Radials

[K9ACT]

I just got my spool of wire to put out radials for the top loaded vertical.

However, I am not sure how long they should be.

1/4 wave on 160 is 40 meters but the flat top is an 80 M dipole, so quarter waves would only be 20 meters.

So, is there any point in making the radials any longer than the top hat?

Next question is regarding the wire itself.

It seems intuitive that the ideal radial would be a disk of copper with the radius of 1/4 wave.

However, most people agree that radials need not be bare wire and that insulation or not, makes no difference.

It seems like bare wire in contact with wet earth would more closely approach the "ideal" copper disk.

So, is there something wrong with the copper disk analogy?

One more question:

I borrowed an ant analyzer from W9AD but it is not clear how to use this on this antenna.

What changes that I can measure as I add radials?

I have a variable cap between the antenna and transmitter.  When this is correctly set, the SWR is 1:1 so there is nothing improve on.

Should I remove the cap when testing ant with analyzer?  If so, what parameter will change with radials?

js

[Steve - WB3HUZ]

Bare or insulated makes little difference. The current they carry is induced.

The longer the radials the better.

Even with the cap in place (and not changing it), as you add radials, you should see the resistive portion of the feedpoint impedance drop. When it stops dropping or the change is very small, you can stop adding radials, assuming your are looking for maximum efficiency.

[K1JJ]

The four-square guys on 75M tell me once they get into the low 30's ohms  range (RF resistive) for a quarterwave vertical, they are close to saturation with the radial laying.  32 ohms is a good number, I understand.   This number will start out much higher with the first few radials.   You get this number by measuring with your analyzer DIRECTLY AT THE BASE of the vertical, between the radial connection and the top section.  (No coax feedline involved or the readings will be off  - unless you make up an exact 1/2 wave feedline inc velocity factor)

You will also find out how good your relative ground is by the number of radials needed to hit the low 30's ohms or saturation.

T

[K9ACT]

Not sure I follow.  Here is what I find:

Cap in at min swr.....

SWR......... 1:1
R..............  43
X.............    2

NO CAP, FEEDLINE* AND GROUND CONNECTED DIRECTLY TO METER

SWR = 12.4
R     =  25
X     =  140

AFTER ADDING TWO RADIALS, THERE IS NO CHANGE

*THE "FEEDLINE" IN THIS CASE IS THE TWINLEAD TIED TOGETHER AT THE METER END.

JS

[Steve - WB3HUZ]

Do not use a feedline. Measure right at the base of the antenna.

Not sure I follow.  Here is what I find:

Cap in at min swr.....

SWR......... 1:1
R..............  43
X.............    2

NO CAP, FEEDLINE* AND GROUND CONNECTED DIRECTLY TO METER

SWR = 12.4
R     =  25
X     =  140

AFTER ADDING TWO RADIALS, THERE IS NO CHANGE

*THE "FEEDLINE" IN THIS CASE IS THE TWINLEAD TIED TOGETHER AT THE METER END.

JS

[K1JJ]

*THE "FEEDLINE" IN THIS CASE IS THE TWINLEAD TIED TOGETHER AT THE METER END.

JS

75 ohm? twin lead is balanced feedline. You would be better off using a coaxial cable the whole run that can lay on the ground and go through metal objects, etc. The twinlead must be isolated from objects for balance it's whole length.  But you probably know this already.

Huz is correct - measure directly at the base of the vert. If there is ANY swr, the feedline readings will be all over the place. Once you get a 1:1 at the base, ANY length feedline will read 1:1 at its other end.  For a perfect match, you will need to step up from 30 ohms to 50 ohms to get the base of the vertical to match the 50 ohm cable. (or whatever feedline type you use)    A 2:1 step up unun would work. There are other ways too, using a 1/12 wave coaxial unun that uses two 75 ohms coaxes in parallel and a 50 ohm coax to match 35 ohms to 50 ohms.

Here is some info on it:
http://amfone.net/Amforum/index.php?topic=5199.msg39428#msg39428

Or, you can eat the 30 ohms to 50 ohms swr and not worry about it using a tuner.


