On 11/4/2012 12:00 PM, Jeff Liebermann wrote:
On Sun, 04 Nov 2012 08:55:50 -0600,
wrote:

I did some searching for info for him but didn't find a lot on just a
loop resonant at 60 Khz. I was looking for turns and capacitance but
found no designs.
So I think he needs to pick a size that he is comfortable with and
calculate turns somewhere between 100uh and 200 uh and find some
polystyrene 10,000 pf caps put 4,5, or 6 in parallel and then find a
large air variable to fine tune it with. (135uh and 5000pf is resonant
at 60khz) Then he needs a matching pickup coil, I've seen a small coil
about 1/5 dia. of the larger coil used. There are other methods.
Should he use coax? Wrap it with foil when complete? (with gap)
You're thoughts? (maybe more uh's?

Ok, let's do the math (for the 2nd time as my PC crashed in the middle
of the first attempt).

The loop antenna is basically an impedance transformer. To feed with
50 ohm coax and no preamp in the antenna, the require turns ratio is:

This may be a wrong assumption. This antenna will be very close to the
receiver and this is a home brew receiver, not a commercial one. I can
make the input impedance anything I want really. I am looking at
directly driving a high impedance input on the ADC. Is there some
reason to work with a lower impedance for a very short run?

BTW, I have seen loop coupled antenna described but one of the pages I
found talked about the complexity of correctly coupling the antenna to
cables, not just impedance, but the balanced/unbalanced issue. This web
page seems to talk about it pretty well.

http://www.w8ji.com/magnetic_receiving_loops.htm

The page I refer to in my original post uses a toroid transformer to
couple the output to the antenna rather than a loop. I like the idea,
but I have no idea how to calculate the number of turns on the toroid.
His app was for 50 ohm cable and a higher frequency than mine.

http://w5jgv.com/rxloop/index.htm

I think this guy did a lot of research in designing his antenna and was
very skilled in building it.


( main-loop-turns / 1-turn-coupling )^2 = Xl-main-loop / 50 ohms
Using your 0.01uF tuning capacitor, which has a reactance of:
Xc = 1 / (2 * Pi * Freq * 0.01uF) = 265 ohms
At resonance, the inductive and capacitive reactances are equal.
Plugging in, I get a turns ratio of:
sqrt ( 265 / 50 ) = 2.3
which isn't very practical for a loop antenna.

I'm a bit unclear on this. Is the 2.3 the number of turns in the
primary loop to one turn on the pickup loop? What is unpractical about
that? Or is this the other way around with 2.3 turns in the pick up to
one turn on the antenna primary?

In case you haven't figured it out, this is my weak suit. I didn't have
a lot of the core courses in various EE disciplines while focusing on
digital. I've picked up a lot over the years in the analog domain and
signal processing, but am still rather weak in E&M. On the other hand,
I got an 'A' in P-chem! In terms of flunking undergrads, that is the
chemistry equivalent of E&M theory in EE.


To get a higher turns
ratio, a smaller cap will be needed. For example, with 1000pF, the
reactances are 2600 ohms for a turns ratio of 7.2, which is somewhat
better. The main loop and the coupling loop would need to be rather
close together. No problem with a ferrite core, but not very
practical with an air core shielded loop. You could use your 1/5th
size coupling loop (as is used in a magnetic loop HF antenna), but at
60KHz, efficient coupling will not happen. The loops need to be
physically close (or wrapped around a ferrite rod). However, a tapped
loop would work if you insist on not using a preamp.

The spacing seems to be a critical factor for optimizing transfer of
energy. One web page talked about spacing the pick up loop, not too far
and not too close.

What exactly is a tapped loop? I'm not picturing this.

I'm thinking of maximizing the voltage out of the antenna into a high
impedance load. Is that not a good thing to do? With the direct RF
sampling by the ADC I need as high an input as I can get.


The more common way is to use a preamp at the antenna. It's purpose
is to amplify the received signal, but also to deal with the loop to
coax impedance conversion. See the schematic at:
http://www.ka7oei.com/wwvb_ant_1_1.gif

Yes, I've seen this page. I noticed that he used steel conduit for the
shield and another page says to avoid magnetic materials for the shield,
although I expect the magnetic shielding of thin wall conduit is minimal.


Note that the amp if connected directly across the main loop. The
0.4uF tuning cap yields a reactance of 6.6 ohms, so the 1K amplifier
input resistor will have no effect on Q. I would have used a smaller
capacitance, but since the author elected to use a bipolar xsistor
instead of a FET, the lower reactance will result in a smaller voltage
swing and will help prevent clipping.

Really? Clipping is a concern? These antennas are typically very
highly peaked at the frequency of interest and I don't expect much
interference from strong sources like AM radio stations. Would the
clipping just be from noise? My understanding is that here, on the east
coast, it is hard to receive WWVB but then that is likely with ferrite
antennas. Will a large loop, say 3 foot diameter, pick up a
significantly larger signal?


Then, there's bandwidth or Q. The main loop will probably require
about 15ft of #26awg wire according to some of the construction
articles. At 0.041 ohms/ft, 15ft = 0.6 ohms.
Unloaded Q = Xl / R = 265 / 0.6 = 440
So, the -3dB bandwidth of the loop will be:
60Khz / 440 = 135 Hz
Since the bandwidth of WWVB is about 700Hz, that should work.

Is the bandwidth that wide? The signal is AM and BPSK modulated at 1 Hz
rate with minimum modulation periods of 0.2 and 0.1 seconds
respectively. I'm surprised the BW is that wide. I understand that the
sharp edge of the modulation requires some bandwidth beyond the bit
rate. But a document I found on a government page says the antennas
only have bandwidths around 260-310 Hz. But they also say the "system
bandwidth" is 5 Hz based on the AM pulse width minimum of 200 ms. I'm
not sure this is a good way to figure it.


It will
never really have a Q of 440 due to loading and losses, but that
should give a clue on how critical the tuning might become with a
huge, single turn loop. Incidentally, I was aiming for a Q of about 5
on the LORAN antennas which was required because the 100KHz LORAN
signal is 20KHz wide. Loops were possible, but amplified whip
antennas were much easier to deal with.

I would be doing simulations on the digital design, but I found my FPGA
software license is out of date. Good thing I didn't get pulled over!
:^) But I'll have to wait until tomorrow to get a new license so I can
work on that.

Rick