<== you are here

DIY: Winding the FCP's Isolation Transformer

If one does not wish to construct their own isolation transformer, it can be procured from Balun Designs. For dual banding use a 160 isolation transformer, DIY or procured.

Balun Designs is the only firm known to us supplying ready-to-use isolation transformers per the specifications on this web site. We will list other suppliers if they emerge. We do not receive royalties from the sale of Balun Designs products.

We frequently get questions, like instead of a balun, why not just any old balun? Why that toroid material, why that size toroid, why that wire, why the PTFE** sleeving? Suspiciously, these questions often occur close to a major contest involving 160 meters, with procurement now unlikely before the contest. So we will take a little aside to outline the decidedly non-theoretical reasons from the original Burn 'em Brothers, the crack them, arc them, melt them, set them on fire wrecking crew, W0UCE + K2AV.

** PTFE is often referred to by the brand name Teflon™, the registered trademark for the PTFE product line of E.I. du Pont.

You can skip those explanations and farther down the page.

How Come that Particular Stuff?

The choice of wire and insulations came at the end of a period of failed transformers subjected to 1500 watts, tough RF situations and some number of close lightning strikes at W0UCE which also took out appliances and other ham equipment. The fix for blown transformers was heavy polyimide insulated wire plus standard wall #12 PTFE sleeving.

#2 powdered iron cores wound with this combination no longer failed unless they were directly struck by lightning or dropped on their windings from height onto concrete garage floors or otherwise damaged from harsh mistreatment. Undamaged, they ran stone-cold in FCP related applications, and are still in use.

The double (heavy) polyimide insulation would not melt, but unfortunately was easily nicked in miscellaneous handling of a core wound with #14 double polyimide only. A nick in the polyimide led to carbon tracks, and burned and shorted devices. However, if the same wire was immediately inserted into PTFE sleeving before exposure to any working or contact, it was protected from nicks and abrasion.

PTFE sleeving also increased the breakdown to 55 kVDC between adjacent wires in the winding and 27.5 kVDC between wires and the core and mounting materials, or 19.5 and 9.75 kV RMS at RF respectively**.

The wire/insulation diameter of PTFE-sleeved #14 double polyimide dictated that the T300 series toroids had the smallest inside diameter that could fit a single 20 turn laid-flat bifilar winding on the inside surface. Two or more layers require a different and more difficult and expensive winding design, usually done by machines.

Balun guru W2FMI(SK), in his detailed book on baluns and ununs***, described an antenna tuner balun he designed to handle "worst cases". Jerry settled on a T400A-2 powdered iron core, and used 21 bifilar turns of the same PTFE sleeved #14 wire for his antenna tuner balun. He had done much testing to arrive at his choice of toroid core material. After our collection of failed attempts with different cores, we settled on W2FMI's choice and sizing of core material. And having already arrived at the same wire and sleeving, we were now synchronized with W2FMI.

There were negative reactions to our using a T400A-2 core, on price and on physical size. The next smaller available diameters were the T300xx toroids. As the T300xx were the smallest usable inside diameter for the windings, and we settled on and tested with the inch-thick T300A-2 core.

Downsizing Components for QRP?

We are often asked about a low power, or smaller core version for QRP. The transformers described below are probably good to 3 kW with an L over FCP. We have steadfastly resisted downsizing. In you will find downsized and miniature tuning devices listed as one of the miscreants. QRP particularly needs minimized losses everywhere in the antenna system, so you go for the low losses of the QRO capable transformer. QRP ops are frequently in the situation where their signal is in the noise at the other end, where all those extra dB shards can make a difference. Save the miniature stuff for back-packing and carry-in sites, where you require minimal size and weight above other considerations.

<== you are here

Isolation Transformer Options

On 160 meters, an isolation transformer wound on a T300xx or T400xx powdered iron core exhibits some non-one-to-one impedance transformation. The "Regular" transformer versions exhibit more than the "Heavy" versions. But the variation in the transformation is usually absorbed without effect in the typical 160 meter pruning and tuning exercise. In many cases the residual inductive reactance is actually helpful for overall tuning of the L/IsoT/FCP system.

