On 12/04/2019 07:05, Ed Stokes wrote:
I’m curious about something I observe about the FT8 signal modulation.
It appears to produce a pronounced spike in the audio at the beginning of each
15 second transmission.
When I operate on 160, 80 and 40 meters especially this spike will cause the
program to stop producing audio. My computer, a mac mini, is relatively close
to the 160m dipole being used on these bands, about 50 feet.
If I throttle back to -41 dB or so I can make the transmission begin without
stopping the audio, although the spike is still evident.
If I then gradually increase the “Pwr” slider I can increase the power without
causing the audio to crash.
If I leave the power increased the audio will crash immediately after the next
transmission period begins.
Is this a known problem?
73, Ed
W1KOK
Hi Ed,
there is no spike in the audio at the start of a FT8 transmission. The
fine detail is that the audio synthesis starts at a zero-crossing so
there should be minimum discontinuity. Having said that there are at
least two ways that the final transmitted signal may not be a faithful
representation of the audio feed to the rig. Firstly many users employ
some sort of VOX switching and unless there is elaborate delaying of
audio going on the start of the signal will be truncated and our efforts
to start at a zero-crossing are defeated. Secondly many rigs suffer from
some sort of ALC overshoot characteristic that causes the output power
to spike, in some cases significantly, for the first few ms of
transmissions. I can't suggest any cure for VOX induced problems, other
than switching to CAT or externally hard switched PTT, but they should
relatively harmless. For the ALC overshoot problems, you can experiment
with leaving the rig at full power and use audio attenuation to limit
the output or vice versa, depending on the nature of the rig's ALC
circuitry one or the other may help. A word of warning, these spikes
caused by ALC problems can be hard to detect as they are often shorter
than a power meter can register so may be far more prevalent than
realized, and can be a common cause of RFI issues. It is worth noting
that a typical rig PA designed for 100W o/p from transistors is usually
capable of ~150W o/p or more without ALC control, and I have seen cases
where a rig throttled back to run a couple of watts o/p will spike right
up to 150W at the start of transmission. This is very bad behaviour an
likely to kill any external PA used to increase o/p beyond 100W. This
was observed on the Icom IC-706 MkI.
On a side note, the waterfall in WSJT-X slow modes like FT8 is optimized
for weak-signal detection and does tend to exaggerate any signal
discontinuities. As a taster, the next feature release of WSJT-X
includes an option to switch to a filtered waterfall, using a
bell-shaped curve window function to minimize spectral leakage (like
narrow filter ringing) rather than the rectangular filter implicit in
the current implementation, that gives a much more faithful
representation of incoming signals at the cost of a small amount of
sensitivity. More interesting changes related to this are also in the works.
73
Bill
G4WJS.
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
