Newsgroups: sci.physics.plasma
From mrv@world.std.com Mon Jan 1 00:13:25 1996
From: Mike R Valente <mrv@world.std.com>
Organization: Abington, MA
Subject: RF and radio applications

Hello to everybody in this news-group.
My name is Mike Valente and I live in Abington, Ma., U.S A..

Recently, I was responding to a message here regarding
RF causing problems by getting into the equipment being used
to monitor an experiment, on the order of 500 milivolts or
so. When I contacted the moderator here, he asked me if I
would like to make my comments public in the conference and
I do think this is a good idea..many of you would benefit at
the same time.
First, let me tell you a few things about myself. I am
one of those people that is REALLY, REALLY into RF and
electronics in general; Now 47 years old, I had my FCC
commercial license when I was 15 and soon afterward was
working on transmitters at the first of many radio stations.
I am also a ham radio operator, Advanced class and have
held my call letters of WA1MNQ for more than 25 years now.
For most of my life, I have been working at making the
RF go where it is supposed to, as well as keeping it out of
where it is not supposed to be. Working in AM Broadcasting
sure saw to it I had plenty of practice in the latter area!
Telephones, intercoms, co-located studio equipment, you
name it, the transmitters would manage to get into it and I
had to go out and clean up the situation; In one case, a
Medical Doctor had set up shop across the street, within 450
feet (137.16 Meters) of the AM Towers for a station I was the
CE of for 6 years. This was in itself not unusual but the RF
problems certainly were. His was a neurologist and he had two
treatment rooms, each one having its own EMG machine..which
also became excellent radio receivers.
I worked with him and after we were done, the signal
strength, as measured at the input terminals of his equipment
was so low that it was the equivalent of moving his office 25
miles (40.234 Kilometers) away from the radio station.
Having worked with RF from SLF to Microwaves, it is my
thinking that for every problem, there is a solution; One
only needs to think creativly and draw on past solutions.
Now, let me move onto a subject near and dear to your
hearts..and mine..RF Generators..You see, I work for a
company that rebuilds them from top to bottom..at power
levels from the one kilowatt desk-tops to one megawatt.
I am one of a crew of 12 and each machine may get one or
two men assigned to the total task, regardless of the power.
Another division of our company rebuilds the tubes for
these and other machines..they are at the other end of the
building. We also custom build generators and some of you may
well be using a machine I built..For plasma or metallurgy
work at many .edu or .gov locations.
So, as you can see, I understand your situation from the
actual generator, matching systems, typical machines (I work
on over 30 models or makes) power control feedback systems
and of course, just like you, trying to keep the RF out of
any system interface. Two of us built a 100 KW machine for
plasma use, custom designed and when that nitrogen feed
plasma ignited (my first plasma) I fell in love! Oh, yes!
I also have my own Lepel 10 Kw machine here at home,
set up on 13.56 Mhz as well as a 2 kw 400 Khz machine.
I am mainly interested in Ball Lightning because it
occurs in nature and I want to understand it..but plasma work
of every kind fascinates me.
It certainly is hard to sum up my interests and
abilities in a page or two but you should have some idea now
what it is I draw upon for knowing how to debug your RFI
problems..I know your machines, I know the impish nature of
RF and I share your interest in plasma..I guess that when it
comes down to it, I am one of you; A curious mind seeking
more...
So then, let's get to some fast and easy ways of keeping
that RF out of your equipment..I will digress occasionally so
that you may grasp an approach taken as used in other fields
but I will quickly return to our special problems.
First, remember that ground is not always ground! It may
be ground at DC or Line Power frequencies but it certainly is
not always ground as far as radio frequency energy is
concerned. Also, there is earth ground and mutual equipment
ground..again two different animals.
Let's say your lab is on the fourth floor and you are
operating at 13.56 Mhz..do not expect that you will have an
earth ground path at that frequency without secondary ground
conductor radiation! It is even possible to WORSEN your
situation of RF leakage this way..In fact, if there are other
labs close to you, you could easily induce energy between the
labs..your "Ground" could become an active antenna! Some of
you could be working above 13 Mhz and so I will throw in
these numbers to convert directly to inches..
Now, don't go away! It's not that hard. Lets take the
frequency of 13.56 Mhz and play with the numbers..the wave
lengths will be in inches AND are for FREE SPACE. The Signals
actually propagate slower in conductors...typically in a coax
line with a plastic core it is .66 the speed of free space
and .88 the speed of free space for foam.
