How to connect antenna base to powder-coated steel?

[Joerg]

Watch out for galvanic corrosion with stainless against pretty much any
other metal. Where you develop more than about 0.15V is a potential
corrosion problem:

http://l-36.com/corrosion.php

With stainless steel studs this client had rather postive experience so far.

[...]

 

[Joerg]

Leave it painted. Mount the antenna INSULATED, and ground the antenna
with a wire, and ground the plate as well. To a ground rod.

A pigtail connection wire is not such a great thing at several GHz. Too
much inductance. It would have to be a metal strip and since nobody
makes them it'll be custom :-(

 

[Joerg]

Wouldn't you be using a COTS NEMA box? Maybe type 6?

No, a big custom steel box with a steel lid on top.

You could simply skip the feedthrough and use a marine grade cable feed?
Then the metal is out of the equation.

Ok, but we do not want the metal to be out of the equation. We'd like a
low-inductance connection to this steel lid.

There are also COTS waterproof cable feeds that I think seal better, but
they aren't stainless.

I've examined a lot of NEMA setups, and they tend to pigtail the coax
through a connector to avoid this issue. I've seen the feedthrough too,
but less often. Think of it this way. That connector is going to need
some weather protection on it. Goop, RTV whatever. The pigtail removes
one point of contamination.

Pigtail isn't a great ground at GHz frequencies. Feedthrough without
connection will work but results in the coax shield picking up noise
from the electronics inside.

 

[tm]

A pigtail connection wire is not such a great thing at several GHz. Too
much inductance. It would have to be a metal strip and since nobody
makes them it'll be custom :-(

There really should not be any RF on the outer surface of the connector if
everything
is done right. You just need a good DC and low HF path for static and
lightning or
other external events.

I would mainly think you want to keep out other RF and EMI from entering the
enclosure.
A large washer should provide capacitance from the connector to the case.
maybe use
a ferrite core around the cable prior to entry to the case.

 

[Joerg]

There really should not be any RF on the outer surface of the connector
if everything
is done right. You just need a good DC and low HF path for static and
lightning or
other external events.

Au contraire. They sometimes call those rubber antennas "dipoles" but
they aren't. They are ground planes, meaning there must be a ground
counterpart. On low frequency antennas that would be a set of radials.
On VHF/UHF/SHF stuff that is usually a metal surface. It has to be there.

If you don't provide a GHz-class ground at the base of the antenna the
system will seek and find other grounds, inside the enclosure. That is
normally not desired for two reasons: RF-wise it's largely uncontrolled
and it'll also pick up any inband noise that might be inside the box.
Range and RX noise immunity might go to pots.

I would mainly think you want to keep out other RF and EMI from entering
the enclosure.
A large washer should provide capacitance from the connector to the
case. maybe use
a ferrite core around the cable prior to entry to the case.

Well, there is a limit on the size of the washer. Plus those tend to
collect all sorts of gunk over time.

 

[Joerg]

epoxy then? or bolt it in before powder coating?

Epoxy is a good idea. Although it does add cost because of the curing
time during which the panel occupies space in production. Bolt-in before
powder-coat, I don't think that'll fly with the guys.

they have some that a 75$ wonder what exotic material they are made of

There is a lot of rip-off going on in the world of connectors.

I'm a bit weary of cheap sma connectors, I have once got some that
were
a bit too cheap, some of the center connector weren't centered and the
isolator
it was definitely not Teflon because you had to be fast with the
soldering iron
or it would melt

In our case none would be soldered, it's all just screwed together.

 

[tm]

Au contraire. They sometimes call those rubber antennas "dipoles" but
they aren't. They are ground planes, meaning there must be a ground
counterpart. On low frequency antennas that would be a set of radials.
On VHF/UHF/SHF stuff that is usually a metal surface. It has to be there.

If you don't provide a GHz-class ground at the base of the antenna the
system will seek and find other grounds, inside the enclosure. That is
normally not desired for two reasons: RF-wise it's largely uncontrolled
and it'll also pick up any inband noise that might be inside the box.
Range and RX noise immunity might go to pots.



Well, there is a limit on the size of the washer. Plus those tend to
collect all sorts of gunk over time.

