Call Search

New to Ham Radio?
My Profile

Friends Remembered
Survey Question

DX Cluster Spots

Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement

donate to eham
   Home   Help Search  
Pages: [1]   Go Down
Author Topic: Dipole on a flat metal roof.  (Read 5987 times)

Posts: 10

« on: May 14, 2009, 06:36:12 AM »

I have installed both a dipole and a G5RV on a flat metal roof.  Due to height limitations I can only get the center about 12 ft. above the roof.  The internal tuner on my TS-480 will not tune it.  I have tried an external tuner as well and it still will not tune on any band.  Does anyone have any ideas as to what I can try.  Also is there a short vertical out there that might work better.  I am needing to work 80-20 Meters.



Posts: 41

« Reply #1 on: May 14, 2009, 07:09:53 AM »

assuming 102ft wire fed with 450ohm (VF.91) your ladder line which provide best match has to be cut as follows:

Freq--------------lengths of 450ohm ladder line in ft
3.8MHz-----------25.88-----143.75    ...
7.175MHz---------37.51-----99.93     ...
14.225MHz--------26.69-----58.18-----89.66   ...

I'll say use ~30 ft of 450 ladder line and converted to coax(any length) with 1:1 W2DU balun outside the shack. The idea is to used the closest common length for different operating bands.

Also read up on this:

Just note Cecil, use 135ft doublet on his web site but you can scale the idea to G5RV.

Good luck.




Posts: 14491

« Reply #2 on: May 14, 2009, 07:18:03 AM »

A dipole, which would be resonant on only one band, could be experimentally shortened in order to achieve resonance in the presence of the metal roof but it still won't work very well. Remember that the antenna is only 0.04 wavelength high on 80M. The metal roof represents a lot of capacitance shunting the antenna.

A vertical would likely work much better over the metal roof, although a 12-foot vertical on 80M won't do too well. One of the motorized mobile screwdriver antennas would probably be the best thing to try. If the metal roof pieces are all electrically common, you can tie the shield of the coax to it to act as the counterpoise.

Bob  AA4PB
Garrisonville, VA

Posts: 14491

« Reply #3 on: May 14, 2009, 07:49:18 AM »

I don't think the ladder line lengths will hold true when the 120-foot doublet is only 12-foot above a metal roof, especially on the lower bands. That's too close for any horizontal 80/40M antenna to be effective anyway.

Bob  AA4PB
Garrisonville, VA

Posts: 10

« Reply #4 on: May 14, 2009, 08:24:13 AM »

Thanks for the reply.  I am thinking about trying a HS-1800/Pro motorized antenna.  Their website mentions ground radials and a capacity hat.  Do you think both would help in my situation?



Posts: 7718

« Reply #5 on: May 14, 2009, 08:37:01 AM »

The input impedance of these antennas will be reduced by proximity to the metal roof. The external tuner should still tune it. What type of external tuner and feedline are you using? And what is the dipole length?

The ground mounted screwdriver antenna will work. I use such an antenna. However, your dipole should work as well or better on 40 and 80 meters.

Having said that, I use a 108" whip atop a Tarheel 200. Sometimes I place a 72" mobile mast section on the Tarheel and the 108" (mfj) whip on this. This gives 2 dB more signal on the low bands.

I work plenty of DX on 80 and 40 meter CW running 500 watts, so it meets my needs.

Posts: 14491

« Reply #6 on: May 14, 2009, 09:34:30 AM »

However, your dipole should work as well or better on 40 and 80 meters.
I would agree if the dipole were at a reasonable height above the metal roof. At 12-feet I don't think so.

The metal roof can probably be used in lieu of the radials if you make a solid connection between the roof and the antenna mount. The capacity hat would help but even better would be to replace the whip on the top of the antenna with a longer one. The total length of the antenna is probably around 8-feet because it is designed for mobile use. If you can increase the whip length by 4-feet you'll get a major improvement in operation on 80M and 40M and still stay at your 12-foot height limitation.

The purpose of the capacity hat is to provide capacitance to ground, thereby reducing the number of turns required in the loading coil, and therefore reduce the loading coil loss. Adding a longer whip instead also increases the capacity as well as increasing the radiation resistance. You get a double benefit that way. As a guess, I'd say that adding 4-feet to the whip will probably increase the overall efficiency by 4 times or more.

