Impedance matching (auto) transformer and common-mode choke for the JPC-7 dipole and other electrically-short (loaded) dipoles and verticals

Figure 1:
The JPC-7 loaded dipole out in the wild!
Click on the image for a larger version.

Loading coils and "electrically-short" antennas

It is well-known that you can make a "short" wire (e.g. one that is significantly shorter than 1/4 wavelength at the operating frequency) resonant by putting in series with it a coil.  There is no "magic" in this as the inductance of the coil, appropriately chosen, can completely cancel out the capacitance of the electrically-short wire, result being that at "resonance" we are left only with a pure resistance.

In an ideal situation, what we would be left with would be just the radiation resistance of this antenna and for such an antenna, this would mean that the feedpoint resistance would be less than 50 Ohms - probably much less!  In reality, the feedpoint resistance would really a combination of "ground" (counterpoise) losses, conductor losses of the antenna, and losses of the coil itself.

What this means is that if you have an electrically short antenna such as a loaded dipole or vertical with only a series loading coil tuned to resonance at the frequency of interest and no other matching scheme, its feedpoint impedance should be well under 50 Ohms on some bands if it is operating efficiently.

This is often not the case with portable antennas!

Figure 2:
The original stainless steel coil (top) for the
JPC-7 (and JPC-12) with the coil rewound with
silver-plated "jewelry" wire (bottom).
Click on the image for a larger version.

The JPC-7

Some time ago I wrote extensively about the JPC-7 (See the article, "Observations, analysis and field use of the JPC-7 portable "dipole" antenna" - LINK) where I discussed the bits and pieces comprising it:  I have used it in the field a number of times, finding it to work as advertised.

In short, this is a loaded dipole - at least on the lower amateur bands (especially 40 and 30 meters) that is intended for portable use:  On these bands (including 20 and 17 meters) it is physically shorter than 1/2 wavelength and it requires the adjustment of its series inductors to resonate.  On the higher bands (15 and above) its overall length approaches and exceeds a half wavelength meaning that it's a full-sized dipole and is (generally) tuned by adjusting the length of the telescoping sections.

Lossy coils!

There is a down-side:  As sold, it has loading coils that are wound with stainless steel:  As noted in the original article, these coils are very lossy, with MOST of the RF power being dissipated as heat on the lower bands (40 and 30 meters in particular - roughly an "S" unit of signal loss) where a fair bit of inductance is required.

Figure 3:
An example of heating of a stainless steel
loading coil on a short vertical - here, made by
Wolf River.  On 40 meters the temperature of
the coil rose by more than 30F (17C) with
60 watts of RF applied for 60 seconds.
Click on the image for a larger version.

The reason for this is that an electrically-short antenna (one that is physically short compared to the wavelength.)  The total length of the telescoping sections alone put together is about 198" (5 meters) - which is about 12.5% of a wavelength at 40 meters implies that the feedpoint resistance would, were there no loss at all, be around 8-10 Ohms, resulting in a VSWR of more than 4:1.

Calculations and measurements indicate that the approximate Ohmic loss of the original stainless-steel loading coil - if we optimistically presume it to have a Q of 47 - would be about 19 Ohms per coil (remember that there are two coils!) and the sum of the two coils would push feedpoint resistance near-ish 50 Ohms.  The result is that roughly 1 "S-unit" (about 6dB) is lost in the coils alone:  Contacts would still be made, but running a "compromised" antenna (e.g. physically small) that already would be less-efficient than its full-sized counterpart and adding another S-unit of loss doesn't sound like an optimal solution!

Using silver-plated copper "Jewelry Wire" (found on Amazon) to rewind the original loading coils dramatically improved the "Q" (approximately 200) and lowering the Ohmic loss to around 4 Ohms.  The result of this is that rather than something in the 40-50 Ohms for the feedpoint resistance, it dropped to "about 15" Ohms on 40 meters - a VSWR of around 3:1 - and even lower impedance than that (higher VSWR) when I reconfigured the antenna for 60 meters (e.g. added extra screw-together sections, moved the coils next to the feedpoint and added extra "drooping" wires to the ends of the dipole).  At the higher bands (20 meters and up) the feedpoint impedance is close enough to 50 Ohms that one can probably forego the auto transformer at all.

For more information about the "Silver-plated versus Stainless Steel" topic, see the blog entry "Rewinding the Stainless Steel coils with Silver-Plated copper wire on the JPC-7 and JPC-12 antennas" - link.

When a worse VSWR is a good thing!

The first thought when being faced with a higher VSWR on an antenna might be that it was made to be worse - but here is a instance where this is not the case.  As noted earlier, an electrically short antenna like a dipole or vertical can be made to be "longer" (from an RF standpoint) with the addition of a "loading" coil - but the job of the coil is to cancel out the capacitance of the, leaving only the resistive portion of the antenna's feedpoint impedance.

For a full-sized dipole or vertical, this resistance is "close enough" to 50 Ohms (perhaps 35-70 Ohms, depending on the antenna and its environment) to provide a decent load to a modern radio - even one without tuner.  But a very small antenna - where a lot more "coil" is required - will have a lower feedpoint resistance unless your coil is very lossy, as was the case with the stainless steel coils on the JPC-7.  With the lower-loss silver plated coil we (mostly) eliminate it as a lossy component - but end up with a different problem.

With a feedpoint resistance of 13-15 Ohms on 40 meters with the JPC-7 and silver plated coil and its resulting 3-ish:1 VSWR one can "fix" this with an antenna tuner to make the radio happy - and I have done this many times, placing the tuner (an LDG Z-11 Pro) right at the antenna (only a few feet/a meter of coax) but almost all common antenna tuners have quite high losses at these low impedances.

Testing with the cover of the tuner removed, I have noted that one ore more of its toroids in particular will run very warm with just 100 watts of power - Figure 4 shows the inside of this tuner showing one of its toroids discolored because of this.  Fortunately, iron-powder toroids are very forgiving of heating with very high Curie temperatures and other than cosmetic (e.g. discoloring the paint) moderate heating won't have any lasting effects as long as it remains intact (e.g. not cracked) and there aren't problems with (possibly-degraded) insulation between turns of the windings.

The other issue is that the balun originally supplied with the JPC-7 - intended for 50 Ohm operation - also got very warm, and after a bit more than a minute of continuous 100 watts at 40 meters the VSWR would start to rise due to its ferrite reaching the Curie temperature, causing the permeability to drop like a rock:  Essentially, the ferrite would "go away" when it got hot - likely not a problem on SSB or CW, but it might be on "key down" digital modes at full power.  This heating seemed to be more severe at the low impedances (below 20 Ohms) than at 50 Ohms.

Eliminating the tuner

Figure 4:
Inside the LDG Z-11 antenna tuner.  The center
toroid shows evidence of have been heated,
apparently due to matching very low "R".
Click on the image for a larger version.

By definition, we can remove the reactive component of the short antenna with the loading coil:  Its inductance will cancel out the capacitance of the antenna at resonance (which is the very definition of resonance) leaving only a pure resistance.  While an antenna tuner is able to cancel out capacitive and inductive reactance - or just pure resistance - we have a situation where, with a properly-tuned loading coil - we have only resistance and for that we don't need a tuner and we can use just a transformer, to change the impedance from whatever it is to 50 Ohms.

An easy way to do this is with an autotransformer.  This is a device with just one winding and in this case - where we are trying to tune to a feedpoint resistance lower than 50 Ohms - we can feed our power across the ends of the entire coil and tap it at various points along the winding to get our desired (lower) impedance.  For my application, having several taps between about 10 and 40 Ohms (plus the natural 50 Ohm feed impedance) would assure the ability to attain a VSWR of better than 1.5:1 for any purely resistive impedance between 7 and 75 Ohms.

The tyranny of the "electrically small antenna" and efficiency

It's worth noting several things about electrically-small low-band HF antennas - which includes portable antennas like the JPC-7, JPC-12 as well as mobile antennas - and how they interact with common antenna tuners (which an autotransformer is not):

• Any efficient, electrically-small vertical antenna will have a very low impedance once it is resonated:  For example, a "perfect", loss-less 1.5 meter (4.9 foot) long vertical antenna system on 40 meters would have a radiation resistance of about half an Ohm.
• Without losses due to the coils and stainless-steel telescoping rods, etc., the feedpoint resistance of the JPC-7 would, at 40 meters, be in the vicinity of 3-5 Ohms, depending on how many screw-together sections are used (e.g. the longer, the higher).
  