A 160 L/IsoT/FCP sometimes gets trimmed/adjusted/tuned to 50 ohms easier with a "heavy" version T400A-2 based transformer. But the T400A-2 will not fit in the 4x4 inch sealing cover plastic electrical box often used for transformer/balun housing. An alternate "heavy" version is a two core, inch and a half stack of a T300A-2 and T300-2, minimally taped together and wound as one core. This is very close to the overall characteristics of a T400A-2 and will still barely fit in a 4x4 inch box. This can also be done with a three-stack of the half inch T300-2.

We cannot specify the benefits of a "heavy" version vs. the "regular" with any certainty due to the huge variation in inverted L characteristics as actually installed due to height, placement, and environment. This subject is treated in depth in We highly recommend that you study the "Taming" section if you are interested in an "advanced" isolation transformer deployment.

Using a larger core is an extra expense that may or may not render an actual benefit. Other than sometimes mildly more monkey business to get a 1:1 SWR to the feed coax, use of the standard single T300A-2 or the equivalent double T300-2 does not in any way effect the distant end signal strength performance of a 160 L/IsoT/FCP.

Micrometals' identical cores for Amidon's inch-high T400A-2 and T300A-2 are the T400-2D and T300-2D respectively. The half inch high versions are labeled the same. The Micrometals cores are frequently listed on eBay.

**See double polyimide coated wire and PTFE tubing specifications. Typical "heavy" or "double" Polyimide (14.6 kV), plus #12 standard wall PTFE tubing 800V/mil times .016 in (12.8 kV) gives a 27.5 kVDC rating for the combination. This results in 27.5 kVDC insulation between the wire and the core and environment, and 55 kVDC between wires in the winding.

***Jerry Sevick, W2FMI(SK) "Understanding, Building, and Using Baluns and Ununs -- Theory and Practical Designs for the Experimenter", pp 58-62, (C) 2003 CQ Communications, Inc, Hicksville, NY

<== you are here

Isolation Transformer Construction

Most of this section on winding an isolation transformer is based on experience and things that have gone wrong. Purchasing a commercial version allows you to skip this section, but if you are going to wind your own, this is must reading.

Transformer** Core Configurations Winding Turns and Lengths
3 Inch Regular (a) qty 1 T300A-2 (T300-2D*) 2 x 7'6" (2 x 2.3 m)
(b) qty 2 T300-2 40 total, 20 bifilar turns
---------- ----------
3 Inch Heavy (c) qty 1 T300A-2 + qty 1 T300-2 2 x 9'2" (2 x 2.8 m)
(d) qty 3 T300-2 40 total, 20 bifilar turns
---------- ----------
4 Inch Heavy (e) qty 1 T400A-2 (T400-2D*) 2 x 12'6" (2 x 3.8 m)
(f) qty 2 T400-2 50 total, 25 bifilar turns

* Micrometals designations for inch thick cores. Amidon and Micrometals designations are identical for the half-inch thick cores
**Discussion of these transformers is above in

Do not bother to cover the toroids with fiberglass tape. As we experienced early on, in several years of outdoor and damp indoors locations, the irregular surface of the fiberglass tape became a point for moisture, mold, and airborne material such as smoke and airborne tree resin particles to accumulate and become a dielectric and carbon track issue. With temperature changes, even a "sealed" box allowed air egress at hardware penetration points. "Project boxes" with doors required weep holes in the bottom of the box to blow out droplets when sunlight struck the box and heated it.

The PTFE sleeved wire should be wound directly on the bare toroid core. In addition to our experience, W2FMI flatly states in his book that fiberglass tape is not needed if PTFE sleeving is used.

If your core consists of multiple toroids, then tape them together at 120 degree intervals with one perpendicular turn of narrow electrical tape plus an inch overwrap. This is just to hold them together to start the winding. After winding, the windings and hold-down tiewraps will hold the toroids together.

Do not tape two toroids together by going "around the equator". This will create a void, inside the tape and between the toroids, which will at some point outdoors will accumulate water, causing sneaky, gradually increasing loss.

