Monday, November 28, 2016

Transmission Line?

A Look At How Short Coax Can Act As Lumped Elements At Low Frequencies

The dilemma of getting a balanced wire or ladderline into the shack has perplexed a few hams over the years. Common solutions to the problem is to run a short piece of high quality coax, such as RG-213, through the wall and attach the ladderline to the coax that goes through the wall. The advice goes on to say to run a balun or just attach the balanced tuner output to the coax that is on the inside of the shack. Hmm, this has always bothered me because of the shortness of the coax not acting like coax at all on low frequencies but a capacitor with low Q. I didn't realize how bad this could be until now, let alone the trickery involved in getting rid of the common mode current.

Some Smithing To Get Us Started

Interesting link on QRZ forum,
Dan shows that the contribution of the lengths of transmission lines affect the transformed impedance in baluns.

One thing that always bothered me is that coax at extremely low frequencies relative to physical length behaves as lumped components with real values of Q causing a shift in the results.

I recreated the Z in SimSmith using Dan’s impedance and 2 sections of TL line just like he has, 2.75 feet total. I then replaced those sections of line with a series L (211.75nH) and C (84.7pF) if I did the maths right.

This creates an input Z of 21.6 - j124 ohms.

Then I switched over to lumped elements with infinity Q,

About the same input Z, 21.4 - j123 ohms.

But since coax exhibits non-ideal Q I did a search for what the Q of coax would be and retrieved this off the internet:

His Q of 25 seems reasonable so I plugged in Q of 25 into the SS model and got this:

A different result, 29.9 - j122 ohms. Meh, not that different.

Try a High Z Load

While the previous example is for terminating impedances near the characteristic impedance of the TL, what happens when we throw a high Z load at the same circuit? Say a really high impedance of 5192 - j2100 ohms, a typical impedance with an 80m doublet fed on 40m at the input to 55 feet of JSC #1318 ladderline. (Okay, I upped the Q to 50).

I set the source Zo(a) to be 100 Watts to show how much power get burned up in each element when you have a high Z load hanging on the end of that 2.75 foot piece of RG-213 coax.

The reveal below shows how pretty much all of your 100Watts gets burned up in the capacitance of the coax.

Zooming in on the details, we see that the capacitance within the coax is burning up 88W, almost all of the power while the inductance within the coax is dissipating little RF power, for a coax Q of 50.

The "up" arrow is mnemonically remembered as "burned up" in that element. If you add up the powers dissipated in each element, it should total the input power from the generator on the right. As well, the "right arrow" in from the left indicates the power coming in from the previous element when you left click on the W symbol you cycle through all the selections.

A Better Approach

Get some brass rod or at least some 3/16 to 1/4 inch threaded rod. Commonly called all thread or reddy-rod and drill two small holes through the shack wall (about an inch or more apart) and continue the balanced line from outside to in by affixing the balanced TL on the outside to the two rods and likewise on the inside. Done.

Printed On Recycled Data

Sunday, November 13, 2016

15/20m Doublet Using 300 TV Twin Lead and a 1:1 Balun

Here is a doublet for 15 and 20m

The antenna length is 31.9 feet long and the twin lead is 31 feet long. The EZNEC screenshot below is the current distribution on 15m. The antenna is up 30 feet in the middle and the ends are at 26 feet. This is a design by my friend John, KN5L. He came up with this idea a while ago and I wanted to share. It's the same concept as optimizing the 88 foot doublet and TL to present 50 ohms to the feedpoint.

The antenna pattern on 15m is this.
The 15m elevation cut is this.

The 20m pattern is this.

And the 20m elevation cut is this.

When you attached 31 feet of TV transmission line I got from and a 1:1 balun you obtain this  SWR and power delivered response. The pink line and left axis represents the power delivered to the antenna (load). The SWR is pretty flat when I build this antenna. I originally had a sloper in mind but of course you can make it a flat top or inverted vee, your choice. It doesn't change the SWR much, if at all.

The measured k-coefficients were obtained from taking a short/open measurement then dumping those parameters into AC6LA's Z-Plots spreadsheet and let it calculate them for you.

You will notice that the power delivered to the load is over 1-dB down on 15m. Given that the gain is up around 8dB, I thought it to be an acceptable trade for this lightweight doublet.

PPD-JR Balanced Line Tuner on this TV TL Doublet

Just for fun I placed the Xmtch on the generator of the SimSmith circuit here. It reveals the loss you would incur for this antenna from 20 through 10 meters using a perfect matching circuit to attach to the input of the antenna line. Again the pink line is the predicted loss given the antenna and TL configuration as previously described.

