Topics

FT8 performance (or, lack of) on 432


Keith Morehouse
 

So far, the 222+ contest has convinced me (or, I have rediscovered) that FT8 on 432 is unusable on paths with lots of Doppler shifted multi-path, such as my paths to DEN and PHX.  These paths are heavily influenced by aircraft scatter which causes various Doppler shifted signals.  The rate of frequency shift I'm seeing is in the range of 25-100 hz over a 15 second FT8 sequence, which is, apparently, too much for the decoder algorithm to handle.  If you calculate how much Doppler shift you are likely to get on 432 MHz from an aircraft traveling at various tangential paths and speeds, it roughly matches what I'm seeing (and what the other end of the path also sees).

I'm thinking a mode such as ISCAT (specifically ISCAT-A, which was designed for microwave rain scatter paths) might be a vastly superior mode for this situation.  FT8 certainly sucks wind under these conditions and proved it repeatedly today.

I have verified to the best of my ability that my system is stable.  My transverters are reference locked to GPS and I have another, independent, reference locked source which I can monitor on the air at whatever frequency I set it to.  I have also completed at least one 432 Q on FT8 that showed no 'drift'.  Evidently there are fewer aircraft on this path (DM58 to SLC area) or they are moving on a course which does not cause Doppler shift on any reflection.

Comments or further discussion welcome !

-W9RM

Keith Morehouse
via MotoG


James Duffey
 

Well, there is a major E-W flight path between you and Phoenix.I guess you probably have planes on that path one way or another every 10 minutes or so. If you look at the slope of the frequency shift and know the frequency, you should be able to calculate the velocity of the aircraft pretty accurately. 

James Duffey KK6MC
Cedar Crest NM

On Aug 1, 2020, at 18:09, Keith Morehouse <w9rm@...> wrote:


So far, the 222+ contest has convinced me (or, I have rediscovered) that FT8 on 432 is unusable on paths with lots of Doppler shifted multi-path, such as my paths to DEN and PHX.  These paths are heavily influenced by aircraft scatter which causes various Doppler shifted signals.  The rate of frequency shift I'm seeing is in the range of 25-100 hz over a 15 second FT8 sequence, which is, apparently, too much for the decoder algorithm to handle.  If you calculate how much Doppler shift you are likely to get on 432 MHz from an aircraft traveling at various tangential paths and speeds, it roughly matches what I'm seeing (and what the other end of the path also sees).

I'm thinking a mode such as ISCAT (specifically ISCAT-A, which was designed for microwave rain scatter paths) might be a vastly superior mode for this situation.  FT8 certainly sucks wind under these conditions and proved it repeatedly today.

I have verified to the best of my ability that my system is stable.  My transverters are reference locked to GPS and I have another, independent, reference locked source which I can monitor on the air at whatever frequency I set it to.  I have also completed at least one 432 Q on FT8 that showed no 'drift'.  Evidently there are fewer aircraft on this path (DM58 to SLC area) or they are moving on a course which does not cause Doppler shift on any reflection.

Comments or further discussion welcome !

-W9RM

Keith Morehouse
via MotoG


John Klem
 

Do you see just one (shifting) FT8 signal on 432, or more than one?  I'm wondering if you also see a stationary signal.  If it's aircraft, I'd also expect that it would eventually disappear, or sometimes you might even see more than one.  There are ways to track planes online so you might be able to correlate what you see with what's flying.

I don't know enough about ISCAT to guess about the utility of that, but...don't shoot...have you considered FT4?  The TX period is nearly 3x shorter, and tone spacing about 4x that of FT8.  This suggests the decoder could be a lot more tolerant of frequency drift, if you can live with the S/N penalty.  At that point you're probably also in the ballpark of being able to use CW, right?

John AA5PR


Keith Morehouse
 

John, sometimes one signal, sometimes 2.  I'm familiar with what aircraft look like on 432 beacons and this is sure what I remember.  Good idea about FT4.  The guys I couldn't work were NOT weak - FT4 would have done it easily.

And yes, the old standby of CW would solve all problems but good luck with getting the 'new breed' to QSY to that mode.  They won't even go to SSB...

-W9RM

Keith J Morehouse
Managing Partner
Calmesa Partners G.P.
Olathe, CO


On Sat, Aug 1, 2020 at 7:56 PM John Klem <klemjf@...> wrote:
Do you see just one (shifting) FT8 signal on 432, or more than one?  I'm
wondering if you also see a stationary signal.  If it's aircraft, I'd
also expect that it would eventually disappear, or sometimes you might
even see more than one.  There are ways to track planes online so you
might be able to correlate what you see with what's flying.

I don't know enough about ISCAT to guess about the utility of that,
but...don't shoot...have you considered FT4?  The TX period is nearly 3x
shorter, and tone spacing about 4x that of FT8.  This suggests the
decoder could be a lot more tolerant of frequency drift, if you can live
with the S/N penalty.  At that point you're probably also in the
ballpark of being able to use CW, right?

John AA5PR