--

Well, you never really completely defined the problem. Maybe you just can't
without a NDA.

Without all the information, we are just shooting ducks in the blind. Just
for the fun of it.

 

[Adrian Jansen]

And that's just the problem. The lock-washer DC connection would be good
enough because it is short. But it'll be eaten away by corrosion. Just
the air is enough to do that out there. Imagine a climate where every
day massive amounts of humidity condensate onto everything.

We have a similar climate here ( coastal Queensland Australia ).
Dissimilar metals making contact are always a problem.
Conformal coatings, or silicone rubbers, are about the only defense
which work reasonably well.

I tried the silver loaded paint, since the version I have is in a very
tough, resistant carrier, and hopefully the silver maintains the contact
even when the paint gets down into the contact surfaces. So it acts
both as a seal and as a conductor. The trick is to coat the surfaces
with the paint, then screw them together while the paint is still
liquid. When it dries, it should form a conductive layer, and exclude
the salt. Whether it works long term remains to be seen.

 

[Joerg]

Well, you never really completely defined the problem. Maybe you just
can't without a NDA.

Without all the information, we are just shooting ducks in the blind.
Just for the fun of it.

It's not really any secret. 2.45GHz has to be radiated off the top of a
large metal lid, connects to a transceiver inside the box. The bulkhead
connector getting the SMA from one side to the other should make
contact. Realistically we have about a 2" radius around it available
before other stuff gets in the way. Box is plain steel, powder-coated in
light gray color. That's about it.

 

[Joerg]

We have a similar climate here ( coastal Queensland Australia ).
Dissimilar metals making contact are always a problem.
Conformal coatings, or silicone rubbers, are about the only defense
which work reasonably well.

I tried the silver loaded paint, since the version I have is in a very
tough, resistant carrier, and hopefully the silver maintains the contact
even when the paint gets down into the contact surfaces. So it acts
both as a seal and as a conductor. The trick is to coat the surfaces
with the paint, then screw them together while the paint is still
liquid. When it dries, it should form a conductive layer, and exclude
the salt. Whether it works long term remains to be seen.

Wish we could do tricks like that. But in our case they'll have a bunch
of sheet metal made at a time and it all sits in the stock room until
needed. Only then is the hardware mounted into it.

 

[tm]

It's not really any secret. 2.45GHz has to be radiated off the top of a
large metal lid, connects to a transceiver inside the box. The bulkhead
connector getting the SMA from one side to the other should make
contact. Realistically we have about a 2" radius around it available
before other stuff gets in the way. Box is plain steel, powder-coated in
light gray color. That's about it.

--

Can you make a 1/4 wave choke joint at 2.45 GHz? Inside or outside the
housing? The electronics inside, if exposed to the corrosive atmosphere,
would
be more of a concern, one would think. Are you using a RpSMA to connect the
antenna? Might want to check the rules for a commercial product in the ISM
band.

Good luck, hope you figure it out.


tm

 

[Joerg]

Can you make a 1/4 wave choke joint at 2.45 GHz? Inside or outside the
housing?

Outside, no. Inside, yes. What I am trying to avoid is any RF pickup
from the inside. That always costs range.

... The electronics inside, if exposed to the corrosive atmosphere,
would

The electronics are mostly coated, I am not so worried about that. We've
got years of experience with that.

be more of a concern, one would think. Are you using a RpSMA to connect the
antenna? Might want to check the rules for a commercial product in the
ISM band.

Regular SMA. We are legally ok if the antenna is not above a certain
gain limit, according to the manufacturer. For the US, that is. The end
customer is not allowed to use any other antennas and would typically
never touch that area anyhow. Most won't even know what an antenna is
because they've all got cable or satellite

This whole reverse connector biz is a joke anyhow, you can buy adapters
almost anywhere.

Good luck, hope you figure it out.

If all else fails well make a custom bracket to bridge over to a
stainless stud with the least amount of inductance.

 

[My Name Is Tzu How Do You Do]

A pigtail connection wire is not such a great thing at several GHz. Too
much inductance. It would have to be a metal strip and since nobody
makes them it'll be custom :-(

If the frequency is that high, it does not need to be tied to the plate
AT ALL. The base of the antenna is 'ground plane' enough for the job. If
the transmitter chassis is grounded (and it damn well should be), then
the antenna already is as well.