(As I sit here and breath a sigh of reflief after spending a scary afternoon at 190' getting ready for the big erection.... )


T

[W2DU]

It seems as if there's always some confusion concerning the appropriate length for radials. For on-ground or in-ground radials the length problem was solved by my colleagues at RCA in 1936, by Dr. George Brown, Bob Lewis (W2EBS), and Jess Epstein, known for decades as BLE. Their findings, after many measured experiments, were published in the Proceedings of the IRE, Volume 25, Number 6, June 1937,

To make a long story short they found that when many radials are used the optimum length is 0.4 wavelengths. The explanation is that beyond that length the returning displacement currents from the vertical are now so weak that longer lengths of radials would result in insignificant improvement to the point of of diminishing returns.

In addition, in-ground radials are not resonant.

Walt, W2DU

[flintstone mop]

I have learnt that 30 ohms at the base is the characteristic impedance of a vertical. So, Tom is saying that a means to match the 30 ohms to the 50 ohm line is important. You might pull that off with a variable coil across the trasnmission line at the antenna end. 75M is pretty easy to get a vertical to resonance without a lot of L and C.
What is the height of your vertical??
Phred

[W2DU]

Worried about feeding a 30-ohm antenna impedance with a 50-ohm coax? Don't be. The result is 1.66:1 SWR, for a power reflection coefficient of 0.0625, 93.75 percent radiated, loss = 0.28 dB. A network to match these two impedances might have an even greater loss--forget it!

Walt, W2DU

[K1JJ]

Worried about feeding a 30-ohm antenna impedance with a 50-ohm coax? Don't be. The result is 1.66:1 SWR, for a power reflection coefficient of 0.0625, 93.75 percent radiated, loss = 0.28 dB. A network to match these two impedances might have an even greater loss--forget it!

Walt, W2DU

Walt,

What do you estimate the matching transformation loss to be using a 1/12 wave coaxial unun for 7.150 mhz.... ie, using two 75 ohm coaxes in parallel (7' long) in series with a 50 ohm coax (7' long) resulting in a 35 ohm to 50 ohm transformation match?

Tnx.

Tom, K1JJ

[K9ACT]

Time to start over.

The feedline IS the vertical component.

The antenna is a "Top Loaded Marconi" in accordance with p.604 of 1973 issue of the ARRL Handbook.  Section titled 160 M antennas, probably in many other editions.

Specifically, it is a ladder line fed, 80 meter dipole, up about 50 ft.

Analogous to a mobile antenna with a loading coil at the bottom but with a cap at the top instead.

This can be tuned with either a coil or cap between the transmitter and the feed line end in the shack.  In my case, about 500 pF produces an SWR of 1:1 as reported in a previous msg.

My interest is in seeing something happen on the ant analyzer when I lay out radials.  So far, with two down, I see no change in SWR, R or X.

[W2DU]

Hello Tom,

If we consider the transmission-line sections as lossless, the mismatch is 1.4286:1, for a power reflection coefficient of 0.0311. The absorbed power is 1/0.0311 = 0.9689 percent, and the mismatch loss is 0.1374 dB.

It should be understood that mismatch loss in non-dissipative, which means only that if a rig is tuned for delivering all the available power into a 50-ohm load at a given drive level, the xmtr simply delivers less power equal to the mismatch loss in dB when re-connected to the mismatched load.

Walt, W2DU

[KL7OF]

A flat top dipole up at 50 ft fed with ladder line?  I'm confused....How is the feedline a  vertical antenna?

[W2DU]

Tom, I just posted another reply, but it got yanked when I tried to post it.

Anyhoo, I goofed in my earlier response to you. I stated that the absorbed power is 1/the power reflection coefficient, rho squared. The correct statement is that 1 - the power reflection coefficient equals the power absorbed. However, the percentage absorbed and the mismatch in dB are correct as stated, I just typed in the wrong algorithm, but performed the calculation correctly.

Walt, W2DU

[K1JJ]

Hello Tom,

If we consider the transmission-line sections as lossless, the mismatch is 1.4286:1, for a power reflection coefficient of 0.0311. The absorbed power is 1/0.0311 = 0.9689 percent, and the mismatch loss is 0.1374 dB.

