[600MRG] Receiver bandwidth for WSPR

[Brian Pease]

Sure enough, I got the expected bogus results with 20Hz BW!  I moved my 
signal to be near the center of the 20Hz BW.  I got -21dB s/n with 400Hz 
BW, as expected, but got -7dB and -4dB at 20Hz BW.  Still at 20Hz BW, I 
then reduced my WSPR signal by 10dB and got no decodes, as expected, 
showing that the s/n is bogus.  I think that WSPR must sampling noise 
over the 200Hz WSPR band.  With a 20Hz filter, most of the band has zero 
noise, so the noise value is way low.
I now conclude, by actual measurement, that it makes no difference what 
BW is used as long as it is >200Hz (as long as there are no strong out 
of WSPR band signals to contend with).
On 3/21/2014 1:35 PM, Brian Pease wrote:
> I have AGC on because it is fast and seems to make no difference with 
> WSPR.  The atmospheric noise experiment is running now.  I averaged 
> ~10 spots for each result.  At 400Hz BW I got -20.4dB avg followed by 
> 3100Hz BW with -20.25dB avg followed by 400Hz BW with -20.0dB.  This 
> shows that the s/n is virtually the same for both bandwidths in 
> real-world conditions.
> My receiver has a 20Hz bandwidth which I will try next, but as it is 
> only 1/10 of the WSPR band, I predict bogus results unless the 
> software is really clever.
> On 3/21/2014 12:58 PM, Edward R Cole wrote:
>> Just for comparison, my 80m noise floor using 2.8 KHz bw with K3 
>> connected to inverted-V is running S3-4.  There are times it reaches S5.
>>
>> This is how WSPR and JT65 provide their super sensitivity.  They run 
>> FSK with a maximum shift of about 180 Hz and use DSP filters to 
>> detect the tones with under 10-Hz RBW.  SNR improvement by reducing 
>> bw from 2.5 KHz to 10-Hz is 24 dB.
>>
>> Reduction of bw from 2.6 KHz to 400-Hz is exactly = 10Log(0.4/2.6) = 
>> 8.1 dB
>> Probably your s-meter lies! ;-)  Did you disable your AGC for this 
>> measurement?
>>
>> But no doubt narrowing bw helps reject noise.  Best SNR will result 
>> when Rx bw is same as signal bw.
>>
>> At 07:44 AM 3/21/2014, Kenneth G. Gordon wrote:
>>> On 21 Mar 2014 at 9:50, jrusgrove at comcast.net wrote:
>>>
>>> > I don't know how Joe
>>> > is making the calculation but I can tell you that bandwidth size 
>>> does matter.
>>>
>>> Most definitely true. Even in the very simple, and MOST annoying, 
>>> case I am
>>> experiencing here: on 80 meters, my constant noise level at the 
>>> moment is
>>> 20db over S-9 when my receiver is in the "SSB bandwidth", about 2.6 
>>> Khz.
>>>
>>> When I drop the bandwidth to around 400 Hz, the noise level drops to 
>>> S-7, a
>>> difference of almost 40 DB.
>>>
>>> So, yes, bandwidth can make a HUGE difference in S/N ratio.
>>>
>>> Ken W7EKB
>>>
>>> _______________________________________________
>>> 600MRG mailing list
>>> 600MRG at w7ekb.com
>>> http://w7ekb.com/mailman/listinfo/600mrg_w7ekb.com
>>
>> 73, Ed - KL7UW
>> http://www.kl7uw.com
>>     "Kits made by KL7UW"
>> Dubus Mag business:
>>     dubususa at gmail.com
>>
>>
>> _______________________________________________
>> 600MRG mailing list
>> 600MRG at w7ekb.com
>> http://w7ekb.com/mailman/listinfo/600mrg_w7ekb.com
>>
>
>
> _______________________________________________
> 600MRG mailing list
> 600MRG at w7ekb.com
> http://w7ekb.com/mailman/listinfo/600mrg_w7ekb.com
>