 

[My Name Is Tzu How Do You Do]

Epoxy is a good idea. Although it does add cost because of the curing
time during which the panel occupies space in production.

4 hours at 80 C for silver filled, conductive epoxy used to attach chip
dies to the heat sink interface all over the world in nearly every chip
made. If you do not know how to inject such a process into a
manufacturing floor, you certainly should not be in charge of producing
anything. "Sapce in production..." You're a goddamned crybaby.

Bolt-in before
powder-coat, I don't think that'll fly with the guys.

Then MASK the powder coat area where the antenna goes. D'OH!

 

[My Name Is Tzu How Do You Do]

We have a similar climate here ( coastal Queensland Australia ).
Dissimilar metals making contact are always a problem.

Ever heard of General Electric anti-oxidant paste?

Dissimilar metals in intimate contact are ONLY a problem for those who
do not spend time resolving them.

 

[My Name Is Tzu How Do You Do]

I tried the silver loaded paint, since the version I have is in a very
tough, resistant carrier, and hopefully the silver maintains the contact
even when the paint gets down into the contact surfaces. So it acts
both as a seal and as a conductor. The trick is to coat the surfaces
with the paint, then screw them together while the paint is still
liquid. When it dries, it should form a conductive layer, and exclude
the salt. Whether it works long term remains to be seen.

Silver filled epoxy, followed by conformal coating over the exposed
epoxy fillet. Job done. Problem solved.

 

[Joerg]

If the frequency is that high, it does not need to be tied to the plate
AT ALL. The base of the antenna is 'ground plane' enough for the job. If
the transmitter chassis is grounded (and it damn well should be), then
the antenna already is as well.

It's 2.45GHz and powder coat is quite thick.

 

[Joerg]

4 hours at 80 C for silver filled, conductive epoxy used to attach chip
dies to the heat sink interface all over the world in nearly every chip
made. If you do not know how to inject such a process into a
manufacturing floor, you certainly should not be in charge of producing
anything. "Sapce in production..." You're a goddamned crybaby.

Four hours at 80C is "nothing"? You've got to be kidding. I ran a place
with production for several years and turned that division around to
profitability. So I know a thing or two about production, and 4h at 80C
does not fly with large sheet metal items.

Then MASK the powder coat area where the antenna goes. D'OH!

That leaves the bare unplated steel. Meaning we'd have to pot that up
later and that is a highly undesirable step in system assembly. Then
we'll rather go the stainless custom bracket route. Less $$$.

 

[My Name Is Tzu How Do You Do]

Four hours at 80C is "nothing"?

Yes, dumbfuck. If it is needed to perform the task, then it has to be
there. You don't take an idiot like you, going around cutting corners and
call it "streamlining". You're a total retard for that mindset alone.

You've got to be kidding. I ran a place
with production for several years

You must have failed at it, since you are not still doing it.

and turned that division around to
profitability.

Sounds like a deeper, stupid Mfgr with stupid processes in place to
start with. Boilerplate stupidity begets boilerplate stupidity.

So I know a thing or two about production,

Not enough to know that you do what must be done to get to the finish
line.

and 4h at 80C
does not fly with large sheet metal items.

Are they going to cry, idiot?

You're an idiot. The US Military uses the very same epoxy on 100 Lb
radio transceivers used in EVERY theater by EVERY ally, and they go
through a 4 hour cure on the epoxy, and another 4 hour cure on the
conformal coating. Those are completed units, fully assembled. They
also go through both cycles whenever they come back in through the depot
repair center, and get this element serviced.

 

[My Name Is Tzu How Do You Do]

That leaves the bare unplated steel. Meaning we'd have to pot that up
later and that is a highly undesirable step in system assembly. Then
we'll rather go the stainless custom bracket route. Less $$$.

You should. You should press a depression in the top of the box for a
stainless plate add-in.

That way, emi gasketing mates the stainless to steel box, and a direct
connection integrates the base plate and antenna. A few screws later,
and you have boxes which can be painted beforehand, and antennas already
mounted on plates that bolt right onto the box. Case closed.