It should be understood that mismatch loss in non-dissipative, which means only that if a rig is tuned for delivering all the available power into a 50-ohm load at a given drive level, the xmtr simply delivers less power equal to the mismatch loss in dB when re-connected to the mismatched load.

Walt, W2DU

Thanks for the info, Walt.

So it looks like the 1/12 wave ununs are reasonable to use. I have them in most all my Yagis, 10-75M and feel better now... :-)


Walt, could you expand on the following some more?:

"It should be understood that mismatch loss in non-dissipative, which means only that if a rig is tuned for delivering all the available power into a 50-ohm load at a given drive level, the xmtr simply delivers less power equal to the mismatch loss in dB when re-connected to the mismatched load."

"non-dissipative"? - Is this just a source and load and no feedline calculation? Is the feedline important?  When I see the graphs of typical feedline losses for 100' and they show the db loss as the swr increases. The lowest loss being when 1:1 is achieved.   What am I misunderstanding in your statement?

Then, I understand that when there is a high swr, the power eventually gets radiated at the load. That is contrary to high swr=high feedline losses.
Please help me clear this up in my head....



BTW, I was thinking of reasons for having a matched antenna. In most cases, it doesn't really matter if there is a 1.6:1 swr mismatch as far as signal strength. As you noted, the loss is only .28db. (Is this .28 db per 100' of coax RG-213 or equiv?)

Anyway, I've come up with a few reasons to try to obtain a good match from the feedline to the antenna:

1) When using a super long feedline of hundreds of feet, the feedline losses can add up.  I have about 450' of 75 ohm aluminum hardline out to the farthest tower and make sure the line is matched well to the load.

2) When using more than one antenna on a given band, it's good to have all antennas at the same impedance in the shack - so when switching, the amplifier doesn't have to be retuned.   I make mine either ALL  75 ohms j0   or 50 ohms J0 for any particular band.

3) When running extreme QRO -  a mismatch will have higher current points along the feedline than when matched 1:1.  When using smaller type coax, these can produce "hot spots"  or even cause feedline damage when rotating the cable on a moving antenna.  A 100w rig has nothing to worry about here.

5) When phasing more than one antenna for stacking, etc, we need a close match (j0 especially) to insure good phase control and power division to produce the pattern we desire in the model.

4) I couldn't sleep at night knowing one of my antennas had anything higher than a 1.5:1 swr....   

Again, most of this does not apply to a 100w station using a single antenna  on a given band.


73,

Tom, K1JJ

[Steve - WB3HUZ]

If I follow correctly, it's a T antenna (more or less). The two-wire balanced line is shorted together at the transmitter end and the  whole thing is fed against ground.

Jack, where are you placing the radials - off the ground lug of the  tuner in the shack, or does the feedline come straight down to the ground where it gets ground and are you placing the radials there?

A flat top dipole up at 50 ft fed with ladder line?  I'm confused....How is the feedline a  vertical antenna?

[W2DU]

Hi Tom,

Thanks for the insightful response. However, it's getting late, so I'll study your response carefully tomorrow and come up with an answer.

CUL, Walt

[KL7OF]

If I follow correctly, it's a T antenna (more or less). The two-wire balanced line is shorted together at the transmitter end and the  whole thing is fed against ground.

Jack, where are you placing the radials - off the ground lug of the  tuner in the shack, or does the feedline come straight down to the ground where it gets ground and are you placing the radials there?

A flat top dipole up at 50 ft fed with ladder line?  I'm confused....How is the feedline a  vertical antenna?

I am curious to see what the modeling shows for the impedance of this antenna with and without radials and the real station experience as well.....Very interesting to me...good luck.

[K9ACT]

If I follow correctly, it's a T antenna (more or less). The two-wire balanced line is shorted together at the transmitter end and the  whole thing is fed against ground.

Jack, where are you placing the radials - off the ground lug of the  tuner in the shack, or does the feedline come straight down to the ground where it gets ground and are you placing the radials there?

The station ground is a buss behind the gear that then crosses the room (20ft) to an outside wall and through the wall to an 8ft ground rod inches from the wall.  The radials radiate from this rod.  As it is on the West side of the building, the East radials will have to go around or over the building which I have not done yet.

