开云体育

ctrl + shift + ? for shortcuts
© 2025 开云体育
  1. QRPLabs
  2. Topics

QMX+ Receiver MDS measurements. (noise bandwidth of a FSK signal) (coherent CW)

 

Hello all,

Note that FT8 vs CW cs SSB bandwidth MDS is a matter of specification.?
Converting from one noise bandwidth to another is a matter of adding 10 log (noise bandwidth ratio).
Similarly for FT8. Here the relevant noise bandwidth is 6.25 Hz (bandwidth of one tone). So you can calculate that for 2500 Hz noise bandwidth, the addition would be 26dB because 10 log (2500 / 6.25) = 26 dB. In other words FT8 has about a 26dB SNR advantage over SSB. All SNR reports in WSJT-X and JS8Call are normalized ro 2500 Hz noise bandwidths.
I would like to confirm that the noise bandwidth in a FSK signal is
referred to a single tone.
From my understanding, If we have 8 tones, to detect the symbol, we
compare 8 magnitudes each other, polluted with 8 sources of noise....
So I was expecting an sensibility improvement of 10*log10(2500/(6.25*8))=17dB
But I never found a clear explanation in textbooks about the topic.

==
A nice idea for these transceivers is to implement coherent CW, as
explained some years ago in QST.
With 6WPM, the BW is as low as 10Hz, and it does not need a computer to operate.
But the lack of interest in these mode against QRSS3..30..60 make me
think twice.
Probably CCW benefits are less than the complexity of sharing a clock
source between TX and RX.

73 dx and happy 25
eduardo ea3ghs
 

Hello all,
?
- With 6WPM, the bandwidth is as low as 10Hz.
?
Using a 10Hz wide op-amp based linear phase filter, one can use normal CW with a speed of 10 WPM.
The only requirement is a stable oscillator at both sides, which is not much of a problem.
?
This was already possible 40 years ago, though a stable frequency source was a bit more difficult at the time.
?
73,
Roelof, pa0rdt?
 

Hello Roelof,
Thanks for your reply and your antenna design. I have one!

* I am not a CW operator yet, but I know ringing is a common problem
with very narrow filters.

In CCW design a different approximation is used, see the attached image.
The ASK detector gives a DC level that drives a local oscillator, to
get receptions free of noises.

* But there is also a second problem .With 10Hz bandwidth it is
difficult to find where everybody is!
We are expecting very week signals.

So a GPS is needed to know your own frequency. With a GPS we can
synchronize also
the start of dot/dash +- error of signal propagation (less than
4000km in first hop, or 13ms t=d/c=4E6/3E8).

* These mechanisms could be integrated in QCX+ as a software. It was
my proposal.
But I am not sure if all these complexities justifies to win....
10*log10(500Hz/10Hz)dB=17dB, or 3 S-units

It is a market problem, not a technical one

73 dx eduardo


On Fri, Dec 27, 2024 at 6:03?PM Roelof Bakker via groups.io
<roelofndb@...> wrote:

Hello all,

- With 6WPM, the bandwidth is as low as 10Hz.

Using a 10Hz wide op-amp based linear phase filter, one can use normal CW with a speed of 10 WPM.
The only requirement is a stable oscillator at both sides, which is not much of a problem.

This was already possible 40 years ago, though a stable frequency source was a bit more difficult at the time.

73,
Roelof, pa0rdt
 

Hello Eduardo,
?
Using a linear phase filter, ringing is not an issue, even with a 10 Hz bandwidth.
?
It can also be done in software, e.g. Linrad can be used for single side-band CW reception in a 6 Hz bandwidth, plus the ability to use as many notch-filters in the same bandwidth as you like.
This is insane, yet I have used it to achieve aural copy of very weak non directional beacons? at long-wave? sitting very close to adjacent carriers.
?
73,
Roelof, pa0rdt
?
?

1 - 4 of 4