Sunday, June 12, 2016

A 1:1 balun was the best overall choice for feeding the horizontal loop...

Years ago I bought a Heathkit SA-2060 (non "A" version) 2kW-rated antenna tuner at a local swap meet for a good price.  While not as heavy-duty as some of the venerable Collins or Viking tuners, it had a nice-sized roller inductor and a pair of large, wide-spaced variable capacitors in a typical "T" ("High-pass") configuration.
Figure 1:
The Heathkit SA-2060 tuner and (now) 1:1 balun feeding the 450 ohm
window line.  It no longer sits, on edge, in the window, as it had previously -
a much more convenient arrangement!

I have used this antenna tuner for years, taking it to Field Day and other than having to tighten some screws and adding thread-locker as well as a bit of lubrication of the moving parts after I got it, it has served me well, (seemingly) capably matching the 300-something foot circumference horizontal "lazy loop" antenna at my QTH that is fed with 450 ohm window line.

Discovering the problem:

A month or so ago I was doing some rewiring after having my main electrical panel replaced in conjunction with the installation of a PV (Solar) inverter system and to do this work I had to "open up" some wall and ceiling spaces in the room containing my ham shack - but this also meant that I had to disassemble and relocate much of what was in the shack to accommodate that which had to be moved out of the way.  While the "radio area" wasn't disassembled for this task, I ended up piling a lot of stuff in that part of the room, making it inaccessible.

One of the things that I did during this work was to pull a brand new 240 volt, 20 amp circuit for my Heathkit HL-2020 linear (really an SB-221 with a brown color scheme) and once I had the room more-or-less back together I reconfigured the amplifier for 240 volts (there were minor complications to this - perhaps another story) and I was ready to get back on the air.  From what I'd read, the combination of the higher mains voltage and the Peter Dahl transformer would provide a higher plate voltage under load along with higher output with slightly less drive - and upon testing with the dummy load, this appeared to be true.

For years my tuner had been sitting on edge in the window with the 450 ohm window line coming through an insulated gap, past the vinyl window frame and connecting directly to the balanced wire connection on the back panel.  In the rearrangement I'd needed to take the tuner out of the window and in the process one of the wires of the window line popped off - something that I noticed as I was preparing to test the amplifier under load.

Happening to have the receiver on at the time, I reconnected the leg of the balanced line and...

There was no difference in the signal strength of the received signals.

Something was definitely wrong here!

I would have expected that with one leg of the balanced line disconnected that I'd get at least an "S" unit or two difference in signal strength, but there was no obvious difference at all.  Grabbing a screwdriver I shorted the balun's balanced line output (with one of the wires of the window line still connected) and, again, could hear no difference, either.  I connected my antenna analyzer and noted that while there was a good match through the tuner, it did not change much if there was one or two wires connected, or if the "balanced" terminals were shorted together.

Hmmm...

What was happening:

Now, I was curious.  It would appear that I'd been actually running the "loop" in a "T" type configuration with the downlead being (more or less) end-fed and the remainder of the antenna being a sort of distributed top hat.  I've never really had trouble working other stations, nor had I really experienced any "RF in the shack" issues as I had a pretty decent, short ground with heavy decoupling of the HF coax feeding to the tuner via a large chunk of ferrite scavenged from an old computer monitor.  In other words, I'd had no reason to question the operation of the balun itself or how it functioned.

The tuner's cover was immediately off and I was comparing the balun connection with that of an SA-2060A manual that I'd found online and the results was inconclusive:  If the wires had been properly identified and taped at the time of initial construction, it looked correct, but if not, the only way to verify this was to remove the balun and check it with an ohmmeter.   

Rewiring the balun to specs - and a problem:

I regret that I didn't make a note of how the balun core was wired, but I do know that it wasn't at all right so I made the necessary changes after "buzzing" the wires with an ohmmeter to identify them and tested the balun on the bench with the antenna analyzer and the other end of the balun terminated with a 200 ohm resistor.  Unlike the original configuration of the balun, according to the analyzer it was now working as it should, having a reasonable match to 50 ohms across the HF spectrum and going to infinity when resistor was shorted or removed.

Putting the balun back in the tuner and reassembling it I had to readjust from my previously-noted tuner settings to find a proper match (a good sign that the settings weren't the same, actually!) and I then checked it out with 100 watts on 40 meters.  Everything appeared to be fine, although the tuner struck me as a bit more "touchy" in its adjustments as compared to before.

Firing up the amplifier I soon discovered that I couldn't tune it up without its "Plate" variable capacitor arcing over noisily.  Grabbing a "Cantenna" dummy load I verified that the amplifier itself was fine at full power, but something else was wrong.  Turning the power all of the way down and then slowly up again I discovered that at around 200 watts of RF output the reflected power went up, suddenly equaling the forward power.  Popping the cover off the tuner again confirmed my suspicion:  The "output" capacitor in the tuner was arcing over.