As you can see, the power applied to the antenna is reduced by the mismatch loss of the TV twin lead. If you wanted to go above 15m, you will suffer more loss.

Myron WV0H
Printed On Recycled Data

Saturday, November 5, 2016

Winter HF Antenna System

Here is my handiwork from last weekend. I put up an 88-foot doublet (20m EDZ at 39ft/ends at 20 feet) with broadside 20m radiation SE/NW and an 89ft Quad with broadside 20m radiation NE/SW, basically orthogonal to the doublet. I attached my SGC-230 tuner to the top of the bottom section (4ft), zip tied it all together so I can take it down in June for field day next year. A good test run. Of course, you can run all HF bands on this get-up with varying pattern shapes but this affords the greatest flexibility given my lot dimensions.

I used Teflon wire for both antennas to keep ice from forming on them. Found some cheap sources on eBay, like 150 feet of #24 (19/36) for $20. I guess that another unknown thing about Quads, you don’t have to have the bottom very far off the ground to keep their performance intact. Sure raising them up to a decent height will improve their gain performance but not substantially, the patterns remain symmetrical. The top of the Quad is at 35 feet and the bottom is at 4 feet with 22.3 foot sides. I use #24 (19/36) PTFE Nickel plated copper wire. It’s just what they had at this eBay store at the time. Turns out Nickel plating helps prevent corrosion and is a hair better in RF performance over Zinc plated copper. Would have preferred silver plated copper though.

I put a DPDT relay in there to allow me to switch between the Quad and the Doublet back in the shack. The DC is sent out to the pole with an electrical extension cord using the third wire to energize the relay and a SPST switch back in the shack to toggle between the Quad or the Doublet. The tuner is made balanced by a nice 4000-ohm choke (7-14MHz) I created from a 25” TV CRT vertical deflection coil on the tuner input. At 3.5MHz the choking action is less, maybe only 1000-ohms but its enough to balance the very low impedance doublet input impedance, around 8 ohms.

The doublet has a tuned (41 foot) JSC-1318 windowline (Universal Radio) I used in my previous blog post about the 80/40/20/17/12/10m 88 foot doublet I also used at 2015 Field Day.

Since the Quad is the ubiquitous 1005/10.1 ratio on 30 meters it also works great on 20 through 12m before getting lobey on 10m, but then I just switch back to the doublet for 10m anyway, if 10 should ever open again.

Since every telescopic pole drains water down into the bottom, I created a drain in the PVC cap at the bottom by drilling many 3/16” holes to make it look like a strainer, so the water can drain out during rainy season. I guyed the bottom section with a green army camo pole coupling nut slipped over the top of the bottom section of the 12m Spiderbeam pole.

I put a SS bolt in the bottom cap to keep the base from kicking out of the area and placed another end cap upside down with a mating hole for the bolt in the end cap to mate up and hold. The ground will get soft and may have to come up with something different than the upside down PVC end cap. May have to use my Teflon puck I used for my previous version of a Quad I had up in the previous years.

The coax running out there is a 55’ piece of ½” Andrew Heliax LDF4-50 I picked up for free at a swap fest because the guy didn’t want it because he had no connectors for it (guess I got lucky because I got connectors I picked up from another guy in the past who didn’t have any Heliax to put them on). I weather-proofed the connections (after this photo was taken) with coax-seal from Radio Shack and claimed victory!

But Wait!

It couldn't tune 3540-3570kHz after installing the relay...Hmmm, too much added shunt capacitance I suspect.

After making an antenna measurement with the Sark-110 and it telling me that the input impedance shifted the conclusion is that the input impedance hit a dead zone of tuning range in the SGC-230 tuner, around 8+j150 ohms at 3560kHz.

Hmmm, what could I do to make the tuner happy again?

Switching the Sark-110 over to single frequency and tuner mode, it revealed a tuning solution of a shunt capacitor of 1000pF and series inductance of 2uH. This is to bring the entire input impedance back to 50 ohms, I just needed to nudge it a bit.

Tuner Helper

So what element is easiest to come up with? You say a capacitor? Yes, probably the right way to think as a shunt capacitor will have less loss than a series inductor. Maybe. But this is based on very large components that yield high unloaded Q. A capacitor that is large enough, say 200pF, with a Qu of 1000 is not going to be a silver mica at 3.5 MHz. Silver micas are all rated at 1MHz and the Q drops off the higher you go in frequency and not wanting a resistor to be across my TL on the higher bands I decided it was way easier to incorporate a series inductor for about the same system loss. It only needed to be about 1uH per leg of the doublet after some clip lead experimentation with some coil stock. 

