Leaky Coax

Unfortunately, at those frequencies, any coax is too lossy to be useful.

I can't parse that, can you try again? How about multiple APs?

Solutions are liquids. I don't think liquids will work.

This has been covered somewhat in the past. See: |

Mitsubishi-cable URL has changed to:

Are you trying to cover the hallways or the rooms? RF doesn't turn corners (through doorways) very easily.

I only have some experience with bench testing leaky coax. If the coax is suffiently low loss to be useful, it will be big, fat, and expensive. Smaller diameter cables might work for short runs, but forget it past about 50ft.

How long are the hallways (in feet or meters)?

Methinks if you can run coax, you can also run CAT5 and use multiple access points. I suspect the 2nd access point will be much cheaper than the coax past about 100ft of cable.

Also, if you have a long thin hallway, shooting down the hallway with a directional (panel) antenna will work. However, if you expect to make the right angle turn through a doorway, it won't work. Another way that might work is to use directional couplers or "taps" along a low loss cable. This is similar to some CATV installations. I don't know of any cheap directional couplers or -10dB taps for 2.4GHz. One could be built from just a circuit board, 2 chip resistors, and three connectors, but I couldn't find any with Google.

William P.N. Smith wrote in news: snipped-for-privacy@4ax.com:

im trying to cover all of the rooms on the floor by an access point in the center of the hallway, and an antenna on each end. the first few rooms from the end of the hallway get the signal fine, but it looks like i'll need another antenna in the center.

as you mentioned, looks like ill need more access points...oh well

smowk

Coax is always Leaky, Drippy, Sneezy and Coughy. Do your networking with Cat5 or better. Does away with that crap.

Hi, What is leaky coax? Holes on the cable or nicked cable? It'll raise SWR on the transmission line ans causes signal loss. (going up as heat)

Tony Hwang hath wroth:

It's coax with either pieces ground off at regualar intervals, or slots die cut into the aluminum shielding. I couldn't find a decent picture. This is as close as I could get with the RCT series:

also makes something called Flexrad-600-DB and Flexrad-600-PVC, but I couldn't find any data sheets (in English) on the stuff.

Andrew also uses air dielectric Heliax, with the corrogated copper outside shield, and just grinds holes into the shielding. It's called RLX series. This worked well, but is a nightmare to handle as any bending would cause the cable to buckle.

As for VSWR, think about it for a moment. If the forward signal is mostly lost to radiation, then there is no RF left inside the coax to be reflected. Therefore, without a reflected signal, there's no VSWR. The coupled mode Radiax is actually a transmission line transformer that matches the 50 ohm coax impedance, to the 377 ohm free space impedance. In other words, it's an antenna.

There's quite a bit on Radiax in the Andrew catalog. Start here:

select "Catalog 38" for links.

Many moons ago, I was playing with some test gear in a screen room. I took a ten meter piece of RG-58A/U at 11.8 GHz. More signal was leaking out of the single braid coax than was making it to the far end. The insertion loss was over 1.0 dB per foot, and the stuff leaking out the side was only down about 25 dB. Changing it to RG-55/U brought the loss down a bit, but improved the shielding to almost 40 dB.

Old guy

Hmmm, Aha, talking about stub kind of things on transmission line. If Z matching is not maintained there will be always higher than desirable SWR. Tony

Hi, Did not measure the SWR there? Standing wave is bad thing. Tony

Tony Hwang hath wroth:

Nope. It's not a stub tuner for matching the line. Perhaps a few analogies might be helpful.

1. Pretend we have a long length of rather lossy 50 ohm coax. Terminate the end with a short or an open. The attenuation of the coax reduces both the forward incident power and the reflected power. Given sufficient coax loss, the VSWR at the source is completely independent of the load and approaches 1:1.
2. Pretend we have a terminated leaky piece of low loss 50 ohm coax cable. The radiated leakage is considered the same as forward loss because it reduces the forward incident power level. The coax is terminated, so there will not be any reflected power. It's the same as #1. Whether all the forward power goes away via dielectric absorption loss, or via radiated leakage, there's nothing left to reflect when it hits the load end. Therefore, the VSWR also approaches 1:1 in this arrangement.
3. Ever peel the jacket off a piece of Radio Shock RG-58a/u junk coax? My guess is no more than 75% coverage. (Real RG-58a/u is 96% coverage). There are HUGE holes between the strands in the braid. Yet, this coax will read almost a perfect 50 ohms on a TDR or VSWR guesser. That's because the degree of coverage is irrelevant in determining the impedance. (Note that I said impedance, not loss). You could peel off ALL the outside braid, and replace it with 3 or 4 straight wires running along the outside of the dielectric, and still have a 50 ohms impedance. The coax would radiate and leak horribly but it would still be a good match.

