Guidance on AM mode Coh2 setup

[rsku...@gmail.com]

I've been using the Coh2 mode and the Coh AGC at times using the AM mode (under Linux) and it amazes me how well it helps separate stations on the channel into a soundscape that my ears can seem to process stations separately.

But I have a few questions on setup. 

I see Leif had a video about zooming way in and then using a large RAT number (9999 I think it was) to lock on one specific carrier on the channel. Initially I was using the right top arrows only to zoom, but I did see that there was an increasing delay in the sound latency and eventually before I got far enough into it to be able to separate the individual carriers Linrad would crash. I later saw that the top left arrows could also zoom, so I was thinking maybe this was how you were zooming the screen Leif.

My question is how much of the zooming should be done using the right top arrows and how much on the left top arrows to get the right mix in performance? And, what about other settings that would help in optimizing this mode?

I am really looking forward to getting the zoom deep enough to get the carrier filter locked on the station of interest.

Thanks in advance for any advice.

Rick Kunath, K9AO

[Leif Asbrink]

Hello Rick,

There are three ways to zoom the baseband.

1) Set the minimum sampling rate for your hardware.
2) Set the first mixer bandwidth reduction large so
the baseband sampling rate will not be larger than
required for the channel of interest. If the hardware
allows 48 kHz (I and Q, 48 kHz bandwidth) and you are
interested in AM you might set the bandwidth reduction
to 3, 8 times for a baseband sampling rate of 6 kHz.
The audio bandwidth will be something like 2.5 kHz.
3) Use the arrows in the upper right corner. The maximum
fft3 size is 15 or 16 depending on other parameters.
With 15 the time span of your fft3 would be 32768/6000=
5.5 seconds with a bin separation of 0.18 Hz. You should
be able to detect on carriers separated by about 0.3 Hz.

The arrows in the upper left corner of the baseband graph
does not affect processing. It just expands the graph
with linear interpolation between fft3 points.

At the moment I am running my Perseus at 48 kHz with the
first mixer bandwidth reduction = 10 and with fft3 window=1
to get the sharpest peak (and a poor filter shape.) The
sampling rate in the baseband is 46.875 Hz and the fft3
size is 15 so one transform spans 712 seconds, 11.8 minutes
and the bin bandwidth is 1.4 mHz (milliHz)

I am comparing the frequencies of different precision
oscillators. As expected not very sucessful, the stability
of the Perseus is far from good enough. This is part of my
latest video nerds 78. How to get an idea about phase
noise of a very good oscillator as random fluctuations of
the frequency.

73

Leif

> --
> There is an excellent Linrad User Guide by Gaetan, ON4KHG, at:
> http://w3sz.com/Linrad%20Installation%20&%20Configuration%20User%20Guide%20-%20V1-0.pdf
> ---
> You received this message because you are subscribed to the Google Groups "Linrad" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to linrad+un...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/linrad/08cfef66-23fa-4741-9260-db5b75ee93afn%40googlegroups.com.