It is important to avoid nicking the transformer wire's polyimide insulation or subjecting it to any kind of abrasion except at the soldering ends of the wire. PTFE sleeving protects the polyimide insulation. Although the polyimide insulation retains its heat resistance and insulating properties beyond common soldering temperatures, we found it easy to nick the polyimide, not so after sleeving. You may wish to do the cutting and sleeving of the wire on a carpet for this reason.

Transformers situated outdoors usually accumulate traces of moisture between the wire and sleeve. RF at a nick in the wire insulation can use the moisture to begin a carbon track. A nick presents a point where only the PTFE has to be penetrated to get to the wire. The strategy is to require penetrating both PTFE and the polyimide to get to the wire.

Do not use any kind of lubricant on the wire. PTFE is already very slippery. It is possible to push 20 feet of #14 heavy polyimide wire inside #12 standard wall PTFE tubing. Some lubricants will make it harder to insert the wire. The lubricant can also be of a composition that deteriorates or introduces loss. To avoid digging into the PTFE during insertion, burnish the wire's insertion end-cut with a fine fingernail file or sandpaper until it is rounded without sharp edges. When done filing be sure to wipe off any grit or other residue from the wire's insertion end.

Before you do anything else with the wire and tubing, measure twice and cut the PTFE tubing to size. To reduce friction, gently straighten the wire just as it is being pushed into the PTFE. Push the #14 polyimide wire all the way into the #12 standard wall PTFE sleeve until the far ends of wire and tubing are even. Then cut the near end of the wire even with the near end of the tubing. PTFE sleeving is now protecting the polyimide insulation from nicks. Do this for two wires, each wire at the length for your choice of core(s) in the table above.

Extremely important: On one of the winding wires mark both ends and the center and the 1/4 and 3/4 points with "flags" of tape, five flags in all. Later, if you don't flag, it will be difficult to correctly modify the turns ratio for Method B impedance matching in Don't use a marker for this as marker ink does not stick to PTFE. Even permanent marker ink simply wipes off as the wire is handled during winding, resulting in one or both ends of the winding unmarked for polarity. Getting the wires flipped changes the isolation transformer into a weak choke. It will add loss to the system by driving power down the feedline shield, but otherwise seem normal. There is a check for this in the construction, but if you have already trimmed leads, you may have to splice wires to repair.

Pair the wires. Think of it as exotic zip cord. Although it is not required, it is easier winding (and looks better) if you tape the wire together roughly every four inches (100mm) with narrow tape. Ordinary width electric tape can be cut in half lengthwise with scissors for this purpose. Wrapping with a turn and a half around the wire pair is sufficient. Once the winding is on the toroid(s) there is no stress on the tape.

The wire lengths in the table include generous lead lengths to reach various points in a project box. Better too long than too short.

One flagged wire end goes to the FCP connection, the other flagged wire end goes to the aerial wire connection. If you shorten a flagged wire to clean up inside the project box, immediately reflag the wire. The flags will be essential if you later want to change the turns ratio.

Extremely important: After winding and preparing to make connections, make sure your connections will follow the schematic below. Perform all four tests in Testing Isolation Transformer Connections below. It is a good idea to label wires with their connections once you have verified them.

<== you are here

Connecting the Isolation Transformer Windings

This section and polarity marks on isolation transformer schematics are new on the web site with V.2019.03.15.

Always connect an aerial wire, FCP and coax to an isolation transformer per the drawing at left.

IF 1) you know or suspect a Balun Designs isolation transformer has been modified from its original wiring, OR IF 2) you are directly wiring any isolation transformer into an installation or enclosure, either user-wound, or a commercial transformer removed from or without an original enclosure, THEN perform all four tests in Testing Isolation Transformer Connections below.

Connecting Unmodified Stock Balun Designs™ Isolation Transformers

All unmodified enclosed Balun Designs isolation transformers are identically wired internally, with identical external connections.

Balun Designs isolation transformers sold January 15, 2019 and later have "FCP" and "Antenna" labels at the correct 1/4 inch stud terminals. Always connect a Balun Designs isolation transformer per labels next to the 1/4 inch studs.

If you have an older Balun Designs isolation transformer with no "Antenna" or "FCP" stickers next to the 1/4 inch studs, then:

Hold the transformer enclosure with the cover side (four screws) facing you, coax connector facing down. Connect the aerial wire to the right side stud and FCP to the left. You might want to add the markings "FCP" and "ANT" with a permanent marker.