This data is good to know when you want to make a phase
delay line for a given frequency..We will get back to that
later...its very useful!
You metric people can use the 300,000 constant or
convert from inches..your option..I grew up with inches, so..
Constant: 11808 divided by frequency in Mhz = One
wavelength in free space, in inches. Let's play.
11808/13.56 = 870.7964 inches full wave length, 653.0973
inches three-quater wave length, 435.3982 inches half wave
length, 217.6991 quarter wave length.
So, at 13.56 mhz the wave lengths are: (Free Space)
1.0 --72.566 feet -- or -- 22.1180 meters
.75 --54.425 feet -- or -- 16.5890 meters
.50 --36.283 feet -- or -- 11.0590 meters
.25 --18.142 feet -- or -- 5.5297 meters
Ahhhh, Your 4th floor lab is HOW far above earth ground?
As you can see, by the time the signal almost gets there, it
is on the way back up your "Earth Ground" line...Oops!!
I can almost hear the collective sigh from those of you
working in those far above earth ground labs...But it is
alright because there are many other ways to solve the
problem.
Before I move onto some of these solutions, I want to
cover the very important area of mutual bonding to all of
your equipment racks or frames AND the equipment in them.
Let's take the standard 19 inch (48.26 cm) rack mounted
equipment cabinet, usually constructed of steel. One would
think that if the equipment in it is screwed onto the rack
by it's mounting ears, that this bonding would be adequate
but this is NOT always so. First, the conductivity of steel
is not all that good, both the cabinet and your equipment
have been painted which will compromise the reliability of
the connections.
What should be done is to place a WIDE copper strap as
a bonding bus (you will note that I am using the term bonding
vs grounding..I do so deliberately) from the top to the
bottom of each and every equipment cabinet and this bonding
strap should have wide strips of copper silver soldered to it
that will reach and be connected to the bonding/grounding
studs provided on each piece of equipment. If no such bonding
bolts are provided, drill a hole in each chassis, make sure
all connections are very good as you install your own
bonding/grounding bolt. This bonding strap usually is run
down the rear and inside of the equipment cabinet and of
course bonded to the cabinet itself, top and bottom. I like
to bond it well to the cabinet at each place there is is a
soldered take-off point for each piece of equipment, in
addition to it's being well bonded at the top and bottom of
each cabinet.
Why copper straping and not just wire? Because of what
is known as "Skin Effect"...DC and Power Line frequencies
will carry through nearly all of the conductor area, but
higher frequencies will not...that conductor could be as
large around as your arm and capable of carrying thousands
of amperes at power line frequencies with very little loss or
heating, yet to the higher RF, conductor usage will only be
a few mills deep and to it, how much surface area is
available to carry the rf power is far, far more important
than how thick it is..once the penetration thickness is
exceeded, any greater thickness is simply not used.
Bonding all equipment to a common point..VERY important!
Let us imagine that you have several equipment racks, perhaps
in two rows with the access to the equipment facing the isle
common to the racks..would you simply run a strap down one
row, cross the isle either overhead or on the floor and jump
the strap onto the other row, picking up each cabinet's strap
(by silver soldering) as you went along? NO!! Why??
Think in terms of distance and wave-length, think of the
path that the RF will take...if care is not taken, the mutual
bonding path will divide unevenly, standing waves will form
on the bonding bus, voltage and current peaks and nulls, too.
At high frequencies not many paces are needed to see .25 W/L.
You need to view the work station as the RF will see it
and knowing what the general highest frequency you will use
is very helpful..you can always go down in frequency and be
safe in later projects..it is the top end you have to design
the bonding system for; That and the maximum current the
bonding system will need to carry in the future. Bonding
strap four to six inches wide is a safe range (10.16 to 15.24
cm) with a thickness of .071 (.18034 cm) as a nice available
and standard size.
When laying it out, view your work station as two sub-
systems; One will be all of the monitoring, recording and
control equipment. In this sub-system, bond all of the racks
together like this: Each bonding strap from each cabinet will
be silver soldered to a central point, so that all the
lengths are as similar as possible. This can be overhead or
on the floor and covered with matts; Keep the lengths as
short as you can because this will raise the useful frequency
range for the future. At the central junction point of this
sub-system also silver solder another bonding strap that will
merge with the second sub-system later on..be sure to have it
heading in the correct general direction of the expected
center of the second sub-system ( It is important that you
avoid radical bends in the bonding straps because RF does not
like going around them and for safety, lightning likes such
bends even less) but do not cut it off the roll yet..leave it
alone and finish the first sub-system first.