• Any automatic (or manual) antenna tuner that you are likely to ever use for portable operation will have rather poor efficiency when trying to match at lower than 20 Ohms or so - which translates to heat as demonstrated in Figure 4.

These facts - among others - conspire against having a small, efficient mobile antenna for the lower HF bands (e.g. 80-40 meters).  In the real world, losses (coil, antenna wire, ground) will conspire to make the feedpoint impedance much higher than the "less than an Ohm" that the would theoretically be - and any difference between the feedpoint resistance at resonance and the predicted radiation resistance is where most of the power in such an antenna system is lost:  In a typical antenna of this sort, the vast majority of transmitted power is lost in heat rather than radiated.

With significant efforts, it may be practical to get the losses of such an antenna system (which includes not just the antenna, but the series matching coil and ground losses an other factors) down to about 10 Ohms - still far above the 0.5-5 Ohms of our "perfect" antennas in the examples above - but as we know, physics conspires against us as trying to force-feed such an antenna with a tuner will probably put it into the impedance range where it is very inefficient.

It's worth noting that many simple and inexpensive mobile antennas achieve at least part of their "matching" to 50 Ohms simply by being lossy:  Most of the power is simply burned up in the coil.  This method is convenient in that it simplifies the problem with matching and is often accompanied by much wider tuning bandwidth (reducing the need to frequently re-tune when one changes frequency) than with our hypothetical "high efficiency" antenna, but the trade-off is poor efficiency.

Auto transformer for impedance matching

Another way to handle this is to simply transform (pun intended!) the impedance downwards from 50 Ohms - and one way that this could be done is with a transformer of some type - and the simplest of these is one with a single winding, called an autotransformer:  Such transformers are commonly used to match a random wire (9:1 matching to about 450 Ohms) and for end-fed half-wave antennas (49:1 matching to about 2450 Ohms) - but we can also efficiently transform the impedance downwards.  By designing appropriately, this transformer can be made to be very efficient.

It would seem that the use of an auto transformer for matching a low-impedance antenna - such as a low-band mobile antenna on a vehicle - used to be more common decades ago, but has fallen out of favor, possibly due to the easy and cheap availability of automatic antenna tuners:  Devices that do this function include the Atlas MT-1 (see Figure 5) and the Swan MMBX, both of which have a number of low-impedance taps. 

Figure 5:
The Atlas MT-1 autotransformer,  The variety of
taps available provide the possibility of achieving a 1.5:1
match to any resistive loads between 9 and 75 Ohms.
Click on the image for a larger version.

My initial thought was to use a ferrite toroid as the core for the auto transformer.  As a general rule of thumb, a transformer should ideally have an inductive reactance of about ten times that of the operating impedance at the lowest frequency (e.g. 500 Ohms for a 50 Ohm system) but, in a pinch, just three times the operating impedance (e.g. 150 Ohms for a 50 Ohm system) was "OK".  With this in mind I wound 7 turns on an FT140-43 toroid with multiple taps.  The inductance of this arrangement was about 45uH which correlates with about 1900 Ohms at 7 MHz - well above the target inductive reactance but it would have been difficult to achieve the multiple taps needed to attain the impedance steps with fewer turns.

This transformer - wound on ferrite - did not work well at all!  When testing it on the antenna, I could not achieve a sensible match and I quickly realized that the problem was due to leakage inductance of the transformer itself.  An ideal transformer would simply transform the voltage according to the tap's turns ratio, but any practical transformer will place some amount of inductance in series with the supposedly ideal tap, and it was likely this spurious series inductance (which needed only to be a few uH to make it "un-matchable") was totally messing up the attempt to tune the antenna, departing far from the ideal transformer at RF.

Measuring the self-inductance of the Atlas MT-1 confirmed this:  Its end-to-end inductance was about 2uH and the inductances between the taps and ground - the results of these measurements made using my HP-4275A LCR Meter (at 4 and 10 MHz - interpolated at 7 MHz) are as follows:

Tap marking (Ohms)	@ 4 MHz Inductance uH (XL Ohms)	@ 7 MHz (Interpolated) Inductance uH (XL Ohms)	@ 10 MHz Inductance uH (XL Ohms)
52	1.87uH (46.6)	1.8uH (79)	1.74uH (116)
23	0.95uH (24.4)	0.95uH (41.8)	0.95uH (61)
18	0.77uH (18.1)	0.75uH (33)	0.72uH (47)
13	0.61uH (14.3)	0.57uH (25)	0.53uH (35.8)

Figure 6:
The impedances (X_L ) of the taps on the Atlas MT-1 auto transformer versus frequency.

While "about 2uH" of inductance at 40 meters (7 MHz) doesn't fit the "3x reactance" rule-of-thumb (e.g. 79 Ohms X_L in a 50 Ohm system) it will still work OK, acting as a parallel inductance across the antenna - but the important part is that there will be a fraction of the leakage inductance compared to the version with the ferrite core mentioned above:  A small amount of this inductance would lower the resonance frequency slightly, but not disastrously so.

Figure 7:
The auto-transformer, wound on a T157-2 iron-powder
toroid with taps terminated with 2.5mm banana plugs.
Click on the image for a larger version.

Replicating the auto transformer

Rather than reinventing the wheel, I decided to (more or less) replicate the electrical properties of the MT-1 (and similar devices) and for this I chose a T157-2 Iron-powder toroid.  With a target inductance of "about" 2uH I wound 13 turns of 16AWG silver-plated PTFE (Teflon) insulated wire which should, in theory yield about 2.4uH - but when compressed together on the core it yielded about 3.6uH which correlates with about 158 Ohm at 7 MHz -  almost exactly 3x the 50 Ohm system impedance.

As can be seen in Figure 7, taps were placed at 6, 7, 8, 9 and 11 turns (from ground) by scraping the insulation off the side if the wire and tack-soldering wires to it providing impedance taps of approximately 11, 14, 19, 24, 36 Ohms - plus another wire across the 50 Ohm feed for the higher bands:  These impedances resulted from where the turns landed and it was convenient to attach taps rather than from any attempts to obtain specific or precise impedances:  After construction, I labeled the leads with the approximate impedances - for obvious reasons!

I used five taps to allow a selection of an impedance to be able to obtain about 1.25:1 VSWR or better, but if I were happy with just 1.5:1, I could have chosen fewer taps in the manner of the Atlas MT-1 discussed, above.

As the impedance of a tap is related to square relation of the number of turns (e.g. twice the number of turns results in 4x the impedance) there's a pretty simple formula to follow to calculate the impedance of a tap:

Z_tap = (Z_sys) / ((Turns_total/Turns_tap)²)

Where:

Z_tap = Impedance of the autotransformer tap

Z_sys = System impedance (typically 50 Ohms)

Turns_total = Total number of turns on the autotransformer (13 turns in our example)

Turns_tap = Number of turns from the bottom (ground) end of the autotransformer to the tap

In other words:

 Z_tap = (50) / ((Turns_total/Turns_tap)²)

Taking our 13 turn autotransformer as an example, we can calculate the impedance at any turn.  Taking the 8th turn as an example:

Z_tap = (50) / ((13/8)²)  therefore,

Z_tap = 18.9 Ohms 

Or, if you know the desired target impedance and want to calculate the turn on which to make that tap, here's the above formula rewritten to solve for it:

Turns_tap = Turns_total / √(Z_sys /Z_tap)

I also included a "50 Ohm" tap (which is connected at the "top" of the transformer, across all of the windings) so that I could still use the common-mode choke (described below) even when operating on the higher bands (20 meters and above) where the natural impedance was close enough to 50 Ohms that I probably wouldn't have needed the autotransformer for impedance transformation, anyway.

At the end of the flying leads are 2.5mm "banana" plugs - which plug in to the feedpoint of the JPC-7.  These allow the selection of taps on the auto transformer which permits the VSWR to be minimized for those bands for which the feedpoint impedance is significantly lower than 50 Ohms:  A bit of care is required to prevent the "floating" banana plugs from touching each other (or anything else metal) but this isn't actually much of a problem.

