[P3] Ability to copy very weak signals

[alorona]

How are your 'ears'-- your ability to copy very weak CW signals?

Here's a way to "measure" your ears. You'll need your panadapter set to a narrow span such as 2 kHz. Turn on averaging so that the noise flattens out to allow you to estimate the true level of the noise. Use a full-screen vertical scale of just a few dB to make it easy to measure small signals just above the noise.

Whatever your noise floor is, when a signal creates a 3 dB "bump" above the noise floor, that means the signal and noise are equal (because if you add two equal powers you get twice the power, which is a 3 dB bump). We call this condition a "signal-to-noise ratio of 0 dB".

Now tune around the band to find very weak signals and see how much of a bump they make. Here are some rough rules of thumb, rounded off for simplicity:

Height of bump above noise        S/N ratio

6 dB                                 5 dB
5                                    3 dB
4 dB                                                        2 dB
3.5 dB                               1 dB
3 dB                                                        0 dB
2.5 dB                                                 -1 dB
2 dB                                                     -2 dB
1.8 dB                                                 -3 dB


I listed some negative S/N ratios because I've heard of folks who can copy signals below the noise floor! That's crazy!

In any case, this is a very approximate measurement, but it's fun having some idea how well you can actually copy CW at phenomenally low levels.



Al  W6LX
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[k6dgw]

"Doing science with your P3"  Nice idea Al! [:-)  What actually gets
delivered to our ears is signal plus noise.  At the "ear end" of the
receive chain, there are things we can do to increase the ratio of
(s+n)/n by either decreasing n or increasing s, or both.

Several techniques in addition to reducing the BW will, to varying
degrees, accomplish this:  The K3 AFX, RIT, RF/AF gain, inverting the
phase of one headphone [my Heil Proset does this], moving one or both
headphones forward on your head, or laying the headphones on the table
face up on the desk can reduce or modify the noise and raise the ratio
(s+n)/n as you perceive it with your ears.

I guess that headphones on the desk rather than on your ears forces the
noise into the background to a greater extent than the desired signal,
but I've never really heard a good explanation of why that works.

My hearing is admittedly significantly compromised but many hams tend to
be in the older rather than younger population, and with birthdays
passing, so does hearing.  On CW, judicious use of RIT can place the
s-component at a peak in hearing acuity without modifying the
n-component, raising the perceived (s+n)/n.

73,

Fred K6DGW
- Sparks NV DM09dn

- Northern California Contest Club
- CU in the Cal QSO Party 7-8 Oct 2017
- www.cqp.org

On 11/26/2016 9:43 PM, Al Lorona wrote:

>
> How are your 'ears'-- your ability to copy very weak CW signals?
>
> Here's a way to "measure" your ears. You'll need your panadapter set
> to a narrow span such as 2 kHz. Turn on averaging so that the noise
> flattens out to allow you to estimate the true level of the noise.
> Use a full-screen vertical scale of just a few dB to make it easy to
> measure small signals just above the noise.
>
> Whatever your noise floor is, when a signal creates a 3 dB "bump"
> above the noise floor, that means the signal and noise are equal
> (because if you add two equal powers you get twice the power, which
> is a 3 dB bump). We call this condition a "signal-to-noise ratio of 0
> dB".
>
> Now tune around the band to find very weak signals and see how much
> of a bump they make. Here are some rough rules of thumb, rounded off
> for simplicity:
>
> ++++TABLE DELETED, DID NOT SURVIVE THE REWRAP ON REPLY++++
>
> I listed some negative S/N ratios because I've heard of folks who can
> copy signals below the noise floor! That's crazy!
>
> In any case, this is a very approximate measurement, but it's fun
> having some idea how well you can actually copy CW at phenomenally
> low levels.

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]

[John Perlick]

There are several factors that make this technique insufficient to "measure" your ears.  

The K3 splits the incoming signal path and sends part of the signal to the P3 and part to the K3 receive circuitry.  The split occurs just after the first mixer so the P3 is sampling a wideband signal (+/- 200 KHz) centered at the 9 MHz IF.  

The K3 passes the same signal through the narrowband IF filters, amplifies the signals and down-converts to 15 KHz, amplifies again, and then uses a TI DSP for filtering, Noise Reduction and detection.

The P3 is an SDR directly operating on the wideband 9 MHz IF input signal.  It immediately converts the signal using a different A/D converter, a digital down converter and digital filters, and a Microchip DSP.  

These two signal paths are dramatically different!  They each have their own imperfections including noise, several kinds of distortion, and differences in digital processing algorithms.  Just because you can see a signal on the P3, it doesn't mean you can hear it through the K3.  And the reverse is also be true--just because you hear it on the K3, it doesn't mean that you can see it on the P3.  There are just too many variables.

The P3 does indeed have an extremely narrow bandwidth--only one pixel wide or roughly SPAN/450.  But, this is after a signal is processed with amplifiers, an A/D converter, a digital mixer and decimation filters, and then converted to a spectrum display with an FFT (with some NR).  All of these processes, whether analog or digital, slightly degrade the S/N ratio of the signal.  

The K3 uses a very modern approach to maximize dynamic range in the signal path.  While it is low noise, it is also highly immune to distortion.  The K3 has a wide variety of selectable signal processing techniques that are not in the P3, including two Noise Reduction systems, Noise Blanking, AFX, EQ, Notch, and both narrowband analog and digital filtering.  You can set up the K3 signal path to operate in many different ways--and these various set-ups will affect noise floor and dynamic range.

There is no guarantee that these two paths maintain the same noise floor and the same distortion characteristics.  Sorry, but you can't use your P3's display to measure your ears.  But, that doesn't mean that some excellent operators can read CW signals seemingly buried in noise.  K0KX comes to mind.

John
K0UM
 


------------------------------

Message: 12
Date: Sun, 27 Nov 2016 05:43:48 +0000 (UTC)
From: Al Lorona <[hidden email]>
To: Elecraft Reflector <[hidden email]>
Subject: [Elecraft] [P3] Ability to copy very weak signals
Message-ID: <[hidden email]>
Content-Type: text/plain; charset=UTF-8


How are your 'ears'-- your ability to copy very weak CW signals?

Here's a way to "measure" your ears. You'll need your panadapter set to a narrow span such as 2 kHz. Turn on averaging so that the noise flattens out to allow you to estimate the true level of the noise. Use a full-screen vertical scale of just a few dB to make it easy to measure small signals just above the noise.

Whatever your noise floor is, when a signal creates a 3 dB "bump" above the noise floor, that means the signal and noise are equal (because if you add two equal powers you get twice the power, which is a 3 dB bump). We call this condition a "signal-to-noise ratio of 0 dB".

Now tune around the band to find very weak signals and see how much of a bump they make. Here are some rough rules of thumb, rounded off for simplicity:

Height of bump above noise        S/N ratio

6 dB                                 5 dB
5                                    3 dB
4 dB                                                        2 dB
3.5 dB                               1 dB
3 dB                                                        0 dB
2.5 dB                                                 -1 dB
2 dB                                                     -2 dB
1.8 dB                                                 -3 dB


I listed some negative S/N ratios because I've heard of folks who can copy signals below the noise floor! That's crazy!

In any case, this is a very approximate measurement, but it's fun having some idea how well you can actually copy CW at phenomenally low levels.



Al  W6LX

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [hidden email]