When the typical 4:1 balun probably isn't appropriate:

What this meant was that the tuner was being asked to match something really awkward - but with my loop and given its length I thought it unlikely that the feedpoint impedance would be really high, but rather it was more likely that it was "low-ish" - probably well below 100 ohms.

The problem with this is that I now had a properly-working balun that provided an upwards impedance transformation.  This meant that if the loop had, for example, a 50 ohm feedpoint resistance, the tuner would be "seeing" around 12.5 ohms.  This is bad news as making a transformation from 50 ohms to 12.5 ohms implies the likelihood of a high-Q configuration of the tuner itself which, in turn, implies high voltage and high current which further implies high losses!

Wielding my antenna analyzer I hooked its BNC connector directly to the balanced line:  Since the analyzer was hand-held and I was checking at "only" 40 meters I didn't think that it would really matter much that it was properly "balanced" or not.  The readings indicated a resistive component of around 10 ohms with a reactance of around 180 ohms inductive, but in tuning around to other amateur bands I couldn't make much sense out of the readings and was particularly suspicious when none of the resistance values seemed to go much above 50-80 ohms.

Suspecting that without the "bandpass filter" effects of the tuner that I was the victim of an AM broadcast station a few miles away being detected by the reflectometer bridge in the analyzer and causing false readings, I dusted off my Heathkit HD-1422 RX noise bridge and connected it to my FT-817, running on battery - this combination being comparatively immune to stray, off frequency RF and like the analyzer configuration that I'd used, more-or-less "balanced" without any obvious ground reference.  With that configuration I got a more sensible resistance reading of around 35 ohms and a reactance measurement in the area of 130 ohms, inductive.  If I took the 35 ohm reading seriously that would mean that the antenna tuner was trying to match something under 10 ohms via the balun!

Most antenna tuners match "up" better than "down":

Figure 2:
The exterior of the Balun Designs Model 1171t 1:1 "ATU" balun.
This model is equipped with studs on the top of its weatherproof case
for connection to a balanced feedline.
This brought to mind a discussion that I'd had with another amateur some time earlier.  He pointed out that it seemed silly that most baluns with tuners offered only a 4-fold impedance up-conversion, but it was likely that a typical antenna fed with balanced line was more likely to see a lower impedance on most bands unless there was a configuration that was particularly prone to high impedance like a 1/2 wave end-fed wire (e.g. a "Zepp" antenna), a rhombic (don't most of us wish we could have one of those?) or a full-wave dipole.

What this meant was that in most cases that the average amateur would encounter, the tuner was going to be matching to substantially lower than 50 ohms resistive through the 4:1 balun - something that is likely to cause problems like loss - which is invariably accompanied by heating - and high voltages, internally.  What had been a reasonable hypothetical scenario was manifesting itself as reality!  In reality, most antenna tuners will more efficiently match to impedances higher than 50 ohms than lower because of these potentially high currents, so having a balun that tried to make my antenna's somewhat low impedance even lower (at least on 40 meters) was probably not a good idea.

The solution:

The clear solution in this case was to use a 1:1 balun, instead to avoid the tuner from seeing an "awkwardly low" impedance.

I had the choice of reconfiguring the existing 4:1 balun - which was now working properly - perhaps by rewinding it with some PTFE 50 ohm coax - but I decided, instead, to get another balun and keep the internal balun intact in the event that it would be needed (it's nice to have options!) as it could be easily inserted or removed from the circuit using a jumper on the rear panel.  Because I was intending to use the tuner/balun combination with my amplifier - which was capable of the full 1500 watts output - I also knew that it needed to be both low loss and capable of handling very high voltages and high currents.

In perusing the various web sites and forums I looked at the possibility of making my own balun - which I could have done for $20-$40 in parts - but ultimately decided that I didn't want to take on yet another project right then and settled on the "1:1 ATU Balun" by a company called "Balun Designs." - link.  The products of this company not only had good reviews, but their web site was also impressive, explaining in good, sensible detail why one balun was better than another for a particular application and it also outlined situations where certain baluns they sold should not be used and why.

Figure 3:
Inside the 1:1 balun.  It is wound with parallel, highly-insulated
enameled copper wire in the "Guanella" fashion described by Joe
Reisert (W1JR) in the September 1978 issue of Ham Radio magazine - that
is, the second "half" of the windings cross over to minimize coupling
between the input and output to provide best isolation and to
minimize the "one turn" effect inherent with "normal" toroid
winding techniques.
The balun that I chose (Model #1171t) is a current balun which effectively operates in series with the signal path (unlike a more common "voltage" balun which typically resembles an autotransformer as in the case of a typical 4:1 balun) and is essentially a common-mode choke that isolates one side of the balun from the other by virtue of the bulk inductance of the core over which a transmission line is wound.  By suppressing the "common mode" aspects of the RF signals with a significant amount of inductance, the windings on the toroids effectively "choke" anything other than differential (balanced) currents and thus isolate one section of the feedline from the other - except for the equal-and-opposite RF that is supposed to be there!