Initial testing using a kludge of clip leads, I verified with my NanoVNA that the auto transformer worked as it should (e.g. I was able to attain less than 1.5:1 VSWR on 60, 40 and 30 meters) and almost as important, the tuning with the auto transformer was only slightly different from that using the original balun indicating that the leakage inductance of the auto transformer was not much different than that of the originally-supplied balun.

Adding a common-mode choke

Feeding a dipole (which is a balanced antenna) with coaxial cable has the inherent hazard of RF appearing on the coaxial cable feedline due to the symmetry of the antenna.  Excessive RF on the feedline can result in a "hot" rig - that is, RF energy appearing on the chassis of the radio as well which can result in distortion (RF getting into the microphone) and/or malfunction of peripherals (outboard keyer malfunctioning, USB interfaces crashing, interference to the sound card) and out "in the field" where one may not have an elaborate ground system already, this may be more likely than at home.

Figure 8:
The auto transformer (left) plus a common-mode coaxial
choke (right).  The choke is wound on an FT140-43 ferrite
toroid.  Both toroids are in the foreground for comparison.
Click on the image for a larger version.

The "input" to the auto transformer is simply the opposite ends of its 13 turn winding which would normally be soldered to an RF connector.  Rather than doing that, I soldered it to a 36" (91cm) piece of RG-316 PTFE coaxial cable - the shield going to the "bottom" (ground) side of the auto transformer, insulating the connections with adhesive-lined heat-shrink tubing.  The rest of this RG-316 was wound on an FT140-43 toroid yielding 13 turns using the "cross-over" technique where about half of the turns are wound on the opposite side of the toroid:  This method is said to (slightly) increase the series choking impedance at higher frequencies (e.g. 15 meters and up).

Not having a UHF connector designed for RG-316 on hand, I used a crimp-type PL-259 intended for RG-58.  I stripped more than usual of the jacket from the end of the coax, folding the shield over the outer sheath.  Using some PTFE tubing and part of the jacket stripped from the coax itself I was able to increase the effective diameter of the inner dielectric.  Assembling the cable - remembering to include the ferrule and pieces of adhesive-lined heat shrink - I was able to fold the outer shield over the ferrule after a bit of tugging on it to increase its inner diameter.  At that point, I was able to crimp the ferrule into place, securing the coaxial cable firmly.

Figure 9:
The auto transformer with the common
mode choke on the JPC-7's feed.
Click on the image for a larger version.

Since RG-316 is fairly small (it's the same size as RG-174) - and because the weight of the connecting coaxial cable and the common-mode choke itself would be hanging from the cable - I protected the connector with several pieces of adhesive-lined shrink tubing - using a smaller piece just behind the connector to increase its outside diameter and then a larger piece over the ferrule, onto the previous piece of tubing.

 

Not content with this, I wound several turns of "miniature" paracord (1.15mm diameter) onto the ferrule and tied it securely, feeding both free ends underneath yet another piece of heat-shrink tubing that was then installed over where I'd tied the paracord - taking careful care not to damage the cord when applying heat to shrink it.

These two strands of mini-paracord were then counter-wound over the RG-316 as can be seen in Figures 8 and 9 and were tied to the ferrite core of the common-mode choke such that when hanging, the weight of the connector was on the cord and not the coaxial cable:  I did a similar thing between the core of the auto transformer and the balun to prevent the cable itself from being pulled.

Putting it on the antenna

Figures 8 and 9 shows the combination auto transformer and common-mode choke at the feedpoint of the JPC-7 loaded vertical.  As noted earlier, testing showed only a slight difference in tuning between the lowest VSWR achieved with the original 1:1 balun and the transformer-choke combination indicating that its effect was minimal:  As figure 10 shows, transmitting 100 watts on 40 meters also resulted in only very slight heating of the auto transformer - certainly a much lower amount of signal loss than that which resulted in the heating and discoloring of the toroid in the antenna tuner pictured in Figure 4.

Figure 10:
Thermal infrared view of the autotransformer
(top) and common-mode choke (bottom)
after 60 seconds key-down with 100 watts
on 40 meters.  The temperature of the
autotransformer increased only by about 2F
(1C) while the common-mode choke got about
10F (6C) warmer.
Click for a slightly larger version.

Testing the common-mode choke

 

The efficacy of the common-mode coaxial choke was also verified:  Without it, grasping the shield of the coaxial cable with one's hand would result in slight detuning of the antenna, but with it, there was no detectable effect - and there was no detectable amount of "hot rig" due to the presence of common-mode currents flowing beyond the choke and onto the radio's chassis - even without the use of a counterpoise/ground wire.

 

The presence or lack of effect of the change of antenna tuning when body capacitance is introduced is a simple - but effective - means of determining the presence of RF current on the feedline at the point where it is grasped.  Figure 10 shows that this core heated only minimally - also indicative of low loss.

Does it work?

I have put this configuration pictured in Figure 9 on the air several times since assembling it on 60 through 15 meters.  As expected, the best match on 60 meters (<1.5:1) required the 11 Ohm tap while 40 meters seemed fine with either the 11 or 14 Ohm tap.  20 meters, on the other hand, found the best match using the 36 Ohm tap while 15 meters worked well with either this or the 50 Ohm tap.  Again, the heating of the autotransformer at 100 watts was also minimal on any band - even on the 60 and 40 meters where the losses would probably have been the highest.

Conclusion

The use of an autotransformer rather than an L/C antenna tuner is a time-honored means of matching an "electrically-short" antenna, so what has been presented is nothing new - but it may be "new" to some of the readers.  For a portable antenna such as this, its size and relative simplicity can't be beat as it's far smaller than any antenna tuner that could handle 100 watts at the low impedances that may be presented - and it's certainly lower loss as well!

The only "complication" is that which is already intrinsic to this type of antenna:  As this is a dipole, there are two elements - each with its own coil and telescoping rod making it a bit "fiddly" to tune, something best done with a VNA or antenna analyzer.  With this antenna I keep a card that is marked with the physical locations of the tap positions of the two coils for the various bands:  These are held up to the coil and the sliders adjusted, quickly getting "close" to a match with the analyzer used to do any final touch-ups on the tuning.

* * * * * * *

Related pages:

• For more information about the "Silver-plated versus Stainless Steel" topic, see the blog entry "Rewinding the Stainless Steel coils with Silver-Plated copper wire on the JPC-7 and JPC-12 antennas" - link.
• Analysis, observations about the JPC-7 loaded dipole - link.
• Analysis, observations about the JPC-7 portable vertical - link.
• Reducing RF susceptibility for the HamGadgets "Ultra Pico Keyer" - and mimizing RF issues on portable HF stations in general - link.

* * * * * * *

This page stolen from ka7oei.com

[END]

 

Hiking and POTA (Parks On the Air) operation from Arches National Park (US-0004)

Figure 1:
Double-O Arch
Click on the image for a larger version.

Earlier this month I had the opportunity to spend nearly a week in Arches National Park in south eastern Utah.  As the name implies, there are a lot of natural arches scattered throughout the area, the result of erosion occurring over millenia, the semi-porous sandstone eventually yielding the the inexorable forces of wind and water.

These trips to red rock country are not new to me:  I've been visiting this part of the state for more than 30 years now, having spent more well over six months cumulative camping, hiking and backpacking in this austere land.  On most of these trips, I have been in the company of other amateur radio operators - and that was the case here, as well.

Compared to other national parks in Utah, Arches is fairly small - on a few 10s of miles long and narrower than this in width.  Compared to some places I go, it's a bit of a "tourist" park meaning that it's fairly crowded with comparatively few developed trails concentrated in a few areas.

Figure 2:
Fins and more fins, backgrounded by the La Sal mountains
as seen along the "primative" trail.
Click on the image for a larger version.

When going to such places, I tend to do about as much hiking as I can - but Arches is comparatively limited, but one of the longer trail systems is that associated with Landscape Arch and Double-O arch.  On this hike I took the "primative" trail, separating me from the madding crowds - a much longer route over occasionally rugged terrain, occasionally requiring a bit of scrambling up or down slick rock:  Just the way I like it!

Over the course of a few hours I made my way from the campground to Double-O Arch where I met the rest of my group who'd taken the other trail where we stayed for a while before splitting again and heading back.  Altogether, I managed about 10 miles (16km) or so by the time I got back to camp.

With temperatures in the mid 80s (about 30C) I set up my radio in the shade of my tent and shade and started operating.

