purpose of a tower-top amplifier

"Dana" hath wroth:

290K is the more commonly used value, but 80F (26.6C) is probably better for warmer climates.

Let's play with the numbers. This is a Wi-Fi newsgroup, so I'll run the numbers for Wi-Fi. 290K temperature. 25MHz bandwidth.

Noise floor = -198.6 + 10*log(290K) + 10*log(25*10^6) Noise floor = -198.6 + 24.6 * 74 = -100dBm

At 54Mbits/sec, the typical receiver sensitivity is -68dBm. Actually, it seems to vary from -66dBm to -74dBm). However, the -68dBm is at a BER of 10^-5. OFDM at 54Mbits/sec requires a minimum SNR of 24.6dB. Therefore, the real noise floor for the receiver front end is: -68dBm -24.6dBm = -92.6dBm

For calculations at other speeds, I created this table. The typical receiver is a DLink DI-624.

Speed Min Modulation Typical Noise Eb/No Sensit Floor mb/sec dB dBm dBm 11 7.0 CCK -82 -89 5.5 6.0 CCK -85 -91 2 1.6 DQPSK -86 -87.6 1 -3.0 DBPSK -89 -86 54 24.6 64QAM/OFDM -71 -95.6 48 24.1 64QAM/OFDM -71 -95 36 18.8 16QAM/OFDM -78 -96.8 24 17.0 16QAM/OFDM -79 -96 18 10.8 QPSK/OFDM -82 -92.8 12 9.0 QPSK/OFDM -84 -93 9 7.8 BPSK/OFDM -87 -94.8 6 6.0 BPSK/OFDM -88 -94

In a perfect world, the noise floor would be exactly the same value for all speeds. However, perfection is almost impossible and variations in demodulator efficiency, digital noise pickup, and test measurement accuracy result in a rather wide (10dB) range. I just hate reality.

So, back to the coax cable noise. I'll use the 54Mbit/sec numbers. With a thermal noise floor of -100dBm, and a receiver noise floor of

-92.6dBm, there is a 7.4dB margin. Since the coax cable contributes the same amount of noise whether it is a few inches or a few miles long, the maximum attenuation that can be tolerated in the coax cable is 7.4dB. Any less attenuation and the coax noise has no effect on the sensitivity. Any more attenuation, and the coax noise exceeds the signal level. By that time, the receiver losses from the coax attenuation will have a larger effect than the added coax noise, so again the coax noise is negligible. Incidentally, there are a few assumptions in this simplification that I'm too lazy to detail. The

7.4dB is not the exact number but close enough.

Incidentally, 7.4dB is about 100ft of LMR-400 plus connector losses.

Noise floor issues yes. Coax generated noise, no. I don't know what he was smoking when he wrote that article.

The actual answer is yes, if you want a reliable link

It is obvious what you know and do not know. Seems you have not been involved in long distance radio links in the VHF, UHF, or Microwave bands. Then you would appreciate what the article was about.

Like most everything, it sure does generate noise. But I do not think you understand what noise is being generated and how it affects receiver sensitivity.

decaturtxcowboy hath wroth:

You're both right. Anything that's not at absolute zero will generate thermal noise. However, the coax cable does NOT generate additional noise that would be additive. For example, if the front end noise temperature were 290K, attaching a 50 ohm dummy load to the front end would not increase the noise level. Putting a piece of coax cable between the front end and the dummy load would also not increase the noise level. Replacing the 50 ohm load with an antenna but burying the antenna inside an RF absorbent box, would also not generate any additional noise. That's because off of these devices (load, coax, antenna) are at the same temperature. The equivalent noise temperature looks exactly like a perfect termination with a 290K noise source in parallel in all cases.

If the cable generated additional noise, that noise level would need to be added to the front end thermal noise contribution. This is not the case and the original article stated the same thing.

A common misconception is that larger objects "retain" more heat and therefore generate more noise. Nope. It's only the absolute temperature that determines the noise level. A tiny 50 ohm chip resistor termination generates exactly the same amount of thermal noise as a much larger dummy load, antenna, or combination of coax and antenna.

