Message-ID: <4BA09E22.9060606@abelian.org>
Date: Wed, 17 Mar 2010 09:17:22 +0000
From: Paul Nicholson <vlf0308@abelian.org>
User-Agent: Thunderbird 2.0.0.14 (X11/20080421)
MIME-Version: 1.0
To: rsgb_lf_group@blacksheep.org
References: <910844.27140.qm@web86505.mail.ird.yahoo.com> <4B9FDFA7.7090208@gmx.com> <38A51B74B884D74083D7950AD0DD85E82A1B5D@File-Server-HST.hst.e-technik.tu-darmstadt.de> <4BA00630.9070300@abelian.org> <20100317073501.185C3F3863@smtps02.kuleuven.be>
In-Reply-To: <20100317073501.185C3F3863@smtps02.kuleuven.be>
Subject: Re: LF: 8.97kHz
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


Rik Strobbe wrote:
 > do you know the difference (signal strength) between Stefan's signal
 > and the 11kHz Alpha signal at your place?

Yes, Alpha amplitude was logged at the same bandwidth during
the test.  The average amplitude of the central 11904.761Hz
component during the 90 minutes analysed was 11.6fT, but there
is probably 50% error on that estimate.

I recommend calibrating your rx directly.  If you are using
a loop antenna you can use a current injection method, as
described in this paper,

https://sakerlagring.telia.se/user/Share.aspx?id=bfe0945f-b99f-4e6d-ae99-5b65c13523dc

section 6.2, or 6.3 if you are unbalanced.   All you need is
a signal source of known amplitude, and measurements of the
loop's inductance and resistance.   Maybe not so straightforward
if you are using a tuned loop.  Assuming a non-tuned loop with
low impedance amplifier input, it is possible to leave the R+C
calibration network in-circuit all the time without affecting
performance  - the noise produced by the calibration network
is a poor impedance match to the amplifier so doesn't make
a difference.  Then you have continuous calibration!

 > This would give me (and maybe others) a target for my 8.97kHz
 > RX setup.

 From what I can see of the spectrograms posted by various
people, the VLF noise floor is clearly dominant on them all.
The diurnal variation is present.   A useful reduction of the
noise floor is obtained by orienting the loop to null as much
of the background as possible.   The null orientation changes
throughout the day.

Further background reduction is obtained by blanking the
stronger sferics.   On my rx, I trigger a 2mS (retriggerable)
blanking whenever the instantaneous amplitude exceeds 10 times
the mean rectified amplitude, where the mean is computed as a
moving average with 15 second time constant.   These settings
were arrived at after some trials against a test signal.
During the day about 3 or 4% of the signal is discarded, at
night it can be 10% or more.   Oh, the blanking takes place
after a high-pass filter to remove the stronger mains hum
components, otherwise they dominate the mean signal amplitude.

I will produce some charts of the average background diurnal
at 9kHz.

--
Paul Nicholson