* * *

Equipment:

Antennas:

The evening before, I had a bit of extra time around dinner and I took that opportunity to set up my portable antennas in the cooling evening air.  For this POTA operation, I eventually set up two antennas - the first one being my JPC-7 loaded dipole.

Figure 3:
Operating CW in the shade, on a portable table, using
a cast iron frying pan to keep the paddle in place.
Click on the image for a larger version.

I've discussed the JPC-7 antenna on this blog before (LINK) - and have used it for several POTA operations already with good results.  Since the last POTA operation I'd rewound the loading coils, replacing the original stainless steel wire with silver-plated copper to reduce the losses - I discuss the details about this HERE.  It's difficult to estimate how much improvement this change made, but it's likely in the general area of 3dB or so - only 1/2 "S" unit or so, but it's certainly worth a bit of hassle to improve efficiency on an already-small antenna.

A day after setting up the JPC-7, I also set up the JPC-12 vertical antenna (described here).  This antenna, too, has been refitted with a silver-plated loading coil as well:  With a few extra mast sections, a top-hat and resonant, elevated radials it also makes for an excellent portable antenna - albeit a bit more complicated to set up than the loaded vertical, particularly when changing bands.

Radio and power:

The radio - an older Yaesu FT-100 (with the CW filter from an FT-100D) which was powered by a 100 amp-hour Lithium-Iron Phosphate battery using a paddle from cwmorse.us - (link).  I've used this particular paddle ("Outdoor pocket double paddle with magnets") for several POTA activations and as before, I've used the same cast-iron fry pan for all of them to keep the paddle from sliding around - often ending up with a bit of soot on the side of my hand and wrist!

Figure 4:
The antennas - and solar panel.
There was no audible interference from the
now-modified solar controller.
Click on the image for a larger version.

Operating (mostly) on 20 meters I managed to make about 285 contacts - all but four of them CW with 277 of them counting as POTA contacts.  The operating position was almost as POTA as one gets:  Sitting in a chair, under a shade, surrounded by sand and red rock.

Mixing antennas with solar - with no QRM!:

 Figure 4 shows the "antenna farm".  In the foreground - just left of center - is the JPC-7 loaded dipole, using a studio tripod for support while in the background - to the right of center - can be seen the JPC-12 vertical with tophat.

Also in the foreground is a 200 watt solar panel - but you may be wondering if this would cause QRM (interference) from its controller:  The answer is NO - but this is only true because I've done previous work to add extra filtering to it.  Even with the antenna (particularly the JPC-7) right next to the solar panel with its controller, I could not hear any interference at all - but this is by design as I have taken steps to make it quiet, and you can read about the details to accomplish this HERE in a previous blog entry.

At this camp site there were two other PV systems in operation located some distance away from the antenna, but I could hear those.  For the one closest, I happened to have an FT240-43 toroid on hand and I was able to cram five turns (with connectors) of the cables from the two panels feeding it:  Predictably, this reduced the QRM somewhat (1-2 S-units) - but as noted in the blog entry noted above, ferrite alone will not likely solve such a QRM issue!

Figure 5:
Red and green auroras backgrounding the big dipper.
Click on the image for a larger version.

The "other" PV system - which was even further away - caused minimal interference so nothing was done about it - but since I'd used my only FT240-43 toroid, I wouldn't have been able do anything about it, anyway.

Red Rock + Aurora = More red!

As it happened, the sun did a bit of burping in the days leading up to and during this trip, the result being the repeated appearance of a visible aurora, the first appearing on October 7 when very visible red pillars appeared in the northern sky:  Scrambling to the top of a nearby bluff, we could see a bit of red and green in the sky along with the Big Dipper.

For the next few days we noticed something else:  On the first night, the sky was spectacularly dark - the Andromeda Galaxy being visible - but on the night of the first aurora and for a few nights thereafter it seemed as though we lost a lot of the "deepness" of the sky.  We also noticed that despite the lack of moonlight, we could see the surrounding landscape and make out large objects on the ground without needing additional light.

Figure 6:
Sky glow, lighting up the camp and environs.
Click on the image for a larger version.

We eventually realized that what we were seeing was sky glow.  In other words, the entire sky was glowing dimly:  Not bright enough to be perceived as color, but the cumulative glow of the entire sky was enough to illuminate the landscape in that odd way.

A few days later the aurora was clearly visible again - and that's when the photo in Figure 6 was taken, showing a bit of red behind the clouds to the north and some green glow on the northern horizon.

* * * * *

This page stolen from ka7oei.blogspot.com

[END]

Rewinding the stainless steel coils with silver-plated copper wire on the JPC-7 and JPC-12 antennas

Portable antennas (verticals, loaded dipoles) typically use coils on the lower HF bands to make them electrically "larger" to alow them to be resonated at frequencies well below their physical size - but what about losses in those coils?

While it's "traditional" to use copper wire wire for these coils, there are a number of modern offerings that use stainless steel - and both types have their cheerleaders and detractors, so what's the deal?

Figure 1:
The JPC-12 vertical in the field.

Note:  This post refers to previous entries on this blog about the JPC-7 and JPC-12 antennas that are relevant to this discussion, namely:

• JPC-7 loaded dipole antenna - link.
• JPC-12 portable vertical antenna - link.
  

While some details in this article are specific to these antennas, the general observations may be applied to any HF antenna using loading coils.  I have not (yet?) done A/B field tests with antennas using different (stainless vs silver plated/copper) coils and/or simulations - perhaps a topic for a future blog entry?

* * * * *

In the previous posts I have discussed the JPC-12 vertical and the JPC-7 dipole:  To make either antenna usable at frequencies lower than their natural resonance, inductance is required (the "loaded" part) to achieve resonance at the desired frequency - and for their lowest operating frequency - 40 meters - it takes a fair bit of "loading", indeed.

For this, the JPC-7 dipole, which has a "coil-less" resonance of around 22 MHz, has two coils with adjustable taps - one for each element - a slider being used to adjust the amount of inductance:  Higher inductance = lower frequency.

The JPC-12 vertical - made by the same folks - unsurprisingly uses the exact same coil as the JPC-7 - and for the same reason:  To add inductance to make the electrically-short element - a radiator of approximately 150" (381cm) total length (resonant around 18 MHz without any added inductance and using the originally-supplied components) offer a semblance of a match on lower bands.

Having the coil in common, they also share the same trait:  Loading coils wound with stainless steel - and since, when running on a lower band like 40 meters - all of these coils run quite warm at nominal transmitter power (100 watts or so) there are definitely power losses in the coil - but how bad is it?

Wanting to answer this question, I ordered an extra coil from the seller from which I'd bought my JPC-7 and JPC-12 antennas and with that - and the three that came with the two antennas originally - I now had four coils - enough to do direct A/B comparisons on both antennas when I rewound two of them with silver-plated wire.

Why stainless?

The coils originally supplied with the JPC-7 and JPC-12 are wound with 1mm diameter (18 AWG) stainless-steel wire.  Fortunately, an austenitic (non-magnetic, as checked with a neodymium magnet) type of stainless steel is used:  If this wire been magnetic at all things would be much worse in terms of loss.  While the 1mm diameter stainless steel wire is very rugged physically, the fact that it is stainless steel means that its resistance is quite high compared to copper - in this case the end-to-end DC resistance is about 4 ohms, but the RF resistance, taking the "skin effect" into account, is likely to be very much higher.

Using Owen Duffy's online skin effect calculator (link to archived page) and assuming 1mm diameter, 316 Stainless, the 4 ohms of DC resistance translate as follows to RF resistance including skin effect:

• 3.5 MHz = 5.2 ohms
• 7 MHz = 7.2 ohms
• 14 MHz = 9.6 ohms
• 28 MHz = 13.6 ohms

These values would be for the entire coil, but since one uses slightly less than the full number of turns of the coil to resonate at 40 meters, the losses should be lower - but the message is clear:  The less of the coil that you need to use, the lower the loss.   The total length of 1mm wire is estimated to be about 180 inches (457cm) and  by comparison, copper wire of this same diameter and length would have a DC resistance of about 0.1 ohm - or, according to Owen Duffy's calculator, a skin effective resistance of 2 ohms at 28 MHz.