You mean indefinite length other wise certain wave length interval and it can be open or close termination. Also feed point has to be matched.

Leakyage is not a function of poor coax, it's rather result of mismatch which causes standing wave(= high SWR) Poor coax causes more loss for sure. Build a simple SWR meter and experriment. At this frequency you can easily make one from copper clad PCB and a few diodes and some sensitive meter.

Junk? I am mil-spec oriented all my life working on military gears.Mostly wave guide/hard line coax. We can even have non-radiating ladder line feeder if it is properly matched and tune(same thing). Just things get smaller in scale when frequency goes up like in Wi-Fi and more affected by environment like weather condition, geography; over jungle, desert, big body of water like ocean, etc. No sense talking because I am an old fart. What would I know any way, LOL!

>

Tony Hwang hath wroth:

No, not infinite length. Just long enough that the loss in the coax cable is sufficiently high that you cannot detect much of a reflection at the generator end of the coax. You can open, short, terminate, or mismatch the coax any way you want, and you won't see much of a change.

Incidentally, that's why I can get away with using RG-6a/u CATV 75ohm coax in a 50ohm system. The attenuation losses are so much larger than the losses from the mismatch.

I beg to differ. Leakage has little to do with VSWR. There is a skool of thought that suggests that a high VSWR system radiates. I don't subscribe to this theory as I've done the measurements and math on various high VSWR systems and there's never any left over signal that could be attributed to leakage or radiation.

Oh, I have, many times, on many frequencies.

Mil-spec coax is far better than what can be bought at Radio Shock. I was horrified when I saw what they were selling, with terrible shielding coverage. Probably adequate for CB (27MHz) but useless for anything higher in frequency. Yet, the VSWR was wonderful, which is my point. Most of the braid could be missing, and the VSWR would still be near 1:1.

Quit complaining. I'm 58 and still learning.

Well John, then you can settle the debate since you were there to witness, who is right, Darwin or the creatationists?

(couldn't resist)

fundamentalism, fundamentally wrong.

rico snipped-for-privacy@hotmail.com (Rico) hath wroth:

Dunno. However, I'm regularly engaged in a similar discussion. Which are better, public product acceptance or the standards committee? This is also known as de facto standards versus de jure standards. The debate has many of the creation versus evolution characteristics including the requisite emotional appeals, pontifications, and dogma. It can also make an otherwise normal participant appear to be demonic possessed. There's even a compromise skool of thought that resembles Scientific Creationism. Darwin had it easy.

Secret to long life...never waste energy resisting temptation.

Let he who is free of doubt, cast the first opinion.

Or if you played with rhombic or long wire transmitting antennas in the HF region - the termination at the far end of the antenna was usually a length of lossy transmission line - one commercial short wave station (WNYW located 30 KM ESE of Boston) was using a 500 foot long section of open wire transmission line - except it was made of #12 AWG steel rather than copper to increase the loss. Or would you rather try to buy a 600 Ohm 50 KW non-inductive resistor, and then shield it to prevent it from radiating? At 5 MHz, the return loss was around 20 dB which is a 1.2:1. (Hey Jeff, is the old ARINC site still on the Cabrillo Hiway between Half Moon and San Gregorio?)

As Jeff says - it's just a length that is sufficient that the return loss is masked by the insertion loss. Assume a short or open circuit and a line that's a very small fraction of a wavelength. Either one is going to reflect all the energy back - but let's just use a 99% reflection as an example. That's a return loss of about 0.09 dB, or a SWR of 200:1. Now, insert a twenty decibel attenuator in between the source and that short/open, and repeat the measurement. The return loss is now 40 dB (20 dB out, 20 dB back), as the reflection is now seen to be 1%. That's a

1.02 SWR. In fact, the SWR of the attenuator is likely to be larger than that caused by the short or open. That attenuator is the equivalent of feet of coax, guide, or whatever transmission line turns you on..

I have no idea where you got that idea - it's flat out wrong. By the way - did you ever work with slotted waveguide antennas?

That's not going to be very accurate, but it should allow you to experiment with some attenuators, and an open or short and discover what we're talking about.

All righty. Grab a pair of water pump pliers and a 1/2 inch diameter ball bearing, and squeeze that ball into the left half of the wide wall of some WR187 (C Band) waveguide carrying 5.0 GHz in TE1,0 mode, until the ball is half way into the guide, but do so without cracking the guide. If you know your waveguide theory, the bump on the inside is going to be a capacitive element. Think that might make for a SWR problem? How much leakage do you get because of it? You can repeat the test using RG-402 (.141 semi rigid), except you don't need the ball.

I started playing with radios in 1948, but didn't get my first commercial license until 1959. I've also worked on transmitters from below 70 KHz on up to 61 GHz, though most of my work was between 1 and 9.5 GHz.

Old guy

snipped-for-privacy@painkiller.example.tld (Moe Trin) hath wroth:

See:

's the old Army MARS station reworked for ham use. One of the antennas (not in the photos) was a truely monsterious rhombic. I think it's made out of copper-clad steel cable, but I'm not sure. It's not in the photos so I don't know if it's still there. I've never actually seen the load, but I'm told it's huge and inside a large steel cabinet.

I guess you mean the /Federal/RCA/AT&T/ITT/Globe HF CW station, KFS. It closed in 1999. Huge antenna farm. KFS was operated remotely from KPH near Bolinas. The histories are mixed together.

is now a museum. I'm not sure of the status of KFS. I think it was torn down and turned into condos or something. I tried to find it with Google Earth but couldn't find the building or antennas. The KFS call sign is still active, but now located in East Palo Alto: |

I like your explanation better than mine. Bottom line is that if you have enough attenuation, it doesn't matter what's at the other end of the coax cable. If the coax leaks signal, that's even better because anything that leaks, can't reflect and therefore can't add to the VSWR.

I would like some clarification of this."Standard Attenuation" is defined as the insertion loss of a linear two port device in a non-reflecting system which is initially connected together at the insertion point by a standard connector pair.The Standard Attenuation is the ratio expressed in decibels of the powers absorbed by the load before and after insertion of the linear two port device being calibrated. Therefore I would assume the return loss to be 20dB with a SWR of 1.222 By placing a 20dB Attenuator in line you have effectively altered the SWR and you would not get a 99% reflection, more like 10%, and the power of any reflected signal at the source would be more like 50dB down on the original signal. , as the reflection is now seen to be 1%. That's a

I've never heard of the term "standard attenuation". Reference please?

Nope. Let's walk through the exercise. See:

formulas.

Start with a lossless 50 ohm transmission line. Terminate it with a load that offers 99% power reflectivity which is a reflection coefficient of 0.99. Grinding: VSWR = (1 + 0.99) / (1 - 0.99) = 199:1 Return Loss is: -20 log (0.99) = -0.087 dB Loss due to mismatch is: -10 log (1 - 0.99^2) = -10 log (0.02) = -17dB That's terrible VSWR and horrible mismatch loss.

Now, we add a 20dB attenuator in series with our perfect transmission line. Since the line is perfect, it doesn't matter which end the attenuator is located. The 20dB attenuator simulates the effects of a lossy cable.

The return loss is now -40dB plus the previous -0.087dB or -40.087dB. Might was well call it 40dB. The reason it's 40dB instead of just 20 is because the attenuator reduces both the forward (incident) power and the return (reflected power).

Grinding some more: Reflection Coef = 10 ^ (-RLdb/20) = 10 ^ (-40/20) = 0.01 VSWR is: VSWR = (1 + .01) / (1 - 0.01) = 1.02:1 Loss due to mismatch is: -10 log (1 - 0.01^2) = -10 log (.999) = -0.0044dB That's almost perfect VSWR and almost no additional mismatch loss.

In other words, if you stuff an attenuator (or lossy coax) between the source and a horribly mismatched load, the mismatch loss will be entirely determined by the loss, and not by the load.

Also, note the graph at Fig 1 at:

shows the effects of attenuation on VSWR. Note that the more attenuation, the better (lower) the VSWR.

Incidentally, the same principle of line loss swamping out the effects to mismatch loss is the reason I can get away with using 75 ohm CATV coax (RG-6/u) in ethernet 10base2 systems. The reflected signal is simply lost in the high line attenuation. Without reflections to cause peaks (reinforcement) and nulls (cancellation) on the line, the VSWR at the source is an almost perfect 1:1. It doesn't work with short lengths because there's not enough attenuation, but I've done some really long runs (i.e. 900ft), and CATV coax works perfectly.