If your metal roof is made of small pieces that do not have a good electrical connection between them then you could add a number of wire radials laying directly on top of the metal roof.

Bob  AA4PB
Garrisonville, VA

Posts: 10

« Reply #7 on: May 14, 2009, 09:39:12 AM »

Great, thanks for the info.  One more thing, the roof also has a lightning protection grid on top of the metal sheeting.  Actually, first you have the metal sheeting, then an asphalt type of material, then the lightning grid.  Would I connect the coax shielding to the lightning grid to use as a counter poise?

Posts: 7718

« Reply #8 on: May 14, 2009, 10:20:44 AM »

mounting a dipole above a metal roof does not affect the radiation angle. It can lower the input impedanace. What the dipole radiation pattern depends on is the height above the ground in the far-field; this is where the reflection takes place from.

The reason I say the dipole will probably outperform the screwdriver on 80 and 40 meters is the low radiation efficiency of the screwdriver when ground mounted. Mine runs 10% on 80 meters. The dipole will run nearly 100%.

Mounting the screwdriver on a large enough ground plane can help considerably.

A top hat on a short vertical serves two purposes: To reduce the loading inductance and to make the antenna current distribution uniform. The latter will often yield more of an advantage than the former. Add a sufficiently large top hat to a 10' 3.5 MHz vertical and the radiation resistance is nearly quadrupled. The GND and loading coil losses are now reduced by a factor of nearly four. And the bandwidth is increased. Top loading is almost something for nothing.

Dave's rule for small antennas:


Pick any two

Posts: 17476

« Reply #9 on: May 14, 2009, 11:54:20 AM »

WX7G wrote:

>>mounting a dipole above a metal roof does not affect the radiation angle. It can lower the input impedance...

The effect on the radiation angle depends on the distance
from the antenna to the edge of the roof.  For low angles
where the antenna can "see" over the edge of the roof there
is no difference, but for higher angles where the reflection
is off the roof rather than the surrounding ground the
effective height of the antenna is measured relative to
the roof.

But in practice the difference is usually small unless
you are on a very large roof such as a warehouse.

It can and does lower the impedance, though, and that is
probably why you are having matching issues.  We had an
ARES station with a 75m dipole mounted 10' over a metal
roof and I had to add a 1uH shunt inductor across the
feedpoint to get a good match.  The antenna worked very
well once I did that.  I think the impedance of the
dipole otherwise was around 12 ohms, so a folded dipole
fed with coax to a 1 : 1 balun may also give a good
match.  The best approach in such a case is to measure
the minimum SWR and calculate the impedance at resonance
assuming it is below 50 ohms.  (So an SWR of 4 : 1 at
the antenna means a 12.5 ohm impedance.)  Then use a
shunt inductor (as described in the mobile chapter of
the ARRL Antenna Book and perhaps the Handbook) or other
method to match the impedance to coax cable.

The roof proximity effect will be different on different
bands because the height in wavelengths changes.  For
this reason multi-band horizontal antennas may be more
difficult to use reliably.  Fortunately the reactance
of a shunt inductor increases with frequency, so a single
shunt coil may work reasonably well with dipoles for
multiple bands, but you'll have to try it and see.

Certainly I would expect the G5RV to be a problem:  the
input impedance is already low on 80m and the low height
off the roof makes it more so.  Low impedances tend to
be more of a problem for tuners to match (especially to
match efficiently) than higher impedances at the same
SWR.  (And this all assumes that the twinlead feeder isn't
laying on the metal roof, which will throw all the bands off.)

So I think you will have better results with a dipole
than with a G5RV, but in either case you will have to
provide some additional impedance matching.  With the
dipole you can connect a coil of just a few turns of
wire across the feedpoint to raise the impedance.  Then
adjust the antenna length for the desired resonant
frequency and adjust the coil for best SWR at that

Good luck!

Posts: 14491

« Reply #10 on: May 14, 2009, 01:00:56 PM »

I question whether the dipole or G5RV will be 100% efficient on 80M when located so close to the metal roof. In addition, you have the difficulty of trying to cover all bands with it.

I also submit that making a significant increase in the length of the whip on the vertical will improve the efficiency much more than adding a top hat to the stock whip.