Why stainless steel, then?  Obviously, stainless steel won't oxidize/corrode like many metals - and it may be that in quantity, stainless steel wire is less expensive than silver plated/copper, but in this case I believe that there's another reason.  Other manufacturers of portable antennas (Wolf River, for example) advertise the use of stainless steel for some their coils as well, extolling the virtues of the material in regards to its inability to corrode - but I'd be surprised if such corrosion is likely to be the main reason for a hypothetical copper coil's losses in an electrically-short antenna that would make it worse than stainless.

I suspect that the "advantage" of a stainless steel coil is, in fact, related to the fact that it is lossy.  As portable antennas - when used on the lower HF bands - are necessarily smaller than their full-sized counterparts, their radiation resistance will be commensurately lower and this means that the feedpoint resistance may be lower as well when fed with simple matching schemes such as a series coil.

What this means is that rather than somewhere "around" 50 ohms, the feedpoint impedance of an "electrically small" antenna (e.g. one that needs a coil) when using a very low-loss coil may be much lower, resulting in an "unacceptable" VSWR (e.g. >2:1) at resonance:  While this would actually imply greater efficiency due to lower loss, it's "inconvenient" to the user.  While a more versatile means of matching the antenna is possible (multiple coil/capacitors such as a simple antenna tuner or the use of an autotransformer) this complicates construction and operation and would surely increase cost.

As implied earlier, another method of dealing with low feedpoint impedances is to add series resistance to raise it to something closer to 50 ohms to make radios (and their operators) "happy" - but an ohmic resistance in the signal path (say, the use of stainless steel) means power loss, and power loss means heat!

How hot is it?

Figure 2:
The original loading coil (lower) wound with stainless wire as
seen with a thermal infrared camera.  After 60 seconds at 75
watts (on 40 meters) the coil temperature rose by 110F (61C)
from the ambient 53F (12C) to about 166F (74C)!
Click on the image for a larger version.

I've operated both the JPC-7 and JPC-12 antenna a number of times in the field on the "lower" bands of 40 and 30 meters at 100 watts, using both CW and SSB, and observed that in each case, the coil gets "hot".  As the coil forms are (apparently) molded nylon, this is nowhere near the likely softening point of more than 300F (150C) - and being open to the air to allow convective cooling, and using a mode where the duty cycle is intermittent certainly helps prevent a "meltdown".  (Compared this to PVC - which has a softening temperature in the area of 140-180F or 60-80C)

As a test, I put both the original stainless steel and the rewound silver-plated coils in series on the JPC-12 vertical, putting a jumper across the coil not under test.  I then transmitted 75 watts into the JPC-12 vertical for 60 seconds and measured the temperature of the coil with  both an infrared thermometer and thermal camera, noting a temperature rise of about  110F (61C) - still not hot enough to risk melting the coil form, but certainly enough to dissuade one from running a 100% continuous mode like SSTV, RTTY or other digital modes on a hot day!  (Note:  On a hot day a temperature rise of 110F/61C may well be enough to soften a PVC coil form.)

The picture in Figure 2 - taken with a thermal infrared camera - shows the heat produced when testing with the JPC-12 vertical.  (Note:  During this test I swapped positions of the two coils to see if there was much difference in the current/heat of the stainless coil owing to differences in current distribution, but as expected, there was not.)  Similar results were observed when operating SSB and CW on the JPC-7 loaded dipole.

At this point I should make something clear:  The reader should not presume that the use of a stainless steel coil is going to result in an antenna that doesn't work, but rather it implies a degree of loss of efficiency.  As I've made many contacts with both the JPC-7 and JPC-12 in their original form, I know that it's perfectly capable of usable performance - but how much better would it be if we were to address coil losses?

Also, once I had seen the loss in the coil, I couldn't "un-see" it and I had to do something about it.

Choice of wire

In order to minimize losses in an electrically-small antenna it is important to reduce resistive losses and the loading coil and reducing the generation of heat produced by it is a good place to start - and copper wire is an obvious choice.  Knowing that the wire used is 1mm diameter - about 18 AWG - there were a lot of choices:  I had some enameled 18 AWG wire already on-hand and I could easily have obtained some tinned 18 AWG "buss" wire as well.  Finding bare copper wire was a bit more difficult, but since we need only make contact on the ends and along the slider, there's no reason for the entire coil to be bare and thus be subject to oxidization:  If I needed to do so, I could have wound the coil with enameled wire and then selectively remove the insulation along the path of the inductor's slider with fine sandpaper.

On a hunch, I did a search and quickly found on Amazon some 1mm (18 AWG) "Silver plated" copper wire of the same diameter described as being used for jewelry - a small spool costing about US$15 with more than enough wire to re-do three of these coils. ^{Footnote 1}

Figure 3:
The coil - still with the stainless steel wire.  On the left end of
the slider (the "top") of the coil can be seen the insulator.
Prior to disassembly move the slider to the end opposite the
insulator (maximum inductance) as shown.  When removing
or installing the Allen screw, keep a firm grip on the end with
the insulator to prevent it from rotating and damaging the
insulator itself or the end of the rod that protrudes into it.
Click on the image for a larger version.

The use of silver-plated wire is traditional in RF devices as it has the advantage over copper wire in that as it oxidizes, the result (e.g. silver tarnish) is still a conductive substance, much better than copper oxide - and compared to bare copper it is less (chemically) reactive overall - plus the coil looks very nice!

Rewinding the coil:

The coil form itself - with molded grooves - is quite rugged and lends itself very well to being rewound by hand.  Using a silver-colored "Sharpie" I noted where the original coil's windings started and ended.  I would also recommend using a ruler to take careful measurements of both the start and end points of the coil as well as taking a photo of it - particularly if you are rewinding the coil of a JPC-12 vertical and do not have a second coil as a comparison.

It is also important to note that one end of the slider is insulated to prevent the shorting the unused turns of the coil itself - something that would surely reduce "Q" and overall efficiency:  It is important to reinstall the slider assembly in the same orientation as before to put the insulated end of the slider rod on the "top" (e.g. the side closest to the top of the vertical or end of the dipole).

When rewinding, first move the slider to the end farthest away from the end with insulator on the rod (e.g. the "bottom" of the coil, with the stud protruding) and cover the spring contact with a bit of tape to keep it with the slider body:  This moves the slider - and the contact spring - well away from the end of the wire that we are going to remove first.  Using an Allen wrench, carefully remove the screw holding the end of the slider bar with the insulator (e.g. the part at the top of the coil, with the female threads):  The end of the wire is tucked under the supporting post and the screw itself goes into the brass slug at the center of the coil with the M10 threads used to assemble the rest of the antenna.  Keep tension on the hardware with a finger as you undo this to minimize the possibility of it being launched across the room.

Figure 4:
This shows the end of the new wire looped around the screw
and the post tightened down to hold it in place as it is wound.
A blade screwdriver is used to push the wire into the groove
below the slider boar to keep it from jumping out of the slot.
Be sure to start the wire in the same place as the original coil.
Click on the image for a larger version.

At some point, the coil of stainless steel wire will unwind itself rather forcefully when it slips out from under the screw (it may be a good idea to wear glasses) as it is under a fair bit of spring tension:  Even if you are prepared for this to happen, it can be startling!  At this point be sure that the contact spring is still on the slider block:  If it is not, look for and find it now!

With the tension released, remove the other end of the slider bar.  At this point, carefully remove the slider bar from the insulated end so that you have just the support post and set the rest of it aside.  At this point you'll have a loose coil of stainless wire to set aside.

Take the end of the new wire and using a pair of needle-nose pliers, bend a loop to go around the screw for the support post and using (just) the support post that was insulated for the slider, secure it in place, under the post.  Lay the wire in the groove and at the point where you marked the coil to begin then push the wire into the shallow slot above which the slider moves to hold it in place.

Figure 5:
As the wire is wound, keep pressure on the wire and coil form
with a thumb while rotating the form itself, forcing the wire to
drop into the molded slots.  Continue winding until you get
to where you had previously marked the end of the original
coil - but there's no harm if you add one extra turn.
Click on the image for a larger version.

Keeping the wire under tension - and using a thumb as necessary to hold that tension and push it onto the form - tightly wind the wire onto the form, making sure that it drops into the wire slots.  When you get to where you marked the end of the coil (you can go one extra turn if you like!) push the wire into the slot again (to help hold it in place) and - leaving enough extra to go around the screw on the bottom of the coil - trim it off.  Before putting a loop in the end of the wire to go around the screw, again use a blade screwdriver to push it into the groove to help hold it into place.

