El 01-09-12 18:13, KBa escribió:
On 31.8.2012 22:04, Wimpie wrote:
El 31-08-12 20:07, Dave Platt escribió:
In ikabel.net,
wrote:

Problem exists how to get reliable insulation to split driven
element
and support rod which is exposed to RF voltage ?
Fiberglass is told to be sensitive creating conductive coal brigdes
which creates true resistive short at driving point ? (at least one
fiberglass tubing is done so... gap was 2 inches )

Assuming power of about 1 kW and a HW dipole (say 60 Ohms), the
voltage will be about 350Vp. I can hardly imagine that a plastic gap
of 0.05m will degrade to failure because of tracking.

On the other hand, if the original poster is attempting to drive this
antenna off-resonance, with open-wire feedline and a wide-
impedance-range transmatch / "antenna tuner", then it's entirely
possible that the feedpoint will sometimes be "looking into" a much
higher impedance, and that a matched drive will result in very high
voltages at the feedpoint. Even a coax-feed antenna might have this
problem, I suppose, if the split element is being driven by something
like a delta match which has a relatively high impedance-
transformation ratio.

From what I see in a short Google-search, fiberglass *can* be
hygroscopic, depending on what resin was used to bind the fiberglass.
A fiberglass rod which was made with a somewhat-hygroscopic resin
(e.g. polyamide) might tend to behave badly in the face of high RF
voltages, whereas a rod made with a non-hygroscopic resin might be
fine unless dirty or wet.


Hello Dave,

Thanks for de addition. During the simple voltage calculation my mind
said: "what if VSWR is really bad".

With 1 kVp over 0.05m I would not expect problems. You are right,
Polyamide (Nylon) is bad for RF and as far as I know, isn't fully UV
resistant without additives.

Once I had a nylon bolt in the field of a tuning capacitor running
close
to breakdown over about 3mmm. The bolt produced lots of smoke within
some seconds and broke down. Changing from PA to PE solved the problem.

Assuming 50% glass fill factor and effective loss factor of 0.2 (so
Q of
the glass/PA composite material is just 5), I would expect a
dissipation
in the range of 4 Watts at 1 kVp and 10 MHz in a 1.5 Inch thick massive
rod. Maybe Kba can provide us some additional info on the working
voltage across the gap (and frequency).

When using epoxy or polyester fiberglass, loss will be significantly
less.


Thanks for all comments, ... some new approaches have arisen... to
eliminate possible risks and troubles before they exist...

This problem is dual, both electrical and mechanical as insulator in the
element gap is under quite high mechanic stress...
This is future OWA antenna split driven element... will be used at HF...

This previous fiberglass problem wasn't instant, it developed slowly
after several years of succesful use ( and weather effects, moisture
etc. ) so probably the reason was only the moisture, dust and dirt
layers in the inner surface of the insulator.
Tracking (forming of conductive (carbon) path) is mostly a slow
process that may take years to develop to failure, so this sound
strange (if creepage is the root cause).


Insulator was round fiberglass tube of 65mm diameter and around one
meter long, used as center of 7MHz antenna feed element ( 60mm diam.),
the outer surface of the tube was weather ( and UV ) protected and
water proof, but the 5 cm gap inside the tube was not entirely though
element ends were not in open air... some moisture do condensate
always inside the element... I guess...

This changes the story when moisture can enter the inner side of the
tube and may stay there. I assumed you used a massive rod for
mechanical strength. At the inner side you don't have the cleaning
effect due to rain, wind, precipitation, etc. Could there be salt
built-up (due to seeping water that evaporates slowly)?

You mentioned it is 1 m long, that is really a large creepage
distance! Assuming about 5mm wall thickness and fully conducting
inner side of the tube, radial field strength (350V) will be in the
50V/mm range, well below permissible levels to avoid insulation
failure at MHz frequencies in FR4 material.

It would be nice to inspect such a failed tube to figure out how the
tracking took place, because I have now idea.




Measured impedance was quite low, 25 ohms resistive near resonance, so
probably voltages have been also low at one kw power level.

I cannot tell the exact fiberglass grade that was used, probably it
was green FR4 or other strong quality as mechanical stress was high in
the insulator and long and heavy 7MHz element halves loaded the
insulator center.
This fiberglass insulator tube finally got to a ohmic short and didn't
recover.

Later this insulator was replaced with another type and problems
disappeared.

. . .

Now one other future split driven OWA antenna has been under plan ...
but there are the same mechanical stress problems as element center gap
should ( ? ) have center support rod but it exposes also to all
weather effects... in this case element diameter will be 25 mm...
So insulation in the gap should be good but material should be also
something capable handling bending moment of the element halves...

Now the latest plan is try to move the mechanic load away from the
center support rod to element support plate which could be much longer
in length... then the element center stress is expected to be lower
across the gap...

Then also the element clamps have more separation as this support plate
is also again FR4...
Secondly the element halves will be insulated from the clamps so there
will be less risk of this FR4 plate becoming conductive...
Insulation between element and clamps could be thin straps of teflon
sheet...

Best might be if the fiberglass center support rod can taken totally
out or be of round teflon bar... good insulator, low friction,
probably air particles etc. don't stick on it easily...

Teflon feels tough but elastic, so in the stress point it's not much
use, but probably functions as a vibration damper in the gap... assuming
the element support plate does not transfer element end bending effects
due to wind load directly to element center gap...

Anyway it's going to be interesting to see how these precautions
will actually work... in both at mechanical and electrical
reliability... hopefully in both.

Besides mechanical cracks that may provoke solid insulation breakdown,
carbon trace forming (tracking) over wheather exposed plastic surfaces
is the most likely problem, so I think you should try to maximize the
creepage path. As E-field strength in the insulation materials will be
low, you don't need low loss materials.

Thanks for sharing the info on the insulator failure!


--
Wim
PA3DJS
www.tetech.nl
Please remove abc first in case of PM