Feeding ground independant loop antennas

Thread Starter

wpri268

Joined Feb 18, 2012
30
I am a BCB DXer phasing a pair of terminated, untuned pennant antennas. The characteristic impedance of ea antenna is ≈ 940Ωs. 75Ω coax is coupled to ea antenna via 12.5:1 isolation xfmr (≈ 88:7 turns) built with stacked (Amidon) FT240-61 cores. Ea coax then goes to a preamp powered from the feedline. Lately I have been wondering how much signal loss there is across those transformers because of the number of turns involved. I understand that a higer μ core would reduce the turns but they all have higher core losses. Higher μ= higher loss!
How can I eliminate the transformers and feed the antenna directly I asked myself? From what I have read single ended and differential jfet amps are often used with loop antenns. Whereas the jfet has an extremely high Zin the amps can be easily set to match the antenna characteristic impedance. The amp then can be isolated from the feedline via a simple 1:1 iso xfmr or a choke balun. The dilemma I face is how to supply DC power to the amp while still maintaining ground independance?
Thoughts?
 
Last edited:

inwo

Joined Nov 7, 2013
2,419
Do you have a diagram of antenna?
I'm not familiar with them.


Wouldn't you isolate the antenna from the amp? Not the feedline.
 

alfacliff

Joined Dec 13, 2013
2,458
loops can be also fed with an inductive link. any good antenna manual (arrl, rsgb and such) will have the plans. the transformer has so many turns because the ratio is so high. plans for a 160 meter loop ham antenna could be used, its not that far off frequency.
cliff
 

Thread Starter

wpri268

Joined Feb 18, 2012
30
Thanks for the thoughts and comments so far. I guess that I needto explain myself a little better

http://www.w8ji.com/k9ay_flag_pennant_ewe.
http://www.bamlog.com/flag.htm

For those not familiar with pennant antennas these are two links that contain lots of technical and construction information about them.

My two pennants are built with the dimensions spelled out in the bamlog link. Note that according to the plans the antenna is not connected to ground anywhere but is galvanically isolated from the feedline by a RF isolation transformer. Typically the xfmr ouput is fed into an RF preamp (as mine are) and they are powered by DC power through the coax.

My decision to use 61 material was a concious one. While it is true that another ferrite materials can be used that will reduce the turns, higher μ materials are very lossy. 61 is virtually lossless below 10MHz. However the high turns count can represent losses too due to skin effect and turn to turn capacitance.

For this reason I would like to eliminate the transformer by placing a high impedance JFET or BJT preamp right at the antenna output. What I am really having trouble with is how to power the amp yet keep tha antenna floating. How do I decouple the DC ground?
 

Tesla23

Joined May 10, 2009
542
Lately I have been wondering how much signal loss there is across those transformers because of the number of turns involved. I understand that a higer μ core would reduce the turns but they all have higher core losses. Higher μ= higher loss!
How can I eliminate the transformers and feed the antenna directly I asked myself?
I'm no specific knowledge of these antennas, but I would try to understand what is limiting your sensitivity before attacking the transformers. I would have thought that unless you are located on an exceptionally quiet site, a dB of loss in a coupling transformer would have minimal effect.

As a general rule, at HF with a normal antenna, your performance is rarely limited by the receiver noise figure. This is because there is so much external noise picked up by the antenna. Linearity is normally a much bigger problem - limiting your ability to detect a small signal in the presence of much larger ones. I guess I am warning that if you insert a low noise amp with modest linearity, you may gain almost nothing on sensitivity but degrade your intermod performance significantly.

I would have thought the termination resistor in these antennas would contribute significantly to the receiver input noise, but I haven't analysed it. You may find that by reducing the loss in the transformer you are able to hear the noise from that resistor more clearly. When you include the external noise picked up by the antenna, if the system is well designed, this will swamp the termination resistor noise.
 

Thread Starter

wpri268

Joined Feb 18, 2012
30
Good point about noise from the system however QRM and atmospheric noise at my QTH would swamp any thermal noise. Sensitivity is not an issue either with my Ten Tec RX340 and DX Engineering RPA-1 preamps.

What I am ultimately trying to improve is antenna phasing.

I believe that the impedance mismatch between the antennas and the coax are adding to the complex impedance of the antennas. The transformers themselves are adding leakage inductance as well as magnetizing inductance. This in turn is making phasing of the antennas with my DXE NCC-1 Antenna Phaser sometimes difficult.

My transformer design was driven by the known Zo of the coax and the approximate characteristic antenna impedance of 940Ω shown in the plans. The actual antenna Z was never measured but they are by nature not purely resistive (non resonant). Recently I measured the feedline impedance at the RX end of the coax with an AIM4170 and found it to be 60ish ohms +/- jXX depending on frequency.. not exactly 75+ j0Ω.