At this point I temporarily wrap a the loose end of the coil with a bit of electrical tape to keep it from unraveling while I loosen the post at the top of the coil and align it carefully so that I can plug the slider bar back in and re-mount it and the other post at the bottom of the coil, torquing the screws firmly and being careful to prevent the post with the insulator from twisting as this is done.

Figure 6:
The finishing end of the coil with the wire looped under the
slider rod support and tightened down.  In this picture you
can see how the wire has been pushed into the groove, under
the slider.  To the left of the end of the wire can be seen the
blob of adhesive used to lock the end of the coil into place.
Click on the image for a larger version.

Now, the coil has been successfully re-wound.  While it may not be strictly necessary, I put a dab of "Shoe Goo" - a thick rubber adhesive - on the top and bottom 2-3 turns of the coil near where the wire drops into the slot and connects to the post to "glue" it into place, making sure that it doesn't jump out of its slot.  If you don't have "Shoe Goo" or something similar, some RTV ("Silicone") can work as can epoxy - but cyanoacrylate and polyurethane glues (e.g. "Super" and "Gorilla" glue, respectively) may not work very well - and "hot melt glue" are definitely not recommended as either will likely break loose their bonds across a wide temperature range and changing mechanical stress. 

The trick here is to bridge several turns of wire with the adhesive to lock them into place together as much as adhere them to the coil form.

Results

Figure 7:
The coil rewound with silver-plated wire (upper), under the
marker.  As can be seen, the temperature rose by about 3F
(less than 2C) above the ambient temperature of 53F (12C)
after 60 seconds of key-down on 40 meters at 75 watts.
Click on the image for a larger version.

As expected, the use of lower-loss wire for the coil results in a dramatic reduction of generated heat which no doubt corresponds with an improvement in overall antenna efficiency - The "after" picture (Figure 7) of the coil using the thermal camera after 60 seconds of transmission on 40 meters with 75 watts shows the difference.  As in Figure 2, the original stainless steel coil is on the bottom, but it is the one that is jumpered, putting all of the RF energy into the upper (silver-plated) coil, instead.

Touching the coil immediately after the 60 second key-down, the loss-related heating of the coil wound with silver-plated wire was barely perceptible - a far cry from the original stainless-steel wound coil that was  "hot"!

Electrical comparison of the stainless and silver-plated coils

For capacitors and inductors, one measurement of their departure from the ideal is their "Q" (e.g. "Quality Factor") and for inductors, the majority of this is likely to be the radio of the inductive reactance of the coil (X_L) to its ohmic resistance.  I decided to measure the unloaded "Q" (Q_u) of the original stainless steel loading coil and the rewound silver-plated coil.

To do this I used a NanoVNA and the method described in W7ZOI's article "The Two Faces of Q" (link) under the section called "Measuring Resonator Q":  I used both methods (#1 using parallel L/C and #2 with L/C in series) to determine the "Q".

Using method #1, for the "C_c " capacitors I used two 1pF NP0 capacitors in series each (0.5pF) which resulted in a 35-45dB through loss at resonance.  I put a high-quality 27pF silver mica capacitor in parallel with the coil under test and measured the -3dB response of the resonance curve.  In this test I set the variable inductor to the mark indicating tuning for 40 meters (around 22 uH) which, with the 27pF capacitor, yielded a resonance in the area of 6.6 MHz for each of the two coils being tested

Assuming that the Q of the series silver mica capacitor (C_o) is 1000 (a mediocre value - it's probably a bit higher) the results were:

• Original stainless steel coil unloaded Q_u:  47
• Rewound coil (silver-plated wire) unloaded Q_u: 199

I then used method #2 (with L/C in series) and got:

• Original stainless steel coil unloaded Q_u:  47
• Rewound coil (silver-plated wire) unloaded Q_u: 221

At the risk of being accused of "cherry picking" my results, I'll note that for high "Q" values and where the value of C_o is quite small, method #1 is less forgiving in terms of variances and minor losses in the test fixture, so we'll use the value from method #2.  The reader should also note that with a higher Q, deficiencies in the test measurement and effects of the coil itself will result in lower than actual Q_u (e.g. you will not get an erroneously higher value of Q) so it is likely that even the higher reading from method #2 on the silver-plated coil is, itself, a bit conservative.

Note:  During testing I observed that just laying the coil on my wooden workbench lowered the Q of the silver-plated coil significantly (15-20%) so all readings were taken with both coils held about 12" (25cm) above it.  I think that there is likely some effect of free-space capacitance that is reducing the reading so I suspect that the "actual" Q_u of the silver-plated coil is higher, still.  This same effect was extremely small with the stainless steel coil, further indicative of its lower Q_u.  

Comment:  It's worth mentioning that with higher "Q" coils, the physical aspects of the coil itself - namely the ratio of the length versus diameter, spacing between turns, material of the coil form, increasingly affect the Q - both for reasons of geometry (which can affect the amount of wire needed and degree of mutual coupling) and less obvious parameters such as distributed capacitance, etc.

Taking these Q_u measurements at face value, we can calculate the approximate "R" (resistive) loss of the two coils using the general formula:

• Q = X_L  / R

Or the more general form, knowing the inductance:

• Q =  2π f L / R
  

And rewriting this equation for R we get:

• R =  2π f L /Q

So, for a frequency of 6.6 MHz (which should be representative of 40 meters) and an inductance of 22uH, X_L is approximately 912 ohms, so for each of the two coils the apparent "R" value - which would be a combination of conductor loss and skin effect resistance we get:

• Original stainless steel coil:  R= 19.4 ohms
• Rewound coil (silver-plated wire):  R=4.1 ohms

The reader should be reminded that for ideal components, at resonance the reactance of the inductor is losslessly canceled out by the reactance of the capacitor so what we are left with - the value "R" mentioned above - will be the ohmic (conductor loss + skin effect) losses of the materials.  This also means that the "R" value will be added to the feedpoint resistance - and the effect of this "R" value is to lose power as heat as we will see below.  It is not lost on me that the loss values appear to be far higher than those obtained from Owen Duffy's calculator if one presumes skin effect to be the main source of loss - which we know is not going to be the case. 

The ohmic loss mentioned above is not going to be the only source of loss in a real antenna system:  In the case of a vertical, the "ground" losses (ohmic loss of radials, dirt, etc.) and with any antenna, the materials from which it is constructed (wire, telescoping rods which are themselves stainless steel, any balun being used, etc.) will come into play - and for an "electrically small" antenna such as either the JPC-7 or JPC-12 on 40 meters, will dominate and probably be the main points of loss besides the coil.

This goes to show how - in either case - doing anything to physically "embiggen" the size of the antenna - such as making the elements longer (adding drooping wires to the loaded dipole, adding a tophat to the vertical) will reduce the amount of inductance needed and increase the radiation resistance - both things that will contribute to improved efficiency.

With the stainless coil, it gets worse the lower you go!

Out of curiosity I re-did the Q_u measurements using a 270pF silver mica capacitor - which lowered the resonant frequency to about 2.2 MHz - and got the following results using method #2: 

• Original stainless steel coil unloaded Q_u: 29
• Rewound coil (silver-plated wire) unloaded Q_u: 277

Given the lower frequency and lower skin-effect losses I fully expected the loaded Q_u to be slightly higher - which is true for the silver-plated coil - but initially I did not expect the Q_u to go down on the stainless steel coil so I re-did the measurement using method #1 and got about the same results (to within a few percent) - but in retrospect, I realized that this was to be expected.

As Q_L can be defined as being the ratio between inductive reactance ( X_L ) and skin effect and ohmic resistance (R), if "R" remains pretty high and X_L lowers with frequency, the "Q" will be lower:  Since the resistance of the stainless steel wire is so high to begin with, it figures significantly in the reduction of Q and thus the losses incurred.

In perusing the forums in the back-and-forth discussions about stainless steel versus silver-plated coils, people have observed a "hotter" coil at the lower frequencies.  At first glance, this makes sense since lower frequency = "more coil" = more lossy wire - but the fact that - at least at HF - the Q of the stainless coil goes down significantly with frequency makes it even worse! 

