the other end undamaged adds the 3 dB
> loss back to the reckoning. Which means that for an adaptive blocklength
> oriented ARQ protocol like pskmail, PSK125 is superior in all cases!!
>
> The principle is very simple. We are talking statistics here. The chance a
> block is hit by QRM or QSB is doubled when you double the exposure time. And
> it takes only half the time to repeat a block compared to PSK63.
>
> PSK125 is less frequency sensitive than PSK63. The time requirement for
> synchronisation is a lot smaller (actually, 50% of the time it syncs on the
> first character of the frame). And many blocks survive a QSB hit which will
> kill it in PSK63.
> In practise in PSK125 you use a 32 character block where in PSK63 16
> characters were used.
>
> This is why I have chosen to make PSK125 the default mode for PI4TUE.
> ===================================================
> If during the session you need to, you can switch to PSK63 on-the-fly.
>
> Rein PA0R
>
--
Pär Crusefalk
Turnévägen 5
14243 Skogås
Tel: +4687717482
Mob: +46768800670
Other related posts:
