# Combining rg58u with rg59 u coax

Discussion in 'Antennas, Feedlines, Towers & Rotors' started by KX2P, Jan 6, 2019. 1. There are a whole lot of factors that come into play here.

First is the theoretical impedance of a half wave dipole vs it's height above ground: Now, let's assume the transmitter sees that impedance, for a fixed tuned transmitter the actual power output will change:

https://duckduckgo.com/?q=ten+tec+the+two+kinds+of+swr&t=fpas&ia=web

Then, there is the issue of measuring the actual vswr:
https://www.birdrf.com/~/media/Bird...-Measurements-Using-In-Line-Power-Meters.ashx

Mixing different impedance lines will cause a impedance transformation, depending on both line impedances, and the length of line in wavelengths:

https://duckduckgo.com/?q=impedance+change+with+a+series+stub&t=fpas&ia=web

And, the vswr will cause more loss in the line:

And you have to consider that 75 ohm coax, even "Hardline" is available for little or no money Considering all the above, there really is not much reason for the average station to run 50 ohm coax other than historical convention.

Rege

Edit: why 50 ohms vs 75? For a given dollar amount of material, 50 ohm coax handles a bit more power, and 75 ohm coax has a bit less loss.

https://www.microwaves101.com/encyclopedias/why-fifty-ohms

Rege

KC8VWM likes this.
2. @KX2P You might actually be able to use this to an advantage if you read this other QRZ forum thread.

3. So I went and modeled a #12 AWG center-fed horizontal wire, L ft long, 50 ft agl, over average dirt.

I used three pieces of coax to get from the antenna to the rig. The coax nearest the feedpoint is X ft long. The middle piece of coax is Y ft long, and the remainder of the coax nearest the rig is (50 -(X+Y)) ft long. In other words, the total length of the feedline is always 50ft.

I used the optimizer in AutoEz to find useful combinations of L, X, and Y.

The blue SWR50 plot is with L=33.45ft, and fed with RG58 all the way, which about the best you can do with a standard dipole.

The green plot is with L=33.1ft, fed with 34.5ft of RG59, followed by 15.5ft of RG58.

A slight improvement is the pink plot, L=33.6ft(the legend on the plot is wrong) , fed with 29.7ft of RG58, 5.1ft of RG59, and finally 15.2ft of RG58.

In all cases, the optimizer was tasked to find combinations of the above that resulted in the lowest SWR50 at 14.0 and 14.35MHz.

The simulator finds insignificant additional loss between the three cases. The simulator considers wire loss, dirt loss, and coax loss, for a -2.06 db overall loss. Connecting the source directly to the wire with zero coax makes the total loss -1.28 db... The 0.78 db difference is attributable to the coax loss. 4. How are you getting 1.28 db of loss with no coax at all?

Rege

5. As i said, the modeler takes into consideration wire losses, and losses in dirt below the antenna. If you add coax/baluns/matching networks,loading elements, etc, losses attributable to those added items all get considered.

The modeler effectively integrates all of the power intercepted by a mythical sphere big enough to encompass the far field of the antenna and compares that to the actual electrical power delivered to the antenna by the "source". The difference is the antenna "loss". This has nothing to do with the antenna's pattern gain in any given direction.

The OP in the this thread was asking about additional loss attributable to adding the RG59, of which, according to the modeler is not enough to even notice...

The NEC code is smarter than you think!

AA7EJ likes this.
6. That's not an answer, if you put 100 watts into the feedpoint of a dipole, it's all radiated.

There are
no

And you simply guessed at ground losses.

Perhaps some energy got dissapated in the dirt, perhaps not!

You can't even assume a transceivers built in "tuner" would have less loss matching a 50 ohm load instead of a 75 ohm load, and if absolute Maximum power from a transmitter is your goal you will find your rig will output more power into a impedance less than 50 ohms (at the risk of final transistor destruction, to be sure)

I believe your 3 transmission line scheme is poor engineering, and makes a lot of assumptions that would be difficult to implement in a lab setting, and would be next to impossible to implement in the field.

