Well, today: The circuit in operation on the INV-L since last night,
100% WSPR @MF. The reports on the local RX in 20m distance were between
0...+7 dB at night and -11...-9 dB in daylight.
On the tree grabber in 3.5 km distance the signal was only detectable at
night, levels reaching about -24 dB in average. Nothing in daylight.
Where does the energy come from? It must be the voltages coming from
other transmitters (DC to HF) which leave voltages up to the range of 1
V at this antenna. This RX voltage spectrum is modulated by the circuit
and radiates a very weak signal. That's why the signal is about 13 dB
weaker in daylight. In this experiment this is an unwanted effect. It
causes some kind of offset, i.e. it lowers the SNR difference to a
signal generated by cloud charges. So the 'sky-current' must be higher
to become clearly observable. And it is easier to detect it in daylight.
The effect should be less expressed at LF because the external voltages
from other transmitters are the same but the expected voltage from a 100
uA 'sky-current' will be 475/137 higher, i.e. the system is 10 dB more
sensitive for the observation of the expected effect.
Unfortunately there is no LF monitor on the tree...
Anyway i need to be more patient. There has been no static rain yet
anyway...
Meanwhile i read a bit about Miller capacity and
https://en.wikipedia.org/wiki/Cascode. I built one out of two BS170 FETs
somehow it does not work yet.
73, Stefan
Am 11.03.2019 17:24, schrieb DK7FC:
...the circuit is completed (attached).
First i used a simple rectangular input signal coming from the source,
over an ferrite core. The gate voltage was +-10V. But then, obviously
due to the FETs internal reverse transfer capacitance i got a 200 mVpp
output across 470 pF (tested in the lab, later to be replaced by the
antenna). I built a test source for the assumed 100 uA 'sky-current'
by using a 1 kV voltage source and 10 MOhm in series. As assumed, a
linear rising voltage was observed during the off-times of the FET.
However, together with the voltage coming over the Crss, the residual
rms value even dropped! Thus i had to reduce the gate peak-peak
voltage and ended up with the circuit shown in the attachment. There
is a DC offset and the AC component is as low as possible. The reduced
the unwanted voltage 8 mV.
Then, 50 mV rms at 475 kHz at 100 uA DC 'sky-current'.
How far will a 50 mV rms WSPR signal be detectable? And will the
'sky-current' reach 100 uA at all? At least the signal strength will
rise with the square of that current, so the report are nearly as good
as a current measurement.
73, Stefan
Am 11.03.2019 13:29, schrieb DK7FC:
Hello Jim,
Yes, indeed.
I'm now preparing the circuit and want to set it up in the evening.
Found an IRF820 in the 'junk box', a good choice. Or, maybe even a
BS170 will work. It handles 60V only but i had the following thought:
The ERP at a constant voltage across the antenna (say 50 V rms at the
fundamental frequency) will rise with the sqare of the frequency (for
electrically short antennas). But that voltage will not be constant
in practice. It will drop linearly with the frequency ( ~ 1/f)
because the sky-current is assumed to be constant and the charge up
time is T/2. And, doubling the voltage at the antenna will also rise
the ERP by a factor of 4. All in all it means the the ERP is
constant, or independend of the frequency!?
That means, it would be wise to select a band where many RX stations
are watching and propagation is promising. The number of RX stations
is higher at MF but (groundwave) propagation is better at LF.
Since the peak voltage will be lower at MF (shorter charge-up time),
MF is less critical.
And which mode?! I think the best choice is a 100% duty cycle WSPR
beacon. It also has the advantage that one can check the results in
the database. And there are many RX stations, even in the groundwave
distance and even in less than 100 km here. And i have my tree which
can detect small differences at very low levels.
At MF, the BS170 is a good choice i think. And it can be driven by a
Raspi directly, maybe with a ferrite transformer in the gate-source
path, to decouple the shack from the antenna...
If that works, one could try LF.
73, Stefan
Am 10.03.2019 20:18, schrieb [email protected]:
All great ideas in the preceding messages.
A steady signal from the 136 kHz sky-current transmitter could
indicate that a data bit might be available later from the 10 Hz
sky-current transmitter.
Agreed, the triggered gap sounds better than an HV vacuum relay (for
jitter and reliability).
Sounds like the voltage divider doesn't need to be more than a
gigohm, but perhaps for future reference: a reliable 50kV 1 terohm
divider can be made very easily: encapsulate ten 100 gigohm
resistors in hard epoxy, being sure not to touch them during
assembly. A single linearity calibration after curing will be remain
valid +/- 10% for years in varying humidity and temperature.
73,
Jim AA5BW
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of DK7FC
Sent: Sunday, March 10, 2019 5:52 AM
To: [email protected]
Subject: Re: LF: Re: Re Loomis?& ... 12.47 Hz
Hi Jacek, DC,
Yes yes, i already see a new project for summer :-)
A floating antenna will charge up to a voltage where the E field
strength is high enough to start partial discharges, something like
20 kV maybe, depending on the wire diameters and homogenity of the
arrangement, also on pressure and humidity. So there will be a
voltage limit given by the arrangement. But a higher field strength
in summer will help to charge up the antenna faster, so higher
switching frequencies are possible.
One could build/use a triggered spark gap,
https://en.wikipedia.org/wiki/Trigatron, this will hold higher
voltages and you can connect the trigger directly to a ublox GPS
module running e.g. at 3 Hz :-) I think this will give a lower
jitter than when using a vacuum relay.
Could this work over my 3.5 km path if i use my inv-L? I don't think
so.
The time to integrate will be to short and there will be much QRN
during such an experiment, so the SNR will be very low. Currently
(at 12.47 Hz) i already have 15 kV DC available...
Anyway, interesting questions, there is something interesting to learn.
Maybe i will do a few experiments in summer. Building a HV divider
and measurement should be no problem...
73, Stefan
PS: Assuming 100 uA constant current 'coming from the sky', my 470
pF antenna would charge up with 213 kV/s, so it would be possible to
switch at 10 Hz and get a 21 kV 'square wave'. Maybe the harmonics
could be detected easier than the fundamental frequency.
At 137 kHz it would be 1.5 V only (0.53 V rms). This can be switched
with a normal FET! Imagine we would switch the antenna permanently
at, say 136.172 kHz, using a FET and a ublox GPS module. Then, as
soon as the field strength is high enough, we would see a carrier on
the grabbers? We need to try that out! Amazing! What a great hobby
we have!?!
PPS: For this experiment, the antenna can be series resonated, which
would help to concentrate the energy to the wanted spectrum...
|
