To: rsgb_lf_group@blacksheep.org
References: <4d96a9d4-5f22-7650-872b-6344e9aea8e3@abelian.org>
 <579355A36AEE9D4FA555C45D556003AB9AAAF572@servigilant.vigilant.local>
 <2166618d-7238-0516-367c-2f917532eed2@abelian.org>
 <5A20A4EC.4060009@posteo.de>
 <c0930471-a16b-48ea-ffcb-536ac6e4b9df@abelian.org>
 <5A2694B9.4050407@posteo.de>
 <21caec0e-1514-3294-89a3-8569f0c6d443@abelian.org>
 <27BE047B-55C1-45E1-9BD8-575D5EF7179C@hurontel.on.ca>
 <4c4d7934-b02f-0c9f-e251-4c5a07086bef@abelian.org>
 <526a4012-19d5-aff0-0745-4b02a3a134e8@abelian.org>
 <Pine.LNX.4.64.1712120859410.14610@plato.ucs.mun.ca>
 <c834e36c-4f87-4d90-be66-fac9c8c1f6f5@n1bug.com> <5A2FD635.8080503@posteo.de>
 <807b7164-c345-9650-0c6b-d813b3304483@abelian.org>
 <0374401c-a9a5-ba61-89e0-15d5ad016ac8@abelian.org>
 <36a938b7-9db4-7417-26d6-5b4c4eac900e@abelian.org>
 <5A35A023.5010707@posteo.de>
 <1f5204c1-f256-f26e-e6dd-a4edc0c8a443@abelian.org>
 <dc56dc4d-031d-1efb-f68b-9fa6163c4c1c@abelian.org>
 <579355A36AEE9D4FA555C45D556003AB9AB1735D@servigilant.vigilant.local>
From: Paul Nicholson <vlf0403@abelian.org>
Message-ID: <efc610b1-06f3-84f9-bc29-8c72ff0f8c64@abelian.org>
Date: Fri, 29 Dec 2017 12:39:56 +0000
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.3.0
MIME-Version: 1.0
In-Reply-To: <579355A36AEE9D4FA555C45D556003AB9AB1735D@servigilant.vigilant.local>
Content-Language: en-US
Subject: Re: LF: RE: Todmorden receiver
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


Luis wrote:

 > From datasheets at Farnell, A262A6E is a 150:150 Ohm transformer
 > and A262A7E is a 600:600 Ohm

 > Is that correct ?

Correct.  The A6E has 125mH windings and A7E has 500mH windings.
The 'impedance' is a rather useless parameter for a transformer.

Both transformers have two primary and two secondary windings,
so there are plenty of combinations to use to set line levels.

 > There is a metal screening housing for the A262 transformers.
 > Is it needed or recommended ?

I don't use the metal screening cans.  I've not found a need for
them.

 > The "shield" pin in the transformer can be left floating

I leave the shield pin open circuit at both ends.  This is
to maximise the capacitive isolation between the house
ground and the rx ground.

 > Assume you are using twisted pair CAT5 cable to send the
 > signal from Rx to house,

Yes, I find twisted pair works very well and is easier to use
and cheaper than coax.  Keep the termination impedances to
ground well matched at both ends to minimise cross talk.

I use 100 ohm resistors in series with each leg of the balanced
circuit to reduce the resonance which occurs between the cat5
pair capacitance and the transformer leakage inductance.

No need to terminate with any particular characteristic
impedance.  You can keep the balanced load impedance quite
high - a few k at least.

Keep the line voltages fairly low with these transformers,
especially at low frequency, eg keep below 200mV at 50Hz,
or less if you can. Here, about 100mV at 50Hz.

No particular reason for A6E and one end at A7E at the other,
they were just what was to hand at the time and I standardised
on those for all three channels.  I'm sure other transformers
would work just as well.  The main thing with any audio
transformer is not to put too high a signal level through
them, to avoid distortion particularly of the ELF signals,
eg mains and large ELF signals from antenna movement. Needs
adequate HP filtering in the rx.   The total RMS
voltage on the signal lines here is about 1V.

 > Do you use the other pairs to provide power supply ?

Yes, one pair provides power, or parallel two pairs if you
find the resistance is too high.    Consider using quite a
high supply voltage.  I've been using 18V for a long time
but am upgrading to 48V.   Not a problem unless the cable is
more than 100m or so.   I use a switched mode PSU to power the
line and isolating DC-DC converters at each rx.  Use plenty of
smoothing at the supply end so you don't have ripple currents
on the cat5 to cross-talk into the signal pairs.

Keep the DC pair floating with respect to house and rx grounds.
You might want a 10M resistor from each pair (signal and
power) to ground just to drain any static charge build-up.
I also use a 60V gas discharge tube at each end of every pair:

 
http://uk.farnell.com/bourns/2020-23t-c2lf/gas-discharge-tube-180v/dp/1961477

Other things...

Spectrogram is now version 4.0d and is doing well with E/H
and polarisation:

 
http://78.46.38.217/fbins3.html#p=1514541600&b=245&s=fd&m=cardioid&w=r&h=42&z1=0.34&z2=0.64&c=1&mb=772,57,839,232,0

Yesterday I checked the arithmetic with some simulated signals
inserted (digitally) into the signal processing.  I discovered
that the blanker increases E/H by 2% for some reason.  It often
shows impedances higher than 376 ohms, that appears to be just
inaccuracy on noisy or unstable signals.  I can't find a fault
in the arithmetic and the calibration is pretty good now.

The polarisation is the ratio of minor to major axis of the
polarisation ellipse.  +/- 1.0 is circular. 0.0 is linear.
Distance signals should be more linear and closer to 376 ohms.
Nearer sources are likely to be more elliptical and have
lower E/H than free space.  Use narrow bandwidth for best
accuracy (wide measurement box) on weak signals.

A bit delayed on the 48V upgrade, I realised I needed to change
the 60V GDTs to higher voltage ones so I had to order those
(Farnell 1961478).  Now they've arrived but the weather is
too bad to bring the receivers in.

--
Paul Nicholson
--