Just measuring vswr is a project in itself, all a 1:1 - a null, or absolutely no indicated reflected power- means on a typical meter is the actual vswr is somewhat less than 1.3:1 , and that's for a very expensive meter!

For the vast majority of hams, using modern autotune transmitters, a simple run of rg59, exactly long enough to reach from the rig to the antenna will give results within a fraction of a db from theoretically perfect.

Rege

7. Except what is dissipated as heat!

The dipole is lossless only if it is in free space, is made of unobtainium lossless wire, and the source driving it is right at the feed-point. NEC predicts that the "average gain = 1.00 = 0db" (zero loss)

Change the wire to #14 AWG Copper, the "average gain = 0.98 = -0.09db" (i^2*R loss, only)

Put the wire 1/4wl above "perfect lossless ground", the "average gain = 1.00 = 0db" (zero loss)

Put the wire 1/4wl above "Real/High Accuracy ground", Conductivity=5mS, Dielec Const=13, the "average gain = 0.766 = -1.16db" (average ground loss)

Put the wire 1/4wl above "Extremely Poor, Cities", Conductivity=1mS, Dielec Const=5, the "average gain = 0.642 = -1.93db" (worst ground loss)

Put the wire 1/4wl above "Salt Water", Conductivity=5mS, Dielec Const=81, the "average gain = 0.991 = -0.04db" (lowest actual ground loss)

All of those negative "gains" (i.e. losses) are attributable to physical things like I^2*R losses in wire, losses in coax, and losses in the earth as the far-field antenna pattern is formed when some of the energy reflects from the dirt within a few tens of wavelengths from directly under the antenna. btw- for the vertical antenna afficionados out there, this type of loss is always present, and occurs beyond the ends of your radials.

Yes, my model has three pieces of coax, where I specified their properties, including their losses. Coax losses are a function of length and SWR, and NEC accurately models coax loss because it computes the actual SWR in real time.

Yes and no. For comparisons between alternative antenna systems, you just pick your best guess and go with that. I showed a comparison between three ways of driving the same wire dipole, at the same height. NEC shows the expected SWR and the predicted losses for each version, the other losses remaining constant. Only the coax losses changed between tests.

NEC computes this when coming up with the "average gain" as I show above. It is only as good as your knowledge of your local conditions. Note that going from the best case (free space=salt water) to the worst case (city, urban) is about 2db in useful power radiated.

This has nothing to do with mixing coax to feed an antenna. A tuner doesn't know what kind of coax is connected to it; all it "sees" is the R and jX at its output terminals. NEC is really good at predicting what the R and jX seen by the tuner.

Why, because you cant cut coax and solder/crimp coax fittings? The OP asked what would happen if he joined dissimilar coax. All I pointed out to him is that if he is going to do that, there might be an optimum way of doing it; namely to flatten the SWR curve of a particular antenna by choosing where to place, and how long to make the various pieces of coax.

That just shows how primitive and old your equipment is.

Last edited: Jan 8, 2019
8. "But that is not what the OP asked about, is it?"

W7ARK
very nice reply , kudos, unfortunately .....

I think some of the participants need to have a discussion on how to calculate length of a diploe in cubits...
Preferably on a different , their own thread.

73 Shirley

9. Not necessarily. What happens if the 75Ω coax is a half wavelength or multiple? It's not always that simple. There's a lot of factors involved as you undoubtedly know, but that's a hard blanket statement to support. Not trying to start a contest here. I'm sure the OP has gotten the idea by now.

AI3V likes this.
10. And EzNec and similar antenna modelers are very good at showing you what happens if you feed an antenna with a complex (non-resonant, different Z than Zo of feedline) feedpoint impedance with various lengths of feedline. The modelers effectively do for you what you would otherwise do using a Smith chart, but they have the advantage of knowing how the actual antenna feedpoint impedance changes as a function of frequency at the same time as being able to model how the feedline transforms the changing antenna Z as a function of frequency. This makes it really simple to find useful combinations when matching antennas to a 50 Ohm source using linear transformers or stubs.