A 1uH inductor is easily formed with #14 enameled wire and 11 - 12 turns on a 1/2" form, close spaced. Makes for about a Qu of 100 at 3.5MHz, and an SRF up around 219MHz, perfect. A trip to Radio Shack for a hobby box and banana jacks, all set.

After that I went through all of the other bands that the doublet has to offer and all life is good again, it tuned nicely.

Printed On Recycled Data

Saturday, October 1, 2016

A New PPD Length To Play With

A Day in the Backyard

New Doublet Lengths

I found some Teflon insulated #26 with 19/38 stranding on an internet shopping site and decided to get some. It was advertised as only being 150 feet for under $20 and I jumped because 100 feet of new stuff will set you back some cash. I wanted to test out a new doublet length that would keep the radiation from 15m broadside to the wire and keep 40m gain respectable.

The 66 foot PPD I had made a 15m pattern that was a cloverleaf due to it's longer than 1.25 wavelength radiator. I settled on two wires 60 feet long each. The length that gets "consumed" as transmission line is dependent upon how tall the mast is and I decided that 31 feet was the shortest I wanted to go. This created a 58 foot doublet. Shortening it caused the doublet to undergo some pattern shifting on 15 and that is what I wanted. Let's see how it did on 40.

Will the MiniBalun BLT and BL2 BLT Still Tune It?

Pretty much any decent length doublet fed with high impedance line such as the PPD will pretty much have an input impedance landing on the right side of the Smith chart. I didn't bother modeling it as I had a pretty good idea that it should work. But I couldn't help it, I modeled it.

Swept Gain

In reducing any doublet or dipole length you will sacrifice some gain when you go shorter than a half-wavelength. There is a point where the gain on the low end will taper off. How much you want to sacrifice is totally up to you. I chose to keep the loss around 0.7dB. But there is very little loss at 30m. This is a function of overall length and I wanted to go as short a possible to keep the pattern decent on 15m but not sacrificing too much gain on 40m.

This puts our doublet at about 58 feet long and a good compromise.

Here is the 15m broadside response when the apex is at 33 feet, ends at 20 feet.

And the 15m gain not counting feed line loss.

Here is the 20m pattern at the same heights.

Right Side Response

This is the response seen at the tuner feed point using SimSmith and the known k1 k2 k3 constants and measured TL parameters from yesteryear.

The antenna looks good so far, now let's set it up and test it out. Here is my backyard experiment.

I attached my KX2 and the MiniBalun BLT co-located at the feed point. I went through the bands and wrote down the tuner settings so could quickly return to any band.

Just a Toy

I acquired a FLIR TG165 the other day and just had to play. It really does well with subtle temperature gradients and changes but lacks a little accuracy is probably wholly dependent on emissivity so I wouldn't place any confidence in the temperature readings.

This is a screen grab of the Elecraft BL2 BLT. Looks weird in infrared.

This is a shot of the telescopic pole and the TL/wire.

One hot radio?

So there you have it, a visual look at what's cooking...not the inductor or balun, quite the opposite.

Myron WV0H
Printed On Recycled Data

Wednesday, September 21, 2016

Side KX Cover Mod for the KX2

Side KX and End Plates

I purchased the KX2 and have since installed the Gem Products Side KX cover and end plates for the rig, but...

Tight tolerances of the injection molded cover allow a perfect fit, trouble is removing it. I mean I had to really press down hard on one side of the cover to get that lid to pop off. Someone with fingers that has less strength may have great difficulty, not to mention the added stress on the end plates that place undo stress on the screws and screw holes.

Too Tight

Well, a quick phone call to Scott at Gem Products revealed possible fixes as we discussed possible solutions. More interested in finding a quick fix I thought of a way to relieve the stress in the cover and make removal easier. Make some relief slits to allow the side portion of the cover to flex inward a little.

Cutting Corners

To allow the side portion of the cover to bend in a little bit more, I relieved the stress in the corners of one side by cutting 45 degree angles at about 9/16" (15mm) deep. You can go deeper but I stopped at this depth because my saw was cutting at an angle towards the front. If you can cut precisely on the corner a deeper cut should allow for easier pushing of the cover tab.

Enter X-acto


A perfect solution with no impact to the functionality of the cover.

It's hard to see but there is just enough flex there to greatly relieve the stress on the cover and end plates.

You can install the cover in either direction, removal and replacement is unaffected by orientation of the cover..

Now, it's relatively easy to get that cover off.

Update October 9, 2016.