However, when we attach an antenna to the system, and point it at various thermal objects, things change. For example, a 0dBi antenna pointed at the ground will have exactly the same thermal noise contribution as the original 290K and will not increase the noise level. The same antenna pointed at the sun will have a 25,000K contribution and will probably bury the signal bandwidth in added noise. Pointing it at empty space will at 3K (cosmic background radiation).

There are even microwave fire detectors, that use the microwave emissions of a fire to detect and find fires.

The nice part is that microwave frequencies go through may objects allowing the detector to find hidden hot spots.

Where this become important is when the front end noise figure of the receiver starts to approach the thermal noise floor. GaAs FET front ends are close. If cryogenically cooled, even closer. Now, it becomes important to consider the thermal noise contributions of the surroundings. However, it's the thermal noise component of whatever the antenna is looking at that's important, not the coax cable. The coax, connectors, and antenna only contribute the usual 290K which is already part of the noise bandwidth equation.

Hi, What is the definition of noise? Like weed in our garden? Unwated signal is considered noise. Unwanted plant is weed. Electrically unlinear joint like coax connector or any joint will generate harmonics. Temperature which is caused by excited electrons will generate thermionic noise; shot noise. In my working days I used to work on weak signal in the range of -80 to 100db and like RX front and gain of 9db wasa BIG boost making the signal more reliable. Also RX antenna up high on the tower can induce static discharge, corona arc, etc. That's why we DC ground antenna.

Never saw a coax radiating or picking up signal like crazy in real world? Want to experiment? Raise the SWR on the transmission line and scan(sniff)the coax along, what do you see?

Tony Hwang hath wroth:

Wrong. I do that experiment as a demonstration of common electronic assumptions (along with the loss through a mess of adapters, water in the coax, and others). Somewhere along the line, someone mumbled that VSWR causes RF to radiate down the outside of the coax shield. It made sense because every time I ran coax to a high VSWR antenna, I had the coax radiate enough RF to light up fluorescent bulbs in the shack.

However, that's not the way it works. The previous example is an uncontrolled environment. What's happening is that the coax cable has become part of the antenna system and will radiate as described. But, what if I replace the high VSWR antenna with a high VSWR dummy load? If the reflected signal comes down the outside of the coax, as is commonly suggested, then it should radiate as badly as the high VSWR antenna. It doesn't.

You can try your RF sniffer experiment the same way. Put a high VSWR dummy load on the end of a piece of coax and go sniffing for RF. You won't find any. As long as the field between the center conductor and the shield is totally enclosed, you can have a high VSWR termination, but no radiation outside the shield.

Water in the coax or feedline, how I hate that, especially here in Alaska where the water then freezes inside the cable. Talk about a gremlin type of problem.

Somewhere along the line, someone mumbled that

Hmmm, How come a guy miles away picked up my signal when I was folling around with matched dummy load? When Z is smatched SWR is low but perfect SWR does not gurantee perfect match. You explained it just now. Yes, coax can be leaky electrically. Hmmm, No wonder mil-spec stuffs have double triple shielding compared to commercial counter part.

No wonder wave guide feeding antenna is usually pressurized with Nitrogen gas.

Tony Hwang hath wroth:

Nope. They use dried compressed air. Dry nitrogen is too expensive and juggling bottles is a maintenance nightmare.

Tony Hwang hath wroth:

Your unspecified radio probably wasn't very well shielded. You'll be amazed at how much RF is conducted out of various holes and cables on the typical ham radio. Need some entertainment value? Build an RF sniffer or use a scope with a small loop at the end of a piece of coax. Put a dummy load on your xmitter and sniff around the case, mic cable, power cord, and tuner connections and see how much RF is leaking out. Then do the math to figure out the field strength. Perverse square law says that the guy that was "miles" away might get as strong a signal as someone across the country if you used a proper antenna. Just a wild guess, but methinks that a really good HF radio can offer about -30dB shielding. 100 watts is +50dBm, so you've got perhaps +20dBm or at least 100 milliwatts for RF. Welcome to QRP.

Do the math. Typical braided coax has about 98% coverage. That means that somewhat less than 2% of the power leaks out. I'll guess 1%. (Actually, it varies with frequency and length but I don't wanna get complicated). 1% = 0.01 = -20dB. If you have 100 watts (+50dBm) floating around the coax cable, we have somewhat less than 50 - 20 =

+30dBm or 1 watt of power leaking out of the coax. Welcome to QRP again.

