[TangerineSDR] VLF Antenna Site Isolation

[Jonathan]

Hi All,

As was seen in the SAQ spectrum plot, there were prominent mains harmonics
in that section of spectrum. Normally that isn't the case, as mains
harmonics typically tail off at around 6-8 kHz. I was showing my spectrum
plot to Paul, the author of vlfrx-tools and a VLF enthusiast for over 20
years with many advancements for VLF reception at home to the VLF
community. His website is abelian.org and runs and publishes data from a
network of VLF receivers for both natural radio events and VLF amateur
transmissions in the Dreamer’s band.

Paul and I had a long discussion about these harmonics and the VLF system
site installation. He asked the initial question:

*Impressed that you're getting mains harmonics so high, have you got
something clipping somewhere?*

Because I was using his VLF preamp design and he added a lot of headroom
for strong mains hum and sferics in the front end, there definitely wasn’t
any clipping. The audio level was quite reasonable and even the strongest
sferics would be no greater than 45% of the soundcard’s input range.
Initially, he wanted me to perform an experiment by wrapping grounded foil
on the antenna, but after looking at some captures of the VLF spectrum, he
decided it wasn’t necessary to perform the experiment. The experiment was
to attempted to eliminate a potential source of harmonics, any clipping in
the VLF preamp /feedline chain. Paul then talked about the issues that
heavy mains harmonics can cause for a VLF receiver:

*You'll find that with a mains harmonic spectrum like this, your system
noise floor is also significantly raised. The mains cross-modulates with
the VLF noise to raise the floor, and you can hear that, as a sizzling
frying sound to the noise rather than a pure hiss.*

This was exactly what was going on. It can be heard very clearly in the
mains filtered stream and the raised noise floor, with mains harmonics
extending beyond 20 kHz, can be seen in the spectrum display. The problem
is, the increased noise floor makes weak signals from natural radio events
and amateurs difficult to detect. This was clearly seen in the SAQ spectrum
plot, as the mains harmonics were much more stronger than the weak SAQ
transmission. I described the site installation and showed him pictures
that included the installed active VLF antenna, conduit with ground wire,
and ground magnetometer. Paul then went on:

*Looks like a fine installation in an enviable location and just about far
enough from the trees, the VLF reception looks good.*
*You have three good reference signals coming in: NAA, NAU, NML.*
*No sign at all of any clipping and the incoming 60Hz and harmonics are
nowhere near strong enough to overload the pre-amp or SP70.*
*Wrapping the antenna doesn't just hinder the E-field pick-up, it also
shunts the ground pick-up too: The pre-amp amplifies the potential
difference between local ground and E-field and the wrap just puts a
capacitive shunt across its input, thus reducing the response to both
'sources', a non-invasive way to temporarily drop the pre-amp input level.
But it seems you're not overloading so no need to do this test.*

Paul makes a good point here which I wanted to emphasize that the earth
ground connection is part of the antenna, so shunting the antenna element
would also shut any pickup from the earth ground connection, as the
potential difference between the antenna element and earth ground is what
the input of the front end is seeing. Paul, continuing, comments about the
source of the heavy mains harmonics:

*We are probably seeing noise on the domestic ground getting into the rx
ground circuit somehow. The spectrum and waveform looks very typical of
switching PSU interference, UPS, chargers, etc, typical domestic ground
noise.*
*I see a ground wire dropping down from the pre-amp into earth, that should
be the only ground connection out at the site. The cat6 screen and the
extra conductor in the conduit are both potential sources of trouble,
providing channels to bypass the isolation and couple the domestic ground
and the noise it carries, out to the receiver site ground.*
*Where/what does the conduit ground wire connect to at each end?*
*I see two conduits, a grey with box which is presumably the cat6
downlink/power, and a white with a single cable, is that the head of a
ground stake*?

Here, Paul reflected my initial fears with adding any grounds and low
impedance paths between the antenna site and residence. On the VLF preamp,
both power and audio lines are magnetically isolated as well as at the
residence inside the Raspberry Pi box. This isolation is critical to
keeping mains hum and harmonics out of the VLF spectrum. I mentioned the
magnetometer’s cable shield was isolated, enclosed in the “white conduit”
he mentions. I told him the shield on the cat6 VLF receiver feedline is
also floating. I also mentioned the ground wire is just coiled up and
laying on the ground at both ends. Paul goes on:

*The likely culprit is that ground wire, outside the conduit, it will be
galvanically grounded all the way along.*
*By connecting the two earth zones you're bringing all the muck on the
domestic ground out to the rx site, bypassing all your DC and signal
isolation efforts, you may as well move all the antennas back to the
residence.*

I seconded his sentiments. The ground wire is the lowest impedance path. I
do believe there is a high likelihood of these harmonics are primarily
coming from this ground wire, seeing as the site is fairly remote with low
hum levels and domestic interference, but even so, without some isolation
from domestic sources, even the quietest of radio-quiet areas can furnish
interference in the form of ground currents. Finally, Paul proposes an
excellent solution, not only for this case, but for future PSWS users who
want to co-site a VLF receiver with the magnetometer and HF antennas:

*That ground wire might also be a good way to channel lightning currents
back to the residence, at least for the few microseconds before the thin
wire vapourises. A better solution would be a good solid earthing
arrangement at the rx site, establishing a firm and clean local ground for
all the electronics and a short low inductance sink for lightning currents.
This, combined with isolation on all DC and signal circuits in the conduit,
and suitable surge protection at key points, ought to protect the
residence.*
*You're going to co-site things like HF loops and magnetometer and their
power/signal links would need to be isolated too, to both offer an
un-inviting path to lightning currents and to keep at bay the residence
ground noise. The I2C link could be a challenge there. At some point it
gets easier to digitise at the remote site and have a single isolated power
and data link.*

I feel Paul’s argument here is a good one, and can be used for the benefit
of the PSWS. Good, solid grounding at the local antenna site with isolation
on all feedline a will not only keep domestic mains harmonics from
polluting the antenna site ground, but offer good lightning protection for
the residence. This also brings up a case for both power and signal
isolation of the magnetometer, on both ends, that could add a lot of
benefit for the magnetometer’s performance. It has been shown by Bill that
the Pi host does create some interference in the HF bands, and isolation
would be a great mitigation measure. TI makes an IC with both I2C and power
isolation that could work quite well in a future revision.

Jonathan
KC3EEY
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.tapr.org/pipermail/tangerinesdr_lists.tapr.org/attachments/20220103/f15f28ee/attachment-0001.html>