Bob  AA4PB
Garrisonville, VA

Posts: 17476

« Reply #11 on: May 14, 2009, 01:09:41 PM »

Whether a short vertical will be better than a dipole
depends on the desired angle of radiation (as well as how
close the dipole is to the edge of the roof.)

For local working on 80 and 40m the dipole has a better
radiation pattern.  For DX the vertical might do better
in some cases.  Short verticals are generally rather
inefficient on 80m, but may do well on 20m.  When the
dipole is properly matched at the feedpoint to give a low
SWR (as described in my previous post) the losses should
be low unless the feedline is too long.  If your building
is high enough and the dipole can "see" off the edge at
perhaps 15 to 20 degrees below the horizon I'd expect it
to work very well for DX on all bands.

So it all depends what distance you want to cover.

Posts: 14491

« Reply #12 on: May 14, 2009, 01:41:18 PM »

So having a large metal plate running close to and parallel to the dipole for its entire length will not affect its efficiency as long as you can match its impedance?

Bob  AA4PB
Garrisonville, VA

Posts: 17476

« Reply #13 on: May 14, 2009, 05:10:38 PM »

If the roof is well-bonded together and a good conductor,
the effect is like an antenna with a close-spaced parasitic
element or in front of a plane reflector.  There is a
lot of current flowing in the system, but not necessarily
a lot of loss.  There is a huge difference between a
metal roof and lossy earth for close-spaced antennas.  
You can model it in NEC if you enter the permitivity
and resistivity of galvanized steel in for the ground

Now, true, galvanized sheet metal is not as good of an RF
conductor as copper (which was used on some older roofs)
and there is no guarantee that the panels are well bonded
together, especially if they are under a layer of waterproofing
material.  But there is nothing about having two conductors
close together that intrinsically causes high losses,
and an antenna over such a roof most likely will be far
more efficient than one at a similar height over even
the best earth ground.

Losses go up as the antenna gets closer to the roof
because currents are higher for the same radiated power.  
I'd try to keep the antenna flat rather than sloping
the ends downwards.  There probably is still some
minimum height above the roof (depending on the roof
conditions) below which the losses get excessive.
One way to check this is to compare the input impedance
of the antenna over perfect ground with the measured
input impedance:  over normal ground the input impedance
of a dipole dips to around 35 ohms then goes back up to
around 50 ohms due to losses.  I don't have the chart
for dipole impedance vs. height over perfect ground, but
if you compare your measured impedance to that you can
determine the efficiency in exactly the same way we do
for ground radial systems on vertical antennas.

I don't remember the exact numbers in my case, but I
remember calculating 1uH for the shunt matching inductance
on 80m.  That information plus the 10' height above the
roof should be sufficient to calculate the actual
efficiency - which would be an interesting data point.


Posts: 17476

« Reply #14 on: May 14, 2009, 10:33:19 PM »

Ok, after a bit of experimenting with EZNEC, I'm guessing
that my 75m dipole placed 10' over a metal roof is about
50% efficient, maybe better.

I have to find my notes, but I remember that the coil I made
initially had 5 turns, and I ended up using 4 turns to match
the antenna.  Using EZNEC and "perfect ground", a 75m
dipole at 10' made using #18 wire would require 0.6uH
and have a loss of 3dB.  The same antenna made of #12
wire would use 0.5uH and have a loss of 2dB.  The losses
in this case are due to the resistance of the antenna
wire, so larger wire is good.  Raising the antenna also
improves the efficiency (as it does over poor ground.)
The feedpoint impedance of the antenna is around 8 ohms,
of which about half is radiation resistance and half is

Even if my estimates are off and the loss was 5dB, that
is about the same as an inverted vee with the top at
20' and the ends at 10' over average soil according to
EZNEC.  A lot of hams are happy with such an antenna
for NVIS coverage.

So I don't claim that the antenna is 100% efficient,
but rather that one must expect low antenna impedances
when an antenna is close to a metal roof, but with
proper matching you can make it work.  This particular
station is about as strong as most others in the county
when checking into the 75m nets.  But with an 8 ohm
feedpoint impedance it is no wonder that some tuners
have trouble matching it.
Pages: [1]   Go Up
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!