Disclaimers: Balun Designs™ is a highly valued commercial resource, providing a source of professionally constructed transformers described on this web site. They are, however, not responsible for the design of said transformers, nor responsible for knowing or explaining how to use them. Balun Designs warrants their workmanship and that the 1142s uses the same design and materials specified on this web site for those winding their own transformers.

This web site is not a commercial enterprise. K2AV receives no royalties or compensation from any source for use of the antenna and transformer designs. The content of this site is published without charge for the furtherance of amateur radio and the benefit of the readership.

Testing Isolation Transformer Connections

PREPARE for the tests below:
Disconnect any feedline from the SO239 coax jack or terminals used for the feedline connection.
B) Disconnect the FCP and aerial wires. These disconnections prevent external connections in the aerial wire, FCP and coax arrangements from giving false readings in 2), 3) and 4) below.

Perform all four tests:

1) There is no ohmmeter test to verify winding polarity. It must be done visually. The cover of an isolation transformer enclosure must be opened to verify.

Examining whichever end of the bifilar winding can be clearly viewed,

Either: Verify that the two wires intended for or wired to the coax center conductor and aerial wire make up one end of the bifilar winding. (Illustrated left in the photo of a stock Balun Designs isolation transformer.)

Or: Verify that the two wires intended for or wired to the FCP and coax shield make up one end of the bifilar winding.

2) Place ohmmeter probes on the wires intended for or wired to the FCP and aerial wire connections. You should see a dead short.

3) Place the ohmmeter probes on the wires intended for or wired to the coax shield and coax center conductor connections. You should see a dead short.

4) Place ohmmeter probes on the wires intended for or wired to the FCP and coax center conductor connections. You should see an open.

If one or more tests fail, you have either a wiring designation error to resolve, or an outright wiring error to correct before installation. Getting wires mixed up is the most common failure related to a user-built or directly-wired or modified isolation transformer.

Your Existing In-Place Transformer is Reverse-Connected?
Your Transformer Wiring Fails Test 1) Above?
What to Do? What Happened? The Gory Details.

Short Answer:

We all wish we had figured out the reverse connection issues years ago. It's particularly obvious now that normal and reversed connections do not behave the same since we now have 2 port VNA measurements and other results for normal and reversed connections.

The simple and cruel answer to what to do: If your existing FCP/aerial wire connection is reversed, and you suspect some performance degradation, or it simply bothers or worries you, then you can un-reverse it and do the tune-up work over again. There is the suspicion that it will be easier to tune with the normal connection but no guarantees, yet. There is the suspicion that un-reversing will remove some degree of loss, or reduce local noise on the coax shield getting back into the RX via the transformer, but no guarantees, yet. We simply lack enough accumulated testing and multi-site experience in 03/2019 to definitively list probable results of un-reversing.

Really Long and Detailed Answer:

The schematic at the left was W0UCE's original 2009 basic drawing of aerial wire, isolation transformer, FCP and coax connection, still embedded in this web site's home page. Little has changed in nine years of publishing, no changes to the schematic per se. Polarity dots were added 03/2019 to explicitly specify the polarity implied by the presentation.

Many simply wound and connected the transformer per the schematic. They read the drawing to mean that the center conductor and aerial wire are supposed to be on the same end of the bifilar winding. And so they conveniently, if unknowingly, avoided the now-known issues with reversed connections.

Pictured left, K8OZ's isolation transformer in the NCJ article clearly shows the correct polarity connections with marked aerial wire and FCP connections. In 2010 Jim would not have received any instruction to that effect, other than W0UCE's drawing.

K2AV admits to a vague memory of issues taming the first isolation transformer at W0UCE. K2AV couldn't get it to work correctly until he "tripped over" a particular wiring that behaved. This solution, however not understood then, is responsible for the orientation in Jack's original drawing. He went out and traced the wires.

In 2010 we were a long way from the personal knowledge to understand that problem. We were also a long way from owning the needed equipment and programs to implement the knowledge. Some of the equipment became available only recently at less than military prices.