At this point, many of you will be thinking that if this
first sub-system consists of your monitoring, recording and
control systems and is so well balanced as far as the bonding
system is concerned, how are you going to get the wiring you
need in and out of this well balanced sub-system and over to
the instruments at the plasma chamber area and RF generator
control systems without disturbing the balance of this
bonding system....That is a very good and valid point to
wonder about..I am very happy to tell you that this very
situation has been taken into account..I do think that you
will like the solution!
Before I do, I should point out that the second sub-
system is the RF Generator(s) (and modulator(s) for those
extra creative readers out there), Generator to plasma
chamber impedance match-boxes, Rf power baluns for balanced
feed to the chamber and raising both sides of the RF feed(s)
above frame ground.
That said, let's move onto the subject of isolating the
many lines going into and out of the main control system,
sub-system one without disturbing the balance there.
You can thank AM radio broadcasting for this trick, used
for many, many years and still is. It is so simple that once
you catch onto it, you are free to slap yourself on the
forehead for not thinking of it sooner on your own..I know I
did just that, over thirty years ago! Remember that TV and FM
only broadcasting towers are usually grounded at the bases so
there are no special conditions about feeding the tower
lights power..its simply feeding right up the grounded tower.
While airplanes and radio towers became popular at about
the same rough time frame in history, the two never really
had a very good relationship. When it was daytime and not
foggy or a bad visibility day, the radio towers and airplanes
seemed to tolerate each other, even if the airplanes seemed
to like buzzing the towers when it was good visibility.
Of course, this practice was not very fair to the towers
because they could not duck when the airplanes came buzzing
yet the airplanes were much faster and agile than the
stationary towers and they could duck if they came too close
to the towers and their guy wires..well, most of the time.
But the towers had a way to get back at the airplanes
and all that was needed was to wait for bad weather or even
more effective, night time. After a while there were even
more airplanes and towers, so people had to make some rules
to break the stalemate.
It was decided that airplanes would no longer be
permitted to buzz the towers and that the towers would have
to have special paint jobs and special lighting that would
turn on automatically when it got dark on overcast days below
a certain light level and of course at night.
So it came to be that the airplanes flew high away from
the towers which could be seen well in the day time and at
night.
In those early days of radio, the stations were AM and
the towers were all series feed, meaning they were very much
above ground, electrically. High voltage was present between
the tower that sat on big insulators and the earth a few feet
below it..it was bad enough that fences had to be built
around the tower bases so people would not get electrocuted.
The engineers at the radio stations went through some
mild inconvenience and wound some special coils made of heavy
enough wire to carry the load of all those tower lights plus
some extra wire size was added so that the voltage drop
caused by all the extra wire on the long coil form would be
acceptable. There were 3 coils, one for the lights that did
not blink to get power, one for the lights that did blink and
one for the neutral or common return wire for the other half
of the power circuit for both types of lights. The coils did
effect the tuning of the tower, the more coils, the more
effect, so in time they were wound differently to a new
standard and they were caled tower lighting chokes but they
still effected the tower, just not as much as initially. The
idea of the tower lighting chokes was to let the ac power go
onto and up the tower but not let the radio energy on the
tower leak to ground, where it would have been wasted.
That was the way it stayed for a long while as radio
became more and more popular; New stations were being built
and put on the air in astounding numbers. Soon, over
population as the frequencies filled up..Interference was
everyplace, no real control. New rules were made, very strict
control..The communications Act of 1934 and a new Federal
Communications Commission was born. At night most stations
were forced to go directional, beaming to avoid interference.
Stations quibbled over coverage areas. The FCC stepped
in and determined who would beam which way and when. They
were tough and the standards were tougher. Going directional
meant adding at least another tower, in some cases several.
The more towers, the more complicated it was to keep the
system in compliance, the tuning equipment got more
complicated in an effort to keep the systems stable and
beaming exactly in the directions issued. At the base of each
tower, an antenna tuning unit (ATU) and in the transmitter
building the main tower phasing unit. Everything interacted;
a small change here meant a big change there, also field
strength measurements at designated places to confirm
readings. A tighter pattern meant more towers, more lights,
more lighting chokes and more problems.