Update² (12/24):

As an update to a previous update, I  recently did very careful measurements using both my HP-4191A RF Impedance Analyzer and my HP-4275A LCR Meter:  These are both the sorts of instruments used by manufacturers of electronic components - including inductors and capacitors - to measure the characteristics to put into data sheets.

Figure 8:
Coil under test on the 4191A
Click on the image for a larger version.

The '4275A only has discrete frequency measurements - 1, 2, 4 and 10 MHz being the relevant frequencies here - while the '4191 is more tunable and designed for for higher frequencies - up to 1000 MHz.  After using both instruments, it's my impression that while both are quite good, the '4275A is a bit better suited for measurements below 10 MHz and unlike the '4191, it can go well below 1 MHz - down to 10 kHz, in fact. For the measurements below I did check at the lower frequencies (1, 2, 4 and 10 MHz) with the '4191 and noted that the inductance values were very close (within a few percent) of those of the '4275A and that the "Q" values were typically within 10-15%.

I used both the HP-4191A and the HP-4275A at frequencies 10 MHz and below and the '4191A up to 30 MHz for both coils and here are the results with both coils set to the paint mark indicating the setting for 40 meters using the stock antenna hardware:

Frequency (MHz)	Inductance (uH) (Ag-plated)	Q (Ag-plated)	Inductance (uH) (Stainless)	Q (Stainless)
1	16.77	128	21.4	41
2	16.70	185	18.9	28
4	17.0	272	18.8	28
10	19.0 *	199 *	21.4 *	41 *

As mentioned earlier, the '4275A only measures at "1, 2, 4, 10" intervals so for 7 MHz we must do a bit of interpolation of the "Q" value of the 4 and 10 MHz values which would put it at around 235 for the silver-plated coil and 35 for the stainless steel coil if we trust the values with the asterisks.  (In retrospect, I could have used the '4191A as well for the 6.6/7 MHz frequency, but did not.)

Resetting the coils for the 20 meter paint mark, we get these results:

Frequency (MHz)	Inductance (uH) (Silver-plated)	Q (Ag-plated)	Inductance (uH) (Stainless)	Q (Stainless)
1	3.71	68	2.08	11
2	3.69	108	2.06	18
4	3.72	115	2.05	27
10	3.96	256	2.15	35
15	4.3 *	92 *	2.4	34
20	5.1 *	63 *	2.9	22
25	-	-	4.1 *	25 *
30	-	-	9.2 *	11 *

From the above measurements it appears that the paint marks may have been 1 turn different from each other - explaining the different inductance values - but the point here is to note the "Q" value, which would be generally representative at that frequency.  As can be seen, the "Q" of the stainless steel coil is still pretty bad - never exceeding about 35 - while the "Q" of the silver-plated coil likely 100 or  better at any frequency at which we might use that tap.  (It's possible that the Q of the silver-plated coil at 15 MHz is higher than indicated due to self-resonant effects - see the discussion below.)

These values generally agree with what was measured using the methods outlined above using W7ZOI's techniques.  Anyone who has attempted to measure the "Q" of an inductor likely noted that when the values exceed 100 or so, even the slightest amount of Ohmic resistance has a large effect on the reading, and - especially as the frequency increases above a few MHz - any object near the inductor under test that has dielectric or conductive properties (plastic, metal, fingers!) will skew the results

Measuring a physically-large coil at these frequencies is awkward:  The test equipment itself is metal, meaning that its proximity affects the measurements but we cannot use long wire leads to space it far enough away to avoid this effect as this would affect inductance and also the Q, so we can only do the best that we can

Figure 9:
Coil being tested on the HP-4275A
Click on the image for a larger version.

You may have noted that some values in the tables have an asterisk (*) next to them:  These indicate that self-resonance may be at play here, skewing the "Q" and inductance values - and for the silver-plated coil, the readings at 25 and 30 MHz were nonsensical and varied considerably when I move my hand anywhere near the set-up (further indicating the probability of self-resonance) which is why they were not included.

For this coil - and almost any coil that is adjusted using an adjustable tap (including switched-tap coils and roller inductors) - an issue can arise when only a small portion of that coil is used to attain a low inductance.  As an example, for this coil - when tapped at the 20 meter point - there is a large percentage of that coil that is "unused" and those turns beyond the tap that are "floating", but still inductively coupled to the part of the coil that is in use, acting as an auto-transformer, potentially increasing voltage at the far end.  Since these unused turns still have inductance and capacitance to free space, there will be a natural resonant frequency.  In some devices the designers choose to short the "unused" portion of the coil to reduce this effect, but this can have the affect of reducing Q and increasing losses as it represents an auto-transformer with a shorted output - but this may be an acceptable trade-off to avoid other issues, such as high-voltage arcing. Of course, one way to mitigate this would be to have another coil with fewer turns to minimize this effect when lower values of inductance were required.

In-circuit (e.g. in the antenna) I would expect that the self-resonance of this coil would be at least somewhat quashed, but the "in-situ" measurement of "Q" is a bit more difficult and beyond the scope of this article.

 * * *

Testing with the JPC-12 vertical and JPC-7 loaded dipole.

As noted earlier, the rewound coil was initially tested on the JPC-12 loaded vertical on 40 meters - mostly because it uses only a single coil and at that time I had rewound only one with silver-plated wire.  While I was at it I decided to see if I could detect any difference in the current flowing through the coil at a given RF power output related with the use of the original (and lossy) stainless steel coil and the silver plated coil.  Again, figure 7 shows this rewound coil with a thermal infrared camera just after a 60 second key-down at 75 watts, the temperature rise being just 3F (<2C).

Let us now consider the measured resistive losses of the coil (let's say 20 ohms for the stainless coil, 4 ohms for the silver-plated one) at 75 watts - the power at which we observed the temperature rise.  As we know the approximate current to be expected (about 600mA at 20 watts as measured with a known-accurate thermocouple-type RF ammeter) we can calculate the apparent losses at 100 watts which would equate to about 40 watts for the stainless coil and 5.7 watts for the silver-plated coil.  What this means is that about half of the power is lost in the stainless steel coil - but this still represents less than 1 "S" unit of loss. ^{Footnote 2}

Note:  Judging by the ratio of the temperature rise between the two coils (3 degrees F for the silver-plated coil and 110F for the stainless) we would expect far greater difference in power loss between the two coils (more than 30-fold difference, so I'm likely missing something here).

Once I had two silver-plated coils and two stainless steel coils, I could do a direct comparison on the JPC-7 loaded dipole. The JPC-7 is more or less a pair of JPC-12 vertical on their sides, fed with a balun - but rather than having the ground (radial) system to "push" against when radiating RF, it - being a dipole - used both elements against each other and the "ground" under - unlike the vertical where the ground/radial participates directly in current flow - is somewhat less affecting of the impedance, although the proximity of the ground does have the effect of lowering feedpoint resistance and resonant frequency.

With the original stainless steel coils, the feedpoint resistance at resonance is "close enough" to 50 ohms to keep a radio without a tuner happy (it's actually lower than 50 ohms as noted below) - but consider that this means that each half of the dipole is closer to 25 ohms, the two being in series with each other:  With two coils' losses now in the mix - and the fact that a given loss of a coil in a 50 ohm circuit as a percentage was about half that of the same amount of resistance in a 25 ohm circuit - the losses are arguably worse, but "split" between the two elements.

While I didn't have the opportunity to use the thermal infrared camera to measure the temperature rise of the stainless coils on the JPC-7, they both got rather hot to the touch after key-down at 75 watts, indicating a roughly comparable amount of loss as did the original stainless steel coil on the JPC-12 vertical:  As with the vertical there was little change in temperature of the silver-plated coils.

Using a NanoVNA and minimal coax length  ^{Footnote 3} I set up the JPC-7 as per the the manufacturer's instructions on 40 meters with the antenna roughly 3 meters above ground - about the limit of stability for a portable tripod:  From the feed point there were two mast sections, the coil and then the telescoping rod on each side.  Carefully setting the coils and the element lengths to yield the lowest "R" value (e.g. at resonance), I then noted the "feedpoint" resistance at resonance (where reactance, or "J" = 0) using the stainless steel and then the silver plated coils:

• Stainless steel coils:  38 Ohms (1.32:1 VSWR)
  
• Silver plated coils:  15 ohms (3.4:1 VSWR)

It's worth noting that these "feedpoint" readings were taken with the supplied 1:1 balun inline along with a short length of coaxial cable so the above readings are NOT precisely those of the actual feedpoint resistance:  There is likely a bit of loss and transformation occurring in the aforementioned set-up (which includes the balun) so the absolute numbers above may not reflect the actual feedpoint resistance itself.  I also observed that on the JPC-7, the (normalized) 2:1 VSWR bandwidth was lower with the silver-plated coil - an expected effect with higher Q resonator coils.