Not until 2018 did we have any reports of issues specifically implicating an FCP/aerial wire connection reversal. A few reported that touching the SO239 shell on the isolation transformer made a change to the SWR. But adding to misinterpretation, for years a far more common report was that using an isolation transformer eliminated the SWR change when touching the shield, often seen when using a balun instead of the transformer.

There was an emerging suspicion that reversed antenna/FCP connections changed how the L/FCP tuned up. This proved difficult to chase down in the field once a given installer had his FCP up, tweaked and working DX and contests. Understandably, satisfied FCP users were reluctant to take apart a working installation for research testing. The conversation usually included some variation on "If it ain't broke, don't fix it."

Back when the "New" inverted L was put up at K2AV in fall 2015, he used an on-hand Balun Designs 1142s instead of winding a new transformer. It had no Ant or FCP stickers marking the studs. It was connected per-the-diagram only by blind luck.

The FCP wire came through the shed wall on the left and the inverted L wire on the right. The 1142s was screwed to the plywood "breadboard" between the two, using the box screw-down tabs on the back of the box. Thus the cover with four screws was facing the viewer. The coax jack pointed down toward the floor where the coax went underground and out of the shed. Thus the FCP was connected to the left stud, and the aerial wire to the right, with the cover facing and the coax jack down. Blind luck.

Had the FCP been routed on the right side of the shed, K2AV's transformer hookup would have been reversed, and we may have discovered the issue in 2015.

The FCP was on the left side because two big tree trunks were nicely in line to suspend the FCP needing only a center post mounted to a corner of the shed in a nicely uncluttered layout. Compliance with the loss list's ten foot diameter tunnel specification only required trimming a few branches off two backyard trees.

The "new" L/FCP gave results with the expected approximate lengths and a feed R in the low thirties. Later, when RBN readings seemed to be better W and NE from the smaller new antenna than from the larger old antenna, there was only a great sigh of relief that the new power lines up on the service road hadn't ruined the only good 160 transmitting site on the property.

Then in fall 2018, partly to analyze the now and then reported SWR change issue, K2AV purchased a cigarette-pack-sized portable two port vector network analyzer, the Mini Radio Solutions "miniVNA Pro2". He also bought a Surface Pro laptop to replace the household's 2004 vintage memory-starved and stuck-at-Windows-XP Compaq laptop.

The VNA is powered by a rechargeable internal lithium ion battery and uses Bluetooth to talk to a Java VNA driver program running on a Windows computer or Android phone. The Surface Pro, with internal Bluetooth capability, can be used as a tablet by unsnapping the keyboard, making the combo very portable and wonderfully handy for antenna tweaking.

Critically, the internally powered, isolated-by-Bluetooth VNA could measure with the isolation transformer as the only attached metal while testing the wiring reversal. Stand-alone two port tests measured the common mode leakage path of a "stock" Balun Designs 1142s/160 isolation transformer. Reflection mode tests checked K2AV's FCP/IsoT/L feedpoint for common mode effects by attaching and detaching the feedline shield using both normal and reversed winding.

To check the effects of introducing a common mode signal, he used his L's isolation transformer, now a hand-wound version,** somehow still using the normal FCP/Aerial connections. He ran VNA scans in reflection mode at the coax jack to record characteristics without any common mode voltage. He then attached the feedline shield with a clip lead to the coax shield of the VNA/transformer connection, introducing common mode voltage to the combination. The common mode voltage shifted the X=0 point lower. The shapes (not the frequency position) of the R and X curves remained the same, shield attached or not.

** The previous Balun Designs 1142s had been given to K4QPL for use at VP5M to improve the 160 setup there with an FCP. K2AV hand wound another transformer to replace it. Note that an old BD 1142s never gets thrown away, just passed around to friends. Neither we nor Bob at Balun Designs have yet seen or heard of an 1142s or spec hand-wound blown except by a direct lightning strike or mechanical damage.

When the FCP/Aerial connection was reversed, attaching the feedline shield significantly changed both R and X curve shapes, along with a large upward shift of the X=0 point indicating a significant feedback effect.

K2AV then tested a brand new, bought for the experiment, Balun Designs 1142s/160 isolation transformer at 1.825 MHz to measure isolation, R, X and |Z| of the common mode leakage path, with normal and reversed antenna connections.