An even bigger problem loomed..directional stations were
required to place on each tower, at certain elevations or
electrical degrees up the tower (a quarter wave tower being
90 electrical degrees tall) a phase sample or pick-up loop
that sent down the tower, by way of an exactly known length
of coaxial cable, crossed the base much like the tower
lighting chokes as a big coil and ended up in the transmitter
building, where a few watts of power and the phase
relationship of that sample line and thus the power to and
phase of that tower could be compared to the other towers
and the directional patern watched for changes on a phase
monitor, all the towers being sampled at once.
This meant that there was another coil across the base
in addition to the tower lighting choke(s). The ideal was to
not cross the base at all and here was now 2 or more coils in
an ever more complicated and interactive system..rain or dry
things would change.
The energy that the phase pick up loop induced was
traveling down the inside of a small coaxial cable, not its
outside. But as the outside was in contact with the tower,
that part of the line on the tower was at the same high
potential as the tower itself..thus the choke.
Then it came to someone that if you took a large and
long flexable tubing piece, placed the 3 wires for the tower
lighting inside it, also placed the sample line coaxial cable
inside of it and what ever other spare cables one wanted to,
coiled it up on a form, put the hot end of the coil on the
tower and grounded the cold or user end at the base of the
tower, all those things would still add up to just one coil.
And so was born, the multi cable single coil tower
choke. And THAT is how the control lines to and from the main
control system can be rf-isolated..Your shielded lines to the
instruments will carry only the information you wish on them.
Once the cables are inside the copper tubing, the coil
is wound with a tubing bender or dressed around a coil form
and the length of the coil is such that it will be highly
reactive at a frequency lower than that which you are using.
Another method is to slightly modify this design by
placing capacitors arcoss the coil and tuning the coil itself
for the best null possible.
There is yet another situation that should be addressed
here. There are many of you that will no doubt suffer with
this RFI problem: The devices you have at the location of the
plasma itself are getting rf on them...you have tried
everything, bypass capacitors..and then the inputs to the
devices are being loaded down too much by the bypassing.
Well, you are not alone in this..in situations where you
just can not get the rf out, without compromising device
sensitivity..the cure is to put more rf into the device!
BUT, we shall do so in a constructive manner..you see,
we will inject the same level of rf into the problem spot and
it will be exactly 180 degrees out of phase! If the levels
are the same and out of phase by 180 degrees, the net effect
will be to totally cancel the rf from both the unwanted
path and the injected path. To the device, the rf will simply
not be there anymore and you can go on with your work.
Back in the begining of this writing, I spoke of wave
lengths, how to determine what they were for a given
frequency and how rf traveled slower in conductors, with a
reference to two types of coaxial lines.
So, you will need some rf but you also want it to be 180
degrees out of phase..where you get it is an option..you
could get it from a sample loop at the rf generator but there
is a question as to what the phase at that point will be as
compared to the point that your device is picking it up..so,
You may pick it up from the place it is entering the device,
pass it through 1/2 wavelength of line BASED ON the velocity
factor of that line (usually .66 for plastic or .88 for foam
line) and re-inject it there. Or, if you have a decent scope,
you can determine what the phase difference is between a pick
up loop in the generator and the point on the device(s) being
bothered by the rf.
Then you just make up a section of line that will reach
the pick up loop in the generator and your device being
bothered, add to the sample line from the generator a pot to
act as a voltage divider so that you can match the offending
signal's level with your 180 degree out of phase one and you
are all set. Since the generator frequency is fixed, the
length of the phase delay line will remain valid. Remember
that it is just fine if you need to make the line longer to
reach everything you need to..if it takes a full wavelength
plus an eighth wavelength with the velocity factors in mind,
it is not important.
Just keep in mind the requirements that both signals be
of the same level but 180 degrees out of phase.
If you do not have access to an EE Department, you can
always contact the CE at a local radio station or ask around
to locate a ham radio operator local to you..most of the hams
that have been around a while will have more than enough savy
to help you out if you are not comfortable with this part of
the project. I am located about 30 minutes south of Boston,
Ma. and my e-mail address is > mrv@world.std.com <
Land mail: Mike Valente, 112 Oak St., Abington, Ma. 02351 USA
Feel free to comment or ask a question.. Good luck! Mike