The ladder line follows the same path but is overhead instead of on the floor.  It passes through the wall and then slopes up to the center of the dipole.  Because of the location of the trees supporting the ant, the center is not directly over the ground rod.  There is about 50 from the transmitter level to the flat top.  This constitutes the vertical component.

It seems to radiate quite well for the few nights I have been using it without radials.  Too noisy tonight to even try it with the new radials.

js

[k4kyv]

I don't worry about SWR in coax if it is less than about 2.5:1 or 3:1.  On 160 and 40, I set my remote tuner to the middle of the band, and work both ends without  retuning.  I can  detect no difference between doing that and tuning for 1:1 for each frequency.  Since I have to  go down to the tower to retune, I usually just adjust the matching network at the transmitter end to load up into the coax.  On 75/80 I have to go down and readjust the tuner if I QSY more than about 75 kHz.

I am entertaining the idea of replacing my coax with balanced open wire line and running that down to the tower as a flat nonresonant line to feed the tuners, and then as a resonant line from the tuner, up the tower to the antenna.

Many hams have fallen  for the fallacy of adding radials until they get a 1:1 SWR and then stopping because adding more radials makes the SWR start to creep back up.  In that case, you will be better off operating with the additional radials and high SWR, than at 1:1 SWR and fewer radials.  If adding radials brings up the SWR, keep on adding them until additional radials cause no additional increase in SWR, and then re-adjust the matching network (or put in one) at the base of the antenna to bring the SWR back to where you want it.

[K1JJ]

Many hams have fallen  for the fallacy of adding radials until they get a 1:1 SWR and then stopping because adding more radials makes the SWR start to creep back up.  In that case, you will be better off operating with the additional radials and high SWR, than at 1:1 SWR and fewer radials.  If adding radials brings up the SWR, keep on adding them until additional radials cause no additional increase in SWR, and then re-adjust the matching network (or put in one) at the base of the antenna to bring the SWR back to where you want it.

Well said.  It's one of those "opposite to logic" thangs we come across in most areas of life....

First, we must assume someone is measuring at the feedpoint OR measuring at the end of  a 1/2 wave feedline.  If not, then the readings are false. If measured correctly, then what is happening is as the radials are added, the real base impedance is slowly sliding from above 50 ohms, to 50 ohms (1:1 reading) and then eventually below 50 ohms. The swr shows high, then perfect, then high again.  More radials are better - then adjust the match to 50 ohms 1:1 as Don says.

T

[W2DU]

Tom, I appreciate your concern about my statement that mismatch loss is non-dissipative. First, consider the mismatch loss with lossless line. With lossless line there is obviously zero loss due to attenuation. Thus, any power reflected at a mismatched load terminating the line is totally conserved, and reaches the source with the same magnitude as that reflected at the load. The result--no dissipation. However, when the reflected power reaches the source, the source then reduces its available power by exactly the same amount as the reflected power that reaches it. The reduced source power and the reflected power then add with the reflected power reversing direction. The result is that the sum of the powers exactly equals the original continues to receive the same amount of power as before, but it then also reflects the same amount as before, and the process continues with no power dissipated.

Now we come to the use of real transmission lines that have attenuation. Consequently, power is lost due to the attenuation, which is at its minimum when the load impedance is matched to the characteristic impedance Zo of the line--no reflections. But then we know that when there are reflections on the line the total attenuation of the power is the sum of the matched loss due to the inherent attenuation of the line, plus the additional loss proportional to the SWR.

We can now deduce that mismatch loss by itself is non-dissipative, and that dissipation results only due to the inherent attenuation in the transmission line.

Tom, I'm attaching some data from Appendices that appear in Reflections which should help in understanding the concepts discussed here.

Walt,W2DU

[W2DU]

Tom, in my previous post I attached three attachments, only one came through, Appendix 8. Appendices 6 and 7 didn't come through, so I'll send them in separate posts.

Walt

[W2DU]

Tom, here's Appendix 6

Walt

[K1JJ]

Thanks for the extra effort, Walt!

I read through the three and will do it again. It goes slowly with the math.

Yes, when the transmission line is out of the picture, I can see how the system is lossless. Interesting on how the reflected wave reduces the source by the same amount and then adds back in when sent back to the load, over and over. 

Thanks again for the info, Walt.

73,
T