Note:  On higher bands (e.g. 20 meters and up) the feedpoint impedance was much closer to 50 ohms with either coil and it's likely that nothing special will need to be done to keep a radio "happy".

One might be tempted at first to think that because of the higher VSWR, the silver plated coil constituted an antenna that was "worse" - but that would be wrong - this actually indicates the opposite.  What this measurement shows us is that the apparent total resistance of the two silver plated coils at 40 meters was 23 ohms less (about 11.5 ohms for each coil) than that of the silver plated coil - and this increased resistance is what accounts for the power being lost as heat.

This realization still leaves us with the problem that if we take away much of the loss of the coils we lower the feedpoint resistance which means that we can end up with a rather high VSWR - of over 3:1 - meaning that many radios won't be particularly happy with the situation without throwing a tuner into the mix - which, itself, could contribute the losses we just worked to rid.  This leaves us with several options:

• Pretend we didn't see this and continue using the stainless steel coils.  This is an obvious choice and I can attest that both the JPC-7 and JPC-12 antennas do work pretty well despite the loss of the coil, but personally, I can't "un-see" the lossy nature of these coils, so that's not an option for me.  As a "portable" antenna is all about compromise of performance, I prefer to minimize the deleterious effects of as many aspects of this "compromise" as I reasonably can.
  

• Use an antenna tuner.  Placing a tuner at the antenna is the preferred choice as it will minimize mismatch losses that will result if the tuner is placed at the far end of the cable feeding the antenna (e.g. in the radio.) Whether the magnitude of mismatched loss of the cable when the tuner is placed at the distal (radio) end of the feedline to match the lower-loss silver-plated coil is worse than using no tuner at all with the stainless steel coil cannot be easily answered without knowing the properties of the coax used and how a specific tuner works under the impedance conditions that it might see.  It's worth noting that as the "R" of a load on a tuner drops much below 50 ohms, a typical antenna tuner rapidly becomes less efficient - and can be quite lossy at 10-15 ohms.
  

• Rework the balun.  The JPC-7 has a 1:1 balun (one that isn't very "balanced" - but that's another topic) but it is clear that you could  choose a balun that inherently provides a suitable transformation - but more than one such balun would be required to cover all bands.
  

• Autotransformer.  A tapped autotransformer used to be a common "thing" many years ago for matching short verticals (e.g. mobile installations) to deal with the low feedpoint resistances at resonance - often well under 20 ohms for a low-loss coil.

These devices seem to be less common these days, but if you look carefully they may still be found on the surplus market - namely the Atlas MT-1 and Swan/Cubic/Siltronix MMBX, both of which offer selections of impedances that will easily yield 1.5:1 VSWR or better at any likely feedpoint resistance at and below 50 ohms.

 

I have tested the Atlas MT-1 (by putting two units back-to-back) and found a single unit to have about 0.2dB of loss on 40 meters which theoretically represents about 5% power loss per-unit - likely a fraction of the loss that would occur even in a good-quality antenna tuner, particularly when it is matching low impedances.  (Useful articles about RF autotransformers may be found in the November 1976 issue of "Ham Radio" magazine - link and the December, 1982 QST - link.)

As mentioned previously, the losses of the stainless steel coil are "lower than about an S-unit" on the lower bands so the user would have to weigh the benefits of the potential losses incurred by matching a silver-plated coil and additional matching versus just using the stainless steel coil and getting a more convenient match and just "eating" the losses.

Conclusion:

The reader should not go away thinking that antennas using loading coils wound with stainless steel wire don't work:  They do - and can be quite effective - but... 

In my measurements, the losses added by the stainless steel coils amounted to roughly "an S-unit" (more or less - mostly less) in a worst-case situation for the vertical antenna and somewhat more loss for the loaded dipole.  I have very successfully used both antennas with their original stainless steel coils for portable, remote and POTA operations with good results.  The difference of "about an S-unit" may be an issue for marginal situations using SSB, but it's less likely to be a problem for CW or digital modes under the same band conditions and distances where the signal margins are lower.

As electrically-small HF antennas will often have lower feedpoint resistance than their full-sized counterparts this means that intentionally using low-loss coils can shift the impedance well below 50 ohms, complicating the matching of the radio to it - particularly in the case of the loaded dipole:  The use of a radio's built-in antenna tuner - particularly with a long length of coax - may well incur losses greater than those of the lossy stainless steel coil without a tuner.

I'm guessing that the use of stainless steel wire for the coils is at least partly a result of it "simplifying" the operation of a portable antenna by resistively (lossily!) providing a feedpoint resistance closer to 50 ohms.  From a standpoint of operational simplicity and cost (both avoiding more complicated matching arrangements) the use of stainless steel - and simply "eating" the power loss - may be a reasonable compromise for most users.

But, it's not as simple as that.  The above is certainly true for the loaded dipole where the feedpoint resistance ends up being quite low (15 ohms on 40 meters) but for the vertical - where more variables are at play (e.g. lengths of radials, length of vertical resonator) one can easily attain a good match (<2:1) to 50 ohms even with the lower loss of the silver plated inductor coming into play.

All of the above should also point to something else:  In my respective articles about the JPC-7 and JPC-12 antennas I noted that performance could be improved by making them electrically "larger" (e.g. the addition of a top hat to the JPC-12 and "droop" wires on the JPC-7) which both reduces the amount of loading inductance and likely increases the feedpoint resistance - both of which contribute to improved efficiency.

Should you toss or rewind your stainless steel loading coil in favor of something using lower-loss material?  If you are trying to eke out every last bit of efficiency from your portable antenna and are prepared to deal with the possibility of slightly more complicated matching requirements (at least on the lower HF bands like 40 and 30 meters) to deal with potentially low feedpoint resistance - then perhaps.  If you operate a lot of SSB, operate using high power (>= 100 watts) and/or high duty cycle, it may well be worth doing what you can to reduce at least one of the sources of loss of these types of portable antenna systems and a potential failure point due to heat.

* * * * *

Footnotes:

1. This silver-plated jewelry wire that I used is varnished, so it's not actually bare - but this poses no problem with this project:  The protective coating is pierced when the new wire is clamped under the posts and the slider easily "bites" through it after having moved across it a few times, so there is absolutely no need to strip it.  The varnish on the rest of the coil offers protection from oxidation and while silver oxide is a reasonably good conductor, unoxidized silver is much better, so the coating is left intact.
2. The term "S Unit" is occasionally used in this article, but always with a bit of "hand waving" indicative of its ambiguity.  An "official" international definition of an S Unit is a 6 dB difference in signal level according to IARU Region 1 Technical Recommendation R.1 (where "S9" = -73dBm into 50 ohms - link).  While U.S.-made radios and many SDR programs use this definition by default, Japanese radios are often calibrated with 3 dB S-units meaning that for these radios, smaller amounts of signal change are more strongly indicated.  The reader should always note that while modern SDR-based receivers often do have reasonably good relative signal indications (e.g. the S-meter moves as it should for given changes in signal level) this is likely not true for older, analog radios.
3. For both transmitter and VNA testing, minimal coax length was used.  For the former, a very short (15cm) coax jumper was used, connected directly between the radio and the antenna feed, the radio being powered by battery.  For the VNA, the instrument was connected similarly - the 15cm coax for the JPC-12 and hanging directly from the JPC-7's balun - to minimize possible effects of common-mode RF currents on the antenna.  In real-world operation this would be emulated by using an effective common-mode choke as close to the antenna feed as possible. 

Related articles:

• Observations, analysis and field use of the JPC-7 portable "dipole" antenna - link.
• Observations, analysis and modifications of the JPC-12 vertical antenna - link.
• "The Two Faces of Q" by Wes, W7ZOI - link.
• About Q-factor of RF inductance coil - link.
• High-Q RF Coil Construction Techniques by Serge Stroobandt, ON4AA - link.
  

   * * * * *

This page stolen from ka7oei.blogspot.com

 

[END]