Now that I can measure the antenna Z, I can massage the transformer turns to make the impedance match better but that wont eliminate the reactive impedance from the transformer. So, why not eliminate the transformers?

This is why I want to place a high Zin (940Ω) JFET preamp at the antenna terminals leaving the preamp output to match the feedline. But that brings me back to the original question-how do I decouple the DC ground from the antenna?
 

PRFGADGET

Joined Aug 8, 2011
53
You might take a look at capacitive coupling in place of the transformer , since you seem to be interested in receiving only .
There are many of the older Antenna Handbooks available on-line that deal with this.
Beyond that, I would suggest that you take a look at "Magnetic Loop" antennas.
One thing to keep in mind (depending on where your situated) is lightening or other sources of high voltage ingress to the antenna / pre-amp's / receiver.
 

Tesla23

Joined May 10, 2009
542
Good point about noise from the system however QRM and atmospheric noise at my QTH would swamp any thermal noise. Sensitivity is not an issue either with my Ten Tec RX340 and DX Engineering RPA-1 preamps.
If noise and interference swamp your thermal noise (which is normal for HF), then why are you trying to reduce the thermal noise added by the Rx with preamps? Without the preamps, your receive chain may have a noise figure of say 15dB, this means that the receive chain adds noise equivalent to 15dB above thermal. By adding preamps you may reduce this to say 4dB, which sounds like a great improvement. But at HF, if your antenna is feeding you noise 20dB above thermal, you won't notice the difference. I'm making numbers up (but not atypical), you need to know how quiet your site is.

What I am ultimately trying to improve is antenna phasing.

I believe that the impedance mismatch between the antennas and the coax are adding to the complex impedance of the antennas. The transformers themselves are adding leakage inductance as well as magnetizing inductance. This in turn is making phasing of the antennas with my DXE NCC-1 Antenna Phaser sometimes difficult.
I just checked out the NCC-1 box, you're playing with some fun stuff!

I can't see how the impedance of the transformer can affect the nulling ability with the NCC-1. There is clearly some benefit in making the antennas as similar as possible, but unless NCC-1 has poor isolation between inputs (not specified but it is hard to see how it can meet the other specs if it not good), then its performance should be fairly independent of the impedance presented by the antennas.

Increasing the spacing between the antennas is a surefire way to improve your ability to discriminate between signals.

Sorry to bang on about linearity, but if you are in a site with lots of interference, I'd start by removing the preamps and see what happens. Really, I'd like to connect a spectrum analyser to your antennas and see just what strong signals you have. If you do have strong local signals, then some form of filter before anything else may be warranted. The intermod performance of the RPA-1 is about 10dB worse than the NCC-1, so it's possible that you are getting limited by intermods in the preamps, that you can't null out. Try sticking some narrow-band filters on the antenna outputs before any active device.

This is why I want to place a high Zin (940Ω) JFET preamp at the antenna terminals leaving the preamp output to match the feedline. But that brings me back to the original question-how do I decouple the DC ground from the antenna?
The easiest would probably to use a low ratio balun to convert the notional balanced output from the preamp to an unbalanced coax feed. You could run the power up the coax, and couple it off around the balun using blocking capacitors and HF chokes.

I just want to make sure that you are thinking of making a preamp with relatively low gain and an intercept point in the high 30dBm range, otherwise you will compromise the performance of all that nice gear you have downstream.
 

Thread Starter

wpri268

Joined Feb 18, 2012
30
Luckily I am not in an area of local high powered broadcast stations so I dont believe IMD is an issue. For that matter, I can turn my preamps on/off with no effect other than a change in signal level. Also my QTH is residential and rural so QRM is not terrible.

Not long ago though I did have a powerline noise problem fixed by National Grid after two years of complaining...go figure.

Most commercial preamps I've seen have Zin= 50Ω so they arent useful for directly feeding a high Z antenna. Since I am not up to the task of designing my own preamp I have done lots of research on them. I found a couple circuits designed for use with loop antennas which are well documented so I feel comfortable with them. Links to them are here: http://w5jgv.com/tree_amplifier/tree_amp.pdf http://w7iuv.com/preamp60/preamp_r60.pdf

Both have hig Z input and 50Ω output which is compatible with my requirements. . Both claim to have low NF and excellent OP3 as well as high gain. I would put a 1:1 line isolating balun at the output.