I prefer foil shielding instead of multiple layers of braid. Braid to too messy to crimp.

That true, but its not the coax (per se) that is creating the noise where the classical definition of noise is any unwanted signal - its the mismatch of impedances creating standing waves.

"Dana" wrote in message news: snipped-for-privacy@corp.supernews.com...

| > > On Jan 6, 10:12 am, "Dana" wrote: | > > > Everything produces noise. | > > > And in the applications I was describing we had anywhere from 100 to | 600 | > > > feet of cable to deal with. | > >

| > > Wrong. Coax cable does not normally generate any noise by itself. | | Here is an article that describes noise in RF systems |

| All matter at temperatures above absolute zero (0K, about -460F) radiates | electromagnetic energy. The amount of energy is related to temperature -- | the hotter the matter, the more energy is radiated. This energy is described | by Boltzmann's Constant, 'k' (k = -198.6dBm/degreesK-Hz). This constant, | multiplied by the temperature of the matter a receiver views and the system | bandwidth, yields an irreducible background noise against which a desired | signal must compete. This is thermal noise. | | In a cellular system, the receiving antennas are designed to view the ground | around the site because that's where the subscribers are. The ground | temperature varies, but at 80F, it's about 300K (T=300K). RF engineers | typically use this number as a rule of thumb. | | The receiver bandwidth varies depending on the technology, but the same | principles hold for all technologies. EAMPS, for example, uses 30kHz-wide | channels. Receiver bandwidth is a bit less than 30kHz, for rejection of | adjacent channels. Assume the typical EAMPS receiver has a bandwidth of | 25kHz (B=25,000Hz). By making this assumption, you can calculate the amount | of noise an EAMPS receiver will have in its passband if it contributes no | noise of its own. | | This receiver thermal noise floor often is referred to as 'kTB.' In the | example, assume consistent units: | | kTB = -198.6 + 10 Log(300) | | + 10 Log(25,000) in dBm | | kTB = -129.8dBm | | Thus, if you build a perfect EAMPS cellular receiver, it would have -129. | 8dBm of noise in its passband competing with the wanted signal. | | CABLE LOSS Cable, filters and other passive elements exhibit a loss and | produce thermal noise. | | If a cable (or other lossy element) has 10dB of loss, it will attenuate the | desired signal as well as the input noise by 10dB. But at the output of the | cable, you will see noise at least equal to kTB because the cable itself | contributes it. | | If you put a signal into the cable at -100dBm over a thermal noise of -129. | 8dBm, you have a signal-to-noise ratio of 29.8dB at the cable input. At the | cable output, the signal has been attenuated by 10dB to -110dBm. The noise | you put into the cable also has dropped the same amount, to -139.8dBm. But | the cable contributes its noise floor of -129.8dBm, so the combined | (uncorrelated) noise terms are -129.8dBm. The resulting signal-to-noise | ratio is only 19.8dB at the cable output. You sacrifice 10dB of | signal-to-noise ratio. This is why you spend money on 7/8", 15/8" or larger | coax at cell sites to reduce this loss.

In the real world such calculations mean little as with most wireless systems the problem is not a noise but interference floor. Much of this is in band interference but out of band can be significant especially in metro areas.

And have dehumidifiers. And that is if the site in question is even using waveguide to begin with.

radiates

temperature --

That is not true at all. We use those calculations when designing our cell sites and microwave shots (10 to 30 mile shots are not uncommon.

as with most wireless

Noise is interference

Yes urban areas create even more issues to overcome.

They most certainly use nitrogen gas in military aircraft. And it's no maintenance problem. The bottles are easily refilled every few days. Takes about

5 minutes.

But then...military may spend five times their flight time having ground maintenance done.

I've seen both compressed dry air cylinders and cryogenic tanks filled with liquid nitrogen next to splice cases.

Hmmm. Let me see, what is typical RF power this COMSATs are using? And distance it covers? 1W is lot of power for sure. 5GHz range M/W links on LOS or refraction path I used to deal with had TX power of

3 to 5W. So 1W is lot of power in my book. Then I used to work on TX power of 100s of Kilo Watts. Ever worked on High voltage in the range of 30KV? Nowadays it's different. Solid state devices don't need such high voltage.