Re: LF: RE: WSPR Reports

[Andy Talbot <andy.g4jnt@googlemail.com>]

Yes, as Paul suggests, it does look as if AGC is the culprit.   Narrowing the Rx bandwidth right down to 7Hz won't make any difference to decoding ability provided there is no AGC pumping from out of (WSPR band) signals.   

 

(But any bandwidth selected below 7Hz will then knock out tone energy in the two outside ones).

Andy
www.g4jnt.com

2010/1/27 Clemens Paul <cpaul@gmx.net>

This phenomen,or rather bug,has been reported here by several list members.

I have made the same observations.

Reducing the RX BW from e.g. 250Hz to something below 100Hz results in

an increase of *reported* SNR of up to around 15dB!

I've done some tests with a friend 20km away on 10m and we've found out that the *real* SNR

is not improved a bit or in other words an extreme weak wspr signal which was just below

the decoding threshold in 250Hz BW coudn't be decoded with 20Hz BW either.

 

73
Clemens
DL4RAJ 

----- Original Message -----

From: Andy Talbot

To: rsgb_lf_group@blacksheep.org

Sent: Wednesday, January 27, 2010 9:44 PM

Subject: Re: LF: RE: WSPR Reports

All the critical filtering is done in the software, and results in a 1.46Hz noise bandwidth - which is what the software measures and reports upon.

 

Your observation that a narrower IF bandwidth is interesting.   Its most likely based on the fact that typical signals are corrupted by pulsed non-gaussian noise which upsets the frequency bin detection process, and upsets sync recovery.

 

Andy
www.g4jnt.com

This email has been scanned for damaging side-effects by the health and safety police, is guaranteed to contain no substances hazardous to health, but may contribute to dissolving the nether and polar regions

2010/1/27 Rik Strobbe <Rik.Strobbe@fys.kuleuven.be>

Andy,

 

maybe the "magic" lies in a narrow filter (either IF or audio).

If I switch from a 2.5kHz (SSB) filter to a 270 Hz filter I also get a significant increase in SNR values reported by WSPR.

Unfortunaely it hardly affects the decoding level (minimal SNR needed to get a positive decode), that rises by almost the same amount.

 

73, Rik  ON7YD - OR7T

---

Van: owner-rsgb_lf_group@blacksheep.org [owner-rsgb_lf_group@blacksheep.org] namens Andy Talbot [andy.g4jnt@googlemail.com]
Verzonden: woensdag 27 januari 2010 21:26
Aan: rsgb_lf_group@blacksheep.org
Onderwerp: LF: WSPR Reports

Has F6CNI managed some sort of magic noise reduction system, or is it a just a case of being in a very quiet location...

These are pretty impressive S/N values

 

 

2010-01-27 20:14	G4JNT	0.503875	+13	0	IO90iv	+23	0.200	F6CNI	JN19qb	391	243
2010-01-27 20:08	G4JNT	0.503875	+10	0	IO90iv	+23	0.200	F6CNI	JN19qb	391	243
2010-01-27 20:02	PA0A	0.503900	+15	0	JO33de	+37	5.012	F6CNI	JN19qb	502	312
2010-01-27 20:00	G4JNT	0.503875	+14	0	IO90iv	+23	0.200	F6CNI	JN19qb	391	243
2010-01-27 19:54	G7NKS	0.503920	-5	0	IO92ub	+20	0.100	F6CNI	JN19qb	422	262
2010-01-27 19:54	G4JNT	0.503875	+14	0	IO90iv	+23	0.200	F6CNI	JN19qb	391	243
2010-01-27 19:44	G4JNT	0.503875	+10	0	IO90iv	+23	0.200	F6CNI	JN19qb	391	243
2010-01-27 19:40	PA0A	0.503900	+15	0	JO33de	+37	5.012	F6CNI	JN19qb	502	312

 

Andy
www.g4jnt.com

---

Eingehende eMail ist virenfrei.
Von AVG überprüft - www.avg.de
Version: 9.0.733 / Virendatenbank: 271.1.1/2649 - Ausgabedatum: 01/27/10 14:08:00