For reference, here are a couple of excerpts regarding ground independance from websites dealing with flag and pennant antennas:

The Flag or Pennant will not be perfectly balanced even if the feed method is, because one side of the antenna is closer to ground. Using an isolated winding allows the feedline connection to look like a perfect ground independent current source, so the voltages from each loop terminal to ground can seek any value they like to cause even current distribution. You may or may not have problems with transformers without very low primary to secondary capacitance, depending on how that capacitance is distributed. If the capacitance is in a ratio that happens to cause the correct voltage division on the output, you'll never have a problem even with a fair amount of coupling. The antenna won't respond to feedline currents, and the feedline won't cause additional unbalance to the antenna.

Important: The terminated diamond loop is a balanced antenna and I am feeding it with an unbalanced coaxial feed line. It is of utmost importance to decouple the coax shield wire from the antenna! The shield must act only as a shield and not as an antenna allowing pick up of noise along the feed-line otherwise the radiation pattern of the terminated loop will be compromised. This is called common-mode pick up and is very undesirable. To RF decouple the shield of the coax feedline I placed 12 ferrite beads, Mix 77, over the RG58u coax feed-line next to the remote pre-amplifier box. This is often referred to as a current balun.

An afterthought about baluns... does the ferrite core material need to be inductive at the frequency of interest or can it be resistive? I ask since I have a large number of Laird/Steward LF beads that are the perfect size for a sleeve balun. The LF material impedance peaks somewhere around 4MHz.
 

Tesla23

Joined May 10, 2009
542
Most commercial preamps I've seen have Zin= 50Ω so they arent useful for directly feeding a high Z antenna. Since I am not up to the task of designing my own preamp I have done lots of research on them. I found a couple circuits designed for use with loop antennas which are well documented so I feel comfortable with them. Links to them are here: http://w5jgv.com/tree_amplifier/tree_amp.pdf http://w7iuv.com/preamp60/preamp_r60.pdf

Both have hig Z input and 50Ω output which is compatible with my requirements. . Both claim to have low NF and excellent OP3 as well as high gain. I would put a 1:1 line isolating balun at the output.
The FET input one will have a high Zin, but the bipolar one is a resistive feedback design that the author claims is designed for 50Ω in/out. I guess I don't really understand your design objectives, I wouldn't put the 30dB preamp in front of a decent receiver and expect to be able to extract weak signals in the crowded broadcast band.

Luckily I am not in an area of local high powered broadcast stations so I dont believe IMD is an issue.
The key is understanding intermodulation. Let's say you have a reasonable receiver with an input 3rd order intercept of 15dBm. If you put 30dB of gain in front of it, then the effective intercept will be no better than -15dBm. I don't know what sort of sensitivity you are aiming for, but with a -15dBm 3IP then two input signals at -46dBm will give you 3rd order intermods at -108dBm. i.e. Two signals at about 1mV each will stop you hearing a 1uV signal. 1mV is not a strong local signal, it's only a bit stronger than the rural service level for AM broadcast.

The bipolar design looks like a more serious HF preamp, the transistor is running at 100mA or so, but I'm not sure it will meet your impedance objectives.


An afterthought about baluns... does the ferrite core material need to be inductive at the frequency of interest or can it be resistive? I ask since I have a large number of Laird/Steward LF beads that are the perfect size for a sleeve balun. The LF material impedance peaks somewhere around 4MHz.
Technically, all that is needed is a high impedance, it doesn't really matter if it is resistive or reactive. Ideally the impedance to ground from each of the two arms of the balun should be the same. In my experience I've used material that was inductive, but a bit of loss is no problem.
 

KL7AJ

Joined Nov 4, 2008
2,229
I am a BCB DXer phasing a pair of terminated, untuned pennant antennas. The characteristic impedance of ea antenna is ≈ 940Ωs. 75Ω coax is coupled to ea antenna via 12.5:1 isolation xfmr (≈ 88:7 turns) built with stacked (Amidon) FT240-61 cores. Ea coax then goes to a preamp powered from the feedline. Lately I have been wondering how much signal loss there is across those transformers because of the number of turns involved. I understand that a higer μ core would reduce the turns but they all have higher core losses. Higher μ= higher loss!
How can I eliminate the transformers and feed the antenna directly I asked myself? From what I have read single ended and differential jfet amps are often used with loop antenns. Whereas the jfet has an extremely high Zin the amps can be easily set to match the antenna characteristic impedance. The amp then can be isolated from the feedline via a simple 1:1 iso xfmr or a choke balun. The dilemma I face is how to supply DC power to the amp while still maintaining ground independance?
Thoughts?
You can tap your feedline across a couple of turns of the total loop. This acts as an autotransformer. However, the presence of the transmission line with this method can "warp" the pattern a little. I always prefer the transformer method. :)

Eric
 

KL7AJ

Joined Nov 4, 2008
2,229
I should also mention that the higher losses associated with higher mu only exist because they saturate sooner....this is NOT an issue with weak signal reception! You can use the highest mu cores available with total abandon for this application.

Eric
 
Top