I had originally ordered my KX2 with the KXPD2 paddles but were backordered. I received them after publishing this blog post and wanted to tell you that the KXPD2 paddles do not fit on the KX2 when the cover is in place or vice versa. In order to remedy that, I took my cover to the 1 inch belt sander and ground a notch about 1/16 inch deep on the cover where the top of the KXPD2 meets the cover. Now I can install the cover and leave the KXPD2 paddles attached.

(I'll attach a photo tomorrow).

Myron WV0H
Printed On Recycled Data

Wednesday, September 7, 2016

A 80/40/10 Meter Doublet Fed With LadderLine

Ever since Field Day 2015 and the incorporation of the 88 foot doublet and tuned feeder giving a 50-ohm match for 40/20/17/12 meters I decided to see if I could make an additional doublet for 80 and 40 meters using the same concept, a tuned feeder with a correct length of deployed wire.

I modeled it set up at the same heights as the 88 footer, 39 feet in the middle and 31 feet at the ends yielding a 159 foot doublet.

Getting this much wire in the air always produces some decent gain on 80m and this example is no exception.

And the 40m gain is a real plus as well.

6 and 8dBi gain is nothing to sneeze at. You would be hard pressed to get as much gain with any vertical system.

You might be thinking you would have to "aim" the antenna for 40m. This is only true if the sun is down and 40m propagation goes from bouncing off the atmosphere in the immediate area to longer skip that exists in the midnight hours. I'm afraid I'm asleep by this time.

Like the 88 foot doublet with 41 feet of JSC #1318 Ladder Line, we can achieve 50-ohm matching  with 88 feet of the same Ladder Line. Just attach a 1:1 balun such as the Elecraft BL2 to the LaderLine feed point to get to coax or whatever.

Max SWR of 1.8 on 40m, good, good and you pick up a bonus band like 10m.

Okay, okay. The 10m pattern is going to be crazy, right? Well, yes but look for yourself.

And the gain plot for 10m.

There are some nice lobes there (pardon the Ferengi slang)...10dBi, okay that has to count for something!

Oh, BTW, I ordered a KX2 and a QRP Guys Mini Tuner Kit today.

Myron WV0H
Printed On Recycled Data

Sunday, August 7, 2016

Tuner At The Antenna Or In The Shack?

Antenna Tuners at the feedpoint or in the shack?

I had a question regarding the placement of an antenna tuner yesterday and was wondering what effect it would have placing the antenna tuner at the feed point of a GAP Titan vertical that is already somewhat tuned. He wanted to know if placing an automatic antenna tuner out at the feed point would lower the transmission line power dissipation. He didn't say what the exact SWR was at the feed point so I assumed two values of 2:1. So which will be more efficient? Instinctively I would say tuner at the antenna. But by how much?

SimSmith Approximations

Choosing two cases involved picking two 2:1 SWR impedances that could be realized at the feed point of any antenna. This installation also involves 140 feet of Davis RF Buryflex coax from shack to antenna as well as 500W input to the coax from an Elecraft KPA-500 amplifier. (20m was assumed for this discussion). Conservative tuner element values were assumed as well as an L network. Most automatic tuners utilize this topology. And we are ignoring the non-perfect characteristics of coax and the stray reactance that is present in real tuners, etc.

I might add, each SimSmith "building block" contain information regarding it parameters. The R and X values are those values looking into the right side looking left. Also, the ^W symbol indicates the amount of power "burnt up" in that block. When you left click on that W symbol it toggles through a bunch of parameters, ^dBW, <-dBw, <-W, ^W.

^dBW is the amount of power in dBW burned up in that block.
<-dBW is the amount of power in dBW entering into that block from the right.
<-W is the amount of power in Watts entering into that block from the right.
^W is the amount of power in Watts "burned up" in that block.

The G block is the generator block, the transmitter, your rig on the right. The xMtch(a) is the setting that allows the generator to produce a constant level of power regardless of the impedance seen downstream, unaffected by reflections from an imperfect load attached to it. The (a) part is telling the source to be the value in Watts in the "a" field just below it. That way we can set the transmit power out with just one field entry.

Baseline Case: Matched Loss to 50 ohm load (85W dissipated in coax, 500W input)

Case 1A: 2:1, High Z (80+j33ohms), Tuner in Shack

Case 1B: 2:1, High Z (80+j33ohms), Tuner at Antenna 

Case 2A: 2:1, Low Z (27-j13 ohms), Tuner in Shack

 Case 2B: 2:1, Low Z (27-j13 ohms), Tuner at Antenna


There is only a 15 Watt savings in power dissipation in the coax by locating the tuner out at the antenna, a 0.15dB difference. So as long as the antenna feed point SWR is somewhat low it's hardly worth the trouble.

Printed On Recycled Data