While many of these current baluns are wound with PTFE coaxial cable to preserve the 50 ohm impedance, this particular balun was wound with what amounts to parallel-conductor transmission line consisting of enameled copper wire covered with PTFE spaghetti tubing - a configuration that has potentially lower loss and higher power-handling capability by virtue of its larger conductors.  What this means is that this "parallel transmission line" winding inside particular balun isn't particularly close to 50 ohms in its natural impedance (it's likely in the 70-100 ohm range) when terminated with a 50 ohm load the apparent match, when viewed with an antenna analyzer seems to degrade as frequency increases - likely a result of the "transmission line" comprising the balun's winding causing some impedance transformation.  This property is of relatively little importance for this application since it is designed to sit on the output of an antenna tuner, anyway:  As long as it has low loss and can withstand the expected voltages and currents, it would have minimal effect on the overall system efficiency despite a mismatch under most conditions.

When this balun arrived I connected it to the output of the SA-2060 tuner with a short (approx. 18", 50cm) RG-8 style jumper to minimize losses and was easily able to tune the antenna with settings radically different than before - another good sign!  Finding that everything looked good on the analyzer, I hit it with 100 watts - and then 1500 watts and had no problems at all.  I did notice that the window line became warm to the touch and the balun core and windings also became perceptibly warm, but by no means "hot" as the thermal image in Figure 4 suggests.

Figure 4:
A thermal image showing the heating of the balun and transmission
line.  As can be seen, the closer to the "output" of the balun, the
warmer the windings got, but after approx 20 seconds at 1100 watts of
RF on 40 meters their temperature stabilized.  The image
above depicts a maximum temperature inside the balun of less
than 120 degrees F (49C) with the feedline at approximately
105 F (41C) both being warm, but not "hot."   (The "warm" UHF
connector at the bottom appears thus as it is reflecting heat from
elsewhere in the shack.)  Considering that over 1kW
of RF is flowing, this amount of heating represents negligible loss - likely
less than that occurring within the tuner itself.
Of course, the amount of heating will depend both on the power level
and the amount of current flowing through the balun, and this depends
on the matching/impedance conditions encountered.
I also observed that if I disconnected just one side of the balanced line, the signals on the band dropped by several S-units and the relative floor relative to signals came up but when shorting the balanced line, this caused signals to all but disappear - exactly what I was expecting to happen in a circuit that properly rejected common-mode signals.  When checking across the bands at different times of day I also observed that the noise floor was 1-2 S-units lower than before and that the previously S-5 noise from the switching supply on the nearby DSL modem (just a foot or so away) was now barely detectable at the S-2 noise floor on 40 meters:  Thus are the benefits of common-mode rejection in the prevention of electrical noises from the shack and the house's electrical system from being conducted onto the feedline/antenna and into the receiver.

As far as the warming of the window line, I did some calculation and determined that the feedline was likely seeing a VSWR somewhere in the range of 8:1 to 20:1 or so, which meant that, when it was dry, it was losing as much as 0.5 dB along its 30 foot (10 meter) run - a worst-case (20:1 VSWR) loss of up to 11%, or in the area of 150 watts at the test power of 1100 watts.  This is a small fraction (approximately 1/12th) of an "S" unit, but it would certainly explain the warmth!

A few days later I had the opportunity to check into a 40 meter round table with a group of friends across the western U.S. and conditions were abysmal, but not only could I hear all of the stations pretty well, one of the stations with the weaker signals reported that they could hear my just fine, with my signals being comparable to another station across town from me running about the same power - a reasonable indication that I wasn't burning up too much power in losses!

Final comments:
  • One of the first things that I do when I get gear at a swap meet - be it commercial, a commercially-made kit or homebrew - is to check it out, making sure that all hardware is tight and electrical connections are solid, but I will admit that it never occurred to me when I got this tuner to check to see that the balun was wired properly!
  • While the implementation of the new balun reduced my 40 meter noise floor from about S-5 to S2, I could still hear some low-level noise related to the AC mains and a small switching supply in my shack.  After a few weeks it finally occurred to me that where the new balun was hanging by the feedline placed it squarely over a small switching supply in a wall-mounted outlet strip.  When I moved the balun away from the power supply the noise floor went from S-2 to S-1 or so and I now hear only atmospheric noise:  For a while I had a metaphorical palm-shaped dent in my forehead!
  • There is an interesting article written by G3TXQ - "Tuner Baluns:  4:1 or 1:1" - link here.  This article analyzes the impedances typically found and makes suggestions about how to determine whether a 1:1 or 4:1 balun is most appropriate when used with a tuner/antenna/band combination.   Spoiler alert:  He suggests a 1:1 balun for the general case - in fact, the same Balun Designs #1171 as depicted above.
    [End]

This page stolen from "ka7oei.blogspot.com".