With the VNA in transmission mode, Port 1 was fed to the antenna side 1/4 inch studs with clip leads so the port 1 high side and ground connections could be flipped. The coax side SO239 was terminated in a closed 50 ohm termination. The port 2 center conductor was connected to the external shell of the termination with a clip lead to sense any leakage from the transformer to the outside of the termination shield (common mode leakage path). The results follow:

Common Mode Leakage Path Testing Results - 1.825 MHz

Tested Unit: New Balun Designs 1142s/160
Field Strength Test: Using 15 Watts to Tuner + Coax at K2AV (Discussion below)
ATR-30 Tuner Set For: 50 + j0 at AIM4170 in Shack at K2AV (Discussion below)
Leakage Path Blocking Test Normal Wire Connections Reversed Wire Connections
dB Isolation 43.6 dB 26.5 dB
Inductive Reactance (X) 5.23 k Ohms 2.07 k Ohms
Resistance (R) 14.2 k Ohms 240 Ohms
Impedance Magnitude |Z| 15.1 k Ohms 2.09 k Ohms
-------- --------
Field Strength Test
Volts/meter at 125 Feet 1.15 V/m 1.14 V/m
-------- --------
ATR-30 Tuner Settings
Trans/Ant/Inductance 2.0 / 4.5 / 18.0 2.5 / 3.2 / 18.0
Resonance, Tuner Bypassed 1.817 MHz 1.838 MHz

Note the large difference in the R values of the common mode leakage path. With reversed connections a small unfortunate series capacity would leave an R of only 240 ohms blocking induced TX power on the coax shield. This would cause a hard-to-predict voltage feedback from the aerial wire that varied considerably with changes in frequency. This explains the shape changes in the R and X curves using reversed wiring when connecting the coax shield.

Reversing the connections also causes a loss of 17 dB feedline common mode noise blocking. Common mode noise can come into the shack on feedlines and be blocked with various devices. But common mode noise can also leak into the antenna itself. Once into the antenna circuit, the noise cannot be separated. The normal connection's 44 dB blocking of common mode voltage explains a common story that changing to an isolation transformer made the antenna quieter on RX.

If you have already successfully matched your L/FCP with a reversed isolation transformer connection, you may not wish to change the connection unless you think you have issues of some sort and you are prepared to redo your match to the L/FCP. If you are having difficulty matching and are reversed, then you have nothing to lose -- change to the normal connections and proceed from there.

To measure field strengths, K2AV used a continuously variable ATR-30 tuner and an AIM4170 VNA in the shack to present exactly 50+j0 to the transmitter for both the regular and the reversed connections. Transmitting at 15 watts, he measured field intensity 125 feet from the vertical. The measuring instrument was an in-certification lab-calibrated Potomac Instruments FIM-41 professional field intensity meter. The field strengths are reported in the table above.

Contrary to expectations, reversing the connections did not produce significant field strength loss. However, this single result at K2AV should be considered preliminary and possibly not representing the majority.

There may be a situational reason for the very small reversal loss at K2AV. The exposure of K2AV's 160/80 hardline shield to induction from the aerial wire is only 25 feet with 8 feet in buried conduit. The coax shield is grounded at the house entrance. There's 12 feet laying on the ground. Then the 8 feet in buried conduit and five feet up to the transformer. A more common 80 or 100 feet of RG-8 style coax laying on the ground outside the house might produce a much higher interaction, and loss.

A reversal loss test is under design, with 80 feet of RG-213 laying on the ground running straight away from the vertical. But we lack some equipment items to pull it off. It will be a while before it can be done.

In the field it will be very hard for most stations to measure loss caused by a reversed connection, even a significant loss. It's tricky. To properly measure requires an accurate power meter, a continuously variable tuner, an RF analyst to display the (R ± jX) presented by the transceiver side of the tuner, an accurate, stable field strength meter, and a working knowledge of the various circumstances that can screw up field strength readings.

Reversing connections probably produces two obviously different (R ± jX) readings at the shack side of the transformer. See the last category in the table above. It is probable that a reversed connection has significantly hampered tuning and matching for installers.


Use your Back Button to Return to the Prior Page